;. n.
‘1
. .u .
9‘1:
..

...

,
r, . 3
ELF... 1 F

3.3.4:

.. 1
n. ..
.5.

2.)}!325
. tfl . .1

greets

” 137:4”)

-
a

This is to certify that the
dissertation entitled

INHERENT AND REGULATED MRNA

Doctoral

 

STABILITY IN A. THALIANA

presented by

Rodrigo Antonio Gutierrez

has been accepted towards fulfillment
of the requirements for the

degree in Biochemistry and Molecular
Biology

Major Pflessor’s Signature
December 16, 2002

Date

MSU is an Affinnative Action/Equal Opponunity Institution

 

 

LIBRARY l
Michigan State
University

 

 

 

PLACE IN 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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIFiC/DateDue.p65-p.15

 

INHERENT AND REGULATED MRNA STABILITY IN A.
T HALIA NA

By

Rodrigo Antonio Gutierrez

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Graduate Program in Biochemistry and Molecular Biology

2003

ABSTRACT

INHERENT AND REGULATED MRN A STABILITY IN A.
T HALIANA

By

Rodrigo Antonio Gutierrez

mRNA degradation provides a powerfirl means for controlling gene expression
during growth, development and many physiological transitions in plants and other
systems. Enormous advances have been made in the understanding of the mRN A decay
process, particularly related to the control of rapid transcript turnover. However there is
still limited knowledge regarding the nature and associations of genes with unstable
transcripts fi'om a genomic perspective, or the physiological significance of rapid mRNA
turnover in intact organisms. To address these questions, cDNA microarray technology
was applied to identify and characterize genes with unstable transcripts in Arabidopsis
thaliana (AtGUTs). At least 1% of the 11,521 clones represented on Arabidopsis
Functional Genomics Consortium microarrays correspond to transcripts that are rapidly
degraded, with estimated half-lives of less than 60 min. Analysis of public microarray
expression data for these genes indicates that mRNA instability is of high significance
during plant responses to mechanical stimulation and is associated with specific genes
controlled by the circadian clock. Control of mRNA stability has often been proposed as
a component of circadian gene expression. However there is no direct evidence that the

rate of mRNA turnover can be regulated by the circadian clock in plants or any other

system. In this dissertation it was shown that mRNA stability for two AtGUTs, Cor-like
and SEN], changes throughout the day and that this change is commanded by the clock.
Furthermore, it was found that DSTI gene function was required for the normal diurnal
oscillation of these genes. The evidence obtained indicates a previously unknown
connection between expression of clock-controlled genes and the DST-mediated mRNA
decay pathway. In contrast to the current understanding of rapid mRNA decay, little is
known regarding the determinants for the long half-life of extremely stable transcripts in
plants or other systems. One of the reasons that the study of stable mRNAs has lagged
behind studies of unstable transcripts in plants, is that current methods are not well suited
for the study of long-lived mRNAs. The significance in plants of the best known stability
determinant, the pyrimidine-rich sequence fiom the human a-globin transcript, was
evaluated. This and a related synthetic sequence were fused to reporter transcripts and
expressed in tobacco cell cultures, maize cell cultures and in Arabidopsis plants. The
results obtained indicate that these sequences are not recognized as stability elements in
plants. To facilitate identification of plant stability elements, and to improve future
studies of stable as well as unstable transcripts in plants, a novel regulated promoter
system that allows for transcriptional-pulse type experiments in Arabidopsis was
developed. This system makes use of the natural transcriptional regulatory properties of
the At-EJG’LI gene. The promoter of this gene is reproducibly and transiently induced
producing a synchronized population of transcripts that can be monitored overtime. This
tool should be useful for the study of specific aspects of the mRN A degradation process

of unstable as well as stable transcripts in Arabidopsis

To Julia, mother of father, my inspiration and model of perseverance.

To my family, thank you all for your patience and support

A Julia, madre de padre, mi inspiracién y modelo de perseverancia.

A mi familia,gracias por la paciencia y el apoyo.

IV

ACKNOWLEDGMENTS

Many thanks are due at the end of this adventure. But because my memory is
fragile and because I want to get rid of this manuscript “asap” I apologize up front for the

names I will never print in these two pages.

