I"

w
m
d
.,.

 

.
.34: =
2‘ 9...
. 33A

mum“.

$3.44.?” 3....

.33“

yuu.i\.%n
. ..

a... M

I
ti,
bk?
3...?
€-
.O:- a.
&!\.3. 51)::
$3.33.“?
3. diurl‘a u. u

.3.
. ..
£1
.1. 1
3.7:
1‘1:

.‘1
59th
“541.54.!
{$3532
it..\’ann
5’22.
.113}!
I?!
:l

s n fit»
31.0.2.3.

lit.
539%.. #5.... 1
1.... :......1-.L\.
cii§ilv

3:2. {-

 

 

LIBRARY
MICI’IIyaI I State
University

 

This is to certify that the
dissertation entitled

IDENTIFICATION OF eQTLs REGULATING EXPRESSION OF
EARLY LOW TEMPERATURE RESPONSIVE GENES IN
ARABIDOPSIS

presented by

Ritu Sharma

has been accepted towards fulfillment
of the requirements for the

PhD degree in Cell and Molecular Biology

 

 

 

' Major Professor’s Signature
37/“ /o 5/
Date

MSU is an affinnative-action, equal-opportunity employer

 

PLACE lN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KzlProleocsPresIClRCIDateDueindd

 

 

 

IDENTIFICA

LOW TEI

 

 

IDENTIFICATION OF eQTLs REGULATING EXPRESSION OF EARLY

LOW TEMPERATURE RESPONSIVE GENES IN ARABIDOPSIS

By

Ritu Sharma

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY
Cell and Molecular Biology

2008

 

 

lDENTEFlCA'

TEN

Many ,:
freezing tern;
acclimation im
as a central :
temperature re
oi the gene net

In this s:
early changes

0f Arab/'dop3;5

quantitative va

DODUIaiIQn Ge”
Segregation. 5

analysis detern
IeQTLSI M

wart

MOSI Of these c

 

Abstract

IDENTIFICATION OF eQTLs REGULATING EXPRESSION OF EARLY LOW
TEMPERATURE RESPONSIVE GENES IN ARABIDOPSIS
By

Ritu Sharma

Many plants have acquired the ability to enhance their tolerance to
freezing temperatures through the process of cold acclimation. Cold
acclimation involves a multitude of temporal transcriptional changes that serve
as a central step to developing freezing tolerance. Several hundred low
temperature responsive genes have been identified, however, the regulation
of the gene networks remains poorly understood.

In this study, natural genetic variation was used to identify regulators of
early changes in gene expression occurring at low temperatures. Comparison
of Arabidopsis accessions from different geo-climatic regions identified
quantitative variation in the expression of some early responsive genes.
Measurement of transcript levels as a quantitative trait in a mapping
population derived from a cross between Cvi-O X Ler revealed transgressive
segregation, suggesting allelic variation at multiple regulatory loci. Further
analysis determined that the heritable component underlying this variation in
gene expression was contributed by several expression quantitative trait loci
(eQTLs). Many transcripts were found to be regulated by multiple eQTLs.
Most of these eQTLs were trans-acting with the exception of two genes, CBF2

and 3, which showed ciS-acting eQTLs. Thus, the variation in gene expression

 

 

observed '~‘-
geoes ther'
posthe a'e
regulators C
regsiarf,’ er;
of the: targ
r gelate tre

pathway. wne
CORSITUCCOT

Fegc7atoz'y p:
chromosome
genes anaIyZe
temperature r
this study I”as
research She;
eQTLS- Which

networks at ,0},

 

observed was not merely due to sequence variation in the promoters of the
genes themselves. Furthermore, for some eQTLs, Cvi-O contributed the
positive allele, suggesting that these may arise from variation in negative
regulators of transcription for which weaker alleles in the accession not
regularly exposed to low temperature may have resulted in higher expression
of their target genes. Some of these eQTLs were identified to commonly
regulate the expression of multiple genes, suggesting a common regulatory
pathway, whereas others regulated unique genes. These findings allowed the
construction of a regulatory network for the genes analyzed with the
regulatory positions of the different eQTLs. Two adjacent eQTLs on
chromosome 1 were identified as the master regulators of seven of the nine
genes analyzed. Underlying these eQTLs could be early regulators of the low
temperature response, such as the elusive thermo-sensor(s). Significantly,
this study has served to narrow down the regions of the genome where further
research should be focused to identify the genes underlying the regulatory
eQTLs, which will facilitate our understanding of the regulation of gene

networks at low temperatures.

This thesis is dedicated to the loving memory of my mother,

Late Mrs. Sushila Sharma.

I would I
after Pam Greé
work on this pr
blurnng the DH
Committee. Ke-
for all their com

The pen.
challenging fOr
have been DOS
and faculty who

The dec:
Sound proleCt ir
chaliengmg DGr
Monaliy, Dan A
have not SUM“

I traVelec

an ellrichmg SC

 

Acknowledgements

I would like to thank Michael F. Thomashow for accepting me to his lab
after Pam Green moved to Delaware. I’m thankful to him for allowing me to
work on this project and trusting my instincts when technical limitations were
blurring the picture. I am also grateful to the members of my Guidance
Committee, Ken Keegstra, Steve Triezenberg, Barb Sears and Gregg Howe,
for all their compassion, guidance and support.

The period in which this study was completed proved to be extremely
challenging for several reasons and without the help of some, it would not
have been possible. I want to express my deep gratitude to family, friends,
and faculty who helped me to reach the end of this project.

The decision to not move with Pam to Delaware after having started a
sound project in her lab was a tough one. Starting afresh in a new lab was a
challenging period and if it were not for the support of my friends Kristin
Mcnally, Dan Appledorn, Gauri Jawdekar and Sarah Fowler I would literally
have not survived it. I can never thank them enough for their support.

I traveled halfway across the globe for my PhD in the hopes of finding
an enriching scientific experience. I would like to thank the Plant Research
Lab as a whole for providing me that experience through interaction with
some distinguished scientists, both internal and invited.

lam grateful to Jan Zeevaart for sharing his time and wisdom for some
very motivating discussions of my data. All throughout my graduate study
motivation and guidance has come to me from Steve Triezenberg. I am

extremely grateful to him for helping me grow as a scientist. I also thank Ken

Keegstra for h
the program.

I want t
Canella, Sara?—
to thank Anne
initiating me int
her workspace
the technical p
overcome then
helped me feel
their SUDpOrt w
Chang-530! an

anall/sis.

Finally I i

 

Keegstra for his patience and guidance during times I felt completely lost in
the program.

I want to thank the members of Thomashow lab, especially Donatella
Canella, Sarah Gilmour, and Marcella Carvallo, for their support. I would like
to thank Annette Thelen at the Research Technology Support Facility for
initiating me into the world of real time FIT-PCB and then letting me take over
her workspace to set up my experiments. She understood my frustration with
the technical problems I was facing and helped me keep faith in my ability to
overcome them. I will never forget her compassion. She and Jeff Landgraf
helped me feel welcome at the RTSF during the long hours I spent there and
their support was invaluable. Thanks are also due to Dr. Curt Wilkerson, Dr.
Chang-Bao, and Dr. Vinod Prabhu for their helpful suggestions during data
analysis.

Finally I would like to thank my family who believed in me until the very
end when I had almost quit. My father, my brother, my husband and my
parents-in-law have been the ones who pulled me through the last stage of

this PhD.

vi

Chapter1.
1.1 Int:
1.2 L0
1.3 Co

1.3.1
1.3.2

1.3.3
1.3.4
1.3.5
1.4 Perr
1-5 Natt
1-6 Und
1.7 OTL
1.8 Statl
1-9 Gen!
Tran:
1'10 Mal

chaPIEr 2.

2.1 "Wed
2'2 I"fesu
2.2.1
2.2.2 A

L
2.23 e

C
2.2;

2.2.3

 

Table of Contents

Chapter 1. Literature Review .............................................................. 1
1.1 Introduction ............................................................................ 1
1.2 Low —Temperature Stress ........................................................ 2
1.3 Cold Acclimation ..................................................................... 5

1.3.1 Biochemical Changes Accompanying Cold Acclimation ......... 6

1.3.2 Changes in Gene Expression Associated with Cold

Acclimation ................................................................. 8

1.3.3 The CBF Gene Regulatory Pathway .................................. 9

1.3.4 Transcriptional Regulation of CBF Genes ......................... 10

1.3.5 Conservation of CBF pathway in higher plants .................. 12
1.4 Perception of the low temperature stimulus ................................. 13
1.5 Naturally occurring genetic variation ........................................... 15
1.6 Understanding Complex Traits using Natural Variation ................... 17
1.7 OTL Mapping ........................................................................ 19
1.8 Statistical methods of OTL Mapping ........................................... 24
1.9 Genetical Genomics: OTL Mapping for Regulators of

Transcription ......................................................................... 25

1.10 Main Objectives of Current Study ............................................. 27

Chapter 2. eQTLs Regulating Low Temperature Responsive

Expression of CBF1 .................................................... 28
2.1 Introduction ........................................................................... 28
2.2 Results ................................................................................ 33
2.2.1 Natural variation in CBF gene expression ........................... 33
2.2.2 Analysis of low temperature responsive CBF expression in
Ler and Cvi-O ................................................................. 36
2.2.3 eQTL Mapping for Regulators of Low Temperature Responsive
CBF Expression .............................................................. 41
2.2.3.1 Identification of a Common Large-effect eOTL Regulating
CBF2 and 3 Expression at Low Temperature ................. 45
2.2.3.2 Identification of Common Small-effect eQTLs for
Expression of CBF 1, 2 and 3 Genes ............................ 45
2.2.3.3 Uncovering Additional Small-effect eQTLs for CBF
Expression ............................................................. 52
2.2.3.4 Use of Additional Normalization Control for Expression
Analysis ................................................................. 55
2.2.4 Identification of eQTLs Regulating of CBF 1, 2 and 3 Expression
using Act1 as Internal Reference ........................................ 57
2.3 Discussion ....................................................................... 68
2.3.1 Identification of a Major-effect eOTL Common for Regulation
of CBF2 and 3 Expression ............................................... 73
2.3.2 Identification of Additional eQTLs Using Actin mRNA for
Normalization ................................................................ 73
2.3.3 Genes Located in Regions Showing QTLS ........................... 74
2.4 Materials and Methods ............................................................. 77

vii

‘V‘M‘j

.
*fl-vm -ww

Cha

(1‘ r.\

Chapp

V-‘-\ 1“

Mg

is.

is

43
ChapIEr

Chapter 3. eQTLs Regulating Expression of Genes Induced Early During

Cold Acclimation ............................................................. 85
3.1 Introduction ......................................................................... 85
3.2 Results ................................................................................ 87
3.2.1 eQTL mapping .............................................................. 92
3.2.2 Identification of eQTLs using 188 rRNA as the Internal
Reference ................................................................... 101
3.2.3 Identification of eQTLs using Act1 as Internal Reference ....... 101
3.3 Discussion .......................................................................... 121
3.3.1 Common eQTLs Regulating Multiple Low Temperature-Induced
Genes ........................................................................ 123
3.3.2 Unique eQTLs Identified for Regulation of Specific Low
Temperature-Responsive Genes ..................................... 124

Chapter 4. Optimization of Taqman® Real Time RT-PCR for Molecular

Trait Measurement ......................................................... 125
4.1 Introduction ........................................................................ 125
4.2 Results ............................................................................. 127
4.2.1 Primer and Probe Optimization for Taqman® Real—time
RT-PCR ...................................................................... 127
4.2.2 Comparison of One- and Two-Step Method Towards
Optimization of Real-Time RT-PCR .................................. 131
4.2.3 Optimization to Remove DNA Contamination ...................... 136
4.2.4 Identification of Outliers and Their Effect on eQTL Mapping..138
4.3 Discussion ......................................................................... 143
Chapter 5. Discussion and Implications for the Future ...................... 146
References ................................................................................... 1 58

viii

VII “'1'.

 

1. Tab5e 2
2. Table 3

3. TabTe 3

 

 

List of Tables

1. Table 2.1 .............................................................................. 62
2. Table 3.1 .............................................................................. 83
3. Table 3.2 ............................................................................. 103

 

 

"I Hi II“

 

 

 

10.

11.

12.

13.

14.

15.

16.
17.

List of Figures

Figure 2.1. Geographical distribution of ArabidOpSis accessions
selected for analysis ............................................................ 34

Figure 2.2. CBF expression in Arabidopsis accessions exposed to
gradual reduction in ambient temperature ................................ 35

Figure 2.3. a. CBF expression in Cvi-O, Ler, and Ws following a

sudden drop in ambient temperature ..................................... 37
b. Histogram Showing CBF transcript levels
normalized with 188 rRNA .................................................. 37
Figure 2.4. a. Expression of CBF genes in a time course of Cvi-O,
Ler, and Ws, at 4°C ............................................................ 39
b. Graph showing CBF transcript levels normalized
with 18S rRNA levels .......................................................... 39

Figure 2.5. Comparison of CBF gene expression in Cvi-O and Ler in
response to sudden change in temperature ............................. 40

Figure 2.6. a. Expression of CBF genes in Cvi-O and Ler following

sudden change in ambient temperature .................................. 42

b. Histogram showing CBF transcript levels in
RILS ................................................................................ 42
Figure 2.7. Taqman real-time RT-PCR amplification profiles for a.
188 rRNA, b. CBF1, c. CBF2, and d. CBF3 transcripts .............. 44
Figure 2.8. eQTLs for CBF2 expression .................................. 47
Figure 2.9. eQTLs for CBF3 expression ................................. 49
Figure 2.10. eQTLs for CBF1 expression ................................ 51

Figure 2.11. eQTLs for CBF2 expression after holding CBF locus
constant .......................................................................... 53

Figure 2.12. eQTLs for CBF3 expression after holding CBF locus
constant .......................................................................... 54

Figure 2.13. eQTLs for CBF1 expression after holding CBF locus
constant .......................................................................... 56

Figure 2.14. Actin expression in response to low temperature....59

Figure 2.15. eQTLs for Act1 expression ................................. 61
Figure 2.16. eQTLs for CBF1 expression ................................ 63
Figure 2.17. eQTLs for CBF2 expression ................................ 65

 

 

i. 1.2-www-

18.

19.

20,

21

22.

23.

24.

25.

26.

27.
28.

29.

30.
31.

32.

33.

34.

3S.

36.

Fl

Fi
as

Fn
70x

FISU
Gene

FISU
8X95

FIGUI

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.
36.

Figure 2.18. eQTLs for CBF3 expression ................................ 67

Figure 2.19. Chromosomal locations of known low temperature
associated genes .............................................................. 76

Figure 3.1. Expression profiles of Zat12, Rav1, and ERF genes at
low temperature ................................................................. 88

Figure 3.2. Expression profiles of CBF1, CBF2, and CBF3 genes at
low temperature ................................................................. 89

Figure 3.3. Expression profiles of Thau-like, CCCH-type Zn Finger
containing transcription factor, and the Dof1-domain containing
genes at low temperature ..................................................... 93

Figure 3.4. Taqman real-time RT-PCR amplification profiles for
transcripts of indicated genes .............................................. 45

Figure 3.5. eQTLs for regulation of CCCH-type Zn Finger
containing transcription factor expression .............................. 88

Figure 3.6. eQTLs for regulation of Thaumatin-like gene
expression ....................................................................... 89

Figure 3.7. eQTLs for Dof1-domain containing gene

expression ...................................................................... 90
Figure 3.8. eQTLs for Rav1 expression ................................. 92
Figure 3.9. eOTLS for Zat12 expression ................................. 93
Figure 3.10. eQTLs for Ethylene Response Factor gene
expression ....................................................................... 94
Figure 3.11. eQTLs for Rav1 expression ................................ 96
Figure 3.12. eQTLs for Zat12 expression ............................... 97

Figure 3.13. eQTLs for Ethylene Response Factor gene
expression ........................................................................ 98

Figure 3.14. eQTLs for CCCH-Type Zn Finger Transcription Factor
gene expression ................................................................ 99

Figure 3.15. eQTLs for Dofi-domain containing gene

expression ...................................................................... 100
Figure 3.16. eQTLs for Thaumatin-family gene expression ....... 101
Figure 4.1. Sequence alignment of CBF1-3 genes .................. 109

xi

 

~ .- “will

 

36.

37.

38.

39.

40

41.

42.

43.

44.

45.

46.

47_

date

FigL
data

Fig‘

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

Figure 4.1. Sequence alignment of CBF1-3 genes .................. 109
Figure 4.2. Cross-reactivity of CBF2 primers with CBF1 and
CBF3 ............................................................................ 110
Figure 4.3. Gradient PCR to determine temperature for specific
amplification of CBF2 ........................................................ 111
Figure 4.4. PCR using genomic DNA of Ler and Cvi-O as
template ......................................................................... 113
Figure 4.5. Taqman real-time RT-PCR amplification profiles ...... 114

Figure 4.6. Technical variation for one-step real time RT-
PCR .............................................................................. 115

Figure 4.7. Technical variation for two-step real time RT-
PCR .............................................................................. 116

Figure 4.8. Effect of genomic DNA contamination on RNA
estimation ....................................................................... 1 17

Figure 4.9. No-RT (reverse transcriptase) samples after DNase I
treatment ........................................................................ 119

Figure 4.10. OTL Cartographer output with background noise in

data ............................................................................... 120
Figure 4.1 1. Removal of outliers reduced noise in the
dataset ........................................................................... 122
Figure 5.1. Model of gene regulatory pathway ........................ 127

xii

 

Dual-5" ~

 

 

CBF
ERF
TF

LOD
cm

RIL

OTL

eQTL

CBF
ERF
TF
LOD
CIM
RIL
NIL
QTL

eQTL

List of Abbreviations

CRT/DRE Binding Factor
Ethylene Responsive Factor
Transcription Factor

Log of Odds

Composite Interval Mapping
Recombinant Inbred Line
Near Isogenic Line
Quantitative Trait Locus

expression Quantitative Trait Locus

xiii

 

 

711 II? t 1".“4 .' {37"
'8“
..

 

1.1 lntrt

P
environm
high light
plants are
coping m4
temperatLJl
YIGIdS in la
I1 is a phys
Solute diff
membrane 1

Most
temperature
12:0 TOmat
(TaYIOr 81 a.
ten'iDEratUreS

alterajjOHS ir

OI iOnS frOm
respiration r

are able 10 I:

Chapter 1

Literature Review

1.1 Introduction

Plants are constantly exposed to numerous abiotic stresses in the
environment such as drought, salinity, chilling, freezing, high temperature,
high light intensity, anoxia, and nutrient imbalances. Being sessile organisms,
plants are unable to escape these stresses and have developed various
coping mechanisms (Korner and Larcher, 1988; Tran et al., 2007). Low
temperature constitutes a significant environmental stress that limits crop
yields in large parts of the world, often leading to significant economic losses.
It is a physical stress that can influence basic aspects of cell sun/ival, such as
solute diffusion rates, enzyme kinetics, macromolecular interactions,
membrane fluidity, and overall integrity of the cell (Bugbee, 1995).

Most plants of tropical and subtropical origin are sensitive to low
temperatures and are injured by chilling temperatures ranging from 0°C to
12°C. Tomato, maize, and rice are some examples of chilling sensitive plants
(Taylor et al., 1974). Other species, such as potato, can tolerate chilling
temperatures but are injured or killed at freezing temperatures (Sukumaran
and Weiser, 1972). Plants sensitive to low temperatures suffer physiological
alterations including changes in cell membrane structure, consequent leakage
of ions from cells (Nagao et al., 2008), and changes in photosynthesis and
respiration rates (Taylor et al., 1972). On the other hand, cold tolerant plants

are able to prevent the physiological damage associated with exposure to low

‘Ali

 

 

 

temperatures
canola. shov
(Gilmour et a
freezing fem.
these plants
However. a p
tolerate free
achieved rim
YamaguchI-S

Cold a.
gene exDress
The process I
prevent low (e
The median:
toleranCe 10 1
”TechaanmS v

 

 

temperatures. Many temperate plants such as Arabidopsis, wheat, rye, and
canola, Show a wide range of cold tolerance ranging from -5°C to -30°C
(Gilmour et al., 1988; Kaye and Guy, 1995; Guy, 1999). This ability to survive
freezing temperatures is not a constitutive property as sudden transfer of
these plants from warm temperatures to freezing conditions causes death.
However, a pre—exposure to low non-freezing temperatures conditions them to
tolerate freezing temperatures. This induction of freezing tolerance is
achieved through the process known as cold acclimation (Shinozaki and
Yamaguchi-Shinozaki, 1996; Thomashow, 1999; Zhu, 2001).

Cold acclimation involves Specific changes in cellular biochemistry and
gene expression that are rapidly triggered in response to low temperature.
The process can be divided into sequential stages that collectively serve to
prevent low temperature induced injury (Guy et al., 1985; Thomashow, 1998).
The mechanisms by which plants perceive low temperature and develop
tolerance to the stress remain far from understood. Elucidation of these
mechanisms will ultimately aid in the development of strategies for improving

crop productivity.

1.2 Low temperature stress

Low temperature causes several kinds of injuries to plants, the type
and severity of which vary with temperature (Shinozaki and Yamaguchi-
Shinozaki, 1996, 2000). Plants are broadly classified according to their level
of tolerance as being chilling sensitive, freezing sensitive, or freezing tolerant.

Those unable to withstand temperatures in the range of 0°C to 15°C are

4
l
,'!"‘

.
'c
(1.“! W” ""J

 

 

regardec
tempera:
E)

causes c
These me
two majo
temperate
growth, an
Free

of Chilling
temperature
CWSIaIS at 5
[Ce CWSIais

concentratio.

Ceils by rUp

regarded as chilling sensitive whereas those susceptible to subzero
temperatures, freezing sensitive (Paull, 1981 ).

Exposure of chilling sensitive plants to temperatures below 15°C
causes chlorosis, wilting, necrosis and subsequent death (Garber, 1977).
These may largely be a consequence of photoinhibition and dehydration, the
two major adverse effects of chilling (Hetherington et al., 1989). Low
temperature stress in these plants manifests as defects in germination,
growth, and development.

Freezing sensitive plants, although able to tolerate the adverse effects
of chilling temperatures, however, are unable to survive sub-zero
temperatures. This is partly due to further injury caused by the formation of ice
crystals at sub-zero temperatures (Yamada et al., 2002; Nagao et al., 2008).
Ice crystals are initiated in the extracellular space due to the lower solute
concentration in this environment. Ice crystal formation physically damages
cells by rupturing the cell membranes and thus leads to leakage of cell
contents and subsequent cell death (Webb et al., 1993; Webb and Steponkus,
1993; Webb et al., 1994). Another damaging effect of ice formation is cellular
dehydration resulting from decreased water potential in the extracellular
space and subsequent movement of water out of cells (Webb et al., 1993;
Webb and Steponkus, 1993; Webb et al., 1994).

Freezing injury has been observed to occur primarily at the level of cell
and organellar membranes. Three types of freezing injury have been well
documented: expansion-induced lysis, hexagonal II lipid formation, and
fracture jump lesions (Webb et al., 1994; Uemura et al., 1995; Uemura et al.,

1996). Expansion-induced lysis typically occurs at -2°C to -4°C and is caused

1‘...-

 

 

 

by repe-
following
freezing
reducfior
water re-
Tsr
adjacent
arrays of
protOpiast
abundanc.
lameiiae) 1
Wide rang
membrane
ChIO’ODIaSt
aDDOSition_
he"agonal l
membrane t
(VI/Ebb 91 al.,
Fnactc
ObSeNed at 1
unclear how
associathnS E
organelles.

HeXaQOr

Interme‘diate D

 

by repeated freeze/thaw cycles. The rapid movement of water into the cells
following thawing leads to cell lysis. Another phenomenon resulting from
freezing is the endocytic vesiculation of plasma membrane that causes
reduction in cell surface area, increasing the probability of cell lysis when the
water re-enters the cells (Uemura et al., 1995).

The hexagonal ll lesions observed at freezing temperatures appear
adjacent to the membrane bilayer and are composed of hexagonally packed
arrays of monolayer tubes (Webb et al., 1994). In non-acclimated oat
protoplasts, these lesions start to appear below -3°C and increase in
abundance at -10°C. The transition of membrane bilayers (gel-phase
Iamellae) to the hexagonal ll phase (non-Iamellar) structures occurs over a
wide range of temperatures depending on the lipid composition of the
membranes. They are usually formed at the interbilayer junction of the
chloroplast and the plasma membrane where the two membranes are in
apposition. When the membranes are returned to warm temperatures, the
hexagonal ll phase structures are thought to form inverted micelles within the
membrane bilayer, resulting in the loss of permeability barrier of membranes
(Webb et al., 1994).

Fracture-jump lesions are aparticulate regions in the membranes
observed at temperatures below -10°C (Uemura et al., 1995). Although it is
unclear how they arise, they are suggested to result from inter Iamellar
associations and fusion of the plasma membrane with membranes of other
organelles.

Hexagonal ll phase lipids or fracture-jump lesions arise from a common

intermediate phase (Uemura et al., 1995). Upon dehydration that occurs at

 

 

freezing
nmawm
mafiaw
groups a
immawn
is deterrr
destabiiiz
curvature
when the
In 1
inside the
to DTOteIn

cells_

1-3 Cold 1
Bio.
achieved
eXDOSore
2007). The

TSSUIIIng ir

freezing temperatures, the head groups of lipids lose the stabilizing water
molecules that maintain the lipid bilayer structure. As a result, there is
excessive pressure in the acyl groups of the lipids as compared to the head
groups and it exerts a bending force. This bending energy causes either the
formation of hexagonal ll phase lipids or fracture-jump lesions. The outcome
is determined by the intrinsic curvature of the monolayers undergoing the
destabilization. Hexagonal phase II lipids are formed when the intrinsic
curvature of the monolayers is high, whereas fracture-jump lesions are formed
when the curvature is low (Uemura et al., 1995).

In addition to the effects on cell membranes, the reduced water content
inside the cell that results from ice formation in the extracellular space leads
to protein denaturation and precipitation of other macromolecules inside the

cells.

1.3 Cold Acclimation

Biochemical and physiological changes during cold acclimation are
achieved through changes in gene expression that begin shortly after
exposure to low temperatures (Thomashow, 1999; Zhu, 2001; Zhu et al.,
2007). The promoters of specific genes respond to the low temperature signal
resulting in transcriptional activation of these genes. Many such early-induced
genes encode transcriptional factors that recognize specific sequence
elements in promoters of other genes to activate/repress these targets.
Additionally, early-induced genes encode common cell signaling components
such as serine/threonine protein kinases, phosphatases, phospholipases,

response regulators, calmodulin related proteins, calcium-binding proteins etc

 

 

1‘1-

 

(Yamagu
these sic
are enzj
membran
cumulativ

protecting

1.3.1 Bloc
Mal
tolerance l
Metabolon
comDositlo
memboiiteg
resDonse
Carbohydra
been gem
be achieve<
Specific mac
Com;
eSSentiajjy C
tileir metabc
achieVed

temDEratUre I

In add:

 

(Yamaguchi-Shinozaki and Shinozaki, 2006; Zhu et al., 2007). Induction of
these signaling genes also results in upregulation of genes whose products
are enzymes in specific metabolic pathways, transporters, chaperones,
membrane stabilizing cryoprotectants, or RNA binding proteins. The
cumulative effects of this cascade are biochemical changes that result in

protecting cells and establishing homeostasis in the altered environment.

1.3.1 Biochemical Changes Accompanying Cold Acclimation

Many plants accumulate osmolytes, which correlate with their freezing
tolerance (Garg et al., 2002; Sakamoto and Murata, 2002; Beck et al., 2007).
Metabolome studies revealed a significant change in the overall metabolite
composition of cells during cold acclimation. Simultaneous assay of 434
metabolites in Arabidopsis accession Ws-2 revealed changes in 325 in
response to low temperature (Cook et al., 2004). These included
carbohydrates, amines, organic acids etc. Increase in soluble sugar levels has
been demonstrated to correlate with enhanced freezing tolerance, which may
be achieved by regulating osmotic balance of cells or affording protection to
specific macromolecules (Stitt and Hurry, 2002).

Comparison of two accessions of Arabidopsis (Ws-2 and Cvi-O) that
essentially differed in freezing tolerance revealed quantitative differences in
their metabolomes at low temperatures, suggesting that metabolic changes
achieved during cold acclimation may play a significant role in low
temperature tolerance (Cook et al., 2004).

In addition to osmoprotective solutes, plants also accumulate proteins

with cryoprotective properties. Proteins accumulated by cabbage and spinach

dunng

 

membl
Voiger
hydropl
tempera
freeze-ii
mduced
irnprovet
whole pl;
chloropla

1996). C1

 

phase les
1996)

In c
temDeratur
resulting i,
2002i. The
deCreaSe ir
BTOWse, 20 C
bOHGS in the
0r fluidity of I
Arabidopsis)
in this profiles

SUCh alterafic

 

during cold acclimation have been shown to protect isolated thylakoid
membranes against freeze-thaw induced damage (Volger and Heber, 1975;
Volger et al., 1978; Hincha et al., 1993). Arabidopsis seedlings accumulate
hydrophilic polypeptides, such as COR15 and COR6.6, in response to low
temperatures. In vitro, these polypeptides demonstrate the ability to decrease
freeze-induced fusion of liposomes but are unable to prevent their freeze-
induced leakage (Uemura et al., 1995). In vivo, over-expression of COR15
improves the freezing tolerance of chloroplasts by 2°C, without affecting
whole plant tolerance. Localization studies revealed COR15 to be targeted to
chloroplasts, which was suggested to be the site of action (Uemura et al.,
1996). COR15 is thought to defer the formation of deleterious hexagonal ll
phase lesions in the membrane to lower freezing temperatures (Uemura et al.,
1996).