First in my mind, I would like to thank my advisor Dr. Pam Green, for her superb
guidance throughout my doctoral program at the MSU-DOE Plant Research Laboratory.
For helping me discern what was most scientifically interesting, for giving me freedom to
pursue all my ideas, and especially for the time spent to improve my presentation and
writing skills. “A gene codes for" or a “Gene encodes ” not a “Gene encode for”. .. just
one of the many suggestions that are now part of my vocabulary (or certainly hope so)

and English thinking process.

I want to thank all past and present members of the Green lab which I was fortunate
to interact with: Dr. Yukako Chiba, Dr. Jay de Rocher, Dr. Scott Diehn, Linda Danhof,
Dr. Michael Feldbriigge, Dr. Mark Johnson, Dr. Jim Kastenmayer, Preet Lidder, Dr.
Nikki LeBrasseur, Dr. Gustavo MacIntosh, Dr. Miguel Perez and Dr. Ambro van Hoof.
For great discussions, comments, and for making the “Green lab” a stimulating and
thought provoking environment for research. I want to especially acknowledge Dr. Scott
Diehn for teaching me the basic RNA techniques when I started in the lab and his
guidance during my rotation. I would also like to thank Dr. Mark Johnson, Dr. James
Kastenmayer, and Dr. Ambro van Hoof for critical comments of the two manuscripts that
were published during this dissertation. I want to thank Linda Danhof, for providing all

the lab basics, keeping everything running smooth, and for excellent technical support. I

should also especially acknowledge the patience of Nikki and Preet, for bearing with the
“drumming” that went on during my incubations, spins, precipitations, reading, writing,
thinking, etc. in the back room. I would also like to thank the members of the Arabidopsis
Functional Genomics Consortium for technical advice during my rnicroarray experiments
and especially Dr. Robert Schaffer, Dr. J eff Landgraf and Dr. Ellen Wisman for great
discussions.

I want to thank the members of my PhD guidance committee: Dr. Christoph
Benning, Kenneth Keegstra, Michael Thomashow and Steven Triezenberg; for great

comments and professional advice during my PhD studies.

I would like to thank Dr. Tom Newman for providing and sequencing EST clones. I
would also like to thank our collaborators, Dr. Rob Ewing and Dr. Mike Cherry for
assistance in the sequence analysis of AtGUTs by the oligomer counting method.

I also want to thank everybody in the Plant Research Laboratory, for sharing your
knowledge and skills whenever I needed it. This is a great scientific community and I
have benefited greatly from being part of it. I also want to thank the Plant Research
Laboratory staff for making paperwork, ordering, mailing, meeting and course
registrations, reimbursements, etc. so much easier.

I also want to thank many people in the Biochemistry and Molecular Biology
Department, for also sharing knowledge and research advice whenever I needed it. And I
want to thank the Biochemistry and Molecular Biology staff for facilitating my visa

paperwork, keeping track of my PhD requirements etc.

VI

I want to thank the Graduate School for the financial support that help me attend the
course “Bioinformatics: Writing Software for Genome Research” at the Cold Spring
Harbor Laboratory.

Last but definitely not least, I want to thank my beautiful wife Maite, for her
tremendous support and patience during these years. Thanks for helping me go through
the difficult times and for celebrating the successes. Thanks also to mom, dad, brothers...
all my family in Chile, which despite disagreeing on the benefits of being so far away,
lend me unconditional support throughout my PhD.

1 would also like to acknowledge the funding agencies that made possible this
research, grants from the DOE (DE-FG02-91ER20021), USDA (9801498), USDA (2000-
01491), NSF (DBN987638), NSF (IBN9408052) and Michigan State University

Research Excellence Fund to Dr. Pam J. Green.

VII

TABLE OF CONTENTS

ABSTRACT ....................................................................................................................... II
DEDICATION .................................................................................................................. III
ACKNOWLEDGMENTS .................................................................................................. V
TABLE OF CONTENTS ............................................................................................... VIII
LIST OF TABLES ............................................................................................................ XI
LIST OF FIGURES .......................................................................................................... XII
ABBREVIATIONS ........................................................................................................ XIV
CHAPTER 1 ........................................................................................................................ 1
mRNA stability in plants and genomic approaches to study mRNA decay
Molecular determinants of mRNA stability ........................................................................ 4
Sequence elements that control inherent mRN A stability ............................................ 6
Differential control of mRNA stability ...................................................................... 11
Stable mRNAs in plants ............................................................................................. 18
Genomic approaches for the study of mRNA decay ......................................................... 19
New insights into mRNA stability derived from microarray studies ......................... 20
Common themes and new trends in mRNA decay ..................................................... 24
Future prospects .......................................................................................................... 27
References ......................................................................................................................... 28
CHAPTER 2 ...................................................................................................................... 33

Identification of unstable transcripts in Arabidopsis by cDNA microarray analysis: Rapid
decay is associated with a group of touch- and specific clock-controlled genes.

Introduction ....................................................................................................................... 34

Materials and Methods ...................................................................................................... 37
Half-life measurements and preparation of RNA samples. ........................................ 37
Hybridization of cDNA Microarrays .......................................................................... 37
Microarray data analysis ............................................................................................. 38
Sequence and gene expression analysis ...................................................................... 39

Results and Analysis ......................................................................................................... 42
Monitoring mRNA stability using cDNA microarrays. ............................................. 42
At least 1% of clones on the 1 1K Arabidopsis microarrays correspond to unstable
messages ..................................................................................................................... 43

General structural features of genes with unstable and stable transcripts are similar.53
AtGUTs are predicted to play a role in a broad range of cellular processes but most

prominently in transcription. ...................................................................................... 54

Rapid mRN A degradation is associated with ArabidOpsis responses to mechanical

stimulation and circadian rhythms .............................................................................. 57
References ......................................................................................................................... 63

VIII

CHAPTER 3 ...................................................................................................................... 67
Circadian rhythms and control of mRNA decay: Oscillation of the Arabidopsis Cor-like
and SEN] transcripts is dependent on normal DST-mediated mRN A degradation

Introduction ....................................................................................................................... 68
Materials and Methods ...................................................................................................... 71
Arabidopsis strains and growth conditions ................................................................. 71
Half-life measurements and preparation of RNA samples. ........................................ 71
Results ............................................................................................................................... 72
Stability of Ccr-like and SEN] mRNAs changes during the day. .............................. 72
Cor-like and SEN] mRNA stability changes are dictated by the circadian clock. ..... 76
DST 1 fimction is involved in the normal oscillatory expression of Cor-like and SEN]
genes. .......................................................................................................................... 80
The effect of the dst] mutation is specific to a subset of DSTl targets ..................... 82
Diurnal expression of other CCGs is not compromised in the dst] mutant. .............. 85
Regulation of SEN] mRNA stability is defective in the dst] mutant ......................... 87
Discussion ......................................................................................................................... 89
References ......................................................................................................................... 95
CHAPTER 4 ...................................................................................................................... 99
Mammalian determinants of long mRNA half-life and their significance to plants
Introduction ..................................................................................................................... 100
Materials and methods .................................................................................................... 103
Plant materials and culture ....................................................................................... 103
Gene constructions ................................................................................................... 103
Protoplast preparation and transformation ............................................................... 104
Plant transformation ................................................................................................. 105
RNA isolation and northern blot analysis ................................................................. 106
Results ................................................................................................................. . ........... 107
or-Globin stabilization element and a related sequence do not increase abundance of a
reporter transcript in maize or in tobacco protoplasts .............................................. 107
a-Globin stabilization element and a related sequence do not increase abundance of a
reporter mRNA in transgenic Arabidopsis seedlings ............................................... 108
Discussion ....................................................................................................................... 1 10
References ....................................................................................................................... 115
CHAPTER 5 .................................................................................................................... 118

Promoter region of the At-EXPLI gene drives transient expression of reporter transcripts:
A new regulated promoter system to study mRNA degradation in Arabidopsis.

Introduction ..................................................................................................................... 119
Materials and Methods .................................................................................................... 123
Arabidopsis strain and growth conditions ................................................................ 123
Plasmid constructs .................................................................................................... 123
Arabidopsis transformation ...................................................................................... 124

Half-life measurements with a transcriptional inhibitor ........................................... 125

Half-life measurements using the At-EXPL] promoter: ........................................... 125
Results ............................................................................................................................. 126
At-EXPL] gene is transiently induced ...................................................................... 126
Transient induction of At-EXPL] is highly reproducible ......................................... 127
At—EXPL] gene transcription is severely down-regulated 45 min after its induction
.................................................................................................................................. 128
Promoter region of At-EXPLI gene drives transient expression of reporter transcripts
.................................................................................................................................. l 30
At-WLI promoter system allows measurement of mRNA stability. .................... 132
Discussion ....................................................................................................................... 137
References ....................................................................................................................... 141
CHAPTER 6 .................................................................................................................... 144

Final remarks and challenges ahead.