In order to maintain the functional roles of cellular membranes at low
temperatures, changes in lipid profiles are made during cold acclimation
resulting in homeoviscous/homeophasic adaptation (Wallis and Browse,
2002). These changes compensate for the lipid rigidification and resultant
decrease in membrane fluidity observed at low temperatures (Wallis and
Browse, 2002). Membrane desaturation is increased by introduction of double
bonds in the acyl chains, which results in increased level of molecular motion
or fluidity of the membranes. Both transcriptional (delta-9 desaturase gene in
Arabidopsis) and post-transcriptional mechanisms were found to be involved
in this process (Wada et al., 1990; Kodama et al., 1994; Upchurch, 2008).
Such alterations of the chloroplast membranes increase the turnover of

components of the photosynthetic electron transport chain, resulting in

 

 

 

resumpflc
and Some

Phi
accompar
due to tra
chlorophyll
carboxytas
Calvin cyclI
(Strand 61 a

2002).

1.3.2 Chang
Siudie
eXi>OSure Io
reDreSSed (Fc
ShinOZalq‘ and
to be involvec

temperamre, ‘

   
  

CIUStered bas .

transiently eXp

resumption of photosynthesis at low temperatures (Raison et al., 1982; Hugly
and Somerville, 1992; McConn and Browse, 1998; Kim and Portis, 2005).
Photosynthetic suppression observed at low temperatures is
accompanied by increase in soluble sugars (Strand et al., 1997). This occurs
due to transcriptional down-regulation of photosynthetic genes, such as the
chlorophyll a/b binding protein (CAB) and ribulose-1,5-bisphosphate
carboxylase (FIBCS). During cold acclimation several genes involved in the
Calvin cycle are induced and some enzymes are post-translationally modified
(Strand et al., 1997; Strand et al., 1999; Hurry et al., 2000; Stitt and Hurry,

2002).

1.3.2 Changes in Gene Expression Associated with Cold Acclimation
Studies elucidating the total transcriptome after different periods of
exposure to low temperature showed hundreds of genes to be induced or
repressed (Fowler and Thomashow, 2002; Vogel et al., 2005; Yamaguchi-
Shinozaki and Shinozaki, 2006). Extensive gene regulatory networks appear
to be involved in adaptation to low temperature stress. In response to low
temperature, 302 genes were found to be upregulated in Arabidopsis
seedlings (Vogel et al., 2005). These differentially expressed genes were
clustered based on their temporal expression profiles. A few genes are
transiently expressed with peak transcript levels 2-4h after a low temperature
stimulus and subsequently decrease to background warm levels. These
genes are thought to be general shock genes, although they may also
contribute to cold acclimation. Another cluster of genes is highly induced early

in response to low temperature followed by a gradual decrease in levels.

 

 

 

’7

If)

{8’

C101

ma)

Tho

 

However, the expression of this gene cluster is maintained at levels above
background warm levels. This cluster includes several transcription factors
genes such as CBF, Rav1, Zat12, MYB, etc. Genes are also induced later
during cold treatment, remaining elevated throughout exposure to low
temperature. In addition to clusters of genes upregulated, 212 genes are
down regulated during cold acclimation. The down-regulation of these genes
may be significant for enhancing the freezing tolerance of plants (Fowler and
Thomashow, 2002; Vogel et al., 2005).

As mentioned earlier, freezing reduces the amount of water available
through formation of ice crystals resulting in an osmotic imbalance inside the
cells, a characteristic shared with drought and salinity stress (Beck et al.,
2007). In addition, considerable overlap is observed in the resulting
transcriptional responses to these stresses (Shinozaki and Yamaguchi-
Shinozaki, 1996, 2000; Zhu, 2001; Shinozaki et al., 2003). Greater overlap is
observed between drought and salinity stress where more than half of the
genes induced in response to drought are also induced in response to salinity
stress. In contrast, only 10% of the genes induced by drought are also
induced by low temperature. Extensive cross talk between the gene networks

is suggested for these stresses (Knight, 2002).

1.3.3 The CBF Gene Regulatory Pathway

Studies of low temperature responsive genes resulted in the
identification of CBF (C-repeat binding factor) genes in Arabidopsis thaliana
(Stockinger et al., 1997). The CBF genes comprise a gene family with six

members (CBF1-CBF6) encoding transcriptional activators belonging to the

 

-{ r i
I‘M—l, ”Km-c. Wm
VI“ ~ -

APZ'EREBF
of DNA-blnl
temperature
of the CBF
peak 2-4h 2
levels abov
temperature
Temperature
transcripts in

CBF ;
The CBF pr
Elemem) eli
EXDTBSSion. (
of the COR
Stimulus and
tolerance (Jag

TranSC
or 3 identifiec
and Thomasi

include 9906

AP2/EREBP (Apetella2/Ethylene Responsive Element Binding Protein) family
of DNA-binding proteins (Stockinger et al., 1997; Haake et al., 2002). Low
temperature activates rapid induction of CBF promoters resulting in detection
of the CBF transcript within 15 min of cold treatment. CBF transcript levels
peak 2-4h after exposure to low temperature and subsequently decrease to
levels above background (Gilmour et al., 1998). The mechanism of low
temperature induction of CBFs appears to perceive and respond to absolute
temperature of the environment (Zarka et al., 2003). The accumulation of CBF
transcripts increases with decreasing temperatures (Zarka et al., 2003).

CBF proteins play an important role in the process of cold acclimation.
The CBF protein binds to the CRT/DRE (C-Repeat/Dehydration response
Element) element in promoters of target genes to up-regulate their
expression. Over-expression of CBF in Arabidopsis plants leads to activation
of the COR (Cold Responsive) genes regardless of a low temperature
stimulus and as a result these transgenic plants display constitutive freezing
tolerance (Jaglo—Ottosen et al., 1998).

Transcriptome profiling of Arabidopsis plants over-expressing CBF1, 2
or 3 identified more than 90 genes showing constitutive expression (Fowler
and Thomashow, 2002). These 90 genes comprise the CBF regulon and
include genes encoding cryoprotective polypeptides such as COR15a,
enzymes involved in synthesis of solutes with cryoprotective properties,
transcriptional activators, etc., which may act in concert to confer cold

tolerance to the plant.

1.3.4 Transcriptional Regulation of CBF Genes

1O

 

 

 

The

CBF and 0‘.
using repel
promoter th
W1thin this :

expression
responsiven
bind any I
Consensus I
(basic Helix-
Mutati
in CBF exp,
appears to h
reduced Cr C
identified in
iChinnUSamy
tra”~°"3rll3ti'c>n 1

[GET IHCrEaSE

 

The mechanism by which the low temperature stimulus is relayed to
CBF and other concurrently expressed genes is not known. Promoter analysis
using reporter gene fusions have identified a 125 bp region in the CBF
promoter that is sufficient for low temperature induction (Zarka et al., 2003).
Within this segment, two regions named lCEr1 and lCEr2 (induction of CBF
expression region 1 or 2) are found to act in concert to impart cold
responsiveness (Zarka et al., 2003). These sites have not yet been shown to
bind any transcription factors, though the lCEr2 region contains the
consensus recognition site for a family of transcription factors, the bHLH
(basic Helix-Loop-Helix) proteins (Massari and Murre, 2000).

Mutational analysis has led to the identification of some genes involved
in CBF expression. The L084 gene encoding a DEAD-box RNA helicase
appears to have a positive role in CBF expression as the [034 mutant shows
reduced or delayed induction of CBF genes (Gong et al., 2002). ICE 1 was
identified in another mutant screen as a regulator of CBF3 expression
(Chinnusamy et al., 2003). The ICE1 gene encodes a MYC-like bHLH
transcription factor that binds to MYC DNA binding sites. Over-expression of
ICE 1 increases the freezing tolerance of Arabidopsis plants (Chinnusamy et
al., 2003). Finally, FRY2 (Xiong et al., 2002) and H081 (Lee et al., 2001),
encoding a novel transcriptional repressor and a novel RING-finger protein
respectively, are thought to play a role in CBF expression through down-
regulation of CBF genes.

In addition to mutants affecting the expression of CBF genes, mutants
altered in freezing tolerance have been found that are unaffected in CBF

expression. Sensitive to freezing-6 (5er mutant, showing lower freezing

11

 

( . I
W

tolerance than
genes in respc
of CBF genes
that the SF RE
to induce tr
constitutive»
gene expres
known to (:0
Suggested
functioning
2007),

In .
TSSanSe
2003) S
TBSanS—‘t
agitation
re's‘pohs:
(zarka e

and r813

1-3.5 C
respon:
Of CB;

Incjudi'

tolerance than the wild type Arabidopsis, shows normal induction of CBF
genes in response to low temperature but is unable to induce known targets
of CBF genes, such as the COR genes (Knight et al., 1999). This indicates
that the SFFI’6 encoded protein acts in conjunction with or downstream of CBF
to induce transcription of COR genes. The eskimo1 mutant, showing
constitutively enhanced freezing tolerance, is unaltered in CBF and COR
gene expression and shows increased accumulation of proline, an amino acid
known to correlate with freezing tolerance in plants. The ESK1 gene has been
suggested to be a novel negative regulator of cold acclimation, which is
functioning independent of the CBF regulon (Xin and Browse, 1998; Xin et al.,
2007)

In addition to low temperature, CBF genes are also induced in
response to mechanical agitation and cycloheximide treatment (Zarka et al.,
2003). Studies of the CBF2 promoter have shown that the elements
responsible for induction of this promoter by low temperature and mechanical
agitation are tightly linked. The lCEr1 and lCEr2 regions described before are
responsive to low temperature, mechanical agitation, and cycloheximide
(Zarka et al., 2003). However, it is not yet known if these stimuli are perceived

and relayed to the CBF2 promoter through a common regulatory pathway.

1.3.5 Conservation of CBF pathway in higher plants

There is considerable evidence to suggest that the CBF cold-
responsive pathway is conserved in a wide variety of plant species. Orthologs
of CBF genes have been identified in cold tolerant and sensitive plants,

including Brassica napus, wheat, rye, rice, barley, tomato, and maize (Jaglo et

12

 

I
ml
.. v \4.‘ “Fur um-
I . M

 

 

al., 2001; Cl”
2004). The 1
domain and
and DSAWF
Dubouzet et

The t
library scree
like their Ara
and 5028, (
CFIT/DRE ei
piOmOIers j,
COHIains 20
ChrOmOSOme

OTL (Quam

Aimiz locUs
al., 2006), T
Mimi and .
SEVeraj 0th,
Ofihologs. 8
three CBF

(Zhang et al.

14 Famepti

The

al., 2001; Choi et al., 2002; Dubouzet et al., 2003; Lee et al., 2004; Qin et al.,
2004). The CBF orthologs display highest conservation in the DNA-binding
domain and its two flanking "signature sequences" (PKK/RPAGRxKFxETRHP
and DSAWR) in Arabidopsis thaliana (Jaglo et al., 2001; Choi et al., 2002;
Dubouzet et al., 2003; Lee et al., 2004; Qin et al., 2004).

The two CBF orthologs in Brassica napus, identified through cDNA
library screening, are induced within 30 mins of exposure to low temperature,
like their Arabidopsis counterparts. Furthermore, two CBF target genes Bn115
and Bn28, orthologs of COR15 and COR6.6 respectively, also contain the
CRT/DRE element in their promoters that is characteristic of the CBF target
promoters in Arabidopsis (Jaglo et al., 2001). Hordeum vulgare (barley)
contains 20 CBF orthologs, 11 of which form two tandem clusters on
chromosome 5H-L (Skinner et al., 2006). In wheat (Triticum monococcum L),
QTL (Quantitative Trait Loci) mapping for frost tolerance identified the Fr-
A(m)2 locus on chromosome 5A(m), which contains 11 CBF genes (Miller et
al., 2006). Two additional CBF genes in wheat are located on chromosomes
7A(m) and 6A(m) (Miller et al., 2006). QTL mapping for frost tolerance in
several other plant species is leading to identification of additional CBF
orthologs. Solanum chopersicum (tomato), a chilling sensitive plant, contains
three CBF orthologs, only one of which is induced by low temperatures

(Zhang et al., 2004).

1.4 Perception of the low temperature stimulus

The mechanisms of low temperature perception are unknown.

However, observations in cyanobacteria and plants point to the role of certain

13

 

 

 

 

signaling c
temperatur
Exposure:
gene. whi
desaturatac
gene in a r

Ana
1W0 histldir
component
et al., 200(
plasma m,
Spanning )—
SUDSeqUen
physical St;
histidine k;
Cytokimn Si

Signaling he

Stud

signaling components. Change in membrane fluidity is thought to initiate low-
temperature responsive signaling in the cyanobaterium Synechocystis.
Exposure of Synechocystis to low temperature leads to induction of the desA
gene, which encodes a fatty acid desaturase. Reduction in membrane
desaturation by Pd-catalyzed hydrogenation also leads to the induction of this
gene in a manner similar to its low temperature induction (Vigh et al., 1993).

Analysis of low temperature response in Synechocystis also identified
two histidine kinases, Hik33 and Hik19, and a response regulator, Rer1, as
components of the signaling pathway functioning at low temperature (Suzukl
et al., 2000). It is suggested that low temperature induced rigidification of the
plasma membrane leads to the phosphorylation of Hik33, a membrane
spanning histidine kinase. The phosphate group is transferred to Hik19 and
subsequently to the response regulator (Rer1), thereby relaying the change in
physical state of the membrane to cell nucleus (Suzuki et al., 2000). In plants
histidine kinases have been shown to play roles in ethylene Signaling,
cytokinin signaling and osmosensing (Urao et al., 2000). Their role in cold
signaling has yet to be established.

Studies in Arabidopsis Show a rapid increase in cytoplasmic Ca“2
concentrations following exposure to low temperatures (Knight, 2002; Kaplan
et al., 2006). This Ca+2 influx appears to provide the secondary messenger
required for expression of certain cold-regulated genes, including COR and
other CBF-targeted genes. Similarly, pharmacologically induced increases in
intracellular Ca+2 levels at warm temperatures also induced COR gene
expression (Tahtiharju et al., 1997). The flux signature (the amplitude and

duration of Ca+2 peak) in response to low temperature is different for different

14

 

L.

 

ceil

becc
the 5
eth8'

2004);.

plants i
and me
it is no:
selective
5:0 (Ding
Ara
phOSDhatit
thrOuch thi
FUfther an.
1e mpera fu re
Identified 5,
phosphohpag

temperatures

 

cell types. The flux also varies through the stages of cold acclimation,
becoming diminished in amplitude and prolonged in duration. This change in
the signature of Ca+2 flux during cold acclimation is suggested to play a role in
either developing a cold memory or serving as a cold memory (Gao et al.,
2004)

Ca+2 influx is also observed in response to mechanical agitation of
plants although it is not yet clear if the increases in Ca+2 brought on by cold
and mechanical stress are mediated through the same mechanism. However,
it is noteworthy that the activity of an anion mechano-sensitive calcium-
selective channel has been shown to increase as temperature is lowered to
6°C (Ding and Pickard, 1993).

Arabidopsis suspension cells display an increase in the levels of
phosphatidic acid within a few minutes of exposure to low temperatures
through the activation of phospholipases C and D (Vergnolle et al., 2005).
Further analysis of the transcriptome of suspension cell cultures at low
temperatures, in the absence and presence of phospholipase inhibitors
identified 58 genes to be regulated by phospholipase C and 87 by
phospholipase D, suggesting involvement of both Signaling pathways at low

temperatures (Vergnolle et al., 2005).

1.5 Naturally occurring genetic variation

Variation is an essential component of natural populations that is
discernible both in and between populations (intra- and inter-species) (Farrall,
2004). Natural selection of variation, determined by the existent environmental

conditions, has given rise to diverse phenotypic traits in organisms.

15

 

 

 

Ade
genome
continuaiiy
selectrve l
polymorpr
Sgnmcant
nature (H
Changes -
accumuia:
and tempo
observed t
variation in
Phenotype
Ilnkage 0r ;
selection Ca
”79 most ac

OVer SpaCe é

ByerS, 2005).,

 

Adaptation of species to their environment is thought to be a result of
genome evolution (Farrall, 2004). The genomes of a population are
continually undergoing change, mostly through spontaneous mutations. The
selective pressures acting on a population are important in determining which
polymorphisms will be maintained. Environmental conditions provide
significant selective pressures that develop and maintain genetic variation in
nature (Hoffmann and Weeks, 2007). Over the course of time, genetic
changes that provide a selective advantage under local conditions are
accumulated with resultant increase in survival rate of a population. Spatial
and temporal changes in environment drive the diversity of survival strategies
observed between populations. However, it is not always the case that the
variation in a phenotype arises from selection of genetic changes affecting the
phenotype directly, in some instances phenotypic evolution results from
linkage or pleiotropic effects (Cork and Purugganan, 2004). Although natural
selection can have a negative effect on variation as it may lead to fixation of
the most adapted genotype in a population, variation in selective pressures
over space and time enhance genetic variation (Gebhardt and Stearns, 1992;
Byers, 2005).

Fisher in 1930 proposed a random mutation model for evolution,
wherein the changes in genome were suggested to be random with respect to
their impact on the phenotypes (Farrall, 2004). This was based on the
premise that large effect mutations would produce deleterious effects and will
be selected against (Farrall, 2004). However, Quantitative Trait Locus (QTL)
analysis showed that, in fact, differences in genomes were localized to a few

regions, which can have impacts as high as 20% on the variation in

16

 

Ill

"-3

 

'_.i ‘--—v V.—

phenotype IF
and Orr, to
proposed the
of large effec
(Kimura. 19E
genome can
model to alio
OTLs of lar-
(Farrall, 200.

OTLS IS expc

1.6 Understa
Seven
910W, deve
““9 establis
KOOrnneef, 2
2005; Shind
understand-I
reQUIEttioj-j (
al" 2004; A
extended 1C
Variation w
environmer

QEneS 0n 1'

 

phenotype (Farrall, 2004). The findings of OTLS lead others, such as Kimura
and Orr, to propose modifications to Fisher's adaptation model. Kimura
proposed that during the process of changes in the genome, random changes
of large effects would have a high probability of fixation if they were favorable
(Kimura, 1969; Ota and Kimura, 1971). These favorable changes in the
genome can then be identified through QTL analysis. Orr further modified the
model to allow the prediction that quantitative variation is determined by few
QTLs of large effect and more OTLS with progressively smaller effects
(Farrall, 2004). Therefore predicting that the distribution of effect sizes of

QTLs is exponential in nature.

1.6 Understanding Complex Traits using Natural Variation

Several independent studies examining different aspects of plant
growth, development, and interaction with the biotic and abiotic environment
have established the vast genetic diversity in nature (Alonso-Blanco and
Koornneef, 2000; Mackay, 2002; Koornneef et al., 2004; Alonso-Blanco et al.,
2005; Shindo and Bemasconi, 2007). This variation proves helpful in
understanding the regulation of complex plant traits that are under polygenic
regulation (Alonso-Blanco and Koornneef, 2000; Mackay, 2002; Koornneef et
al., 2004; Alonso-Blanco et al., 2005; Shindo and Bemasconi, 2007). When
extended to the study of plants inhabiting diverse environments, this genetic
variation would allow identification of genes that help plants adapt to different
environments. Furthermore, it would allow quantification of the effects of these

genes on the variability in the trait (Alonso-Blanco et al., 2006).

17

 

 

~ .
9; 14h

 

 

V
discrete
Contgnuc
Change
unoerlyzr
The gent
et al., 2C
trait. stat
of the Inc

Tc
quantitati
variatIOn
total phe
genetic a
info add;

Variability of a trait that can be measured on a continuum rather than in
discrete classes is called continuous variation (Alonso-Blanco et al., 2006).
Continuous variation makes possible the gradual changes underlying adaptive
change in nature. For a population showing such variation in a trait, the
underlying genotypic variation can be identified through crossing experiments.
The genetic loci contributing to the variation are termed QTL (Alonso-Blanco
et al., 2006). In addition to identifying genetic loci controlling a quantitative
trait, statistical analysis can be performed to understand the behavior pattern
of the individual loci.

To understand the regulation of a quantitative trait, traditional
quantitative genetics defines the trait in terms of variances. The continuous
variation for a phenotypic character is measured on a quantitative scale. The
total phenotypic variance observed in the population is first partitioned into
genetic and environmental variances. The genetic variance is further divided
into additive, dominant and epistatic effects. Analysis of this information
allows estimation of heritability of the trait (Kearsey and Farquhar, 1998).

Additive genetic effects are the sum of mostly small effects of alleles of
many genes influencing the phenotype. The selection for additive genetic
effects can accumulate over time, as and when mutations contributing small
selective advantages occur. Dominance genetic effects result from the
interaction of alleles at the same locus, where one allele exhibits a dominant
influence on the phenotype by masking the contribution of the recessive allele
at that locus (Frascaroli et al., 2007; Melchinger et al., 2007; Melchinger et al.,
2007). Dominance effects are the basis of hybrid vigour (heterosis) in

breeding programs (Melchinger et al., 2007). Epistasis refers to the

18

 

 

t
I
.5
2!:

 

nonaddfl
to an effl
the loci v
the influe
Epistatic

individual

expected I

1.7 OTL M
The
(Restriction
Length Pol:
SSLPS (Sir
AraDIdOpst
quantitative
unaffected I

Chron'lOSOm;1

nonadditive interactions occurring between alleles at different loci, which lead
to an effect on the phenotype that is different from what would be expected if
the loci were influencing it independently. This makes it difficult to separate
the influence of a single locus on the phenotype (Juenger et al., 2005).
Epistatic interactions can be synergistic (greater than that expected from
individual loci) (Bonhoeffer et al., 2004) or antagonistic (lower than that

expected from individual loci) (Sanjuan and Elena, 2006).

1.7 OTL Mapping

The availability of a large number of molecular markers such as RFLPS
(Restriction Fragment Length Polymorphism), AFLPs (Amplified Fragment
Length Polymorphism), RAPDS (Random Amplified Polymorphic DNA), and
SSLPS (Single Strand Length Polymorphism) dispersed all along the
Arabidopsis genome has made it possible to map genes that contribute to
quantitative variation. These markers segregate as single genes, and are
unaffected by the environment. The genes are identified indirectly as
chromosomal regions contributing to the phenotypic variation between
ecotypes, i.e. the OTLS.

Major techniques of QTL mapping are based on the same basic
principle, which is to establish statistical association between two characters
based on co-segregation in a segregating population (Loudet et al., 2002;
Perez-Perez et al., 2002; Harada et al., 2004; Keurentjes et al., 2006). In this
mapping the statistical association of a phenotype is tested with the molecular
marker derived chromosomal maps of the individuals. The power of QTL

detection depends on several factors including the population-Size and -type,

19

 

 

 

OTL
recom
and tin

and We

display
lines to 5
contain a
other (Ki
descendai
chances o:
segregants
Phenotype t
were“) of S
Used to dete

high Significa

 
   

0f Several cc

the accuracy of trait measurement, and the statistical methods used for the
correlation analysis (Knapp and Bridges, 1990; Mackinnon and Weller, 1995;
Rebai et al., 1995; Belknap et al., 1996; Zou et al., 2005). The resolution of
QTL analysis is affected by few major factors, namely, the number of
recombinant events that are being represented in the segregating population
and the number of markers examined (Knapp and Bridges, 1990; Mackinnon
and Weller, 1995; Rebai et al., 1995; Belknap et al., 1996; Zou et al., 2005).
The general approach for mapping OTLS is to select two lines that
display contrasting phenotypes for a quantitative trait and interbreed these
lines to generate homozygous descendants (Recombinant Inbred Lines) that
contain a small number of segments from one line in the background of the
other (Kearsey and Farquhar). The combination of genomes in the
descendants is a random process and several lines are chosen to increase
chances of finding maximum combinations of the two parent genomes in the
segregants. Such hybrid individuals are then assessed for the quantitative
phenotype being studied, and estimates are made of the contributions (or lack
thereof) of specific segments to the observed variation. Statistical analysis is
used to determine which genetic loci are segregating with the phenotype with
high significance, i.e. the QTLs (Kearsey and Farquhar, 1998). QTL analysis
of several complex traits has revealed their polygenic regulation (Gora-Maslak
et al., 1991; Cardon et al., 1994; Plomin et al., 1994; Frankel et al., 1995;
Rodriguez et al., 1995; Fijneman et al., 1996; Young, 1996; Fisler and
Warden, 1997; Li et al., 1997; Pomp, 1997; Yano and Sasaki, 1997; Vallejo et

al., 1998; Page et al., 2000; Consoli et al., 2002).

20

g . ". kl" f'fl

 

l
.7: 1‘ r313: .5“

 

 

 

inc:

cieri

ECVE
Recc
Doub
Lines
et al-,

differ

homoz

Darenta
(K80. 2c
“mmh e
IndiVidua
fleqUGnci
homozygc
heterozyg‘
between n
F2 Can be
DOpU/afion c
an 5 RIL l
ileferOZngt

01 (1/2).(1/2:

QTL analysis requires genotypic and phenotypic information of
individuals in a given population. Although these populations are usually
derived through an experimental cross, natural populations may also be used
in such studies. The experimental populations used include backcross,
advanced backcross, F2, Recombinant Inbred Lines (RILS), lntermated
Recombinant Inbred (IRI) populations, Advanced lntercross (Al) populations,
Double Haploid (DH) populations, lntrogression Lines (lLs), Near Isogenic
Lines (NILs), etc. (Comstock and Robinson, 1948; Stuber et al., 1992; Beavis
et al., 1994; Chen et al., 2007; Keurentjes et al., 2007). These populations
differ in the characteristics that define population structure, such as
homozygosity, genotypic frequencies, and linkage disequilibrium.

To better understand the population structure of RILS, consider two
parental inbred lines, P1 and P2, which differ in the quantitative trait of interest
(Kao, 2006). A cross between these lines would result in an F1 population in
which every individual contains the same heterozygous genotype. The F1
individuals are selfed to produce an F2 population, in which the genotypic
frequencies of the P1 homozygote, the P1P2 heterozygote, and P2
homozygote at a single locus are 1/4, 1/2, and 1/4, respectively and the
heterozygosity of this population is 1/2. The frequency of recombinants
between two loci is equivalent to the recombination fraction. Individuals from
F2 can be selfed for several generations to derive an RIL population. An RIL
population generated by selfing the F2 individuals for t-2 generations is called
an Fi RIL population. The Ft population contains P1 homozygote, P1P2
heterozygote, and P2 homozygote genotypes with the expected frequencies

of (1/2)-(1/2t), (1/2t-1), and (1/2)-(1/2t), respectively and the heterozygosity of

21

 

<4-
I ._ . I .3
I ' \ (,z‘ mm

 

the poput
populatior
popuiatior
It is
obtained 1
using RIL:
QBHOWDBC
increases
a design .1
the experi
time the I
Statistical
bIOIOQlcaI r
K«102005)
FOr 1
different se
SI’UCIUre I
ad‘v’antages
estimatiOn o
m9iier ”Umt
Imogfféssror
small ”Umbe
isogeniC bag
The

the population is (1/2t-1) (Kao, 2006). Therefore, the homozygosity of such a
population increases with higher number of generations to which the
population is selfed.

It is generally assumed that the data for genotype and phenotype are
obtained from the same population. However this is not always the case when
using RILS. Phenotypic trait values are often obtained for the progeny of the
genotyped population. This is possible because of the high homozygosity as it
increases the correlation of t generation with t + 1 generation. The use of such
a design in QTL mapping has several benefits. It reduces the overall cost of
the experiment as it eliminates the need to generate genotypic data every
time the phenotypic data is collected (Kao, 2006). It also improves the
statistical power of the experiment through sampling several progeny in
biological repeats for each genotyped individual (Knapp and Bridges, 1990;
Kao,2006)

For the purpose of obtaining maximum information it is helpful to use
different segregating populations as they display differences in the population
structure. Different aspects of the population structure provide distinct
advantages. For instance, the increase of homozygosity may help in the
estimation of additive effects due to accumulation of homozygotes, whereas a
higher number of heterozygotes allow estimation of the dominance effects.
lntrogression Lines, derived through repeated backcrossing of RILS, contain a
small number of specific genomic regions from the donor parent in a nearly
isogenic background of the other parent.

The number of crossover events (pairing and synapsis) that occur

during meiosis determines the recombination frequency in a population. In

22

1 - “nu-:- -.

 

i: a AID-"11"";

 

 

case of
chromos
chromos
represer.
recombir
further bi
al., 2005)
Lines (Alt
The
individual
(50%). wt
DODUIation
both payer
However, ti
deVeIODmer
analysis of
Siege of 9,0
muitiple IOci
allelic interac

In the

 

case of Arabidopsis, a maximum of 2 chiasmas are formed on any one
chromosome, thus allowing only two recombination events, at the most, per
chromosome (Sanchez-Moran et al., 2002). Therefore, an RIL population
represents a low number of recombination events. To increase the
recombination frequency within the population, the lines can be intercrossed
further before inbreeding them to achieve homozygous populations (Zou et
al., 2005). Populations derived in this manner are called Advanced lntercross
Lines (AlLs).

The advantage of using RIL populations is that in such a population, an
individual contains several introgression fragments from the donor parent
(50%), which allows analysis of more genomic regions using a smaller
population and detection of inter-allelic interactions like epistasis. Moreover,
both parental genotypes are almost equally represented in the population.
However, the RIL population also presents some limitations, such as diverse
developmental timings of different lines. This poses a significant challenge in
analysis of traits requiring phenotype information of individuals at the same
stage of growth and development. Furthermore, simultaneous segregation of
multiple loci in the population may hinder QTL detection because of inter-
allelic interactions, making it important to test for epistatsis.

In the absence of epistasis, the detection of small-effect QTLs can be
masked in an RIL population by the presence of large-effect OTLS. Detection
of small effect OTLS can be improved by using a different population, such as
the Near Isogenic Lines (NILs) or lntrogression lines (lLs), because
individuals in these populations contain a single introgression of the donor

parent. The NIL population also shows more homogeneous growth and

23

 

 

 

 

devel

popul.

CVi ar
popuia:
12 smal
deveiopr
branch n

RIL popul

1.8 Statist

OTL

Choi et al.)

“33* the null

 

developmental characteristics since most of the genetic background in the
population is identical (Swarup et al., 1999; Keurentjes et al., 2007).
Keurentjes et al compared populations of RILS and NlLs derived from
Cvi and Ler and reported discovery of additional OTLS using the NIL
population compared to the RIL population (Keurentjes et al., 2007). In total
12 small-effect OTLS were detected for different aspects of plant growth and
development (flowering time, length at first silique, total plant length and basal
branch number) that were masked by presence of large-effect OTLS in the

RIL population (Keurentjes et al., 2007).

1.8 Statistical methods of QTL Mapping

QTL mapping procedures such as Interval Mapping (Jansen, 1993;
Choi et al.) and Composite Interval Mapping (CIM) (Jiang and Zeng, 1995)
test the null hypothesis that a OTL is absent for the given trait based on the
data obtained from a given population. A putative OTL is assumed to exist at
every position throughout the genome and multiple tests are performed
across the whole genome by analyzing small portions at a time, thus
eliminating regions that lack statistical support for containing a QTL. To
control the error rate for this genome-wide analysis and improve the
confidence of results, the critical threshold value of the test statistic is
selected. The methods used today fall mainly into four classes:
Nonparametric methods, Least-squares based regression methods,
Likelihood based methods and Bayesian mapping. These analyses are
computationally demanding due to many repetitive tests during the search

procedure. Therefore, several computer softwares such as MAPOTL, BQTL

24

 

 

and

list ca

1.9 Gt

in mar.
activity
variatior
due to d
Furtherm.
appear as
Altr
10 identify
Understano
‘0 9am insi
this metho
transcriptOml

variah‘on e m

   
  

ITIiiuQnCeS
result frOm v.
its e”(Messie-
from sequent
SEQUGnCe Va
Qene, terms.

DeIrOrmed t

and QTL Cartographer have been developed for performing QTL analysis (a

list can be found at www.stat.wisc.edu/biosci/linkage.html).

1.9 Genetical Genomics: QTL Mapping for Regulators of Transcription
The molecular complexity of quantitative variation has been analyzed
in many studies. The analysis of alcohol dehydrogenase (Adh) enzyme
activity in Drosophila revealed a complicated molecular basis of allelic
variation (Stam and Laurie, 1996). Allelic variation was found to arise both
due to differences in the coding as well as non-coding regions of the gene.
Furthermore, it was noted that several allelic effects were combining to
appear as "super-alleles" which were segregating as a single major QTL.
Although QTL analysis has been primarily employed in plant breeding
to identify regulators of commercially interesting traits, it has also helped
understand the mechanics of evolution. QTL mapping is now being extended
to gain insights into regulation of differential gene expression, thus bringing
this methodology to the genotype-phenotype interface that is the
transcriptome. Variation in gene expression can result from a DNA sequence
variation existing in the gene, for example in its promoter region, that directly
influences expression of its transcript. Variation in gene expression can also
result from variation in the DNA sequence of another gene that is involved in
its expression. Therefore it is reasonable to imagine that eQTLs could arise
from sequence variation of the gene itself, termed as cis-acting eQTLs or from
sequence variations that would affect levels or functional efficacy of a different
gene, termed trans-acting eQTLs. In the recent past, several studies were

performed to understand the heritability of differential gene expression.

25

 

 

 

Micrc
expre
propc
Studie
compc
anas'yz
betwee
to the
approxi
used a,
further .
Positive:
acting (E
higher In
acting e(
eQTLs (E
Tr,
different (
Genes the
by the s.-

reQUlaIOrS

 

Microarrays were used in most of these studies to compare genome-wide
expression profiles of different eootypes/strains of organisms to determine the
proportion of genes that show genotype specific variation in expression.
Studies in yeast, Drosophila, mice, and humans show a strong heritable
component of differential gene expression. Doss et al used microarrays to
analyze genome-wide expression profiles of F2 mice generated from a cross
between two common inbred strains of mice and identified eQTLs contributing
to the variation in several thousand genes (Doss et al., 2005). They found
approximately one-third of these eQTLs to lie within 10 cM of the gene being
used as the trait. They classified these eQTLs as cis-acting and through
further analysis reported at least 64% of these cis-acting eQTLs as true
positives, whereas the remaining 36% could not be confirmed as truly cis-
acting (Doss et al., 2005). They also noted that the cis-acting eQTLs showed
higher linkage statistics. Interestingly, several other groups also reported cis-
acting eQTLs to be supported with higher linkage statistics than trans-acting
eQTLs (Bystrykh et al., 2005).

Trans-acting eQTLs identified in yeast were shown to map to several
different classes of genes besides transcription factors (Yvert et al., 2003).
Genes that were clustered based on similar function often revealed regulation
by the same eQTL. Therefore, this approach can be used to identify
regulators of differential gene expression that lie further upstream of the
process of transcription. In fact, in the recent past this approach of genetical
genomics has helped dissect several complex traits (Brem et al., 2002;
Wayne and McIntyre, 2002; Yvert et al., 2003; Ron et al., 2007; Ronald and

Akey, 2007).

26

.. ”mud

 

 

 

1.10 Ma
lr‘.
understa
plants. A
undergoe
CBF regi.
only set c
regulon 9
response 1
regUlators. I
expression
to identify
res.Dorise it
include tho:
Signal frOm
approach w
early Signal
to prOVlding

reQUlation_

 

 

1.10 Main Objectives of Current Study

In the present study, a QTL mapping approach was chosen to
understand the regulation of low temperature responsive transcription in
plants. As mentioned above, the transcriptome of Arabidopsis thaliana
undergoes extensive change at low temperature (Vogel et al., 2005). The
CBF regulon, although significant for low temperature tolerance, is not the
only set of genes induced in response to low temperature. In fact, the CBF
regulon genes comprise only about 33% of genes found to change in
response to low temperature. In an attempt to determine how many early
regulators of low temperature induced gene expression exist, change in gene
expression at low temperature was treated as a molecular trait. The goal was
to identify QTLs regulating expression (eQTLs) of genes induced early in
response to low temperature. Regulators of early responsive genes would
include those that encode the thermosensor(s) and protein(s) that relay the
signal from the thermosensor to the upregulated genes. Therefore this
approach would potentially allow identification of the thermosensor(s) and
early signal transduction components of low temperature signaling in addition
to providing an insight into the complexity of low temperature responsive gene

regulation.

27

 

 

.. .....i’

 

2.1

pt:

act

eie
tar!
aCc
tOle

Chapter 2

eQTLs Regulating Low Temperature Responsive Expression of CBF1-3

2.1 Introduction

Cold acclimation involves several changes in the biochemistry and
physiology of plants (Thomashow, 1998; Zhu, 2001; Yamaguchi-Shinozaki
and Shinozaki, 2006). Changes in gene expression have been established as
a central step translating into the biochemical and physiological adaptations
observed in cold acclimated plants (Yamaguchi-Shinozaki and Shinozaki,
2006; Zhu et al., 2007). Transcriptome studies using microarrays have
identified nearly 700 genes that were differentially expressed during cold
acclimation, at least some of which may function in a concerted manner to
generate cold tolerance (Fowler and Thomashow, 2002). These differentially
expressed genes were grouped into distinct classes based on their temporal
expression profiles (Vogel et al., 2005).

Among the genes rapidly upregulated by low temperature are the CBF
family of transcription factors, which recognize the CRT/DRE sequence
element in the promoters of COFI genes, resulting in the induction of these
targets (Zarka et al., 2003). CBF genes play an important role during cold
acclimation as over-expression of CBF 1, 2, or 3 confers constitutive freezing
tolerance to transgenic Arabidopsis (Jaglo-Ottosen et al., 1998). Nearly 90

low temperature-induced genes are found to be constitutively expressed in

28

 

 

tra
po
lev
lov.

DO!

evi

feECj

doll/r

plants over-expressing CBF 1, 2, or 3 and are classified as the CBF regulon
(Fowler and Thomashow, 2002).

Studies indicate complex regulation of the CBF genes at low
temperatures involving positive and negative regulators (Zarka et al., 2003).
Sequence elements required for low temperature responsiveness of the CBF2
promoter have been identified, however, the transcription factors interacting
with these elements remain unknown (Zarka et al., 2003). Several genes that
appear to play a role in the regulation of CBF expression at low temperatures
have been identified through mutant screens. L084 and ICE 1, encoding a
DEAD-box RNA helicase (Gong et al., 2002) and a MYC-like bHLH
transcription factor (Chinnusamy et al., 2003), respectively, appear to
positively regulate CBF expression. The Ios4 mutant displays decreased
levels of CBF1, 2, and 3 transcripts whereas the feet mutant demonstrates
lower levels of CBF3 transcripts only (Chinnusamy et al., 2003), suggesting
possible differences in the regulation of individual members of the CBF family.

In addition to positive regulators of CBF expression, genes showing
evidence for a role in negative regulation of CBFs have also been identified.
These genes include FRY2 and H081 that encode a novel transcriptional
repressor and a novel RING-finger protein, respectively (Xiong et al., 2004),
(Lee et al., 2001). HOS1 has been proposed to function by targeting certain
positive regulator(s) of CBFs for ubiquitination and subsequent degradation
(Lee et al., 2001).

Additionally, there is evidence to suggest CBF gene regulation by a
feedback repression mechanism, either involving the CBF proteins or a

downstream target. The [031 mutant, disrupted in low temperature specific

29

‘II‘- ‘

 

 

“MB, WK“? A'FF

 

"f

DIE

Bil

protein synthesis, shows super-induction of CBF genes (Guo et al., 2002). It is
proposed that inhibition of protein synthesis in this mutant would prevent
accumulation of CBF and downstream target proteins, rendering them
incapable of feedback repression, leading to super-induction of CBF
transcripts (Guo et al., 2002).

However, the mechanism by which CBF genes are regulated at low
temperatures remains poorly understood. Several approaches can be
employed to identify regulators of CBF gene expression; biochemical
approaches designed to identify proteins binding to promoters of these genes
(Baker et al., 1994), genetic approaches involving mutational analysis (lshitani
et al., 1998; Xin and Browse, 1998; Knight et al., 1999; Lee et al., 2001), or
molecular approaches studying the effect of over-expressing these genes in
planta (Jaglo-Ottosen et al., 1998). Such approaches, however, are limiting
when studying complex traits, which involve polygenic regulation and gene X
environment interactions, as they generally focus on individual genes or
genotypes.

An alternate approach to understand the regulation of complex traits
involves the use of naturally occurring genetic variation, which is widespread
in nature (Alonso-Blanco and Koornneef, 2000). Gradual genetic change,
occurring over long periods of time, contributes to the ability of populations to
adapt to their environment (Farrall, 2004). Genetic variation, when coupled
with environmental heterogeneity, results in the fixation of different genotypes
in different environments. Ecotypes of plants growing in diverse geographical

regions that experience different ambient temperatures may contain genetic

30

 

l
‘-

 

 

mi
al.

lra

variation that allowed them to adapt to local temperatures. Comparison of
such ecotypes may reveal the genetic differences that led to this adaptation.
Complex traits, both in plants and animals, tend to show continuous
variation, which can be contributed by heritable and/or non-heritable factors
(Orr, 1999). The non-heritable component results from the effect of the
environment on trait development. The heritable component, on the other
hand, arises due to the effect of several genes on trait development (Orr,
1999). These genes can be identified through QTL mapping, which narrows
down a chromosomal region that may contain one or more genes affecting
trait development (Kearsey and Farquhar, 1998). This approach is limited to
detecting only those genes that show allelic differences among the parents.
Use of QTL mapping has recently been extended to provide insights in
to the regulatory mechanism of transcription in yeast, C. elegans, plants,
mice, and humans (Bing and Hoeschele, 2005; Deutsch et al., 2005; Kirst et
al., 2005; MacLaren and Sikela, 2005; Li et al., 2006). In these studies
transcript levels of genes were considered as a molecular trait in segregating
populations to identify QTLs, which were termed eQTLs (expression QTLs)
(Zou et al., 2007). Comparison of genome-wide expression profiles in various
systems revealed a strong heritable component of differential gene
expression (Bing and Hoeschele, 2005; de Koning et al., 2005; Deutsch et al.,
2005; Yamashita et al., 2005). Alterations in transcript levels were reported to
result in tighter regulation of genes than mutations in gene sequence
(Wittkopp et al., 2004). Microarray analysis of two ecotypes of Arabidopsis,

Bay-0 and Sha, revealed that 32.2% genes demonstrate heritable differences

31

in expression (Vuylsteke et al., 2005). Variation in gene expression was found
to occur due to cis- and trans-acting eQTLs (Vuylsteke et al., 2005).

The present study was undertaken to explore the potential of using
natural variation in CBF gene expression to identify regulators of low
temperature associated transcription of these genes. CBF transcript
abundance at low temperature was considered as a molecular trait to identify
eQTLs controlling CBF expression. Identification of common eQTLs for CBF 1,
2, and 3 genes would indicate that these genes are induced through a
common pathway, whereas identification of different eQTLs would suggest
involvement of distinct pathways. Characterization of the genomic region
underlying such eQTLs would lead to identification of genes regulating CBF
expression at low temperatures. These regulators could potentially include
genes encoding the thermosensor and downstream signal transduction

components, amongst others.

32

 

m:-

 

 

WI

In:

2.2 Results

2.2.1 Natural variation in CBF gene expression

In addition to the availability of vast genetic and genomic resources,
Arabidopsis thaliana grows across a wide range of geographical regions with
considerably diverse environmental conditions, thus making this species an
ideal candidate for the current study (Shindo and Bemasconi, 2007). To
identify eQTLs regulating low temperature responsive expression of CBF 1, 2,
and 3 genes, Arabidopsis accessions collected from warm and cold climates
were compared to identify variants in low temperature responsive CBF
expression. The accessions Col-0, Ct-1 (081094), Cvi-0 (08902), No-O
(CS1394), Stw-O (CS1538), and Tsu-1 (CS1640) were compared for
differences in CBF transcript accumulation at low temperatures. These
accessions differed with respect to the temperatures (and other factors) they
were regularly exposed to in their natural habitats (Figure 2.1).

As described earlier, exposure to low temperature leads to the rapid
induction of CBF1-3 genes (Gilmour et al., 1998). In seedlings exposed to
4°C, CBF transcripts can be detected as early as 15 min, with highest
expression levels observed at around 2h followed by decrease to levels about
three-fold above those in control plants (Gilmour et al.). In a preliminary
screen for natural variation in low temperature responsive CBF expression,
seedlings of each accession were subjected to a gradual decrease in ambient
temperature (starting from 21°C, with decrease in set temperature point by
2°C every 4h) and total RNA of each accession, incubated for 1h or 4h at

17°C, 15°C, 13°C,

33

 

to

di‘.

tel

WC

 

Temperature January Temperature July

 

3"" 1' .. .r-é""—_,:.
a. .o' 0“ _. I .
' T. ‘ as “31753 .
. is“- .
a
r!"
N13539::
Symbol Awesslon Temperature
0c OF
C “3"" 32 i 90
21 J 70
O 3"” 10 so
-1 p, 30
O "0'0 -11 f.“ 10
-23 '34 -1o
. CV“) -34 ‘ -3o
. Col-0 '46 '50

Figure 2.1. Geographical distribution of Arabidopsis accessions selected
for analysis. Colours indicated in the legend mark the original habitats of the
different Arabidopsis accessions. The heat map represents average
temperatures across different geographical regions during winter and summer
months with the colour scheme depicted in the legend (The Reader’s Digest

World Atlas, 1991 ;ISBN 0-89577-388—0;pg 11).

Temperature (°C)/Time (hours)

.c

11l1h

111%

is

17!"!
15I1h
15/4h
13l1h

1:
2
h.
._

    

Col-0
Cvi-O
Stw-O
Tsu-1
No-O

Figure 2.2. Expression of CBF genes in Arabidopsis accessions exposed
to gradual reduction in ambient temperature. Northern blot showing total
CBF transcript accumulation in seedlings of Col-0, Cvi-O, Stw-O, Tsu-1, and
No-O after reaching indicated temperature for varying periods of time (1h or

4h). A probe that hybridized to CBF 1, 2, and 3 was used for detection of total

CBF transcripts.

35

11°C, 9°C, 7°C, 5°C, 3°C, and 1°C, was analyzed using Northern analysis
(Figure 2.2). Although most accessions demonstrated low-level induction of
CBF transcription at temperatures as high as 15°C, there was a marked
increase at 5°C. Of the accessions tested, Cvi-0 exhibited lowest levels of
CBF transcript at 5°C. Collectively, these results indicated natural variation in

CBF expression amongst different accessions of Arabidopsis.

2.2.2 Analysis of low temperature responsive CBF expression in Ler and
Cvi-0

Analysis of a segregating population derived from a cross between two
parents showing quantitative variation in CBF transcript accumulation could
be performed for the identification of loci regulating CBF expression. Since a
segregating population for the Cvi-0 X Ler cross was available (Alonso-Blanco
et al., 1998), quantitative variation in CBF transcript accumulation was tested

in Cvi-0 and Ler.
Considering the marked difference in CBF expression at 5°C (Figure
2.3), it was assumed that a single temperature point would be sufficient for
comparison of CBF expression. To confirm this and select a temperature for
comparison, seedlings growing in 12h photoperiod at 22°C were exposed to
different temperatures for 2h. Northern analysis was performed to compare
the response of accessions We (previously known to accumulate high levels
of CBF transcripts at 4°C (Cook et al., 2004)), Cvi-0, and Ler, following a
reduction in ambient temperature from 22°C to 10°C, 8°C, 6°C, 4°C, or 2°C

for 2h (Figure 2.4). Although both Cvi-0 and Lerinduced CBF expression at

36

 

Tem erature ° for 2h

W10864
.

Ler F
WS oi
Cvi-0 q

2

 

 

 

 

 

 

 

 

 

 

b.
500 v_ _
.1?
c
D
2‘
E
.1:
8
s
E
o
_I
<
Z
I
E
123 123 123 123 123 123
Warm 10°C 8°C 6°C 4°C 2°C

Figure 2.3. Expression of CBF genes in Cvi-0, Ler, and We following a
sudden drop in ambient temperature. a) Northern blot showing total CBF
transcript in short-day photoperiod grown seedlings following 2h incubation at

the indicated temperature. A common probe was used for detection of CBF 1,
2, and 3. b) Histogram showing CBF transcript levels normalized with 18S
rRNA. The y—axis depicts transcript level in arbitrary units and the x-axis

shows the accessions Cvi-0, Ler, and W3 (1, 2, and 3 respectively) after 2h at

depicted temperatures.

37

the same temperature, Cvi-0 accumulated lower levels of CBF transcript, thus
confirming quantitative variation between the two accessions. Furthermore,
the quantitative difference in transcript levels among the accessions increased
with decrease in temperatures. The accession Ws showed highest levels of
CBF transcript.

To determine if Cvi-0 was able to accumulate CBF transcripts to levels
comparable with Ler following an extended period of exposure to low
temperature, a time course experiment was performed at 4°C. As shown in
Figure 4, both Cvi-0 and Ler showed similar kinetics of CBF induction, with

the transcript peaking at 4h and decreasing to lower levels by 24h.
The CBF genes have been shown to be influenced by the circadian
clock; basal levels of CBF3 were found to undergo circadian cycling (Harmer
et al., 2000) and the induction levels of CBF 1, 2, and 3 were dependent on
the time of day at which they were subjected to the low temperature stimulus
(Fowler et al., 2005). To determine if the exclusion of circadian effects would
abolish the variation in CBF transcript accumulation, Cvi-0 and Ler seedlings
grown under continuous light were analyzed after 2h incubations at 12°C,
10°C, 8°C, 6°C, 4°C, 2°C, and 0°C (Figure 2.5). Cvi-0 was found to
accumulate lower levels of CBF transcript compared to Ler even in the

absence of a photoperiod.

Further, to determine if quantitative variation in CBF expression existed
in a segregating population derived from Cvi-0 and Ler, an RIL population
derived from the Cvi-0 X Ler cross was analyzed for CBF transcript

accumulation after 2h at 4°C in the absence of a photoperiod. The use of

38

 

‘91“!

Cvi-0 Ler Ws
2h 4h 24h 2h 4h 24h 2h 4h 24h

[ .. to» -04

 

 

 

 

 

 

 

 

A -A—Cvi-0

‘ -l- Ler
*Ws ‘

 

 

 

 

}

 

 

2h 4h 24h

 

Time (hours)

Figure 2.4. Expression of CBF genes in a time course of Cvi-0, Ler, and

Ws, at 4°C. a) Northern blot showing total CBF transcript accumulation with a
common probe for detection of CBF 1, 2, and 3, in short-day photoperiod
grown seedlings after 2h, 4h, or 24h at 4°C. b) Graph showing CBF transcript
levels normalized with 183 rRNA levels. The y-axis depicts transcript level in
arbitrary units and the x-axis depicts periods of 2h, 4h, or 24h at 4°C for the

accession mentioned.

39

 

 

Ler

188 rRNA

Cvi-0

18S rRNA

Figure 2.5. Comparison of CBF gene expression in Cvi-0 and Ler in
response to sudden change in temperature. Northern blot showing total
CBF accumulation in continuous-light grown seedlings after 2h at indicated

temperatures, using a common probe for CBF 1, 2, and 3 detection. 18S rRNA

levels in the samples are

Tem erature ° for2h
W 1210 8 6 4 2 0

.~ new-ll

 

IQOQQOéfi ..

 

 

 

Milt”

 

51;; 3‘: egg- (, “1g .4»:-

sdq-QQ 9.0

 

 

 

shown below.

40

continuous light allowed identification of eQTLs associated with variation in
low temperature associated transcription of the CBF genes, eliminating
eQTLs that may be associated with photoperiod mediated regulation of these
genes. Additionally, continuous light conditions simplified the experimental
design for simultaneous comparison of large number of lines. The RIL
population comprises 162 lines inbred to F3 generation that have been
comprehensively characterized for their genotypes using AFLP and CAPS
markers distributed across all five chromosomes (Alonso-Blanco et al., 1998).
The results of Northern analysis indicated considerable variation in CBF
transcript accumulation among the RILs (Figure 2.6). Furthermore, many RILs
expressed the CBF transcript to significantly higher levels than Ler,

demonstrating transgressive segregation of this molecular trait. This

suggested allelic variation in genes involved in low temperature-regulation of

CBF in the parents.

2.2.3 eQTL Mapping for Regulators of Low Temperature Responsive CBF

Expression

As mentioned previously (see Section 1.8), accuracy of trait

measurement significantly impacts the power and resolution of QTL detection
thereby necessitating minimal experimental variation in CBF transcript
quantification. The high level of homozygosity of the RIL population used in
this study allowed biological replicates for trait measurement, which provided
greater statistical reliability. Initially, Northern analysis was used for transcript

quantification; however, significant experimental variability was observed

41

a.
RILs after 2h at 4°C
1 2 3 4 5 67 a 9101112131415161718192021222324252627282930

CBF1-10”.. I... 111...... 4.. . «a
meABoo-eogioogggeogog...o........

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

b.
250
200
’0?
g
C
3150
2‘
g
E
S,
5100
> I!
3 l n
2 ~ -I it 1
m 50 1." £ 4 t" 1t 31'
E i 1 I: 1’. i
11.3.. -31 . 1 31313.1.
llaizli' 22131. rail-l -.=‘~I-:i:.-ii it 'i 2
JIHLIQHEJ .1111 111111.11 .111’1’1 1.! A)”; ,

 

 

 

 

 

 

 

 

 

Figure 2.6. a) Expression of CBF genes in Cvi-0 and Ler following sudden
change in ambient temperature. Northern blot showing CBF mRNA levels in
continuous-light grown seedlings incubated for 2h at 4°C, using a common

probe for detection of CBF1, 2, and 3 transcripts. 18S rRNA levels in the

samples are shown below.

b) Histogram showing CBF transcript levels in RlLs. CBF levels
normalized with 188 rRNA levels in 77 randomly selected RILs and the
parental lines Cvi-0 and Ler. The y-axis depicts transcript level in arbitrary

units and the x-axis shows different RILs. CBF transcript levels in Ler and Cvi-

0 are numbered 1 and 2 respectively.

42

when analyzing larger sample sets by this approach, which was not suitable
for further analysis. Hence, to improve the accuracy of trait measurement,
Taqman® real-time RT-PCR (Applied Biosystems Inc., ABI) was chosen for
transcript quantification. This method was optimized for quantification of CBF
and 183 rRNA transcripts (see Chapter 4, section 4.2). The 18S rRNA gene
was selected as the internal reference to normalize expression data for CBF
genes. Two-step Taqman® real-time FIT-PCR was performed for estimation of
transcript levels in 155 RlLs exposed to 4°C for 2h in two independent
experiments. Transcript levels for CBF 1, 2, 3, and 188 rRNA genes were
derived from the 01 values determined from real-time amplification plots of the
samples (Figure 2.7). To achieve accuracy and reproducibility of transcript
measures, 01 values were determined at the exponential stage of
amplification. The amount of transcript present in each sample was estimated
as 2'ACt values and outliers were removed before further analysis (see section
2.4 Material and Methods).

Average expression levels from the two biological replicates were used
for Composite Interval Mapping (CIM) to identify eQTLs for CBF 1, 2, or 3
expression at low temperature. Data analysis was performed using the QTL
Cartographer program available at www.statqen.ncsu.edu/qtlcart/. The critical
threshold level for identifying the eQTLs was determined to be at a LOD score
of 2.5 (see section 2.4 Materials and Methods). Mapping of phenotype data
obtained using real-time RT-PCR led to identification of eQTLs for CBF 1, 2
and 3 expression (Figures 2.8, 2.9, and 2.10). The eQTLs were named on the
basis of their chromosomal location, for instance eQ C1-3 denotes the third

eQTL present on chromosome 1 (Table 2.1).

43

a 1g rBNA b.At§g25490: QRTLQRE Binding Proteln 1
— F—

 

 

 

ARn—v
ARn—>

 

 

 

 

 

 

 

   

c. At 70: RT RE Bindln Protein 2 d.A_g_t4 2548019RT1DRE Binding Prgtfln 3

 

 

 

 

 

ARn—>
ARn—>

 

 

 

 

 

 

 

 

 

Cycle —>

 

Figure 2.7. Taqman real-time RT-PCR amplification profiles for a. 188
rRNA, b. CBF1, c. CBF2, and d. CBF3 transcripts. The y-axis represents
fluorescence emission and the x-axis represents cycle numbers. Each
amplification curve represents a single well, and every RIL is measured in
triplicate wells. The horizontal red line depicts the threshold at which the 01

value for each amplification curve was determined.

2.2.3.1 Identification of a Common Large-effect eQTL Regulating CBF2
and 3 Expression at Low Temperature

This analysis identified large-effect eQTLs for differential expression of
CBF2 and 3 genes co-localizing on chromosome 4, which were supported by
LOD scores of 19.5 and 17.8, respectively (Figures 2.8 and 2.9). The positive
allele leading to higher CBF2 and 3 accumulation was contributed by Ler for
both genes. Additionally, this eQTL, named eQ C4, was coincident with the
location of CBF genes in the genome. The overlap with the genomic location
of CBFs suggested that this might be a cis-acting eQTL. However, it might
also be a trans-acting eQTL, located in physical proximity to the CBF genes,
overlying a common regulator of CBF2 and 3 transcription. In contrast, low

temperature responsive expression of CBF1 gene did not identify a regulatory

eQTL at this location (Figure 2.10).

2.2.3.2 Identification of Common Small-effect eQTLs for Expression of

CBF1, 2and 3Genes

The low temperature responsive transcription of CBF 1, 2, and 3 genes
identified shared small-effect eQTLs on chromosome 5, with LCD scores of
2.7, 4.4, and 5.5, respectively (Figures 2.10, 2.8 and 2.9). An additional set of
eQTLs on this chromosome was identified in common for CBF2 and 3
expression (3.9 and 2.7 LODs, respectively), which were contributed by Ler

for both genes (Figures 2.8 and 2.9). CBF1 expression did not reveal an

eQTL at this location through this analysis.

45

Figure 2.8. eQTLs for CBF2 expression. CIM analysis of 183 rRNA
normalized CB21 transcript levels in RlLs treated at 4°C for 2h, using OTL
Cartographer. In the top panel, the y-axis represents the LCD scores and the
x-axis represents the five chromosomes of Arabidopsis. The area between
each set of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive

value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, which is Ler.

46

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

om
zlowo em xx Mel). C as . We

V IVV: m: 51 1: tilt/x

11k .

cg 5.15. NV mm

an mam mV_
A: 35 mV
6.. m? .15 «V 80..
3.50qu0

m. FN-
wdT

8.0

. ace
1 QFN

Nw

@—

Figure 2.8. eQTLs for CBF2 expression.

47

Figure 2.9. eQTLs for CBF3 expression. CIM analysis of 188 rRNA
normalized CBF1 transcript levels in RlLs treated at 4°C for 2h, using QTL
Cartographer. In the top panel, the y-axis represents the LCD scores and the
x-axis represents the five chromosomes of Arabidopsis. The area between
each set of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive
value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, which is Ler.

48

 

 

 

 

8.0
- v. N
1 New».

 

3

 

 

3m New .3

 

8.: 5.8 .aV

 

 

 

1?.
8.23.5113

1 mm.

1 mdw

 

 

000;

mp

 

«duo EMU

Nd?
g fivw-

Figure 2.9. eQTLs for CBF3 expression.

49

Figure 2.10. eQTLs for CBF1 expression. CIM analysis of 188 rRNA
normalized CBF1 transcript levels in RILs treated at 4°C for 2h, using QTL
Cartographer. In the top panel, the y-axis represents the LCD scores and the
x-axis represents the five chromosomes of Arabidopsis. The area between
each set of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive

value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, which is Ler.

5O

 

 

ll

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:

 

 

 

 

 

 

4V

5.. 5.8 .mV

 

 

 

 

 

 

 

 

md

80.. V

 

 

 

“dam nkmo

mud-
omd-
cod
00.0
mud

wd

N...

w...

YN

51

Figure 2.10. eQTLs for CBF1 expression.

In addition to being regulated by eQTLs common to CBF1 and 3
expression, the low temperature-induced expression of CBF2 gene was found
to be regulated by a unique small-effect eQTL on chromosome 2 (Figure 2.8).
This eQTL was identified with a LOD score of 2.7 and the positive allele at this
location was contributed by Cvi-0 (Figure 2.8).

Therefore, these analyses led to the identification of regulatory loci for
CBF 1, 2, and 3 genes, revealing similarities and differences in their regulation
based on the assumption that eQTLs overlapping in chromosomal location
result from the same underlying gene (Table 2.1). Although CBF1 expression
was found to be regulated by trans-acting eQTLs, CBF2 and 3 expression
revealed a potential cis-acting eQTL in addition to the trans-acting eQTLs
identified. Additionally, both parents, Ler and Cvi-0, contributed alleles leading

to higher accumulation of transcripts.