References ....................................................................................................................... 149

LIST OF TABLES

Table 2.1. Arabidopsis thaliana genes with unstable transcripts (AtG UT s) ................ 44
Table 2.2. Arabidopsis genes with unstable messages that belong to the MIPS

transcriptional category (04) as of May 2002 ................................................... 56

Table 2.3. Comparison of gene expression data for all genes in the AF GC 11K
microarray and AtGUTs in selected treatments ................................................ 58
Table 4.1. Examples of Arabidopsis genes with stable transcripts ......................... l 13
Table 4.2. Sequence elements located downstream of the translation stop codon can
increase mRNA abundance of endogenous as well as reporter genes in different plant
systems ............................................................................................. 114
Table 5.1. Summary of mRN A stability measurements with regulated promoter

system .............................................................................................. 136

XI

LIST OF FIGURES

Figure 1.1. A conceptual framework of mRNA stability in eukaryotic cells ..................... 3
Figure 1.2. Summary of sequence elements that have been demonstrated to control
mRNA stability in plants. .................................................................................................... 6

Figure 2.2. Confirmation of the instability of transcripts identified by microarray analysis.

........................................................................................................................................... 53
Figure 2.3. Instability is associated with a broad range of plant processes ....................... 55
Figure 2.4. Cluster analysis indicates that a set of AtGUTs is induced by mechanical

stimulation (touch) and another is controlled in a diurnal fashion. ................................... 59
Figure 3.1. Ccr-like mRN A stability is regulated during the day. .................................... 73
Figure 3.2. SEN] mRNA stability is regulated during the day. ........................................ 74
Figure 3.3. NIA2 mRNA stability is comparable in the morning and in the afiemoon ..... 75

Figure 3.4. LHY mRNA is highly expressed in the morning and decreases to background
levels in the afiemoon. ...................................................................................................... 77
Figure 3.5. Cor-like mRNA stability is regulated by the Arabidopsis circadian clock ..... 78
Figure 3.6. SEN] mRNA stability is regulated by the Arabidopsis circadian clock. ........ 79
Figure 3.7. LHY mRN A is highly expressed in the subjective morning and decreases to

background levels in the subjective afiemoon. NIA2 mRN A stability is comparable in the

two conditions. .................................................................................................................. 81
Figure 3.8. Diurnal oscillation of Cor-like mRN A is altered in the dst] mutant. ............. 83
Figure 3.9. Diurnal oscillation of SEN] mRNA is altered in the dst] mutant. ................. 84

Figure 3.10. Diurnal oscillation of SA UR-A CI, AtGRP7/CCR2, CCAI and LHY in dst]

mutant and 1519 parental plants ........................................................................................ 86

XII

Figure 3.11. Regulation of SEN] mRNA stability is altered in the dst] mutant ............... 88
Figure 4.1. Reporter system for testing putative stabilization elements derived from
mammalian transcripts. ................................................................................................... 104
Figure 4.2. The oc-globin stabilization element and the synthetic poly(CCCU) sequence
do not increase the abundance of a reporter transcript in maize or in tobacco cells. ...... 108
Figure 4.3. The a-globin stabilization element and the synthetic poly(CCCU) sequence

do not increase the abundance of a reporter transcript in transgenic Arabidopsis plants.

......................................................................................................................................... 109
Figure 5.1. Transient induction of the At-EMDLI gene. .................................................. 126
Figure 5.2. At—EXPLI gene expression is highly reproducible. ...................................... 128

Figure 5.3. At—EXPL] mRNA disappears with similar speed in the presence or absence of
cordycepin. ...................................................................................................................... 129
Figure 5.4. At-EXPL] promoter region drives transient expression of two globin reporter
mRNAs in transgenic Arabidopsis plants. ...................................................................... 131
Figure 5.5. At-EXPL] promoter system can be used to study mRNA degradation in

Arabidopsis ...................................................................................................................... l 34

XIII

AF GC

ARES

At-EXPL]

AtGUT

CCG

cDNA

CT

DNA

DST

GTB2

MEME

mRNA

ORF

SMD

UTR

ZT

ABBREVIATIONS

: Arabidopsis Functional Genomics Consortium
: Adenylate/uridylate-rich elements

: Arabidopsis thaliana expansin-like gene 1

: Arabidopsis thaliana gene with unstable transcript

Clock-controlled gene
Complementary DNA
Circadian time
Deoxyribonucleic acid
Downstream element
: quality control parameter that indicates the fraction of pixels in one spot

in a microarray slide that have intensity values 1.5 times the background.