2.2.3.3 Uncovering Additional Small-effect eQTLs for CBF expression
Co-localization of the large effect eQTL on chromosome 4 with the
CBF genes and its very strong statistical support suggested it to be a cis-
acting eQTL. If this were true, eliminating sequence variation at the CBF locus
by holding it constant in the population as Ler or Cvi-0 would eliminate this
eQTL. Moreover, this approach could potentially uncover additional small-
effect eQTLs that were being statistically masked by this eQTL. To determine
if this was the case, the RIL population was divided into two groups, one
comprised of lines containing the CBF locus from Cvi-0 and the other from Ler

(see Materials and Methods for details).

52

Cvi at CBF locus

 

31LOD

2.4 .

o... . ,1 t
1 / r \ P re
A tot/541,4“ \l’ \ firth

 

 

 

 

 

o .

 

12.2 30 ,

00 .\ Air—”r1 A; r l

-6.10 .W/ ”“4 U \~ 1,. M4 1W” W
-12.2 .

 

 

 

 

 

 

 

 

 

 

Ler gt QBF Iocgs
3 tLOD

 

2.4 . ll

0.6

 

 

333 "30 p 1 Mt
0.0 ‘T fl“ /\‘ “le t‘lv \J\’\’\ 'F\ ~ A\ \ Vq v/h\\1_\j‘u
-1.12 M f U (j \ l
-2.25 -1

 

 

 

 

 

 

 

 

 

Figure 2.11. eQTLs for CBF2 expression after holding CBF locus
constant. CIM analysis of CBF2 transcript levels in RlLs containing the CBF

locus from either Cvi-0 or Leras indicated.

53

Cvi at CBF locus

 

3LLoo

2.4 .

 

 

 

 

 

 

0.: iii-1mm WU MN I VM 21% ”\1

11.8 f" a0

5.92 . 1 l
0.0 1N\/U%«L .efv‘wl \WH my

-5.92 . Tv” MUM V l . Nb
-11.a . v‘

 

 

 

 

 

 

 

 

 

Ler gt QBF locus

 

3 to

I: 1 l
flb rm Wit ll! lN / ‘ UV
giflleJr/

 

 

79.6 -
39.9

 

-39.9
-79.9

 

 

 

 

 

 

 

 

Figure 2.12. eQTLs for CBF3 expression after holding CBF locus
constant. CIM analysis of CBF3 transcript levels in RlLs containing the CBF

locus from either Cvi-0 or Leras indicated.

 

The two subpopulations of RlLs thus created were analyzed by CIM. The

division of the RIL population into smaller subgroups increased the critical

LOD score required to determine eQTLs. As would be expected, elimination

of sequence variation around CBF locus led to disappearance of the eQTL on

chromosome 4 from both subpopulations for CBF2 and 3 expression (Figures

2.11 and 2.12). The graphs do not show a threshold line because it is higher
than the maximum LOD value represented on the y-axis. The subdivision of
the population resulted in identification of an additional transacting eQTL for
the regulation of CBF1 expression (Figure 2.13). This eQTL was located on
on chromosome 3 and was supported by a 7.3 LCD (Figure 2.13). Cvi-0
contributed the positive allele for this eQTL. Furthermore, none of the small-
effect eQTLs identified for the CBF genes using the entire population were
detected using the subpopulations generated in this analysis (Figures 2.11,
2.12, and 2.13). Division of the RIL population into subpopulations for
elimination of sequence variation at the CBF locus may have inadvertently
abolished variation at additional genomic regions. This would occur as a result
of the smaller number of RlLs in each subpopulation, which would lead to
reduction in overall genomic variation represented in each subpopulation. If
these additional regions overlapped with the genomic locations of the small-
effect eQTLs identified using the entire population, they would no longer be

detected in the analysis of the subpopulations.

2.2.3.4 Use of Additional Normalization Control for Expression Analysis
The choice of internal reference gene to normalize loading differences

between samples determines the accuracy of transcript quantification using

55

 

 

Cvi at CBF locus

 

 

a ) LOD
/3, 15.3, 73
6.4 -
l
4.8 " 1T

l

1.6 '
1 .Nfl J/\"~—-l Li‘s/MNV-v—‘N‘ Krk lflvfll. .m/F JV] V

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0
1.40 l 80
0.70 L 1:. A JV“
0.0 V .w ». [V V‘s—x, j\A {V J. \K
v w; \K/V ~WV/v VI / \W \r’
on I ‘ l Vt“
-1.40 l W/
Lgrg QBFIOCQ
3 f
‘ l
2.4 '1‘ 1
1.8 i
l
l
l
l
‘2 1 1 1 . l l
..1 will l .1 l
l ‘ . l
O .K/rm \Alh (k Lit/1 rfz ‘ TN‘ 1} My x. If, \a Ml NA
1.58 a0
0.79 \ . l *1 My
0.0 «bwmfi 4" .__ 'I . i:— ‘” 51 7 7 A W
-0.79 -1 vf * :3 1 W 1 ”VA
-1.58 ‘

 

Figure 2.13. eQTLs for CBF1 expression after holding CBF locus
constant. CIM analysis of CBF1 transcript levels in RlLs containing the CBF

locus from either Cvi-0 or Leras indicated.

56

real time RT-PCR. There is no single reference gene considered to be a
universal standard for all systems. Thus, it is necessary to determine the gene
that would serve as an ideal internal reference for the system being studied
(De Boever et al., 2008; Lyng et al., 2008; Flemans et al., 2008). The choice
of reference genes is based on two main factors, the relative abundance of
the reference gene and its unaltered expression under the experimental
conditions being tested (De Ketelaere et al., 2006; Maccoux et al., 2007;
Skovgaard et al., 2007; Hoogewijs et al., 2008). Although a reliable reference
for normalization, 18$ rRNA is not ideal for quantification of low abundance
mRNAs. The relative abundance of this transcript in the RlLs was, of course,
significantly higher than CBF transcripts since cDNA samples had to be
diluted 2000-fold to quantify 18S rRNA levels. Studies investigating changes
in gene expression have also demonstrated that using different reference

genes can uncover additional differences in regulation of transcription

(Deutsch et al., 2005).
The actin-encoding gene, Actf, was selected for use as a second

internal reference as transcript levels of A011 are more comparable to CBF
transcript levels than 188 rRNA and remain unaltered at low temperatures

(Figure 2.14). Furthermore, analysis of Act1 expression in the RIL population

did not reveal any eQTLs for Act1 (Figure 2.15).

2.2.4 Identification of eQTLs Regulating of CBF1, 2 and 3 Expression

using Act1 as Internal Reference

Using Actf normalized transcript levels, eQTLs were identified for

CBF 1, 2, and 3 genes (Figures 2.16, 2.17, and 2.18). The eQTL identified for

57

 

I. all

 

CBF
rRNi
data
and

Funr
dvor
Howe
CBF 1

00me

genes
adoht

CBF1

CBF2 and 3expression overlying CBF locus on chromosome 4 using 18S
rRNA normalized dataset was reproducibly detected in the Act1 normalized
data (Figures 2.17 and 2.18). Ler contributed the positive allele for both CBF2
and 3and the LCD scores of the peaks were 7.8 and 6.8, respectively.
Further in agreement with the 18S rRNA normalized dataset, this eQTL on
chromosome 4 was not found to regulate CBF1 expression (Figure 2.16).
However, this analysis was not able to identify the small-effect eQTLs for

CBF1, 2, and 3 expression identified on chromosome 5 using 188 rRNA

normalization.
Additional information was obtained for regulation of CBF1 and 2

genes by using Act1 as the internal reference. A common pattern of two

adjoining eQTLs at overlapping genomic locations was identified for both

CBF1 and 29enes on chromosome 1 (Figures 2.16 and 2.17).

58

 

FIQUre
Showln

and {re

Time at4°C
W 1h 4h 8h 24h

Cvi-0

 

Figure 2.14. Actin expression in response to low temperature. RT-PCR
showing Actin transcript levels in Ler and Cvi-0 seedlings grown in warm (W)

and treated for 1h, 4h, 8h, or 24h at 4°C.

59

 

figure 2.15. t
normalized A01
Canogtapher- I
x-aXlS represer
each set of van
ch'omosome. T
95°: confidence
chromosome m
The lower pane
value indicates '

11h higher level:

 

 

Figure 2.15. eQTLs for Act1 expression. CIM analysis of 18S rRNA
normalized Act1 transcript levels in RILs treated at 4°C for 2h, using QTL
Cartographer. In the top panel, the y-axis represents the LCD scores and the
x-axis represents the five chromosomes of Arabidopsis. The area between
each set of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive

value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, which is Ler.

6O

 

 

 

I 73-9 ‘A"; ‘QI-i W‘s.

 

 

 

 

FigUre

 

 

.2 NIH

2 7?.) .P .i......

 

 

 

 

 

 

 

 

 

 

 

 

 

1,. it i: 2 <4 VS. . M...
.225 .H\il J/JW 5V; -2 , as )4 fll. o
V :3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mN

 

 

 

 

 

 

 

oooucm

 

._. :3

Sad.

Figure 2.15. eQTLs for Act1 expression.

61

‘ u ‘c'l-N" M

 

 

Figure216.
CBF1 transc:
h the top 1
represents tl
oi verocal c
chromosome
95”: confide
chromosome
he lower p;
value indicaz

'lh'th higher le

Figure 2.16. eQTLs for CBF1 expression. CIM analysis of A011 normalized
CBF1 transcript levels in RILs treated at 4°C for 2h, using QTL Cartographer.
In the top panel, the y-axis represents the LCD scores and the x-axis
represents the five chromosomes of Arabidopsis. The area between each set
of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive
value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, which is Ler.

62

 

 

 

TL

BF

1:ng

 

:1:

7.25.

L:—

 

it
4::

 

 

 

 

 

 

 

 

 

k

<<

 

om

 

 

‘1
.280—

-- ms

amp

 

 

V V .

 

 

 

 

 

 

 

 

 

 

 

 

C 3

ad {8 ._V
as .Ev .FV 1

 

0001.

 

“firms sham

 

md

m...

VN

Nd

Figure 2.16. eQTLs for CBF1 expression.

63

- . ’I‘sawn—Il

 

Figure 2.17
CBF2 trans
Ir. the top
represents
of vertical
chromoson
95c: conflc
ct'omosor
The lower

lace ihdlc

1111". higher

Figure 2.17. eQTLs for CBF2 expression. CIM analysis of Act1 normalized
CBF2 transcript levels in RlLs treated at 4°C for 2h, using QTL Cartographer.
In the top panel, the y-axis represents the LOD scores and the x-axis
represents the five chromosomes of Arabidopsis. The area between each set
of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive
value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, which is Ler.

 

 

 

 

 

 

 

 

 

 

 

 

-mufldlllu: Nth

Flg

 

 

l

l

 

r“
i

 

 

 

 

 

 

 

 

 

l

 

 

 

 

i
<
é.
3

an

em:

 

 

 

 

as 6.2 .aV

 

 

as .8 .eV.

 

 

 

as 6.8 .eV

 

 

 

 

 

 

. .. -_
mam .55. .FV
cstmsmsV

ooouLm

l

 

iv-
N.N-
0.0
N.N

codes

0.?

N.N

wé

Yo

 

 

m :Go Nkmu

Figure 2.17. eQTLs for CBF2 expression.

65

. v - Pfi'fl‘w-n-I”

 

 

Fgure 2.1E
CBF3 trans
to the top
'epresents
oi vertical
chromosor
95°: conti
Cll'ol'li-OSO
The lower
‘v’aloe 10111

W highe

Figure 2.18. eQTLs for CBF3 expression. CIM analysis of A011 normalized
CBF3 transcript levels in RlLs treated at 4°C for 2h, using QTL Cartographer.
In the top panel, the y-axis represents the LOD scores and the x-axis
represents the five chromosomes of Arabidopsis. The area between each set
of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive
value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, Which is Ler.

66

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

._. km

 

ww-
1. _.1
I... 11 {5951/1 7- l-2?rltii7§).\(\/} l.\r/ )\>\1 o
i < /\r\ t. 2 CK (ll/xKv
< -..
9......N
‘ ..
.282 1.. E .. 1 we we .me .E
3v
flfl mm .5
{NF
59
6.94.8 .eV 89 2N

Figure 2.18. eQTLs for CBF3 expression.

67

 

«_- ~11

Ll

 

2.3

acc
Eva
(All
res
acc
the

819!

Var
ide
Kl1e
exr
ide

the

2.3

CB

2.3 Discussion

Comparison of different accessions of Arabidopsis thaliana indicated
natural variation in low temperature responsive CBF expression among
accessions collected from different environments (Figures 2.3, 2.4, and 2.5).
Evaluation of an RIL population derived from a cross between Cvi-0 and Ler
(Alonso-Blanco et al., 1998), accessions from warm and cold environments,
respectively, revealed transgressive segregation of CBF transcript
accumulation (Figure 2.7). This suggested presence of allelic variation(s) in
the parent accessions that were independently segregating to result in
significantly higher levels of CBF transcripts observed in some RlLs.

In past two years several studies have demonstrated that quantitative
variation in gene expression is modulated by heritable factors that can be
identified as eQTLs (Bing and Hoeschele, 2005; Carlborg et al., 2005;
Kliebenstein et al., 2006). QTL mapping performed using CBF1, 2, or 3
expression as the molecular phenotype in the Len'Cvi RIL population led to
identification of eQTLs regulating low temperature responsive expression of

these three genes.

2.3.1 Identification of a Major-effect eQTL Common for Regulation of
CBF2 and 3 Expression

A major-effect eQTL eQ C4, with the highest statistical support (upto
LOD score ~20), was detected on chromosome 4 for CBF2 and 3 expression
(Figures 2.8, 2.9, 2.16, and 2.17). This eQTL overlapped the chromosomal

location of CBF genes. The Ler allele at this eQTL was found to contribute to

68

 

 

 

Te

 

 

 

 

eQTL
Gene
188 rRNA Act1
90 C1 -3
CBF1 (At4925490) 90 05-1
00 C1-4
'30 02.1 00 C1-3
00 C4
CBF2 (At4925470) 90 C1 -4
00 05-1
00 C4
90 05-2
90 C1 -1
60 C4
CBF2 (At4925480) 90 C4
90 C5-1
eQ C5-2

 

 

 

 

 

Table 2.1. List of eQTLs identified using 188 rRNA and Act1 normalized gene

expression data. The nomenclature is based on the location of the eQTL.

69

 

higher
Alonsc
was Sr
2005';
demo
CBF.
resu
loler
con.
Chrr

ext

higher expression of both CBF2 and 3 genes. Using the same RIL population,
Alonso-Blanco et al reported a freezing tolerance QTL at this location, which
was suggested to arise from variation in the CBF2 gene (Alonso-Blanco et al.,
2005). Sequence comparison of this locus between Cvi-0 and Ler
demonstrated variations at the CBF locus, including a 1.6kb deletion of the
CBF2 promoter in Cvi-0. Transformation of Cvi-0 with CBF2 gene from Ler
resulted in higher levels of CBF2 transcript and restoration of the freezing
tolerance of transgenics to levels comparable to Ler. It was reasonable to
consider the possibility that the eQTL observed for CBF2 expression on
chromosome 4 was a cis-acting eQTL arising from the sequence variation
existing in the promoter.
In several systems cis-acting eQTLs, which arise due to allelic variation
in the gene itself, were consistently identified with higher linkage statistics
than trans-acting eQTLs, which arise due to allelic differences in regulatory
factors involved in the expression of the gene (Bystrykh et al., 2005). In mice,
analysis of eQTLs contributing to variation in several thousand genes
revealed about one-third of the eQTLs lying within 10 cM of the gene being
used as the trait and were classified as cis-acting eQTLs (Doss et al., 2005).
Further analysis was able to show that, out of these cis-acting eQTLs 64%
were true positives (Doss et al., 2005). The remaining 36% genes that could
"Qt be confirmed as truly cis-acting may arise due to trans-regulatory factors
IOcated in physical proximity of the gene. Subsequent studies have addressed
e(DTLs lying within 5cM of the gene as local acting eQTLs, which could be cis-

0r trans-acting (West et al., 2007).

70

The eQTL identified for CBF2 expression on chromosome 4 was
supported by a high LOD score (Figures 2.8 and 2.15), which is regarded a
characteristic of cis-eQTLs. An eQTL at the same location, with similar high
LOD score, was also identified for CBF3 expression (Figures 2.9 and 2.16).
However, analysis of CBF1 expression did not reveal an eQTL at this position
(Figures 2.10 and 2.14). The CBF genes are located adjacent to each other
on chromosome 4 in the order CBF 1, 3 and 2. These transcriptional activators
display high sequence conservation (CBF1 and 2, 86% identity; CBF1 and 3,
86% identity; CBF2 and 3, 87% identity) (Liu et al., 1998) and a significant
overlap in their target genes, suggesting functional redundancy (Gilmour et
al., 2004).

When comparing the sequence of the CBF locus from Cvi-0 and Ler,
the CBF1 promoter has been found to exhibit highest conservation with only
seven SNPs in 1kb region upstream of the transcriptional start site, whereas
CBF2 and 3 promoters were found to be more divergent with a 1.6 kb deletion
and 23 SNPs, respectively (Alonso-Blanco et al., 2005). A study in mice
comparing the frequency and density of SNPs in the promoters of genes
identifying cis-eQTLs and those not regulated by cis-eQTLs (total of 4752

genes) found that the cis-eQTL genes have a significantly higher frequency of
Cis-SNPs compared to non-cis-eQTL genes. Furthermore, these SNPs are
present in the phylogenetically non-conserved regions. These SNPs were
also found in regions predicted to affect predicted transcription factor binding
sites, a likely consequence of the higher cis-SNPs density in these genes
(GuhaThakurta et al., 2006). In studies using microarrays, sequence variation

in coding regions can manifest falsely as cis-acting eQTLs since the probe is

71

usually based on one parent and has a considerable influence on the
hybridization efficiency with the other (Alberts et al., 2007). However, this
possibility can be ruled out in the present analysis as the coding regions of
CBF1-3 genes were determined and primers were designed based on
conserved sequence.

It has also been reported that Cvi-0 shows reduced CBF2 transcript
accumulation and increased levels of CBF1 and 3 transcripts (Alonso-Blanco
et al., 2005). Furthermore, CBF2 has been suggested to have a negative
regulatory effect on CBF3 (Alonso-Blanco et al., 2005). In a separate study,
analyzing publicly available microarray data for low temperature response of
wild type Arabidopisis and cold tolerance mutants, suggested CBF2 and 3 to
negatively regulate each other (Benedict et al., 2006).

The eQTL on chromosome 4 identified for CBF2 could be attributed to
the higher sequence divergence of its promoter, rendering it less active in Cvi-
0. Therefore, supporting the possibility of this being a cis-acting eQTL for
CBF2 expression. If CBF2 exerts a negative regulatory effect on CBF3
expression, the eQTL identified for CBF3 on chromosome 4 could arise due to
altered levels of CBF2 transcript, therefore making it a local-acting eQTL.
CBF2 has also been suggested to have a negative regulatory effect on CBF1
expression; however, an eQTL overlying CBF2 was not identified for CBF1
expression. This suggests involvement of distinct regulatory mechanisms for
low temperature responsive expression of the different members of CBF gene
family. Analysis of RIL subpopulations with CBF locus held constant, as either
Ler or Cvi-0, resulted in disappearance of the eQTL on chromosome 4, thus

suggesting it to be a local-acting eQTL.

72

2.3.2 Identification of Additional eQTLs Using Actin mRNA for
Normalization

As discussed in Section 2.2.3.3, the selection of internal reference
genes for normalization of expression depends on the system under study.
Although 18$ rRNA is routinely used as a reference, it is not ideal for
normalization of transcripts that are much lower in abundance, such as the
CBF mRNAs. It was therefore important to use another reference gene for this
analysis and the Act1 gene was selected as a second normalization control.
The validity of Act1 for use as the second reference gene was established in
two ways. Firstly, it was demonstrated that the Act1 gene did not show
differential expression at low temperature (Figure 2.14). Secondly, eQTL
analysis was performed to ascertain that this gene did not show heritable
variation in expression in the RIL population (Figure 2.15).

Consistent with the 188 rRNA normalization, CBF2 and 3 expression at
low temperature identified common eQTLs on chromosome 4 overlapping
with the genomic location of the CBF locus (Figures 2.8, 2.9, 2.17, and 2.18).
Furthermore, CBF1 expression analysis did not reveal any eQTL at
chromosome 4 (Figure 2.16). Moreover, use of A011 as the internal reference
to quantify CBF transcript levels in the segregating population further
expanded the search for regulatory loci controlling CBF1 and 2 expression.
Two adjoining trans-acting eQTLs at overlapping genomic locations on
chromosome 1 were identified for both CBF1 and 2 genes (Figures 2.16 and
2.17).

Thus, natural genetic variation found in CBF expression at low

temperatures was successfully used to identify regions in the Arabidopsis

73

genome (eQTLs), characterization of which will lead to identification of genes
whose products are involved in regulating low temperature responsive
transcription. The independent detection of freezing tolerance QTLs in
chromosomal regions that were found to contain eQTLs for CBF expression
substantiates the evidence for a regulatory role of these eQTLs during cold
acclimation (Alonso-Blanco et al., 2005). In addition to the eQTL on
chromosome 4, other freezing tolerance QTLs were found to overlap with
eQTLs identified for regulation of CBF expression in the present study.
Overlap was observed between a short-day photoperiod specific freezing
tolerance QTL and eQTLs for CBF 1, 2, and 3 expression found through this
analysis. The finding that none of these eQTLs were detected for regulation of
a gene non-responsive to low temperature, Act1, provides further support for
the association of these eQTLs with cold acclimation.

Future fine mapping studies in the future would lead to identification of
the sequence variations underlying these eQTLs and thus identify the genes
responsible for CBF transcript regulation at low temperatures. Regulatory
factors with pleiotropic effects on the low temperature response of plants may
be uncovered in these regions. These may include potential thermosensor(s),
signal transduction components, transcriptional activators or repressors,

regulators of transcript stability, etc.

2.3.3. Genes Located in Regions Showing eQTLs
Genes previously reported to show co-regulation with cold tolerance
(identified through mutational analysis) and genes analyzed in this study were

mapped on the chromosomes to compare their location with the eQTLs

74

observed in the current study (Figure 2.19). The ICE 1 gene, located in the
upper arm of chromosome 1, was found to be coincident with an eQTL
identified for regulation of CBF3 expression. This eQTL however, was not
identified for CBF1 or 2 expression, which is in agreement with the finding that
icef mutants are altered in expression of CBF3, but not in CBF1 and 2
expression (Chinnusamy et al., 2003). This suggests that this eQTL regulating
CBF3 may result from variation in the ICE 1 gene.

Although the H089 gene, shown to have an effect on freezing
tolerance in Arabidopsis, is co-localized with the eQTL for CBF2 expression
on chromosome 2, it can be ruled out as a candidate since hosQ mutants do
not display altered CBF gene expression (Zhu et al., 2004).

There were no other genes, previously identified to play a role in low
temperature stress tolerance, found to overlap in chromosomal location with
the eQTLs identified in this study. Thus, the eQTLs will lead to novel genes

regulating the low temperature responsive expression of CBF1-3 genes.

75

Chromosomes

* «~H089 '
__ICE1
"‘RAV1
_Thau-like
_Gl
~Dof1-Iike “‘3'
Gene
Names ‘
BF1-3
NAC027
CCCH-type
Zn Finger
- H081 Protein
»~LOS4

A—ZA'I'1 2

~ERF

—FRY1I
H082

Figure 2.19. Chromosomal locations of genes known to be linked with
cold tolerance, identified either through mutational studies or

transcriptional profiling.

76

2.4 Material and Methods

2.4.1 Plant Growth and Treatments

Seeds were surface sterilized by incubating with 1 ml of 50% bleach +
0.1% Triton X-100 (Sigma-Aldrich, St Louis, MO) solution with shaking for 7
mins followed by four washes with sterile distilled water. They were re-
suspended in sterile 0.1% agar and cold hardened at 4°C for 3-4 days
following which they were spread (density of approx 120-150 seeds/100 x 20
mm Petri plate) on solid Gamborg’s B5 (Caisson Laboratories, Logan, UT) or
MS nutrient medium (Sigma-Aldrich, St Louis, MO) with 0.8% (w/v) phyta
agar (lnvitrogen, Carlsbad, CA). Seeds on nutrient media were grown in
controlled environment chambers (Conviron, Winnipeg, Canada) maintained
at 22°C illuminated with cool-white fluorescent lights at the intensity of approx
100-150 pmol m'2 s". For the gradual temperature march down experiments,
10 to 12-day-old seedlings grown in continuous light were moved to controlled
environment chambers maintained at 21°C with cool-white fluorescent light.
The temperature of the chamber was decreased by 2°C, at a rate of
0.5°C/15mins, after every 4h. Tissue was harvested in tubes maintained on a
liquid nitrogen bath after 1h and 4h at 17°C, 15°C, 13°C, 11°C, 9°C, 7°C, 5°C,
3°C, and 1°C and was immediately transferred to -80°C. For the cold
treatments involving drop to specific temperatures, 100-150 seedlings of
ecotype were grown as described above. At 10-12 days of age, the seedlings
were moved to controlled environment chambers maintained at either 10°C,
8°C, 6°C, 4°C, 2°C or 0°C with cool-white fluorescent lights. After 2h

incubation at the given temperature, tissue was harvested as described

77

 

 

above. For the analysis of kinetics of gene expression, time course treatment
was performed at 4°C. Seedlings grown as described above were moved to
controlled environment chambers maintained at 4°C, with cool-white
fluorescent lights. After selected time intervals, a plate of seedlings for each
ecotype was removed from the cold chamber and tissue was harvested in

liquid nitrogen.

2.4.2 RNA Isolation

Total RNA was extracted from tissue using Qiagen RNeasy Mini RNA
extraction kits (Qiagen lnc., Valencia, CA) by following a modified protocol of
the manufacturer’s instructions. Briefly, tissue was ground to a fine powder
using liquid nitrogen and 900 pL to 1 mL of BLT extraction buffer (+10 pL B-
mercaptoethanonL buffer) was added to it. 700 piL of this sample suspended
in the RLT buffer was pipetted into the Qiashredder columns provided in the
kit and centrifuged at 14,000 rpm for 2 minutes. The clear supernatant from
the flow through was transferred to fresh tubes and 1/2 volume of 100%
ethanol was added to the sample. Following a thorough mixing of ethanol with
the sample, 700 pL was transferred to the Qiagen spin columns and
centrifuged for 15-20 sec. The flow-through was discarded and the column
was washed with 700 pL RW1 buffer by centrifuging for 15-20 sec. This was
followed by two washes with 500 ptL of RPE buffer. The RNA was finally

eluted in 60pLs of RNase free H20.

78

2.4.3 Northern Blot Analysis

10 pg of total RNA was fractionated on a denaturing formaldehyde
agarose gel and transferred to a positively charged nylon membrane (Hybond
H*, Amersham Pharmacia Biotech, CA) using the downward transfer method.
Blots were prehybridized for 2 h with prehybridization buffer (5X SSC buffer,
10X Denhardt’s reagent, 0.1% SDS (WM, 0. 1M KPO4 pH 6.8, 100 pg/ml
denatured herring sperm DNA) and hybridized for 16 h with radioactive probe
in hybridization buffer (5X SSC buffer, 10X Denhardt’s reagent, 0.1M KPO4
pH 6.8, 100 pg/ml denatured herring sperm DNA, 10% dextran sulfate (WM,
30% (v/v) deionized formamide). 30-50 ngs of probe DNA was labeled with
dCT32P using the Random Priming Kit (lnvitrogen lnc., Carlsbad, CA) with
manufacturer’s instructions. Unincorporated nucleotides were removed by
passing the labeling mix through a sephadex G-50 column. Hybridization of
membranes with the labeled probe was performed in hybridization chambers
maintained at 49°C. Membranes were washed three times with 1XSSC and
0.1% SDS (w/v) solution at 55’C. Phospho-screens were exposed overnight
to the radioactively labeled membranes and the signal was developed using

the Phospho lmager.

2.4.4 RT-PCR

2 pg of total RNA was reverse transcribed using random hexamers with
the Superscript II Reverse Transcriptase Kit (lnvitrogen lnc., Carlsbad, CA)
following manufacturer’s instructions. The cDNA was diluted to a total volume
of 200 uls with sterile distilled water. Equilibration was performed using 188

rRNA transcript levels, as this gene does not respond to low temperature.

79

 

 

 

Primers were designed for the RT-PCR of all genes using the Primer 3
available at http://www.genome.wi.mit.edu/cgibin/primer/primer3_www.cgi.
For 188 rRNA equilibration, fonrvard (5’ GAO GAA CAA CTG CGA AAG CAT
3’) and reverse (5’CCC CCA ACT TTC GTT C'l‘l' GA3’) primers were used to
obtain similar amplicon levels for all samples after 18 cycles. 2 pls of the
equilibrated cDNA was then used for PCR using Arabidopsis Acti forward (5’
GGT CGT ACT ACC GGT ATT GTG CT 3’) and reverse (5’ TGA CAA 'l'l'T
CAC GCT CTG CT 3’) primers using 28 cycles. For PCR, each sample was

tested in technical triplicates.

2.4.5 Taqman Real-Time RT-PCR for Quantification of Molecular
Phenotype

Gene expression levels for CBF 1, 2, and 3 genes were measured in
the RlLs following exposure of 10-day old seedlings to 4°C for 2h. Expression
values were obtained for two independent biological experiments. For each
experiment, the RNA used for measurement of transcript levels was obtained
from a pool of 70-100 seedlings of individual RIL. After removing
contaminating genomic DNA using Turbo DNA free kit from Ambion (ABI, CA)
treatment, 10 pg of total RNA was used for total cDNA synthesis (Archive
cDNA Synthesis Kit). Random primers were used for cDNA synthesis to allow
detection of 188 rRNA, which was used as the normalization control. The
cDNA synthesis was performed for RNA samples from both biological
replicates using reaction components from a common master mix to perform
the reactions in parallel. This helped minimize differences in the efficiency of

the cDNA synthesis reaction in different samples. Similarly, components of

80

the qPCR step were also prepared as a common master mix to quantify

samples from both biological replicates in parallel. The table below shows the

primer and probe sequences that were designed using the Primer Express®

(ABI) program and were subsequently used for quantification of the respective

 

 

genes.
Forward Primer Reverse Primer
Gene Sequence Probe Sequence 8 9%“ c e
188 rRNA 5’ GAO GAA CAA CTG 5' FAM-TGC CAA GGA 5’ CCC CCA ACT
(At2916590) CGA AAG CAT 3' TGT T'l'l' C 3’ TI'C GTT CTT GA 3'
5' AAC GGT 'l'l'A
Rav1 5' GGC GGA G60 5' PAM-AAA GCG GTA
(At1g13260) ACT GTI' TGA 3' ACC CCA AG 3' GGT T0039“ CGT
CBF2 5' GGA ch TCA TGG 5' FAM-AGA CCT TGG STKugAGEaCTICI-TA
(At4g25470) TCT TGA CAT G 3' TGG AGG C 3' TG 3.
Zarrz 5' 060 ATC C'l'l' GT0 5' FAM-TGG AGT TTC Cichégflchgr
(At5959820) CCA TAT GTG 3' CGA TGG GA 3' 3.

 

Primer and probe mixes for the remaining genes were ordered from

ABI (Applied Biosystems Inc., CA). Due to company policy only the probe

sequences for these genes were provided, which are shown in the table

below.

 

Gene

Probe Sequence

 

ACT1
Thaumatin Family Protein
(At1920030)

Dof-1 Domain Protein
(A11 926790)

CBF3
(At4925480)

CBF1
(A14925490)

CCCH-type Protein
(At4929190)

ERF Subfamily 133
(At5961600)

5’ FAM-TGG TCG TAC TAC CGG TAT TGT GCT C 3’

5’ FAM-CCA ACA CCA GCC TAA AAT CAG CTG A 3’

5’ FAM-GAC CAC ACA TGA TGA ATA ACC AAT C 3’

5’ FAM-TGA AGT CGA CGG CGA TGA TGA CGA C 3’

5’ FAM-CGC CGC CGT CTG TTC AAT GGA ATC A 3’

5’ FAM-AAT TAT GGA 'lTl' TGG GGC CTG ATT T 3'

5’ FAM-ACG TTT TCA CTI' GCG TCA CTC TCA T 3’

 

81

2.4.6 MADs calculation
Median Absolute Deviation (MAD), an alternate method of determining

the variability of data, is defined as

N

MAD = Constant * Median ( l vi - Y | )

N

where Y is the median of the data. The value of constant is 1.4826. The MAD
value was calculated for each expression data value. The median of these
values was then multiplied with ( I Y. -— I)?" ), which was calculated for the
expression value of each RIL. RlLs showing MAD values >+3 and <-3 were
considered as outliers and were removed from the dataset. These RlLs were
represented as missing values in the phenotype data used for Composite

Interval Analysis (CIM) using the QTL Cartographer package.

2.4.7 eQTL Mapping

Information for genotype of each RIL was downloaded from
www.arabigopsis.org and was prepared for QTL cartographer by manually
eliminating the redundant marker information along the length of the
chromosomes. MAD values were calculated for the expression levels of
genes for 155 RlLs and those showing MAD values >+4 and <-4 were
considered as outliers. These outliers were removed from the dataset to
reduce noise and it was re-analyzed using QTL cartographer. In the CIM
analysis, the walk speed, which is a parameter that determines the interval
along the genetic map at which the QTL calculations are performed, was set
at 20M. Permutation analysis using 1000 permutations was performed to

determine the critical LOD threshold. A confidence level of 95% was

82

determined for a LOD score of 2.4 and thus the critical threshold levels for

determining the eQTLs was set at the LOD score of 2.5.

2.4.8 Sub-grouping RILs Based on Composition of CBF Locus

The composition of the CBF locus in the lines was determined first by
examining the marker data available for these lines on the TAIR website. Two
chromosomal markers, 94539 and GB.490 C, flanking the CBF locus were
selected and each line was scored for the presence of Cvi-0 or Ler marker on
these two positions. The genetic positions of 94539 and GB.490 C were 44
CM and 73 CM respectively. If these flanking markers were Cvi-0, the CBF
locus was considered as Cvi-0 and if both the markers were Ler, it was
considered as Ler. Using this method, 114 lines were determined to belong to
one group or the other, given the caveat that error in sorting the individuals
can arise if two recombination events occurred within the region being
examined such that the CBF locus was not from the same parent as the
flanking markers. For 48 lines that displayed the presence of different parental
markers in the flanking positions, PCR was performed to determine the
composition of the locus. Two primer pairs were used for the differential PCR,
both based in the promoter region of CBF2. The first primer pair (sequence
generously provided by J Salinas) amplified a ~2 kb fragment from the
promoter of Ler and ~200 bp fragment from the Cvi-0 promoter. The second
primer pair, also based in the promoter region of CBF2, amplified a ~500 bp
region from Ler and did not amplify anything from the Cvi-0 promoter. Thus,

lines showing a 2 kb fragment and a 500 bp fragment were genotyped as Ler

83

at the CBF locus and those containing a 200 bp fragment were genotyped as

CVi-O.

84

Chapter 3
eQTLs Regulating Expression of Genes Induced Early During Cold

Acclimation

3.1 Introduction

In addition to the CBF1-3 genes, hundreds of other genes are induced
in response to low temperature (Fowler and Thomashow, 2002). Many of the
early expressed genes encode transcription factors, suggesting the existence
of multiple regulatory networks at low temperatures (Fowler and Thomashow,
2002). However, apart from the CBF regulon, which is activated by the CBF1-
3 genes, the regulatory pathways active during cold acclimation are poorly
understood.

Multiple functionally related genes are often co-regulated through a
common regulatory gene. For instance, the COR genes, which are induced
with similar kinetics at low temperature, are collectively activated by the CBF
genes. In this instance, mutations affecting the CBF genes will affect
expression of multiple COFl genes. Analogously, allelic differences in a
common regulator will affect the expression 01 multiple genes, leading to co-
localization of their eQTLs to this regulator. Thus analysis of multiple low
temperature-regulated genes for eQTLs will provide the opportunity to identify
common regulatory pathways.

Many early expressed genes at low temperature display temporal
expression profiles similar to the CBF1-3 genes. Furthermore, some of these

genes, including CBF1-3, are also induced in response to mechanical

85

agitation and cycloheximide treatment, suggesting shared regulatory
pathways (Vogel et al., 2005). It was sought to determine if, using eQTL
analysis, common trans-acting eQTLs could be identified for this group of low
temperature-induced genes. Genes showing common eQTLs could possibly
constitute branches of the same regulatory pathway. Finally, to broaden the
scope of the study and include different regulatory pathways, genes
suspected not to be co-regulated were also included for identification of

eQTLs.

86

3.2 Results

In order to identify additional genetic loci containing regulators of low
temperature responsive transcription, six candidates were chosen for eQTL
mapping. The Lerini RIL population was evaluated for transcript
accumulation of Rav1 (Atlg13260), Za112 (At5959820), Ethylene Response
Factor (ERF; At5961600), CCCH-type Zn-Finger transcription factor
(At4929190), Dot-1 domain containing transcription factor genes (At1926790),
and Thaumatin-like protein encoding gene (At1920030).

The expression profiles of these genes at low temperature were
determined through investigation of publicly available microarray data
(compiled from several research groups) accessible through the
Genevestigator database (nt_tp://www.qenevestiqator.ethz.ch/at). Examination
of AtGenExpress data for low temperature stress treatments of Arabidopsis
plants showed ERF, Rav1, and 23112 to be induced early in response to low
temperature, with peak levels of expression at 1h (ERF) or 3h (Rav1 and
23112) and subsequent decline in expression by 24h (Figure 3.1). These
profiles were similar to the CBF1-3 genes, which also peaked after 3h of
exposure to low temperature (Figure 3.2). In addition, all of the above genes
were induced by cycloheximide treatment (Table 3.1) and mechanical
agitation (Vogel et al., 2005) suggesting shared regulatory pathways.
Identification of common eQTLs for regulation of these genes would provide
additional evidence to support existence of shared pathways.

Analysis of the Genevestigator data also revealed a group of low

temperature-induced genes that were not induced by cycloheximide

87

Log, of
signal

Intenslty

 

 

16

14

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12 c
J ’ \ \
a o
’ l C . Q
o I
’ i V 0’ ’
C
10 a ' fl
0 / ‘
8 ‘ l ‘I I - I%
Q
0’ ‘ e
1
6
4
—0 - Zat12
—l—Rav1
2 - n - ERF —
0 s
on 0.511 111 an 6h 12h 24h
Time of exposure at 4°C

Figure 3.1. Expression profiles of Zat12, Rav1, and ERF genes at low

temperature. Digital Northern from Genevestigator database showing

expression levels of ERF, Rav1, and Zat12 genes after 0.5, 1h, 3h, 6h, 12h,

and 24h at 4°C in aerial parts of seedlings of Col-0, measured using

microarrays. The y-axis displays the 1092 of signal intensities obtained for the

gene.

88

 

 

 

16

14

 

12

 

 

10

 

Log, of

 

 

 

 

 

 

 

 

 

 

signal
intenslty '
6 .__‘
/
’ I
—O-CBF1
—I—CBF2
2 - t - CBF3 T
0 1 . . . . .
0h 0.5n 111 an 6h 12h 24h

 

 

 

Time of exposure at 4°C

Figure 3.2. Expression profiles of CBF1, CBF2, and CBF3 genes at low
temperature. Digital Northern from Genevestigator database showing
expression levels of ERF, Rav1, and Zat12 genes after 0.5, 1h, 3h, 6h, 12h,
and 24h at 4°C in aerial parts of seedlings of Col-0, measured using
microarrays. The y-axis displays the logz of signal intensities obtained for the

gene.

89

 

Cold Treatment CYCIOhGXimide

 

AGI Numbers Gene Name . Treatment

0092 Of srgnal) (I092 of signal)
At5959820 Zat12 10.913 91.445
At5961600 ERF 2.054 78.742
At1g26790 Dof1-domain containing 5.853 -0.003
At4g29190 CCCH-type Zn Finger 7.115 -0.611
At1g13260 Rav1 6.578 6.872
At1g20030 Thaumatin-like 14.382 -0.804
At4925470 CBF2 144.658 7.997
At4925490 CBF1 142.391 8.870
At4925480 CBF3 78.764 1.721

 

Table 3.1. Expression of selected genes in response to cycloheximide
treatment and cold. Heat map from Genvestigator database showing
response of Flavf (Atfg13260), Zat12 (At5959820), Ethylene Response
Factor (At5g61600), CCCH-type Zn Finger transcription factor (At4929190),
Dof-1 domain containing transcription factor genes (At1926790), and
Thaumatin-like protein encoding gene (At1920030) upon cycloheximide and

cold treatment.

90

treatment, suggesting distinct regulatory pathways for these genes. This
group included Thaumatin-like protein, CCCH-type Zn-Finger transcription
factor, and Dof-1 domain containing transcription factor genes (Table 3.1).
The gene encoding a Dof-1 domain containing protein was an
interesting candidate for analysis because of its potential to be involved in
conveying the low temperature signature to proteins or DNA. Dof (DNA-
binding with one finger) domain proteins are plant-specific transcription factors
with a highly conserved DNA-binding domain containing a single Cz-Cz zinc
finger. This Dof domain has been shown to possess DNA-binding as well as
protein-protein interaction abilities (Yanagisawa, 2004). Their ability to interact
with protein as well as DNA gives them the ability to act at the intermediate
step of conveying the signal from a protein to DNA. Furthermore, of all the
abiotic stress conditions for which data was available at the Genevestigator
site (http://wwwgenevestigator.ethz.ch/at), the DOM domain-containing gene
was induced only by low temperature and hydrogen peroxide treatment
(Genevestigator), suggesting specific functional role at low temperature and
oxidative stress. The other candidate, the Thaumatin like protein, also
provides an interesting low temperature-regulated gene as it is predicted to be
a GPl-anchored protein based on sequence homology data (TAIR). Both
these genes are of further interest as they do not encode usual transcription
factors like the other candidates included in this analysis, thus furthering the
objective of broadening the investigation of low temperature-induced
regulatory pathways. The CCCH-type Zn-Finger transcription factor-encoding
gene could be involved in regulation of a subset of genes that are differentially

expressed later in response to low temperature.

91

If they were indeed regulated through distinct pathways, eQTL
mapping for these genes would potentially expand the search for loci involved
in regulation of transcription at low temperature. Using the Genevestigator
database, expression profiles of these genes were found to be similar to
Rav1, Zat12, and ERF genes, showing early induction in response to low
temperature treatment (Figure 3.3). Thus the candidates represented groups
of genes that were possibly regulated through common and distinct pathways

acting at low temperatures.

3.2.1 eQTL mapping

Using the Taqman® real-time RT-PCR method, transcript levels for
Rav1, Zat12, ERF, Thaumatin-like protein, CCCH—type Zn-Finger transcription
factor, and Dof-1 domain containing transcription factor genes were estimated
in 155 RlLs that were treated at 4°C for 2h. The amplification profiles
generated during qPCR for different transcripts are shown in Figure 3.4. 2”“
values were determined using 188 rRNA and Act1 as the internal reference
genes (see Materials and Methods).

Transcript accumulation for the six selected low temperature
responsive genes, measured through independent normalizations with 188
rRNA and Act1, was analyzed using CIM to identify eQTLs. Analysis of the
188 rRNA normalized expression data led to identification of eQTLs for two
genes that are not induced by cycloheximide treatment, CCCH-type Zn-Finger
transcription factor and the Thaumatin-like gene (Figures 3.5 and 3.6,

respectively). However, using this normalization, no eQTLs were detected for

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

16
14 ’
0 Q
I
.0
12 ’
O
/ i
o ..
6‘
O
10 9 / / \
r ,q
1 z /’ \
z 'A \
Log; of 8 J’ — - ’1.__
slgnal / ' ’
lntenslty '
_/ '
3‘ ‘ . o - -‘
6 r ’
o
4 - O - Thau-like Gene _
—I—CCCH-Type Zn
Finger Gene
2 —s — Dof1-like Gene —
o i l i T l T
0h 0.5h 1 h 3h 6h 12h 24h
Time of exposure at 4°C

Figure 3.3. Expression profiles of Thau-Iike, CCCH-type Zn Finger
containing transcription factor, and the Dof1-domain containing genes
at low temperature. Digital Northern from Genevestigator database showing
expression levels of Thau-Iike, CCCH-type Zn Finger containing transcription
factor, and the Dof1-domain containing genes after 0.5, 1h, 3h, 6h, 12h, and
24h at 4°C in aerial parts of seedlings of Col-0, measured using microarrays.

The y-axis displays the 1092 of signal intensities obtained for the gene.

93

At 291 .CCCHT ZnFl er At§958920: Zat12

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

: c ,
o: p a:
< <I
At At13260:Rav1
T .
E 11: ..
q <
j :1: 1
Cycle—>
. At1 :T umatln-Il
1% 6E
<1 <1 _ ' .
' "ll
Cycle—>

 

 

 

 

Figure 3.4. Taqman

transcripts of indicated genes. The y-axis represents fluorescence emission

and the x-axis represents cycle numbers. The horizontal red line depicts the

threshold at which the C1 value for each amplification curve

94

 

real-time RT-PCR amplification profiles for

was determined.

 

 

 

coo:
hori;
leve
hum
Dane
110?:

hi
gh

Figure 3.5. eQTLs for regulation of CCCl-l-type Zn Finger containing
transcription factor expression. CIM analysis of 18S rRNA normalized
‘ transcript levels in RlLs treated at 4°C for 2h, using QTL Cartographer. In the
top panel, the y-axis represents the LCD scores and the x-axis represents the
five chromosomes of Arabidopsis. The area between each set of vertical
double lines represents the length in cMs of the indicated chromosome. The
horizontal red line represents the threshold LOD score for 95% confidence
level. The numbered values above the peaks represent the chromosome
number, the position in cMs, and the LOD score, respectively. The lower
panel shows a plot of estimated additive effects where a positive value
indicates that the allele from the high expressing parent is associated with

higher levels of transcript, which is Ler.

95

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5‘ it
i3
5? 60‘
v AN'
3? 8e“
N 0:58
to“ <0 “-
v :05;
,, a
\(4;{_
A‘ I
s:
(1‘ 5
V'
E00
\
co 2‘
N
i 4;:
U)
E“ ‘5
;l
>
o
o it; E
.3. I IN I J l g V
1.0 v co N 1- :: coooqooco
001-me

Figure 3.5. eQTLs for regulation of CCCH-type Zn Finger containing

transcription factor expression.

 

 

Fgu
anal
2h, l

530'.

the i

Figure 3.6. eQTLs for regulation of Thaumatin-like gene expression. CIM
analysis of 18S rRNA normalized transcript levels in RlLs treated at 4°C for
2h, using QTL Cartographer. In the top panel, the y-axis represents the LCD
scores and the x-axis represents the five chromosomes of Arabidopsis. The
area between each set of vertical double lines represents the length in cMs of
the indicated chromosome. The horizontal red line represents the threshold
LOD score for 95% confidence level. The numbered values above the peaks
represent the chromosome number, the position in cMs, and the LOD score,
respectively. The lower panel shows a plot of estimated additive effects where
a positive value indicates that the allele from the high expressing parent is

associated with higher levels of transcript, which is Ler.

97

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. can-
- m5-
1 1f - 1.. Ki 1%, \K/ > .1 I... )7 8.0
1 Ed <. \ /\ <1 I\ alrds. J. .
. «5
289 /;/\/V\,. we ms 8. 1 o
i . F
1 m
om
. m
as 3 PV . e
:1. .2: .FV
80.: .1.

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.6. eQTLs for regulation of Thaumatin-like gene expression.

98

Figure 3.7. eQTLs for Dof1-domain containing gene expression. CIM
analysis of 188 rRNA normalized transcript levels in RILs treated at 4°C for
2h, using QTL Cartographer. In the top panel, the y-axis represents the LCD
scores and the x-axis represents the five chromosomes of Arabidopsis. The
area between each set of vertical double lines represents the length in cMs of
the indicated chromosome. The horizontal red line represents the threshold
LOD score for 95% confidence level. The lower panel shows a plot of
estimated additive effects where a positive value indicates that the allele from
the high expressing parent is associated with higher levels of transcript, which

is Ler.

99

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 ragl<i<is >571 six/cw E \1_>.S\.<\«J/
oi
Home <<< .8 -- on :3 -83 .
2...... a.
L
we
80.;

 

 

 

1 5. T
. med-

cod
mmd
5%

$6

N4

0...

YN

Figure 3.7. eQTLs for Dof1-domain containing gene expression.

100

the Dof1-like domain containing transcription factor, 28112, and ERF genes

(Figures 3.7, 3.9, and 3.10, respectively).

3.2.2. Identification of eQTLs using 188 rRNA as the Internal Reference
The eQTLs identified for CCCH-type Zn-Finger containing transcription
factor and Thaumatin-like gene co-localized with eQTLs identified for
regulation of CBF genes. Differential expression of the CCCH-type Zn-Finger
containing transcription factor gene at low temperature was found to be
regulated by eQTLs on chromosome 5 that were overlapping in position with
eQTLs identified for CBF 1, 2, and 3 expression (Figure 3.5). The first eQTL
on chromosome 5 identified for CCCH-type Zn-Finger containing transcription
factor gene, with a strong statistical support of 4.9 LODs, was common with
the eQTL detected for all three CBFs at the same location (see Chapter 2
section). The second eQTL identified with a LOD score of 2.8 was shared with
CBF2 and 3 genes. The location of eQTL identified for Thaumatin-like gene
on chromosome 1 overlapped with an eQTL identified for CBF3 expression
(Figure 3.6). Cvi-0 contributed the positive allele for this eQTL, which was

supported with 4.4 LODs.

3.2.3. Identification of eQTLs using Act1 as Internal Reference

Expression data obtained for the six genes using Act1 levels as internal
reference identified eQTLs for additional genes. This normalization uncovered
eQTLs for the Rav1, Zat12, ERF, and Dof1-like domain containing

transcription factor genes (Figures 3.11, 3.12, 3.13, and 3.14 respectively),

101

 

 

 

Ca'
x-a)
eac

chrc

.‘3

(O
()1

site

exp:

Figure 3.8. eQTLs for Rav1 expression. CIM analysis of 188 rFlNA
normalized transcript levels in RlLs treated at 4°C for 2h, using QTL
Cartographer. In the top panel, the y-axis represents the LOD scores and the
x-axis represents the five chromosomes of Arabidopsis. The area between
each set of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The lower panel shows a plot of estimated additive
effects where a positive value indicates that the allele from the high

expressing parent is associated with higher levels of transcript, which is Ler.

102

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m Dd.
- mm..-
A g p, - . 0? > 1 y 3&2 - t<><> 00.0
)2? R1). (5 4 /\,\< .3...
om“ t m
‘
.298. em on 2 SF . o
. 9o
. N;
. m;
r . 3..
Ga .1: .5 mm
a.“ .08 .5 111180.. ! m
alga h>flt

Figure 3.8. eQTLs for Rav1 expression.

103

 

 

T‘EOW

Cart

tit".

Figure 3.9. eQTLs for Zat12 expression. CIM analysis of 188 rRNA
normalized transcript levels in RlLs treated at 4°C for 2h, using QTL
Cartographer. In the top panel, the y-axis represents the LCD scores and the
x-axis represents the five chromosomes of Arabidopsis. The area between
each set of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive
value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, which is Ler.

104

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5.
1 8.—.I
n. K E p 2 a) s . \f > cod
1 8...
0a.. mud
I.
.289 3 me me our . o
g / \ E? fl
.. Yo
._ md

 

 

 

 

 

 

 

 

 

 

til
8.. .3: .5

 

 

 

 

 

 

 

 

 

 

 

\5

3; 5.4.0.5

 

 

80: m

 

._. 2N

Figure 3.9. eQTLs for Zat12 expression.

105

Figure 3.10. eQTLs for Ethylene Response Factor gene expression. CIM
analysis of 188 rRNA normalized transcript levels in RlLs treated at 4°C for
2h, using QTL Cartographer. In the top panel, the y-axis represents the LCD
scores and the x-axis represents the five chromosomes of Arabidopsis. The
area between each set of vertical double lines represents the length in cMs of
the indicated chromosome. The horizontal red line represents the threshold
LOD score for 95% confidence level. The numbered values above the peaks
represent the chromosome number, the position in cMs, and the LOD score,
respectively. The lower panel shows a plot of estimated additive effects where
a positive value indicates that the allele from the high expressing parent is

associated with higher levels of transcript, which is Ler.

106

 

 

Es 7

New?
-mts

 

it
7
.g-F'”

 

 

 

 

 

 

 

 

 

. oQo
lK)S(\/Q4K/g-m5o
-nwo

 

om,

 

 

 

 

 

1mm

 

 

 

 

 

 

 

 

(1

mm —

 

 

 

 

 

 

 

 

80.20.N

 

 

Figure 3.10. eQTLs for Ethylene Response Factor gene expression.

107

which were not identified using 188 rRNA for normalization of expression
data. Additional eQTLs were also identified for CCCH-type Zn-Finger
transcription factor gene (Figure 3.15). The expression of Thaumatin-like
gene, however, did not identify any eQTLs through this analysis (Figure 3.16).

Act1 normalized data revealed common eQTLs with similar patterns for
Rav1, Zat12, Dof1-like domain containing transcription factor, and CCCH-type
Zn-Finger transcription factor genes (Figures 3.11, 3.12, 3.14, and 3.15
respectively) on chromosome 1. Furthermore, eQTLs at the same location
were also identified for the regulation of CBF1 and 2 expression using Act1 as
the internal reference gene (see Chapter 2, Section 2.2.3.3).

An additional eQTL for CCCH-type Zn-Finger containing transcription
factor was identified on chromosome 1, followed by eQTLs regulating ERF
and the Dof1-like domain containing transcription factor (Figures 3.13, 3.14

and 3.15).

108

Figure 3.11. eQTLs for Rav1 expression. CIM analysis of Act1 normalized
transcript levels in RlLs treated at 4°C for 2h, using QTL Cartographer. In the
top panel, the y-axis represents the LOD scores and the x-axls represents the
five chromosomes of Arabidopsis. The area between each set of vertical
double lines represents the length in cMs of the indicated chromosome. The
horizontal red line represents the threshold LOD score for 95% confidence
level. The numbered values above the peaks represent the chromosome
number, the position in cMs, and the LOD score, respectively. The lower
panel shows a plot of estimated additive effects where a positive value
indicates that the allele from the high expressing parent is associated with

higher levels of transcript, which is Ler.

109

 

 

 

 

 

 

 

 

 

 

 

 

 

: ma
AWN 6.5. .5
_ m 6% .5
ad .35 .5

 

000..

 

 

 

3km gut

de-
- NYC-

ood

- hfld

vmd

md

w.—

YN

Figure 3.11. eQTLs for Rav1 expression.

110

Figure 3.12. eQTLs for Zat12 expression. CIM analysis of Act1 normalized
transcript levels in RlLs treated at 4°C for 2h, using QTL Cartographer. In the
top panel, the y-axis represents the LCD scores and the x-axis represents the
five chromosomes of Arabidopsis. The area between each set of vertical
double lines represents the length in cMs of the indicated chromosome. The
horizontal red line represents the threshold LOD score for 95% confidence
level. The numbered values above the peaks represent the chromosome
number, the position in cMs, and the LOD score, respectively. The lower
panel shows a plot of estimated additive effects where a positive value
indicates that the allele from the high expressing parent is associated with

higher levels of transcript, which is Ler.

111

 

1 50.0-

> x; > 00.0

 

 

 

 

mt—

:;a<-<a

 

 

 

 

 

CM
59 a on 2 t 11581 . o

. mo
. N.
. 3
. am

\Y md

3343 .5

 

 

 

 

 

 

 

_ .
a.“ .9915 80.; on

 

3km NEGN

112

Figure 3.12. eQTLs for Zat12 expression.

 

nu

‘\
(I)

Figure 3.13. eQTLs for Ethylene Response Factor gene expression. CIM
analysis of A011 normalized transcript levels in RlLs treated at 4°C for 2h,
using QTL Cartographer. In the top panel, the y-axis represents the LOD
scores and the x-axis represents the five chromosomes of Arabidopsis. The
area between each set of vertical double lines represents the length in cMs of
the indicated chromosome. The horizontal red line represents the threshold
LOD score for 95% confidence level. The numbered values above the peaks
represent the chromosome number, the position in cMs, and the LOD score,
respectively. The lower panel shows a plot of estimated additive effects where
a positive value indicates that the allele from the high expressing parent is

associated with higher levels of transcript, which is Ler.

113

 

no.0-

 

 

 

 

 

\ < c A 00.0
< a <12;

1 50.0

 

 

 

 

 

 

 

 

 

 

i
F

 

 

 

 

 

 

 

 

 

in $8 .F

<: ma
5_

as .58 .5

 

 

 

 

 

 

80.; m

 

 

114

Figure 3.13. eQTLs for Ethylene Response Factor gene expression.

Figure 3.14. eQTLs for CCCH-Type Zn Finger Transcription Factor gene
expression. CIM analysis of Act1 normalized CBF1 transcript levels in RlLs
treated at 4°C for 2h, using QTL Cartographer. In the top panel, the y-axis
represents the LOD scores and the x-axis represents the five chromosomes
of Arabidopsis. The area between each set of vertical double lines represents
the length in cMs of the indicated chromosome. The horizontal red line
represents the threshold LOD score for 95% confidence level. The numbered
values above the peaks represent the chromosome number, the position in
cMs, and the LOD score, respectively. The lower panel shows a plot of
estimated additive effects where a positive value indicates that the allele from
the high expressing parent is associated with higher levels of transcript, which

is Ler.

115

 

‘

l
:3
3 .

 

a
9
E

 

 

 

 

 

 

 

 

 

 

l
o F-
o

 

O

(U
N
O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S - LE... ; R
. e
. m
5
mam
—\ l m
.2 ad 6.? .5
. ._ . .
3m .35 .5 . v
_
A. £4.18. .5 804cm

 

 

116

Figure 3.14. eQTLs for CCCH-Type Zn Finger Transcription Factor gene

expression.

Figure 3.15. eQTLs for Dof1-domain containing gene expression. CIM
analysis of Act1 normalized transcript levels in RlLs treated at 4°C for 2h,
using QTL Cartographer. In the top panel, the y-axis represents the LCD
scores and the x-axis represents the five chromosomes of Arabidopsis. The
area between each set of vertical double lines represents the length in cMs of
the indicated chromosome. The horizontal red line represents the threshold
LOD score for 95% confidence level. The numbered values above the peaks
represent the chromosome number, the position in cMs, and the LOD score,
respectively. The lower panel shows a plot of estimated additive effects where
a positive value indicates that the allele from the high expressing parent is

associated with higher levels of transcript, which is Ler.

117

 

 

m

a. 15?-

 

 

 

 

 

 

 

 

3:6

 

xii/x -

0a.,

 

 

 

-. on

:MN .
234$

.mms .

_ S:

 

 

 

 

 

 

 

 

 

 

0.4 .08, .5

0N

 

A

 

 

Em 6.5 .5 u
,

............: .
_
:3 .52 .5

 

 

89 A

 

 

 

50.0.
v0.0-

00.0
3.0
no.0

N4

vN

0.0

06

Figure 3.15. eQTLs for Dof1-domain containing gene expression.

118

Figure 3.16. eQTLs for Thaumatin-family gene expression. CIM analysis of
Act1 normalized CBF1 transcript levels in RlLs treated at 4°C for 2h, using
QTL Cartographer. In the top panel, the y-axis represents the LOD scores and
the x-axis represents the five chromosomes of Arabidopsis. The area between
each set of vertical double lines represents the length in cMs of the indicated
chromosome. The horizontal red line represents the threshold LOD score for
95% confidence level. The numbered values above the peaks represent the
chromosome number, the position in cMs, and the LOD score, respectively.
The lower panel shows a plot of estimated additive effects where a positive
value indicates that the allele from the high expressing parent is associated

with higher levels of transcript, which is Ler.

119

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fl 2 ?<\>\ (>3... $6.32). 614 - “M.”
< /\ < (t E - 5o
o3 mod

5.38. .«m .. ow :2 :09 _. o

, 5 £35 2 5 572.? K
_ f .
5 ,
_ . ed
< .5 .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0N

 

 

 

80.. A

 

 

 

N...

0%

YN

Figure 3.16. eQTLs for Thaumatin-family gene expression.

120

3.3 Discussion

Cold acclimation involves a network of regulatory pathways that
ultimately result in developing tolerance to low temperatures. Genes
upregulated upon exposure of Arabidopsis to 4°C include several transcription
factors that reach peak expression levels between 1-4h and subsequently
decrease to lower levels (Vogel et al., 2005). A subset of these genes is also
induced by mechanical agitation and cycloheximide treatment, suggesting
shared regulatory pathways. This subset includes CBF1-3, Rav1, Zat12, and
ERF genes (Table 3.1). A different subset of genes, including Thaumatin-like
protein, CCCH-type Zn-Finger transcription factor, and Dof-1 domain-
containing transcription factor, are only induced in response to low
temperature but not to mechanical agitation or cycloheximide treatment
(Figure 3.3). The expression of the Dof—1 domain containing gene was found
to be unaffected by cycloheximide treatment, whereas Thaumatin-like and
CCCH-type Zn-Finger transcription factor genes were downregulated by the
treatment. These differences in expression present two interesting groups of
genes, the analysis of which may lead to identification of different pathways
responsive to low temperature.

Altogether, using 188 rRNA and A011 normalization for quantification of
phenotypic data, eQTLs were detected for all genes analyzed in this study
(Table 3.2). The eQTLs identified were all trans-acting and were supported by

Statistically significant LOD scores.

121

 

Gene

eQTL

 

 

 

 

 

 

 

 

 

18$ rRNA Act1

90 01-3
Rav1 (At1g13260)

90 61-4

90 C1-3
Zat1 2 (At5959820)

90 01-4

90 01-3
CCCH-Type Zn Finger TF gene 90 05-1

90 C14
(At4929190) eQ C5-2

9Q 6145

9Q C1-3
Dof1-domain containing gene

90 C1-4
(At1926790)

eQ 01-7
Thau-like gene (At1920030) 90 01-1
ERF gene (At5961600) 90 01-5

 

 

Table 3.2. List of eQTLs identified using 188 rRNA and Act1 normalized gene

expression data. The nomenclature is based on the location of the eQTL.

122

 

'1

3.3.1. Common eQTLs Regulating Multiple Low Temperature-Induced
Genes.

Several eQTLs were found to commonly regulate more than one gene.
Regulation of functionally related genes through a common gene is a widely
employed strategy (Keurentjes et al., 2007). Genes suggesting co-regulation,
based on similar induction responses to low temperature, mechanical
agitation, and cycloheximide treatment indeed identified common eQTLs on
chromosome 1. These genes included CBF1, CBF2, Rav1, and Zat12.
Identification of a common eQTL regulating the expression of these genes
strengthened the evidence for a shared regulatory pathway.

In addition, through identification of common eQTLs at this position, the
CCCH-type Zn-Finger transcription factor and Dof-i domain-containing
transcription factor genes were also found to share regulatory pathways with
the above group. Therefore, this analysis helps bin CBF1, CBF2, Rav1,
Zat12, CCCH-type Zn-Finger transcription factor and Dof-1 domain-containing
transcription factor genes into a common regulatory pathway. Co-localization
of eQTLs controlling more than one functionally related gene would indicate
coregulation by a common gene underlying the eQTL. It is possible that the
trans-regulatory factor underlying the common eQTL identified for the above
genes functions upstream of a branch point in the regulatory pathway, which
is manifested as the difference in response to mechanical agitation and
cycloheximide treatment. Therefore, the CCCH-type Zn-Finger transcription
factor and Dof-l domain-containing transcription factor genes may be
regulated through an arm downstream of the branch-point that is not

responsive to mechanical agitation and cycloheximide treatment.

123

 

The grouping of the CCCH-type Zn-Finger containing transcription
factor gene with this group is further supported by common eQTLs shared

with CBF1, 2, and 3 genes on chromosome 5.

3.3.2. Unique eQTLs Identified for Regulation of Specific Low
Temperature-Responsive Genes.

In addition to common eQTLs, the low temperature responsive genes
analyzed also identified distinct eQTLs. The ERF, CCCH-type Zn-Finger
containing transcription factor, Dof1-like domain containing transcription factor
and Thaumatin-like genes identified eQTLs that were not shared with any
other gene, indicating regulation by different pathways or different arms of a
branched pathway. The expression of the ERF, Dof1 domain-containing
transcription factor, and the CCCH-type Zn-Finger containing transcription
factor genes revealed distinct eQTLs on chromosome 1 (Figures 3.13, 3.14

and 3.15).

124

 

Chapter 4
Optimization of Taqman® Real Time RT-PCR for Molecular Trait

Measurement

4.1 introduction

QTL mapping experiments estimate the effect of genotype on
phenotypic variance observed in a population. Low environmental variance
improves such correlation of genotype with phenotypic data. Using an RIL
population allows reduction in environmental variances by allowing biological
replicates. In addition, technical variance may contribute to increased noise
that needs to be minimized to reveal the genetic variation in a population.

In order to improve the power of eQTL detection, accuracy and
reproducibility of RNA quantification is essential. This has been facilitated by
the development of highly sensitive and reproducible methods of RNA
quantification. Microarrays are employed when simultaneous analysis of
several hundred genes is desired and have been used extensively to identify
eQTLs for multiple genes (Carlborg et al., 2005; de Koning et al., 2005; Rosa
et al., 2006; Alberts et al., 2007; Shi et al., 2007). Real-time RT-PCR is a
quantitative method that provides an accurate means to determine the levels
of specific DNA and RNA templates for a smaller set of genes (Reist et al.,
2003; Huggett et al., 2005; Luu-The et al., 2005; Stahlberg et al., 2005;
Wacker and Godard, 2005; Botteldoorn et al., 2006; Wells, 2007). The latter
method provides a broad dynamic range—at least five orders of magnitude—

for detecting specific DNA or RNA templates. This technique can produce

125

 

data for high throughput samples with a short turn around time. The high
sensitivity of real-time RT-PCR allows detection of subtle differences in
transcript levels between samples, which may have arisen due to variation in
genetic regulators. However, the inherent sensitivity of this technique also
leads to the detection of even minute amounts of contaminating RNA (fewer
than 5 copies) (Valasek and Repa, 2005). Optimization of this technique is

therefore necessary to ensure reproducibility.

126

 

4.2 Results

4.2.1 Primer and Probe Optimization for Taqman® Real-time RT-PCR

Coding regions of CBF2, Fiav1, and Zat12 were PCR amplified from
Cvi-0 and Ler and were cloned to detect sequence variation, if any, between
the two genotypes (see Materials and Methods). Pairwise alignment of the
consensus sequence obtained for coding regions of these genes from Cvi-0
and Ler displayed high sequence identity. This indicated that common primers
could be used to detect transcripts arising from both genotypes. Candidate
primer and probe pairs for CBF2, Flavf and Zat12 were obtained using Primer
Express® (ABI). Using additional criteria, optimal primer and probe
sequences were selected.

CBF 1, 2, and 3 genes show a high degree of sequence identity often
resulting in cross-reactivity of primers (Figure 4.1). The 3’UTRs of genes often
show higher sequence variation among different members of a family and
were therefore chosen as the template for primer design. PCR was performed
using selected primers and plasmids containing coding regions of CBF1, 2,
and 3 genes (Figure 4.2). The primers were found to cause non-specific
amplification of CBF1 and 3 genes. Optimization experiments revealed the
requirement of higher annealing temperature to achieve specificity for CBF2
amplification (Figure 4.3). Additionally, these primers amplified a fragment of
expected size using coding region of CBF2 gene from Cvi-0 (Figure 4.1).
Since the CBF genes are intron-less it was not possible to design primers that

would distinguish between genomic and cDNA template and they detected an

127

 

 

55:853.“....52235622283.ceszscscgeesfi.Sausage:5358.......8355EEGEmcgésgasagug 88880
SEEM.5:8.5835.5Easea:aH.as.m.=5.czEBEESEBEEEBE::EEEEEEEEE5:558.Eazsasgcmumo
585855573225EgageH.5EEEEE5:52.558.889.85...E58:55.52:.SEEEGEEEEE.358%me
Eggs:EEISEE_aecwgcssssczgaaggESEEBQSUSEEE:3aEaESE:ESS5:§88§8:5555..Emu

30 03 0mm 80 30 08 8m 80 on... 03 emm 3m 8... «mm
a.caucuEEEE5053....»0..cacuccoauuuucafitcpuaucg50533305553.F3085... .3305053533333503.”:530358.00 mzmcmmcoo
0.2.5.5555:Sufiiccgucacuczgugcfihh$53805350:030....05000050....E01508caucoquEEEE5508025000—500509.50
05.05035zCtuugcuccuuwcozucg80855:...03:85.:5053053555050Eugpgozgcgsc0555.53$5.0:55m:500809.50

_ . 3.3 . Jazz—3.552.525.5554: . z. _ .2. =1... . 0...... . =0 FguucuucwucuuchcuhwpuucwcuhcwEUPth—m—mbbwuwuuu «END

- - h b n u .
u I I I I I I I

8m 3m 00... Sc 89 03. 0% SE 03. 00v ewe 09 has «an
czeuumuuwuwmcxmgflc.cuwccuumemp.0..cH:3...u000..$50..pumuouufiuuuuBcaBSCéS5520055000....50950005...uuuuuupuucuucuzgufimamcmmcoo

 

E.E0..5..8..Ea2505;00:5083530.808005500500qu5.505....hH.m.0.5m555503358058“...0000:58508505035050» . 1.50

 

can 08 one can 2....” 03 can 8m 3m 8..” 3... 03 EN .Em

 

005....aEm=0pumuucccuutficuugswgExam....0......cm.c5cccuuucucacuhuucgmzhm.SEcc__.500BEEfiBEuCuEuEEE .EE.......E..?FE.:.. L

0....5...:cm.$5..00....50555085053:00:00::cc”ESE.cuuucmcozguifiutfiuwacapouuuficccwccsgC385...oar—ESSSEQEESE WED

00m 00m .weu emu cum 3... SN 03 03 “Kn :3 0m.“ 3.. «ma
0Eaczuccdo.u«.8030ccmgr—80050330803395”... ES:Ecuduuaucogoflafiuou.05:00:550050820:55.5:Cdmdttcusccgz mzwcmwcoo
H:Cuccaccpwuhaugwguccg5.00005005880....0..0..a00CE:hEm.BESS—03:80:05;E58058:GEES“:5:033:50.uccscmlmo
H:..baccmczggag2.00::am.can—00030385035003”...5:9:EwESwucupaEuSEupwcwuctcmuufiugtgzz033:me52.EuBccuENkmO
0:h0ccm...c50H.a8008cccmccuuuhu—EEEH...aficucuufiutfihcwfiouwucliuhuuucgfi5...ququEpSEE“:CESSEESEEE... HBO

and 8a 3H :3 3 8 ..K 8 3. 3 8 a 3 «

 

 

Figure 4.1. Sequence alignment of CBF1-3 genes.

128

Ler CBF1 in plasmid
Ler CBF2 in plasmid
Ler CBF3 in plasmid
Cvi-0 CBF2 in plasmid

h
d)
U
‘0
(B
.J
D.
.0
O
O
‘-

 

Figure 4.2. Cross-reactivity of CBF2 primers with CBF1 and 3. PCR
showing amplification of pGEM T-easy plasmid containing CBF1, 2, and 3
coding regions from Ler and CBF2 coding regions from Cvi-0 using with
primers designed to be specific for CBF2 coding region. The PCR conditions
used were as follows: denaturation at 94°C for 5min, 28 cycles of 94°C for 30
sec, annealing at 58°C for 1min, extension at 72°C for 1min, and a final

extension at 72°C for 5 mins.

129

 

 

60°C 62°C 64°C 66°C 68°C 70°C

Annealing Annealing Annealing Annealing Annealing Annealing
222% 222% 222% 222% 222% 222%
@5880 eggo gggo égeeo EEEO Egeo
Dramas-9 $<um2 595952-0013132 3853-9 8:11:32
Saaag aaaE 000E Saaae aaag tang
Q222c 2228 2228 Q2228 222: 222:
.9“me swmm “NW‘” QFNWQ’ v—Nmm some
asas§ sssg asst gass§ ass; see?
.0002 0002 0002 -0002 0002 0002

 

.1 I ‘
---- --- - -- 1 - —

Figure 4.3. Gradient PCR to determine temperature for specific

amplification of CBF2. Gradient PCR performed with indicated annealing
temperatures to check primers for amplification of CBF1, 2 and 3 coding
regions and genomic DNA from Ler, respectively. The PCR conditions used
were as follows: denaturation at 94°C for 5 min, 28 cycles of 94°C for 30 sec,

annealing at 58°C for 1 min, extension at 72°C for 1 min, and a final extension

at 72°C for 5 min.

130

"'I

 

amplicon of expected size using genomic DNA from Cvi-0 and Ler (Figure
4.4).

Primers for the 18S rRNA gene (ATFiHN13, At2916590), which was
selected as the normalization control, were also designed using the Primer
Express® program. Primers and probes were further tested in the ABI Prism
7900 HT with serial dilution of plasmid DNA containing the target gene. The
real-time amplification plot of technical triplicates from a 1 in 10 dilution series
for gene 188 rRNA gene is shown in Figure 4.5. The amplification curve of
each sample is approximately 3 cycles apart as would be expected for the

dilution series.

4.2.2 Comparison of One- and Two-Step Method Towards Optimization
of Real-Time RT-PCR

cDNA synthesis and quantitative PCR (qPCR) in real time can be
performed in a single step or in two steps, where cDNA synthesis and qPCR
are performed separately (Wang et al., 2004; Psallida and Argyropoulos,
2005; Wacker and Godard, 2005). Both methods were tested in technical
triplicates to assess and compare reproducibility. Estimates of standard error
with multiple runs of the one-step method revealed huge variation between
different samples analyzed in the same light cycler run (Figure 4. 6). Two-step
RT-PCR, on the other hand, showed uniformity in the standard error for
different samples, indicating lower variation within a single light cycler run
(Figure 4.7). The variation between samples measured in the same light

cycler run has been previously reported to be lower than that observed for

131

 

 

 

WA
Ler Cvi-0 '7'"
l

 

 

Figure 4.4. PCR using genomic DNA of Ler and Cvi-0 as template. The
PCR conditions used were as follows: denaturation at 94°C for 5min, 28
cycles of 94°C for 30 sec, annealing at 58°C for 1min, extension at 72°C for

1min, and a final extension at 72°C for 5 mins.

132

Amplification Plot
1,000 E+1

 

1,000
1,000 E-1 P

1,000 E-2 '

ARn
g_/

 

1,000E-3 -‘- *1 ’.'

1,000 E-4

 

 

 

1,000 E-5 .. - 2-..- .- -. _. - _-.._-._.....:

J 510152025303 40

Cycle

Figure 4.5. Taqman real-time RT-PCR amplification profiles. Profiles
obtained for a 1 in 10 dilution series of the 188 rRNA gene cloned in a
plasmid. The x-axis represents fluorescence emission and the y-axis
represents cycle numbers. Each amplification curve represents a single well,
and every dilution is measured in triplicate wells. The horizontal red line
depicts the threshold at which the C. value for each amplification curve was

determined.

133

 

 

 

 

 

 

 

fi-

 

 

T

RIL

 

1P

 

 

 

 

 

 

 

 

 

 

 

 

 

 

888888880
V'V’COCO

N N 1- !- ln 0

CBF2 Transcript Level

Figure 4.6. Technical variation for one-step real time RT-PCR. One-step
real time RT-PCR estimates of CBF2 transcript levels with standard error

obtained from three technical replicates.

134

 

 

 

1 TL—

 

 

 

 

 

 

.Trf
RlLs

 

 

 

Iv-

 

 

 

 

 

1.

 

 

 

 

 

és‘o‘

P

150 r

O O O O
22 s 8 § :43 §
CBF2 Transcript Level
Figure 4.7. Technical variation for two-step real time RT-PCR. Two-step
real time RT-PCR estimates of CBF2 transcript levels with standard error
obtained from three technical replicates. The RlLs analyzed are different from

those depicted in Figure 4.6

135

different light cycler runs. Therefore, the two-step method revealed higher
accuracy and reproducibility. There were several other factors that needed
careful attention in order to optimize detection and reproducibility of template
RNA. Use of random primers for cDNA synthesis allows the use of 18S rRNA
and is more helpful for detection of mRNAs with significant secondary
structure. Variable enzymatic efficiencies between different samples during
the cDNA synthesis step were avoided by using a common reaction mix for all

the samples of two biological replicates.

4.2.3 Optimization to Remove DNA Contamination

Genomic DNA contamination poses a significant challenge during
analysis of transcripts arising from intron-less genes. The primers designed
for real-time RT—PCR were shown to not differentiate between genomic DNA
and cDNA templates and therefore could lead to inaccurate estimation of
gene expression. Genomic DNA contamination in the extracts leads to
inaccurate trait measurements.

PCR of total RNA from RlLs using primers designed for real-time qPCR
showed significant amplification without reverse transcription of these
samples, revealing genomic DNA contamination. Comparison of real-time
qPCR amplification profiles pre- and post-DNase 1 treatment for CBF2 and
188 rRNA revealed the effect of contamination on overestimation of transcript
levels (Figure 4.8a. and b., respectively). A prominent effect of genomic DNA
contamination was observed for 188 rRNA quantification, which amplified the

overall inaccuracy of target mRNA estimation due to the comparative method

136

 

 

Amplification Plot

 

 

 

 

 

 

 

 

 

 

1,000 E+1
1,000 _ .4»
9":- /'l/
1’m0 E‘1 Pro-0N8” I + JAY,
P fix"
/ "’
1.000 E-2 ,rL/
c A .' ‘ "
m 1 ,I r / "if .A, 4
.~- ' ‘- 1,. , 3f - , V,
<1 1.000 ES (,1 ,3 .31.- ,1 l
I i l: "r .1
\ "(,3 i! ,.
1,000 E-4 '
1,m0 E's A L
0 5 10 15 20 25 30 35 40
b- 1,000 E+1

1,000 /

P DNase i —>.-—'
1,000 E-1 '°' / f

I!

I
.21

Post-DNase I —>/ 5:5:

.1:

1,000 E-2 ,f/ ,3

.1
.Ill
‘/

 

ARn

a x l -
I. I I 5 v30- ‘ "-
1.000 E-3 i *1 k I 1. l 121' l
f _/' , ‘~ 1 ' 1

1.000 E-4 , .

 

 

 

 

 

1,000E-5 A ‘ 2
0510152025303540

Cycle
Figure 4.8. Effect of genomic DNA contamination on RNA estimation.

 

Taqman real-time RT—PCR amplification before profiles pre- and post-DNAse l
treatment of samples for a. CBF2 and b. 18S rRNA, in technical duplicates.
The y-axis represents fluorescence emission and the x--axis represents cycle
numbers. Each amplification curve represents a single well, and every RIL
sample is measured in triplicate wells. The horizontal red line depicts the

threshold at which the C. value for each amplification curve was determined.

137

 

a???

of RNA quantification. Different methods of DNase i treatment, including the
On-column® DNase I treatment kit from Qiagen, DNase l (Roche) treatment
followed by heat inactivation of enzyme at 65°C, and Turbo DNA free kit from
Ambion®, were compared to remove genomic DNA contamination without
compromising the RNA integrity and quality of sample for subsequent real-
time RT-PCR analysis. The Turbo DNA free kit from Ambion® (ABI) was
found to be most effective in removing genomic DNA contamination, following

which no amplification was observed for CBF2 and 18S rRNA in samples

before reverse transcription (Figure 4.9a. and b., respectively).

4.2.4 Identification of Outliers and Their Effect on eQTL Mapping
To quantify template levels following real-time qPCR, the cycle number
at which the reporter dye emission intensities rise above background noise,
called the threshold cycle (C.), is determined for the exponential phase of the
amplification plot. The amount of transcript present in each sample is
calculated using the mean C. values of technical triplicates of each cDNA to
account for technical error. AC. values are then determined as the mean of the
triplicate C. value of target transcript minus the mean of the triplicate C. value
for the internal reference transcript (18S rRNA or Act1 mRNA) and 2'ACt is

determined to represent normalized transcript levels in each sample.

initial analysis of 2'“ values of RlLs, performed using QTL

Cartographer, produced peculiar outputs where the program was unable to
separate the eQTL signal from the noise (Figure 4.10a). In order to reduce

noise in the expression data, it was important to determine the nature of the

noise being generated. A normal plot of gene expression values of the entire

138

 

 

 

 

Amplification Plot

 

a, 1,000 E+1

1,000

 

1.000 E-1

ARn

' 1‘ l I. ,t" \' : . .
. ‘ , . . .1
1,000 E-3 . " . ., .9 , , .l '
: " . :

1,000 E-4 ' 1 “

 

 

 

1,000 E-5

 

 

0 5 10 15 20 25 30 35 40

 

1,000 E+1

 

 

1,000

 

1,000 E-1

l/l"
1,000 E-2 , I,

ARn

1,000 E-3 ' '.

1,000 E4 ‘1‘

 

 

 

 

 

1,000 E-S

 

0510152025303540
Cycle

Figure 4.9. No-RT (reverse transcriptase) samples after DNase l
treatment. Taqman real-time RT-PCR amplification profiles for a. CBF2 and
b. 188 rRNA transcripts in total RNA samples without reverse transcription in
samples treated with Turbo DNA free kit (Ambion). The x-axis represents
fluorescence emission and the y-axis represents cycle numbers. The
horizontal red line depicts the threshold at which the C. value for each

amplification curve was determined.

139

 

ii i

4.8 .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 2 3 4 5
Chromosome Number

 

RILS

Figure 4.10. OTL Cartographer output with background noise in data.
a. eQTL profile generated by QTL Cartographer v2.0 using the complete

expression dataset.
b. Scatter plot of expression values for different RlLs. The x-axis depicts

the RiLs and the y-axis shows 2'“ values (measure of transcript

levels)

140

 

population, depicting the variability of data, revealed a long tail suggesting the
presence of outliers, which was also evident from a scatter plot of the values
(Figure 4.100). Data points showing very high expression levels were omitted
and subsequent analysis resulted in elimination of the false eQTL signal
(Figure 4.11a. and b.). This confirmed presence of outliers that were
introducing noise in the phenotype data.

In order to remove this noise, statistical analysis of expression data
was performed to identify the deviants. Standard deviation is the most
commonly used measure for assessing the variability of a dataset. it is
defined as the square root of variance, which is roughly the arithmetic
average of the squared distance from the mean. As expected, squaring the
distance from the mean would result in higher impact of values that are further
from the mean. Therefore, the standard deviation can be greatly affected by
outliers. To avoid the effect of outliers on the estimate of deviation, Median
Absolute Deviation (MAD), an alternate method of determining the variability
of data was used to detect outliers in the expression data. MAD is the

variation of the average absolute variation, which is based on the median

value and therefore is not affected by outliers.

141

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 [L000
1.6 .
1.2 .
D
O
-' 0.8
0.4 h M
, - 3 JW
1 2 3 4 5
Chromosome Number
b.
100 —
90 ®
80 -Q
70 ~7—
60 ‘_ .-—©__
6 50 '1’—‘ ..-_
‘i 40 fi‘ A
N LO)
30 22 -
20 O 9% ‘ .,_..-_ o
10 O Q C 0'“: 2 i A v 0
oz. 9 . c Q 0". .‘R.o‘g{§ J
0 50 100 150

 

 

 

RILS
Figure 4.11. Removal of outliers reduced noise in the dataset.
a. eQTL profile generated by QTL Cartographer v2.0 using the
expression data lacking information form three RlLs.
b. Scatter plot of expression values for different RlLs depicting, in circles,
the values that have been ommitted. The x-axis depicts the FilLs and

the y-axis shows 2'ACt values (measure of transcript levels).

142

4.3 Discussion

Taqman® real-time RT-PCR was an ideal method for measuring
molecular phenotype with the accuracy desired for eQTL analysis. it is a rapid
and high-throughput method that allows simultaneous analysis of hundreds of
samples at a cost lower than the alternate method using DNA microarrays. An
advantage of real time RT-PCR assay particularly significant for this study is
the wide dynamic range of this method, which permits comparison of samples
containing hugely different levels of RNA. Furthermore, the sensitivity of this
technique allows detection of poorly transcribed mRNA and subtle differences
in RNA levels amongst samples.

Several steps were taken to deal with the challenges that can arise
while using this method and to reduce variability of the experimental
procedure. Sequence comparison of CBF2, Rav1, and 23112 genes from the
parental accessions, Cvi-0 and Ler, was performed to ensure that common
primers and probes could be used to detect cDNAs arising from both
genotypes. This comparison revealed high sequence conservation of the
coding regions of these genes between the two accessions, allowing design
of primers and probes demonstrating equal efficiency of amplification and
detection of transcript arising from both genotypes. The specificity and optimal
conditions for the primers and probes were established using regular PCR as
well as real-time PCR in the ABI Prism 7900 light cycler.

The high sensitivity of this technique resulted in its susceptibility to
minute amounts of contaminating RNA that could be introduced through

aerosol contamination of the surrounding work area, amongst others. Another

143

 

 

source of contamination that is intrinsic to this technique is the carryover
contaminating DNA in the recombinant Taq polymerase commercial enzyme
preparations. inclusion of no-tempiate controls along with each run was
important to preclude contamination of samples. No-tempiate controls using
18S rRNA specific primers and probes revealed amplification signals that
were significantly above the background noise. Furthermore, detection of
such signals varied with the batch of enzyme used. This contamination was
found to be specifically problematic for 188 rRNA estimations. It was critical to
resolve this problem due to the impact of the internal reference on the overall
accuracy of transcript measures for the genes of interest. The contamination
problem evident in no-template controls was overcome by the use of two-step
method for real time RT-PCR.

The most critical consideration for accuracy of transcript measurement
is the quality of RNA sample. This relates to the integrity of the RNA (free of
degradation) and purity of sample. inhibitors that tend to co-purify during RNA
extraction and are carried over from reverse transcription of the sample can
have a significant impact on the method. To obtain high quality RNA,
extractions were performed using a commercially available silica column-
based method (RNeasy RNA extraction kit from Qiagen, lnc.). However, this
resulted in purification of contaminating DNA in total RNA extracts of RlLs.
This method has been previously reported to contain contaminating DNA, with
RNA constituting only 50-80% of total nucleic acid. The genomic DNA
contamination detected in the samples was interfering with the accuracy of

transcript measurement, necessitating use of DNase i treatment before real-

time RT-PCR.

144

 

u
“H.

Consistency in the reagents used for different samples and transcripts
was essential to make data comparable across the biological replicates and
different genes analyzed in this study. To achieve consistency, cDNA
synthesis for RlLs from the two independent biological experiments were
performed in parallel, using a common reagent mix. Furthermore, for qPCR of
each gene, common reagent mix was used for both biological replicates and
each cDNA sample was measured in triplicates within each light cycler run.

Although real-time RT-PCR provides several benefits for use in
analysis involving large sample sizes, it is critical to optimize the method at
several levels. The need for high technical reproducibility of transcript
measurement in this study required careful consideration of factors mentioned
above. Ultimately, through these optimization studies, this method was able to

achieve high reproducibility for identification of eQTLs for the different genes

included in this study.

145

 

 

 

Chapter 5

Discussion and implications for the Future

Cold acclimation has been the subject of intense research for several
years. Despite considerable success in understanding various biochemical,
physiological and gene expression changes occurring during cold acclimation,
it is not yet clear how plants perceive the low temperature signal and regulate
early changes in gene expression that ultimately lead to freezing tolerance.
Whole genome transcriptional analysis has led to the identification of genes
that are differentially expressed in response to low temperature. Clustering of
these genes based on common expression patterns suggests existence of
several gene regulatory pathways at low temperatures.

in this study, natural variation observed in low temperature responsive
gene expression was successfully used to identify genetic loci that regulate
the differential expression of early responsive genes. Analysis of nine low
temperature responsive genes identified ten regulatory eQTLs involved in
both common and unique regulation of different genes. Collectively, the data
allowed construction of a pathway depicting the regulatory positions of the
identified eQTLs (Figure 5.1).

As mentioned previously, extensive “cross-talk” is known to occur
between different abiotic stresses such as cold, drought and salinity. Due to
significant overlap in the gene regulatory networks, many regulatory
components may be shared for drought and cold. Although these genes
would lack a selective pressure with regard to low temperature, they would

experience selective pressure due to other stresses such as drought and

146

 

Low Temperature

1 e0 C1-3, 4

CBF1

CBF2

CBF3

Rav1

Zat12

CCCH-Type Zn Finger TF

Dof1 Domain Containing Gene

 

 

(FTO3)leO 01-? led cs-1 l?
CBF1 Rav1
Dof1 Domain
Containing CBF2 28172
Gene car-'3
CCCH-Type Zn Finger TF
leQ cs lea cs-2
CBF1 CBF2
CBF3

CCCH-Type Trxn Factor

1

 

 

 

(FTQ4)leO C4 (FI'O2) led C1-6
( FT 01) e0 C1 -2
CBF2 CCCH-Type
CBF3 Zn Finger TF
(FTQ1) eQ 01-1 1
i i l leO 02
Than-Like Gene CBF3 CBF2

Figure 5.1. Model of gene regulatory pathway. The regulation of nine low
temperature responsive genes has been proposed to be to regulated by the
eQTLs (denoted in blue text) identified in the present study. Overlapping
freezing tolerance QTLs, identified by Alonso-Blanco et al. (2005), have been

indicated within brackets next to the corresponding eQTL.

147

 

 

 

salinity. Consequently, regulatory genes shared between cold and drought
stress would not be identified using this population. Therefore, eQTLs
identified in this study are likely to contain genes whose products are
specifically involved in regulation of the low temperature response in
Arabidopsis. This is supported by the observation that a limited number of
eQTLs were identified in this study, which were commonly regulating several
low temperature responsive gene. The independent identification of these
eQTLs for more than one low temperature responsive genes and not for the
housekeeping gene Act1 raised the confidence in their regulatory role at low
temperatures. Additional confidence in the eQTLs comes from the observation
that some of them coincide with freezing tolerance QTLs reported by Alonso-
Blanco et al (Alonso-Blanco et al., 2005). The QTL contributing to freezing
tolerance on chromosome 4, FTQ4, was found to overlap with eQ C4
identified for CBF2 and 3 regulation (Figure 5.2, 2.8 and 2.9, respectively).
Two photoperiod-dependent freezing tolerance QTLs on chromosome 1,
FTQZ and FTQ3, were coincident with eQTLs identified for ERF and Dof1-like
domain containing genes (Figure 5.2, 3.13, and 3.14). The short-day
photoperiod specific FTQZ was co-incident with eQ C1-5 identified for ERF,
whereas the long-day photoperiod specific FTQ3 was co-incident with eQ C1-
7 identified for Dof1-like domain containing transcription factor gene. Finally,
the freezing tolerance QTL FTQ5 on chromosome 5, specific for short-day
photoperiod, was found to overlap with eQ C5-1 identified for CBF1, 2, 3 and
CCCH-type Zn finger containing transcription factor gene expression. in
addition to identification of eQTLs in regions overlapping with freezing

tolerance QTLs, this study

148

 

uncovered eQTLs for regulation of low temperature responsive gene
expression, which were not detected in the study by Alonso-Blanco et al.

The above-mentioned study on freezing tolerance QTLs demonstrated
that higher freezing tolerance was attributable to the Ler allele at the QTLs,
whereas mostly Cvi—0 alleles were found to lead to higher expression of low
temperature responsive genes in the current study. This could be explained,
firstly, by the physical proximity of the genes giving rise to freezing QTLs or
eQTLs for low temperature responsive gene expression, where one may be
contributed by Ler parent and the other by Cvi-0. in fact, several low
temperature regulatory genes may be located in clusters, causing certain
regions of the genome to be hotspots for genes involved in different aspects
of cold acclimation. Secondly, it is possible that QTLs identified through
analysis of freezing tolerance and low temperature responsive gene
expression are due to the same gene.

Although induction of specific genes is a common strategy employed to
combat stressful environmental conditions, negative regulation of genes plays
an equally important role in the adaptation process. The transcripts analyzed
in this study are undetectable under normal growth conditions, quickly
increasing in levels in response to low temperature followed by a subsequent
decrease. The adjustment in expression levels of these genes may involve
the role of transcriptional repressors, which may be underlying the eQTLs and
freezing tolerance QTLs. Negative regulators that play a role in ensuring
specific transcript levels of target genes during cold acclimation would lack the
selective pressure to maintain highly effective alleles in Cvi-0 (because of

warm temperatures in its natural habitat). Weaker alleles of these genes could

149

 

'1.

result in increased expression of their targets and therefore result in Cvi-0
contributing a positive allele for gene expression. However, the ultimate effect
of an inability to decrease expression of some genes could have a negative
effect on cold tolerance. Thus, the same allele could be identified as having a
positive effect on gene expression, while having a negative effect on cold
tolerance.

Previous studies indicate that plants appear to sense the temperature
shifts in their environment and fine-tune gene expression levels through the
different stages of cold acclimation (Zarka et al., 2003). Zarka et al
demonstrated the elegance of the CBF transcriptional response to changes in
environmental temperature. The study showed that repeated transfer of plants
between warm and cold environments de-sensitized CBF genes such that the
levels of CBF transcripts obtained in subsequent cold shock were significantly
diminished. Cold-acclimated seedlings did not show a detectable increase in
CBF transcript levels after a short (<8h) shift to warm temperatures and return
to cold, however, longer (>8h) exposure to warm temperatures caused
induction of CBF genes. These observations indicate tight regulation of CBF
expression through a refined cold-sensing mechanism that must use
transcriptional activators as well as repressors. Other low temperature
responsive genes may also experience such tight transcriptional regulation
through roles of positive and negative regulators as suggested by the finding
that certain Cvi-0 alleles result in higher expression of genes.

Identification of eQTLs in this study was achieved by using two
different internal references for the normalization of expression data. Analysis

of expression data normalized with 18S rRNA and A011 mRNA levels resulted

150

 

in identification of common and distinct eQTLs. For instance, the eQ C4 found
to regulate CBF2 and 3 expression was detected using both normalization
genes. However, eQ C1-3 and C1-4 were identified for CBF1, 2, Rav1, Zat12
etc only in the Act1 normalized dataset. Use of more than one reference
genes is ideal to obtain maximum information from a gene expression study.
There is still debate in the scientific community about whether the data from
different normalization controls should be combined and how best to do so. it
has been suggested that geometric mean of multiple normalization controls
be used for accurate RT-PCR expression analysis (Vandesompele et al.,
2002)

For the purpose of this study, the expression data normalized with the
two reference genes was analyzed separately and the results presented
independently. Some differences observed in the eQTL profiles generated
using 188 rRNA and Act1 mRNA normalized data can be explained in two
ways. Firstly, the vast diference in transcript abundance for 188 rRNA and the
genes being analyzed in this study would limit the accuracy of normalization.
Secondly, the rRNA genes, which are present as arrays in the nucleolus
organizer regions (NOR) in the genome, show variable cytosine methylation in
different accessions of Arabidopsis thaliana (Riddle and Richards, 2002). This
variation in cytosine methylation of NOR has been attributed to five QTLs
using the RIL population from Cvi-O X Ler (Riddle and Richards, 2002).
Cytosine methylation is used to achieve heritable transcriptional silencing of
genes and variable levels of it in the segregating population would imply
variable 18S rRNA levels (Riddle and Richards, 2002). Therefore, it was

considered likely that some information on variation in low temperature

151

 

responsive gene expression was not recovered using 18S rRNA as a
normalization control. To reduce the error rate in identification of eQTLs using
the normalized gene expression data, it was necessary to select a
normalization gene that was not differentially expressed in the segregating
population. Thus, for the second reference gene, Act1, expression data from
the RIL population was analyzed using CIM to ensure there were no eQTLs
leading to variable expression of this gene in the population.

A common approach used to identify genes contributing to the eQTLs
involves analysis of select candidate genes located in the regions containing
the eQTLs. The genomic regions underlying each eQTL contained hundreds
of genes in the support interval of 2 LODs. Based on functional predictions, a
list of putative candidate genes encoding regulators of transcription at low
temperature was generated. This list included several signaling proteins such
as GTPases, mitogen-activated protein kinases, receptor-like kinases,
tyrosine protein kinases, protein phosphatases, phosphatidylinositol-4-
phosphate 5-kinases, two-component response regulators, receptor kinases,
Ca2+ binding proteins, Calcineurin B-like proteins and Ca2+-dependent
kinases. Additionally, these regions also contain transcription factors,
chromatin-remodeling factors, histone deacetylases, miRNAs, ankyrin-repeat
containing proteins, and ubiquitin ligases.

Low temperature stress is known to elicit spatial and temporal Ca+2
signaling in plants and Ca+2 has been shown to act as a second messenger
mediating low temperature response (Monroy and Dhindsa, 1995; Monroy et
al., 1998), a Ca+2-dependent protein kinase (CDPK) has been shown to

upregulate a low temperature responsive promoter in barley (Sheen, 1996).

152

 

The Ca2+ signal in plants is relayed by several known classes of Ca2+-binding
proteins, including calmodulins, Ca2+-dependent protein kinases (CDPKs),
and calcineurin B-like proteins. CDPKs comprise a class of Ca2+ sensors,
possessing both protein kinase and Ca2+-binding domains in a single
polypeptide, allowing them to detect and relay the Ca2+ signal (Roberts and
Harmon, 1992). The Arabidopsis genome is predicted to contain 34 CDPK
genes, some of which are induced in response to abiotic stresses, such as the
AtCDPK1 and AtCDPK2 genes, which are induced in response to drought and
salt stress (Urao et al., 1994). Furthermore, AtCPK10 and 30 have been
demonstrated to activate a stress-inducible promoter in protoplasts (Sheen,
1996). in rice, OsCDPK13 responds to low temperature stress and its
overexpression results in enhanced cold tolerance of the transgenics (Abbasi
et al., 2004). The genomic regions encompassing the eQTLs identified in the
present study contain CDPK2, 6 and 9, which may play a role in low
temperature signaling mediated by a Ca”‘2 signal.

in addition to the CDPKs, the list also contains Ca2+ sensing proteins
such as calmodulins (CaM) and Calcineurin B-Iike (CBL) proteins. CaM is a
ubiquitous calcium-binding protein found to regulate diverse cellular functions
by modulating the activity of a variety of proteins (Kim et al., 2007). CBL
proteins are similar to the regulatory B subunit of calcineurin protein of
animals (Kudla et al., 1999), that can bind Ca2+ and the CBL interacting
protein kinases (ClPKs) to relay the signal in cell (Shi et al., 1999). in rice,
calreticulin CRT1, is suggested to play a role in low temperature stress (Li et

al., 2003). The region encompassing the eQTLs contain CAM1, 4, and 6,

153

 

_ I
ha;-

CBL3, 7, and 8, CRT2 and 3, AtCIPK6, 8, and 25 and several Ca2+ binding
proteins.

A Ca2+ sensing GTPase in Arabidopsis was found to play a role in
stress-related Ca2+ signal transduction. Its activity was regulated by changes
in Ca2+ concentration, and the gene was found to be induced by ABA and salt
stress and knock-out mutants were found highly sensitive to ABA and salt
treatments (Jayasekaran et al., 2006) GTPase mediated signal transduction in
Arabidopsis thaliana has been reported to be involved in diverse
developmental processes and hormone signaling and the genome is
predicted to encode 93 small GTPases (Li et al., 2001; Yang, 2002).

Mitogen-activated protein (MAP) kinase cascades have been shown to
regulate several signaling pathways in diverse organisms (Teige et al., 2004).
The MAP Kinase Kinase 2 (MKK2) is specifically activated by low temperature
in Arabidopsis (Teige et al., 2004). Furthermore, plants overexpressing MKK2
exhibit increased freezing tolerance and mkk2 null mutants show
hypersensitivity to cold stress. In relevance to this study, the MKK2
overexpressors revealed upregulation of the low temperature responsive
genes CBF2, CBF3 Rav1, ZAT12, and several ERF genes (Teige et al.,
2004). The eQTL regions contain several components of the MAPK pathway,
including ATMKK3 and 10, ATMPK1 and 13, MAPKKK7, 13 and 18, and
MEK1.

Members of another family of signaling proteins, protein phosphatases,
were detected in the regions of interest. PP2A is suggested to function in low
temperature signaling in alfalfa (Monroy et al., 1993; Monroy and Dhindsa,

1995) and AtPP2CA gene is proposed to negatively regulate cold acclimation

154

 

 

in Arabidopsis since antisense inhibition of the protein caused inhibition of
cold acclimation in transgenic plants (Tahtiharju and Palva, 2001 ).

Additional signaling genes that may be involved in the low temperature
responsive gene expression and were found to be present in the genomic
regions identified using eQTL analysis encode receptor protein kinases such
as ARK1 and 2, histidine two-component response regulators such as ARR4
and 24, and inositol polyphosphate 5 phosphatase such as lP5P1 and other
members of this family.

The transcription factor encoding genes present in the chromosomal
regions underlying the eQTLs include members of the MYB, NAC, and PHD
family of proteins. The nascent polypeptide-associated complex (NAC)
domain-containing proteins constitute one of the largest families of
transcription factors in Arabidopsis with more than a hundred members. These
proteins are characterized by a conserved N-terminal NAC domain that can
bind DNA as well as proteins. Five Arabidopsis NAC proteins are predicted to
contain two NAC domains in tandem, a transmembrane region and nuclear
import and/or export signals (At1g60280, Atig60300, At1g60340, At1960350
and Atig60380). These proteins were found to exist as membrane-bound
precursors that are proteolytically cleaved following a stimulus, allowing
nuclear localization of the transcription factor (Olsen et al., 2005).

miRNAs (microRNAs) comprise a class of noncoding RNAs with
regulatory roles in gene expression (Vella and Slack, 2005; Gottwein et al.,
2007; Guarnieri and Dileone, 2008; Mattes et al., 2008; Negrini and Caiin,
2008; Schultz et al., 2008; Tang et al., 2008). Several miRNAs are located in

the regions encompassing the eQTLs. MIR319 has been reported to be

155

 

 

 

 

upregulated in response to abiotic stress (Sunkar and Zhu, 2004). MIR319,
along with MIR159, is known to target several members of the MYB
transcription factor genes and five TCP gene (Palatnik et al., 2007). However,
its role at low temperatures is not yet established. Sequence homology of the
miRNAs with low temperature regulated genes would suggest direct
regulation of the homologous targets.

in conclusion, using natural genetic variation, this study narrowed down
the genomic regions of Arabidopsis containing regulatory genes with a role in
low temperature responsive gene expression. These regions, except eQ C1-1
that coincides with location of ICE1, contain novel regulatory genes since
none of the known low temperature associated genes were located in these
regions. These previously known low temperature associated genes have
mostly been identified using mutational approaches, which are unable to
detect subtle changes in genome that may not have profound effects on gene
expression despite their occurrence in the regulatory genes. Using the eQTL
approach it was possible to identify genomic regions responsible for the subtle
differences in low temperature responsive gene expression. Several putative
candidates are present the in the regions underlying the eQTLs. Ultimately,
the genes underlying the eQTLs would have to be identified to determine the
genetic pathways functioning at low temperature. Fine mapping of the eQTLs
and confirmation of their regulatory effect through generation of NlLs (near
isogenic lines) will lead to identification of the underlying genes. Efforts are
ongoing to develop a resource for a tilling array of NiLs along the entire length

of Arabidopsis chromosomes. Using NlLs containing regions identified in this

156

 

 

study, the process of identifying the underlying sequence variations will be

greatly hastened.

 

157

References

Abbasi F, Onodera H, Toki S, Tanaka H, Komatsu S (2004) OsCDPK13, a
calcium-dependent protein kinase gene from rice, is induced by cold
and gibberellin in rice leaf sheath. Plant Mol Biol 55: 541-552

Alberts R, Terpstra P, Li Y, Breitling R, Nap JP, Jansen RC (2007)
Sequence polymorphisms cause many false cis eQTLs. PLoS ONE 2:
e622

Alonso-Blanca C, Gomez-Mena C, Llorente F, Koornneef M, Salinas J,
Martinez-Zapater JM (2005) Genetic and molecular analyses of
natural variation indicate CBF2 as a candidate gene for underlying a
freezing tolerance quantitative trait locus in Arabidopsis. Plant Physiol
139: 1304-1312

Alonso-Blanca C, Koornneef M (2000) Naturally occurring variation in

Arabidopsis: an underexploited resource for plant genetics. Trends
Plant Sci 5: 22-29

Alonso-Blanca C, Koornneef M, van Ooijen JW (2006) QTL analysis.
Methods Mol Biol 323: 79-99

Alonso-Bianca C, Mendez-Vigo B, Koornneef M (2005) From phenotypic to
molecular polymorphisms involved in naturally occurring variation of
plant development. int J Dev Biol 49: 717-732

Alonso-Blanco C, Peeters AJ, Koornneef M, Lister C, Dean C, van den
Bosch N, Pot J, Kuiper MT (1998) Development of an AFLP based
linkage map of Ler, Col and Cvi Arabidopsis thaliana ecotypes and
construction of a Ler/Cvi recombinant inbred line population. Plant J
14: 259-271

Baker SS, Wilhelm KS, Thomashow MF (1994) The 5'-region of Arabidopsis
thaliana cor15a has dis-acting elements that confer cold-, drought- and
ABA-regulated gene expression. Plant Mol Biol 24: 701-713

Beavis WD, Smith 08, Grant D, Fincher R (1994) identification of
Quantitative Trait Loci Using a Small Sample of Topcrossed and F4
Progeny from Maize. Crop Science 34: 882-896

Beck EH, Fettig S, Knake C, Hartig K, Bhattarai T (2007) Specific and
unspecific responses of plants to cold and drought stress. J Biosci 32:
501-510

Belknap JK, Mitchell SR, O'Toole LA, Helms ML, Crabbe JC (1996) Type l

and type Ii error rates for quantitative trait loci (QTL) mapping studies
using recombinant inbred mouse strains. Behav Genet 26: 149-160

158

 

Benedict C, Skinner JS, Meng R, Chang Y, Bhalerao R, Huner NP, Finn
CE, Chen TH, Hurry V (2006) The CBF1-dependent low temperature
signalling pathway, regulon and increase in freeze tolerance are
conserved in Populus spp. Plant Cell Environ 29: 1259-1272

Bing N, Hoescheie i (2005) Genetical genomics analysis of a yeast
segregant population for transcription network inference. Genetics 170:
533-542

Bonhoeffer S, Chappey C, Parkin NT, Whitcomb JM, Petropouios CJ
(2004) Evidence for positive epistasis in HIV-1. Science 306: 1547-
1550

Botteldoorn N, Van Coillie E, Grijspeerdt K, Werbrouck H, Haesebrouck
F, Donne E, D'Haese E, Heyndrickx M, Pasmans F, Herman L
(2006) Real-time reverse transcription PCR for the quantification of the
mntH expression of Salmonella enterica as a function of growth phase
and phagosome-like conditions. J Microbiol Methods 66: 125-135

Brem RB, Yvert G, Clinton R, Krugiyak L (2002) Genetic dissection of
transcriptional regulation in budding yeast. Science 296: 752-755

Bugbee B (1995) The components of crop productivity: measuring and
modeling plant metabolism. ASGSB Bull 8: 93-104

Byers DL (2005) Evolution in heterogeneous environments and the potential
of maintenance of genetic variation in traits of adaptive significance.
Genetica 123: 107-124

Bystrykh L, Weersing E, Dontje B, Sutton S, Pletcher MT, Wiltshire T, Su
AI, Veilenga E, Wang J, Manly KF, Lu L, Chesler EJ, Alberts R,
Jansen RC, Williams RW, Cooke MP, de Haan G (2005) Uncovering
regulatory pathways that affect hematopoietic stem cell function using
'genetical genomics'. Nat Genet 37: 225—232

Cardon LR, Smith SD, Fulker DW, Kimberling WJ, Pennington BF,
DeFries JC (1994) Quantitative trait locus for reading disability on
chromosome 6. Science 266: 276-279

Carlborg 0, De Koning DJ, Manly KF, Chesler E, Williams RW, Haley CS
(2005) Methodological aspects of the genetic dissection of gene
expression. Bioinformatics 21: 2383-2393

Chen W, Zhang Y, Liu X, Chen B, Tu J, Tingdong F (2007) Detection of
QTL for six yield-related traits in oilseed rape (Brassica napus) using
DH and immortalized F(2) populations. Theor Appi Genet 115: 849-858

Chinnusamy V, Ohta M, Kanrar S, Lee BH, Hong X, Agarwal M, Zhu JK

(2003) iCE1: a regulator of cold-induced transcriptome and freezing
tolerance in Arabidopsis. Genes Dev 17: 1043-1054

159

 

Choi DW, Rodriguez EM, Close TJ (2002) Barley be3 gene identification,
expression pattern, and map location. Plant Physiol 129: 1781-1787

Choi HI, Park HJ, Park JH, Kim S, im MY, Seo HH, Kim YW, Hwang l, Kim
SY (2005) Arabidopsis calcium-dependent protein kinase AtCPK32
interacts with ABF4, a transcriptional regulator of abscisic acid-

responsive gene expression, and modulates its activity. Plant Physiol
139: 1750-1761

Comstock RE, Robinson HF (1948) The components of genetic variance in
populations of biparental progenies and their use in estimating the
average degree of dominance. Biometrics 4: 254-266

Consoli L, Lefevre A, Zivy M, de Vienne D, Damerval C (2002) QTL
analysis of proteome and transcriptome variations for dissecting the
genetic architecture of complex traits in maize. Plant Mol Biol 48: 575-
581

Cook D, Fowler 8, Fiehn O, Thomashow MF (2004) A prominent role for the 3;
CBF cold response pathway in configuring the low-temperature
metabolome of Arabidopsis. Proc Natl Acad Sci U S A 101: 15243-
15248

 

11'

Cork JM, Purugganan MD (2004) The evolution of molecular genetic
pathways and networks. Bloessays 26: 479-484

De Boever S, Vangestel C, De Backer P, Croubels S, Sys SU (2008)
identification and validation of housekeeping genes as internal control
for gene expression in an intravenous LPS inflammation model in
chickens. Vet lmmunol lmmunopathol 122: 312-317

De Ketelaere A, Goossens K, Peelman L, Burvenich C (2006) Technical
note: validation of internal control genes for gene expression analysis
in bovine polymorphonuclear leukocytes. J Dairy Sci 89: 4066-4069

de Koning DJ, Carlborg O, Haley CS (2005) The genetic dissection of
immune response using gene-expression studies and genome
mapping. Vet lmmunol lmmunopathol 105: 343-352

Deutsch S, Lyle R, Dermitzakis ET, Attar H, Subrahmanyan L, Gehrig C,
Parand L, Gagnebin M, Rougemont J, Jongeneel CV, Antonarakis
SE (2005) Gene expression variation and expression quantitative trait
mapping of human chromosome 21 genes. Hum Mol Genet 14: 3741-
3749

Ding JP, Pickard 36 (1993) Modulation of mechanosensitive calcium-
selective cation channels by temperature. Plant J 3: 713-720

Doss S, Schadt EE, Drake TA, Lusis AJ (2005) Cis-acting expression
quantitative trait loci in mice. Genome Res 15: 681-691

160

Dubouzet JG, Sakuma Y, Ito Y, Kasuga M, Dubouzet EG, Miura S, Seki M,
Shinozaki K, Yamaguchi-Shinozaki K (2003) OsDREB genes in rice,
Oryza sativa L., encode transcription activators that function in drought-
, high-salt- and cold-responsive gene expression. Plant J 33: 751-763

Farrali M (2004) Quantitative genetic variation: a post-modern view. Hum Mol
Genet 13 Spec No 1: R1-7

Fijneman RJ, de Vries SS, Jansen RC, Demant P (1996) Complex
interactions of new quantitative trait loci, Sluc1, Siuc2, Sluc3, and

Sluc4, that influence the susceptibility to lung cancer in the mouse. Nat
Genet 14: 465-467

Fisler JS, Warden CH (1997) Mapping of mouse obesity genes: A generic
approach to a complex trait. J Nutr 127: 19098-19168

Fowler 8, Thomashow MF (2002) Arabidopsis transcriptome profiling
indicates that multiple regulatory pathways are activated during cold "
acclimation in addition to the CBF cold response pathway. Plant Cell
14: 1675-1690

 

Fowler 86, Cook D, Thomashow MF (2005) Low temperature induction of
Arabidopsis CBF1, 2, and 3 is gated by the circadian clock. Plant
Physiol 137: 961-968

Frankel WN, Valenzuela A, Lutz CM, Johnson EW, Dietrich WF, Coffin JM
(1995) New seizure frequency QTL and the complex genetics of
epilepsy in EL mice. Mamm Genome 6: 830-838

Frascaroii E, Cane MA, Landi P, Pea G, Gianfranceschi L, Villa M,
Morgante M, Pe ME (2007) Classical genetic and quantitative trait loci

analyses of heterosis in a maize hybrid between two elite inbred lines.
Genetics 176: 625-644

Gao D, Knight MR, Trewavas AJ, Satteimacher B, Plieth C (2004) Self-
reporting Arabidopsis expressing pH and [Ca2+] indicators unveil ion
dynamics in the cytoplasm and in the apoplast under abiotic stress.
Plant Physiol 134: 898-908

Garber MP (1977) Effect of Light and Chilling Temperatures on Chilling-
sensitive and Chilling-resistant Plants. Pretreatment of Cucumber and
Spinach Thylakoids in Vivo and in Vitro. Plant Physiol 59: 981-985

Garg AK, Kim JK, Owens TG, Ranwala AP, Choi YD, Kochian LV, Wu RJ
(2002) Trehalose accumulation in rice plants confers high tolerance
levels to different abiotic stresses. Proc Natl Acad Sci U S A 99:
15898-15903

161

Gebhardt MD, Stearns SC (1992) Phenotypic plasticity for life-history traits in
Drosophila melanogaster. lll. Effect of the environment on genetic
parameters. Genet Res 60: 87-101

Gilmour SJ, Fowler SG, Thomashow MF (2004) Arabidopsis transcriptional
activators CBF1, CBF2, and CBF3 have matching functional activities.
Plant Mol Biol 54: 767-781

Gilmour SJ, Hajela RK, Thomashow MF (1988) Cold Acclimation in
Arabidopsis thaliana. Plant Physiol 87: 745-750

Gilmour SJ, Zarka DG, Stockinger EJ, Salazar MP, Houghton JM,
Thomashow MF (1998) Low temperature regulation of the Arabidopsis
CBF family of AP2 transcriptional activators as an early step in cold-
induced COR gene expression. Plant J 16: 433-442

Gong Z, Lee H, Xiong L, Jagendorf A, Stevenson B, Zhu JK (2002) RNA
helicase-like protein as an early regulator of transcription factors for :
plant chilling and freezing tolerance. Proc Natl Acad Sci U S A 99: "-
11507-11512

 

Gora-Maslak G, McCiearn GE, Crabbe JC, Phillips TJ, Belknap JK, ‘7"
Plomin R (1991) Use of recombinant inbred strains to identify

quantitative trait loci in psychopharmacology. Psychopharmacology
(Berl) 104: 413-424

Gottwein E, Mukherjee N, Sachse C, Frenzel C, Majoros WH, Chi JT,
Braich R, Manoharan M, Soutschek J, Ohler U, Cullen BR (2007) A

viral microRNA functions as an orthologue of cellular miR-155. Nature
450: 1096-1099

Guarnieri DJ, Dileone RJ (2008) MicroRNAs: A new class of gene
regulators. Ann Med 40: 197-208

GuhaThakurta D, Xie T, Anand M, Edwards SW, Li G, Wang SS, Schadt
EE (2006) Cis-regulatory variations: a study of SNPs around genes

showing cis-linkage in segregating mouse populations. BMC Genomics
7: 235

Guo Y, Xiong L, lshitani M, Zhu JK (2002) An Arabidopsis mutation in
translation elongation factor 2 causes superinduction of CBF/DREB1
transcription factor genes but blocks the induction of their downstream
targets under low temperatures. Proc Natl Acad Sci U S A 99: 7786-
7791

Guy C (1999) Molecular responses of plants to cold shock and cold
acclimation. J Mol Microbiol Biotechnol 1: 231-242

Guy CL, Niemi KJ, Brambi R (1985) Altered gene expression during cold
acclimation of spinach. Proc Natl Acad Sci U S A 82: 3673-3677

162

Haake V, Cook D, Riechmann JL, Pineda O, Thomashow MF, Zhang JZ
(2002) Transcription factor CBF4 is a regulator of drought adaptation in
Arabidopsis. Plant Physiol 130: 639-648

Harada H, Kuromori T, Hirayama T, Shinozaki K, Leigh RA (2004)
Quantitative trait loci analysis of nitrate storage in Arabidopsis leading
to an investigation of the contribution of the anion channel gene,
AtCLC-c, to variation in nitrate levels. J Exp Bot 55: 2005-2014

Harmer SL, Hogenesch JB, Straume M, Chang HS, Han B, Zhu T, Wang
X, Kreps JA, Kay SA (2000) Orchestrated transcription of key
pathways in Arabidopsis by the circadian clock. Science 290: 2110-
2113

Hetherington SE, He J, Smillie RM (1989) Photoinhibitlon at Low

Temperature in Chilling-Sensitive and -Resistant Plants. Plant Physiol
90: 1609-1615

 

Hincha DK, DeVries AL, Schmitt JM (1993) Cryotoxicity of antifreeze
proteins and glycoproteins to spinach thylakoid membranes--
comparison with cryotoxic sugar acids. Biochim Biophys Acta 1146:
258-264

 

Hoffmann AA, Weeks AR (2007) Climatic selection on genes and traits after
a 100 year-old invasion: 3 critical look at the temperate-tropical clines
in Drosophila melanogaster from eastern Australia. Genetica 129: 133-
147

Hoogewijs D, Houthoofd K, Matthijssens F, Vandesompele J, Vanfleteren
JR (2008) Selection and validation of a set of reliable reference genes

for quantitative sod gene expression analysis in C. elegans. BMC Mol
Biol 9: 9

Huggett J, Dheda K, Bustin S, Zumla A (2005) Real-time RT-PCR
normalisation; strategies and considerations. Genes lmmun 6: 279-284

Hugly S, Somerville C (1992) A Role for Membrane Lipid Polyunsaturation in
Chloroplast Biogenesis at Low Temperature. Plant Physiol 99: 197-202

Hurry V, Strand A, Furbank R, Stitt M (2000) The role of inorganic
phosphate in the development of freezing tolerance and the
acclimatization of photosynthesis to low temperature is revealed by the
pho mutants of Arabidopsis thaliana. Plant J 24: 383-396

lshitani M, Xiong L, Lee H, Stevenson B, Zhu JK (1998) HOS1, a genetic
locus involved in cold-responsive gene expression in arabidopsis. Plant
Cell 10: 1151-1161

Jaglo KR, Kleff S, Amundsen KL, Zhang X, Haake V, Zhang JZ, Deits T,
Thomashow MF (2001) Components of the Arabidopsis C-

163

repeat/dehydration-responsive element binding factor cold-response
pathway are conserved in Brassica napus and other plant species.
Plant Physiol 127: 910-917

Jaglo-Ottosen KR, Gilmour SJ, Zarka DG, Schabenberger O, Thomashow
MF (1998) Arabidopsis CBF1 overexpression induces COR genes and
enhances freezing tolerance. Science 280: 104-106

Jansen RC (1993) interval mapping of multiple quantitative trait loci. Genetics
135: 205-211

Jayasekaran K, Kim KN, Vivekanandan M, Shin JS, Ok SH (2006) Novel
calcium-binding GTPase (AtCBG) involved in ABA-mediated salt stress
signaling in Arabidopsis. Plant Cell Rep 25: 1255-1262 F

Jiang C, Zeng ZB (1995) Multiple trait analysis of genetic mapping for
quantitative trait loci. Genetics 140: 1111-1 127

Juenger TE, Sen S, Stowe KA, Simms EL (2005) Epistasis and genotype-
environment interaction for quantitative trait loci affecting flowering time
in Arabidopsis thaliana. Genetica 123: 87-105

 

Kao CH (2006) Mapping quantitative trait loci using the experimental designs
of recombinant inbred populations. Genetics 174: 1373-1386

Kaplan B, Davydov 0, Knight H, Galon Y, Knight MR, Fluhr R, Fromm H
(2006) Rapid transcriptome changes induced by cytosolic Ca2+
transients reveal ABRE-related sequences as Ca2+-responsive cis
elements in Arabidopsis. Plant Cell 18: 2733-2748

Kaye C, Guy CL (1995) Perspectives of plant cold tolerance: physiology and
molecular responses. Sci Prog 78 ( Pt 4): 271-299

Kearsey MJ, Farquhar AG (1998) QTL analysis in plants; where are we
now? Heredity 80 ( Pt 2): 137-142

Keurentjes JJ, Bentsink L, Alonso-Blanco C, Hanhart CJ, Blankestijn-De
Vries H, Effgen S, Vreugdenhil D, Koornneef M (2007) Development
of a near-isogenic line population of Arabidopsis thaliana and
comparison of mapping power with a recombinant inbred line
population. Genetics 175: 891-905

Keurentjes JJ, Fu J, de Vos CH, Lommen A, Hall RD, Bino RJ, van der
Plas LH, Jansen RC, Vreugdenhil D, Koornneef M (2006) The
genetics of plant metabolism. Nat Genet 38: 842-849

Keurentjes JJ, Fu J, Terpstra iR, Garcia JM, van den Ackerveken G,

Snoek LB, Peeters AJ, Vreugdenhil D, Koornneef M, Jansen RC
(2007) Regulatory network construction in Arabidopsis by using

164

genome-wide gene expression quantitative trait loci. Proc Natl Acad
Sci U S A104: 1708-1713

Kim HS, Park BO, Yoo JH, Jung MS, Lee SM, Han HJ, Kim KE, Kim SH,
Lim CO, Yun DJ, Lee SY, Chung WS (2007) identification of a
calmodulin-binding NAC protein as a transcriptional repressor in
Arabidopsis. J Biol Chem 282: 36292-36302

Kim K, Portis AR, Jr. (2005) Temperature dependence of photosynthesis in
Arabidopsis plants with modifications in Rubisco activase and
membrane fluidity. Plant Cell Physiol 46: 522-530

Kimura M (1969) The rate of molecular evolution considered from the
standpoint of population genetics. Proc Natl Acad Sci U S A 63: 1181- 1'
1188

Kirst M, Basten CJ, Myburg AA, Zeng ZB, Sederoff RR (2005) Genetic
architecture of transcript-level variation in differentiating xylem of a .
eucalyptus hybrid. Genetics 169: 2295-2303 E

 

Kliebenstein DJ, West MA, van Leeuwen H, Loudet O, Doerge RW, St
Clair DA (2006) Identification of QTLs controlling gene expression
networks defined a priori. BMC Bioinformatics 7: 308

Knapp SJ, Bridges WC (1990) Using molecular markers to estimate
quantitative trait locus parameters: power and genetic variances for
unreplicated and replicated progeny. Genetics 126: 769-777

Knight H, Veale EL, Warren GJ, Knight MR (1999) The ser mutation in
Arabidopsis suppresses low-temperature induction of genes dependent
on the CRT/DRE sequence motif. Plant Cell 11: 875-886

Knight MR (2002) Signal transduction leading to low-temperature tolerance in
Arabidopsis thaliana. Philos Trans R Soc Lond B Biol Sci 357: 871-875

Kodama H, l-iamada T, Horiguchi G, Nishimura M, lba K (1994) Genetic
Enhancement of Cold Tolerance by Expression of a Gene for

Chloroplast [omega]-3 Fatty Acid Desaturase in Transgenic Tobacco.
Plant Physiol 105: 601-605

Koornneef M, Alonso-Blanca C, Vreugdenhil D (2004) Naturally occurring
genetic variation in Arabidopsis thaliana. Annu Rev Plant Biol 55: 141-
172

Korner C, Larcher W (1988) Plant life in cold climates. Symp Soc Exp Biol
42: 25-57

Kudla J, Xu Q, Harter K, Gruissem W, Luan S (1999) Genes for calcineurin
B-like proteins in Arabidopsis are differentially regulated by stress
signals. Proc Natl Acad Sci U S A 96: 4718-4723

165

Lee

Lee

Lil

Li‘

Li

Li

Lil

Lc

 

Lee H, Xiong L, Gong Z, lshitani M, Stevenson B, Zhu JK (2001) The
Arabidopsis HOS1 gene negatively regulates cold signal transduction
and encodes a RING finger protein that displays cold-regulated nucleo-
-cytoplasmic partitioning. Genes Dev 15: 912-924

Lee SC, Huh KW, An K, An G, Kim SR (2004) Ectopic expression of a cold-
inducibie transcription factor, CBF1/DREB1b, in transgenic rice (Oryza
sativa L.). Moi Cells 18: 107-114

Li H, Shen JJ, Zheng ZL, Lin Y, Yang Z (2001) The Rop GTPase switch

controls multiple developmental processes in Arabidopsis. Plant
Physiol 126: 670-684

Li Y, Alvarez OA, Gutteiing EW, Tijsterman M, Fu J, Riksen JA,
Hazendonk E, Prins P, Plasterk RH, Jansen RC, Breitling R,
Kammenga JE (2006) Mapping determinants of gene expression
plasticity by genetical genomics in C. elegans. PLoS Genet 2: e222

Li Z, Onodera H, Ugaki M, Tanaka H, Komatsu S (2003) Characterization of
calreticulin as a phosphoprotein interacting with cold-induced protein
kinase in rice. Biol Pharm Bull 26: 256-261

Li Z, Pinson SR, Park WD, Paterson AH, Stansel JW (1997) Epistasis for
three grain yield components in rice (Oryza sativa L.). Genetics 145:
453-465

Liu Q, Kasuga M, Sakuma Y, Abe H, Miura S, Yamaguchi-Shinozaki K,
Shinozaki K (1998) Two transcription factors, DREBl and DREBZ,
with an EREBP/AP2 DNA binding domain separate two cellular signal
transduction pathways in drought- and low-temperature-responsive
gene expression, respectively, in Arabidopsis. Plant Cell 10: 1391-
1406

Loudet O, Chailiou S, Camilleri C, Bouchez D, Daniel-Vedele F (2002)
Bay-0 x Shahdara recombinant inbred line population: a powerful tool

for the genetic dissection of complex traits in Arabidopsis. Theor Appl
Genet 104: 1173-1184

Luu-The V, Paquet N, Calvo E, Cumps J (2005) improved real-time RT-PCR
method for high-throughput measurements using second derivative
calculation and double correction. Biotechniques 38: 287-293

Lyng MB, Laenkholm AV, Pallisgaard N, Ditzel HJ (2008) Identification of
genes for normalization of real-time RT-PCR data in breast
carcinomas. BMC Cancer 8: 20

Maccoux LJ, Clements DN, Saiway F, Day PJ (2007) Identification of new

reference genes for the normalisation of canine osteoarthritic joint
tissue transcripts from microarray data. BMC Mol Biol 8: 62

166

 

Mackay TF (2002) The nature of quantitative genetic variation for Drosophila
longevity. Mech Ageing Dev 123: 95-104

Mackinnon MJ, Weller JI (1995) Methodology and accuracy of estimation of

quantitative trait loci parameters in a half-sib design using maximum
likelihood. Genetics 141: 755-770

MacLaren EJ, Sikela JM (2005) Cerebellar gene expression profiling and
eQTL analysis in inbred mouse strains selected for ethanol sensitivity.
Alcohol Clin Exp Res 29: 1568-1579

Massari ME, Murre C (2000) Helix-loop-helix proteins: regulators of
transcription in eucaryotic organisms. Mol Cell Biol 20: 429-440

Mattes J, Collison A, Foster PS (2008) Emerging role of microRNAs in

disease pathogenesis and strategies for therapeutic modulation. Curr
Opin Moi Ther 10: 150-157

McConn M, Browse J (1998) Polyunsaturated membranes are required for
photosynthetic competence in a mutant of Arabidopsis. Plant J 15:
521-530

Melchinger AE, Piepho HP, Utz HF, Muminovic J, Wegenast T, Torjek O,
Altmann T, Kusterer B (2007) Genetic basis of heterosis for growth-
related traits in Arabidopsis investigated by testcross progenies of
near-isogenic lines reveals a significant role of epistasis. Genetics 177:
1827-1837

Melchinger AE, Utz HF, Piepho HP, Zeng ZB, Schon CC (2007) The role of
epistasis in the manifestation of heterosis: a systems-oriented
approach. Genetics 177: 1815-1825

Miller AK, Galiba G, Dubcovsky J (2006) A cluster of 11 CBF transcription
factors is located at the frost tolerance locus Fr-Am2 in Triticum
monococcum. Mol Genet Genomics 275: 193-203

Monroy AF, Dhindsa RS (1995) Low—temperature signal transduction:
induction of cold acclimation-specific genes of alfalfa by calcium at 25
degrees C. Plant Cell 7: 321-331

Monroy AF, Sangwan V, Dhindsa RS (1998) Phenol-extracted plant proteins

can be renatured and assayed in gel for protein kinase activity. Anal
Biochem 265: 183-185

Monroy AF, Sarhan F, Dhindsa RS (1993) Cold-Induced Changes in

Freezing Tolerance, Protein Phosphorylatlon, and Gene Expression
(Evidence for a Role of Calcium). Plant Physiol 102: 1227-1235

167

Nagao M, Arakawa K, Takezawa D, Fujikawa S (2008) Long- and short-term

freezing induce different types of injury in Arabidopsis thaliana leaf
cells. Pianta 227: 477-489

Negrini M, Calin GA (2008) Breast cancer metastasis: a microRNA story.
Breast Cancer Res 10: 203

Olsen AN, Ernst HA, Leggio LL, Skriver K (2005) NAC transcription factors:
structurally distinct, functionally diverse. Trends Plant Sci 10: 79-87

Orr HA (1999) The evolutionary genetics of adaptation: a simulation study.
Genet Res 74: 207-214

Ota T, Kimura M (1971) Amino acid composition of proteins as a product of
molecular evolution. Science 174: 150-153

Page RE, Jr., Fondrk MK, Hunt GJ, Guzman-Novoa E, Humphries MA,
Nguyen K, Greene AS (2000) Genetic dissection of honeybee (Apis
mellifera L.) foraging behavior. J Hered 91: 474-479

Palatnik JF, Woilmann H, Schommer C, Schwab R, Boisbouvier J,
Rodriguez R, Warthmann N, Allen E, Dezuiian T, Huson D,
Carrington JC, Weigel D (2007) Sequence and expression
differences underlie functional specialization of arabidopsis microRNAs
miR159 and miR319. Dev Cell 13: 115-125

Pauli RE (1981) Temperature-Induced Leakage from Chilling-Sensitive and
Chilling-Resistant Plants. Plant Physiol 68: 149-153

Perez-Perez JM, Serrano-Cartagena J, Micol JL (2002) Genetic analysis of
natural variations in the architecture of Arabidopsis thaliana vegetative
leaves. Genetics 162: 893-915

Plomin R, Owen MJ, McGuffin P (1994) The genetic basis of complex
human behaviors. Science 264: 1733-1739

Pomp D (1997) Genetic dissection of obesity in polygenic animal models.
Behav Genet 27: 285-306

Psallida C, Argyropouios D (2005) Optimized mRNA isolation and

assessment methodology with two-step real time PCR. Prep Biochem
Biotechnol 35: 273-282

Qin F, Sakuma Y, Li J, Liu Q, Li YQ, Shinozaki K, Yamaguchi-Shinozaki K
(2004) Cloning and functional analysis of a novel DREB1/CBF
transcription factor involved in cold-responsive gene expression in Zea
mays L. Plant Cell Physiol 45: 1042-1052

Raison JK, Pike C8, Berry JA (1982) Growth Temperature-induced

Alterations in the Thermotropic Properties of Nerium oleander
Membrane Lipids. Plant Physiol 70: 215-218

168

 

 

 

Rebai A, Goffinet B, Mangin B (1995) Comparing power of different
methods for QTL detection. Biometrics 51: 87-99

Reist M, Pfaffl MW, Morel C, Meylan M, Hirsbrunner G, Blum JW, Steiner
A (2003) Quantitative mRNA analysis of eight bovine 5-HT receptor
subtypes in brain, abomasum, and intestine by real-time RT-PCR. J
Recept Signal Transduct Res 23: 271-287

Remans T, Smeets K, Opdenakker K, Mathijsen D, Vangronsveld J,
Cuypers A (2008) Normalisation of real-time RT-PCR gene expression
measurements in Arabidopsis thaliana exposed to increased metal
concentrations. Planta

Riddle NC, Richards EJ (2002) The control of natural variation in cytosine
methylation in Arabidopsis. Genetics 162: 355-363

Roberts MR, Harmon AC (1992) Calcium-modulated proteins; targets of

intracellular calcium signals in higher plants. Annu Rev Plant Physiol
Plant Mol Biol 43: 375-414

Rodriguez LA, Plomin R, Blizard DA, Jones BC, McClearn GE (1995)
Alcohol acceptance, preference, and sensitivity in mice. ll. Quantitative

trait loci mapping analysis using BXD recombinant inbred strains.
Alcohol Clin Exp Res 19: 367-373

Ron M, israeii G, Seroussi E, Weller JI, Gregg JP, Shani M, Medrano JF
(2007) Combining mouse mammary gland gene expression and
comparative mapping for the identification of candidate genes for QTL
of milk production traits in cattle. BMC Genomics 8: 183

Ronald J, Akey JM (2007) The evolution of gene expression QTL in
Saccharomyces cerevisiae. PLoS ONE 2: e678

Rosa GJ, de Leon N, Rosa AJ (2006) Review of microarray experimental
design strategies for genetical genomics studies. Physiol Genomics 28:
15-23

Sakamoto A, Murata N (2002) The role of glycine betaine in the protection of
plants from stress: clues from transgenic plants. Plant Cell Environ 25:
163-171

Sanchez-Moran E, Armstrong SJ, Santos JL, Franklin FC, Jones GH
(2002) Variation in chiasma frequency among eight accessions of
Arabidopsis thaliana. Genetics 162: 1415-1422

Sanjuan R, Elena SF (2006) Epistasis correlates to genomic complexity. Proc
Natl Acad Sci U S A 103: 14402-14405

169

 

Schultz J, Lorenz P, Gross G, Ibrahim S, Kunz M (2008) MicroRNA let-7b
targets important cell cycle molecules in malignant melanoma cells and
interferes with anchorage-independent growth. Cell Res

Sheen J (1996) CaZ+-dependent protein kinases and stress signal
transduction in plants. Science 274: 1900-1902

Shi C, Uzarowska A, Ouzunova M, Landbeck M, Wenzel G, Lubberstedt T
(2007) Identification of candidate genes associated with cell wall
digestibility and eQTL (expression quantitative trait loci) analysis in 3
Flint x Flint maize recombinant inbred line population. BMC Genomics
8: 22

Shi J, Kim KN, Ritz O, Albrecht V, Gupta R, Harter K, Luan S, Kudla J F
(1999) Novel protein kinases associated with calcineurin B-like calcium "
sensors in Arabidopsis. Plant Cell 11: 2393-2405

Shindo C, Bernasconi GaH, CS. (2007) Natural Genetic Variation in

Arabidopsis: Tools, Traits and Prospects for Evolutionary Ecology.
Annals of Botany 99: 1043-1054

 

Shinozaki K, Yamaguchi-Shinozaki K (1996) Molecular responses to
drought and cold stress. Curr Opin Biotechnol 7: 161-167

Shinozaki K, Yamaguchi-Shinozaki K (2000) Molecular responses to
dehydration and low temperature: differences and cross-talk between
two stress signaling pathways. Curr Opin Plant Biol 3: 217-223

Shinozaki K, Yamaguchi-Shinozaki K, Seki M (2003) Regulatory network of

gene expression in the drought and cold stress responses. Curr Opin
Plant Biol 6: 410-417

Skinner JS, Szucs P, van Zitzewitz J, Marquez-Cediilo L, Filichkin T,
Stockinger EJ, Thomashow MF, Chen TH, l-layes PM (2006)
Mapping of barley homologs to genes that regulate low temperature
tolerance in Arabidopsis. Theor Appi Genet 112: 832-842

Skovgaard K, Mortensen S, Poulsen KT, Angen O, Heegaard PM (2007)
Validation of putative reference genes for qRT-PCR normalization in
tissues and blood from pigs infected with Actinobacillus
pleuropneumoniae. Vet Immunol lmmunopathol 118: 140-146

Stahlberg A, Zoric N, Aman P, Kubista M (2005) Quantitative real-time PCR
for cancer detection: the lymphoma case. Expert Rev Moi Diagn 5:
221-230

Stam LF, Laurie CC (1996) Molecular dissection of a major gene effect on a
quantitative trait: the level of alcohol dehydrogenase expression in
Drosophila melanogaster. Genetics 144: 1559-1564

170

Stitt M, Hurry V (2002) A plant for all seasons: alterations in photosynthetic
carbon metabolism during cold acclimation in Arabidopsis. Curr Opin
Plant Biol 5: 199-206

Stockinger EJ, Gilmour SJ, Thomashow MF (1997) Arabidopsis thaliana
CBF1 encodes an AP2 domain-containing transcriptional activator that
binds to the C-repeat/DRE, a cis-acting DNA regulatory element that
stimulates transcription in response to low temperature and water
deficit. Proc Natl Acad Sci U S A 94: 1035-1040

Strand A, Hurry V, Gustafsson P, Gardestrom P (1997) Development of
Arabidopsis thaliana leaves at low temperatures releases the
suppression of photosynthesis and photosynthetic gene expression
despite the accumulation of soluble carbohydrates. Plant J 12: 605-614

Strand A, Hurry V, Henkes S, Huner N, Gustafsson P, Gardestrom P, Stitt
M (1999) Acclimation of Arabidopsis leaves devel0ping at low
temperatures. Increasing cytoplasmic volume accompanies increased
activities of enzymes in the Calvin cycle and in the sucrose- j
biosynthesis pathway. Plant Physiol 119: 1387-1398 '

 

Stuber CW, Lincoln SE, Wolff DW, Helentjaris T, Lander ES (1992)
Identification of genetic factors contributing to heterosis in a hybrid from
two elite maize inbred lines using molecular markers. Genetics 132:
823-839

Sukumaran NP, Weiser CJ (1972) Freezing injury in Potato Leaves. Plant
Physiol 50: 564-567

Sunkar R, Zhu JK (2004) Novel and stress-regulated microRNAs and other
small RNAs from Arabidopsis. Plant Cell 16: 2001-2019

Suzuki I, Los D, Murata N (2000) Perception and transduction of low-
temperature signals to induce desaturation of fatty acids. Biochem.
Soc. Trans. 28: 628-630

Swarup K, Alonso-Blanco C, Lynn JR, Michaels SD, Amasino RM,
Koornneef M, Millar AJ (1999) Natural allelic variation identifies new
genes in the Arabidopsis circadian system. Plant J 20: 67-77

Tahtiharju S, Palva T (2001) Antisense inhibition of protein phosphatase 2C
accelerates cold acclimation in Arabidopsis thaliana. Plant J 26: 461-
470

Tahtiharju S, Sangwan V, Monroy AF, Dhindsa RS, Borg M (1997) The
induction of kin genes in coid-acclimatlng Arabidopsis thaliana.
Evidence of a role for calcium. Planta 203: 442-447

Tang Y, Banan A, Forsyth CB, Fields JZ, Lau CK, Zhang LJ,
Keshavarzian A (2008) Effect of alcohol on miR-212 expression in

171

intestinal epithelial cells and its potential role in alcoholic liver disease.
Alcohol Clin Exp Res 32: 355-364

Taylor AO, Jepsen NM, Christeiler JT (1972) Plants under Climatic Stress:

lll. Low Temperature, High Light Effects on Photosynthetic Products.
Plant Physiol 49: 798-802

Taylor AO, Slack CR, McPherson HG (1974) Plants under Climatic Stress:
Vi. Chilling and Light Effects on Photosynthetic Enzymes of Sorghum
and Maize. Plant Physiol 54: 696-701

Teige M, Scheiki E, Euigem T, Doczi R, lchimura K, Shinozaki K, Dangl
JL, Hirt H (2004) The MKK2 pathway mediates cold and salt stress
signaling in Arabid0psis. Mol Cell 15: 141-152

Thomashow MF (1998) Role of cold-responsive genes in plant freezing
tolerance. Plant Physiol 118: 1-8

Thomashow MF (1999) PLANT COLD ACCLIMATION: Freezing Tolerance
Genes and Regulatory Mechanisms. Annu Rev Plant Physiol Plant Mol
Biol 50: 571-599

Tran LS, Nakashima K, Shinozaki K, Yamaguchi-Shinozaki K (2007) Plant
gene networks in osmotic stress response: from genes to regulatory
networks. Methods Enzymol 428: 109-128

Uemura M, Gilmour SJ, Thomashow MF, Steponkus PL (1996) Effects of
COR6.6 and COR15am polypeptides encoded by COR (cold-
regulated) genes of Arabidopsis thaliana on the freeze-induced fusion
and leakage of liposomes. Plant Physiol 111: 313-327

Uemura M, Joseph RA, Steponkus PL (1995) Cold Acclimation of
Arabidopsis thaliana (Effect on Plasma Membrane Lipid Composition
and Freeze-induced Lesions). Plant Physiol 109: 15-30

Upchurch RG (2008) Fatty acid unsaturation, mobilization, and regulation in
the response of plants to stress. Biotechnol Lett

Urao T, Katagiri T, Mizoguchi T, Yamaguchi-Shinozaki K, Hayashida N,
Shinozaki K (1994) Two genes that encode Ca(2+)-dependent protein
kinases are induced by drought and high-salt stresses in Arabidopsis
thaliana. Moi Gen Genet 244: 331-340

Valasek MA, Repa JJ (2005) The power of real-time PCR. Adv Physiol Educ
29: 151-159

Vallejo RL, Bacon LD, Liu HC, Witter RL, Groenen MA, Hillel J, Cheng HH
(1998) Genetic mapping of quantitative trait loci affecting susceptibility
to Marek's disease virus induced tumors in F2 lntercross chickens.
Genetics 148: 349-360

172

 

Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe
A, Speleman F (2002) Accurate normalization of real-time quantitative
RT-PCR data by geometric averaging of multiple internal control
genes. Genome Biol 3: RESEARCH0034

Vella MC, Slack FJ (2005) C. elegans microRNAs. WormBook: 1-9

Vergnolie C, Vaultier MN, Taconnat L, Renou JP, Kader JC, Zachowski A,
Rueliand E (2005) The cold-induced early activation of phospholipase
C and D pathways determines the response of two distinct clusters of
genes in Arabidopsis cell suspensions. Plant Physiol 139: 1217-1233

Vigh L, Los DA, Horvath I, Murata N (1993) The primary signal in the
biological perception of temperature: Pd-catalyzed hydrogenation of
membrane lipids stimulated the expression of the desA gene in
Synechocystis PCC6803. Proc Natl Acad Sci U S A 90: 9090-9094

Vogel JT, Zarka DG, Van Buskirk HA, Fowler SG, Thomashow MF (2005)
Roles of the CBF2 and ZAT12 transcription factors in configuring the
low temperature transcriptome of Arabidopsis. Plant J 41: 195-211

Volger H, Heber U, Berzborn RJ (1978) Loss of function of biomembranes

and solubilization of membrane proteins during freezing. Biochim
Biophys Acta 511: 455-469

Volger HG, Heber U (1975) Cryoprotective leaf proteins. Biochim Biophys
Acta 412: 335-349

Vuylsteke M, van Eeuwijk F, Van Hummeien P, Kuiper M, Zabeau M

(2005) Genetic analysis of variation in gene expression in Arabidopsis
thaliana. Genetics 171: 1267-1275

Wacker MJ, Godard MP (2005) Analysis of one-step and two-step real-time
RT-PCR using SuperScript ill. J Biomol Tech 16: 266-271

Wada H, Gombos Z, Murata N (1990) Enhancement of chilling tolerance of a

cyanobacterium by genetic manipulation of fatty acid desaturation.
Nature 347: 200-203

Wallis JG, Browse J (2002) Mutants of Arabidopsis reveal many roles for
membrane lipids. Prog Lipid Res 41: 254-278

Wang C, Gao D, Vagienov A, Kaltenboeck B (2004) One-step real-time
duplex reverse transcription PCRs simultaneously quantify analyte and
housekeeping gene mRNAs. Biotechniques 36: 508-516, 518-509

Wayne ML, McIntyre LM (2002) Combining mapping and arraying: An

approach to candidate gene identification. Proc Natl Acad Sci U S A
99: 14903-14906

173

 

Webb MS, Hui SW, Steponkus PL (1993) Dehydration-induced Iamellar-to-
hexagonal-ll phase transitions in DOPE/DOPC mixtures. Biochim
Biophys Acta 1145: 93-104

Webb MS, Steponkus PL (1993) Freeze-Induced Membrane Ultrastructurai
Alterations in Rye (Secale cereale) Leaves. Plant Physiol 101: 955-963

Webb MS, Uemura M, Steponkus PL (1994) A Comparison of Freezing
injury in Cat and Rye: Two Cereals at the Extremes of Freezing
Tolerance. Plant Physiol 104: 467-478

Wells D (2007) Use of real-time polymerase chain reaction to measure gene
expression in single cells. Methods Moi Med 132: 125-133

West MA, Kim K, Kliebenstein DJ, van Leeuwen H, Michelmore RW,
Doerge RW, St Clair DA (2007) Global eQTL mapping reveals the
complex genetic architecture of transcript-level variation in Arabidopsis.
Genetics 175: 1441-1450

Wittkopp PJ, Haerum BK, Clark AG (2004) Evolutionary changes in cis and
trans gene regulation. Nature 430: 85-88

Xin Z, Browse J (1998) Eskimo1 mutants of Arabidopsis are constitutively
freezing-tolerant. Proc Natl Acad Sci U S A 95: 7799-7804

Xin Z, Mandaokar A, Chen J, Last RL, Browse J (2007) Arabidopsis ESK1
encodes a novel regulator of freezing tolerance. Plant J 49: 786-799

Xiong L, Lee H, Huang R, Zhu JK (2004) A single amino acid substitution in
the Arabidopsis FlERY1/HOSZ protein confers cold signaling specificity
and lithium tolerance. Plant J 40: 536-545

Xiong L, Lee H, lshitani M, Tanaka Y, Stevenson B, Koiwa H, Bressan
RA, Hasegawa PM, Zhu JK (2002) Repression of stress-responsive
genes by FIERY2, a novel transcriptional regulator in Arabidopsis. Proc
Natl Acad Sci U S A 99: 10899-10904

Yamada T, Kuroda K, Jitsuyama Y, Takezawa D, Arakawa K, Fujikawa S
(2002) Roles of the plasma membrane and the cell wall in the
responses of plant cells to freezing. Planta 215: 770-778

Yamaguchi-Shinozaki K, Shinozaki K (2006) Transcriptional regulatory
networks in cellular responses and tolerance to dehydration and cold
stresses. Annu Rev Plant Biol 57: 781-803

Yamashita S, Wakazono K, Nomoto T, Tsujino Y, Kuramoto T, Ushijima T

(2005) Expression quantitative trait loci analysis of 13 genes in the rat
prostate. Genetics 171: 1231-1238

174

 

Yanagisawa S (2004) Dof domain proteins: plant-specific transcription factors

associated with diverse phenomena unique to plants. Plant Cell Physiol
45: 386-391

Yang Z (2002) Small GTPases: versatile signaling switches in plants. Plant
Cell 14 Suppl: 8375-388

Yano M, Sasaki T (1997) Genetic and molecular dissection of quantitative
traits in rice. Plant Mol Biol 35: 145-153

Young ND (1996) QTL mapping and quantitative disease resistance in plants.
Annu Rev Phytopathol 34: 479-501

Yvert G, Brem RB, Whittle J, Akey JM, Foss E, Smith EN, Mackelprang R,
Krugiyak L (2003) Trans-acting regulatory variation in Saccharomyces
cerevisiae and the role of transcription factors. Nat Genet 35: 57-64

Zarka DG, Vogel JT, Cook D, Thomashow MF (2003) Cold induction of
Arabidopsis CBF genes involves multiple lCE (inducer of CBF
expression) promoter elements and a cold-regulatory circuit that is
desensitized by low temperature. Plant Physiol 133: 910-918

Zhang X, Fowler SG, Cheng H, Lou Y, Rhee SY, Stockinger EJ,
Thomashow MF (2004) Freezing-sensitive tomato has a functional
CBF cold response pathway, but a CBF regulon that differs from that of
freezing-tolerant Arabidopsis. Plant J 39: 905-919

Zhu J, Dong CI-l, Zhu JK (2007) interplay between cold-responsive gene
regulation, metabolism and RNA processing during plant cold
acclimation. Curr Opin Plant Biol 10: 290-295

Zhu J, Shi H, Lee BH, Damsz B, Cheng S, Stirm V, Zhu JK, Hasegawa
PM, Bressan RA (2004) An Arabidopsis homeodomain transcription
factor gene, H089, mediates cold tolerance through a CBF-
independent pathway. Proc Natl Acad Sci U S A 101: 9873-9878

Zhu JK (2001) Cell signaling under salt, water and cold stresses. Curr Opin
Plant Biol 4: 401-406

Zou F, Gelfond JA, Airey DC, Lu L, Manly KF, Williams RW, Threadgill
DW (2005) Quantitative trait locus analysis using recombinant inbred

intercrosses: theoretical and empirical considerations. Genetics 170:
1299-131 1

Zou W, Aylor DL, Zeng 28 (2007) eQTL Viewer: visualizing how sequence
variation affects genome—wide transcription. BMC Bioinformatics 8: 7

175