Multiple Expectation Maximization for Motif Elicitation
Messenger RNA

Open reading frame

Ribonucleic Acid

Stanford Microarray Database

Untranslated region.

Zeitgeber (time-giver) time.

XIV

CHAPTER 1

mRNA stability in plants and genomic approaches to study mRNA

decayI

 

' Part of this chapter was published in “Gutierrez, R.A., MacIntosh, G.C, and Green R]. (1999). Current
perspectives on mRNA stability in plants: multiple levels and mechanisms of control. Trends Plant Sci
4:429-438”.

Normal growth and development as well as the ability to adjust to ever changing
environmental conditions requires the carefully regulated expression of many genes.
Although much of this regulation is exerted at the transcriptional level, post-
transcriptional mechanisms also play a fundamental role. For some genes, post-
transcriptional mechanisms constitute the predominant form of control in response to a
given stimulus. In other cases, an extra level of modulation is provided by post-
transcriptional control that increases the flexibility and speed of responses beyond what
could be achieved through transcriptional regulation alone. The control of mRN A
stability is one of the most prominent forms of post-transcriptional regulation in
eukaryotic cells. The stability of a particular mRN A influences its steady-state levels and
directly affects the rate of its induction or repression following a change in transcription.
Thus, a thorough understanding of how mRNA stability is controlled is essential to
elucidate how the abundance of endogenous mRNAs is governed and to optimize the
accumulation of transgene mRNAs in plants for biotechnological applications.

As illustrated in Figure 1.1, the molecular components that control mRNA
stability can be considered in three layers. Recent work in yeast indicates that eukaryotic
cells contain RN A-degrading activities and protein cofactors that appear to constitute the
general/basal mRNA decay machinery, responsible for the degradation of most stable and
unstable mRNAs. Superimposed on this basal machinery are the sequence-specific
controls that dictate the inherent stability of various mRNAs, the half-lives of which can
vary over a wide range. For those transcripts whose stability changes in response to
exogenous or endogenous stimuli, a third layer of control must be evoked. This last layer

would transducc the signals elicited by various stimuli into changes in mRN A turnover.

Theoretical Hierarchy

Differentially regulated mRNA stability

Stimulus —> —> -> ——> —> —> stabilization or destabilization

 

Inherent mRNA stability

Sequence-specific recognition

Ultra stable Stable Unstable
(days) (hours) (minutes)

 

General mechanisms

Basal decay machinery

Figure 1.]. A conceptual framework of mRNA stability in eukaryotic cells. The molecular
components that control mRNA stability can be considered in three inter-related layers.
According to this framework, the underlying layer contains RNA-degrading activities and protein
cofactors that constitute the general/basal mRN A decay machinery, responsible for the
degradation of most mRNAs. Superimposed on this basal machinery are the sequence-specific
components, represented by the second layer, that dictate the inherent stability of different
mRNAs. mRNA half-lives (indicated in parenthesis) can vary over a wide range, with the average
on the order of hours. The third layer of control would facilitate the transduction of signals into
changes in mRNA turnover to adjust the stability of transcripts in response to exogenous or
endogenous stimuli.

In this conceptual framework, investigation of all three layers is critical because of their
individual importance and the probable inter-relationships among them. For example,
differential control of the stability of a particular mRNA could be mediated by
modulating the activity of a sequence-specific recognition factor, which interacts with the
basal decay machinery. In this dissertation however, attention will be devoted to the two

uppermost layers.

Most recent progress in our understanding of mRNA stability in plants has
emerged from studies of sequence-specific recognition of transcripts for rapid and/or
regulated decay. This chapter will highlight current knowledge for those nuclear encoded
transcripts that have been studied in the most detail. Readers are referred to several
previous reviews for more comprehensive presentations of the mRN A stability and post-
transcriptional control literature in plants (Gallic, 1993; Abler and Green, 1996; Johnson
et al., 1998) and other eukaryotes (Ross, 1995; McCarthy, 1998; Mitchell and Tollervey,

2000; Tucker and Parker, 2000; Wilusz et al., 2001 ).

Molecular determinants of mRNA stability

The decay rate of transcripts in plants seems to be similar to those observed in
other multicellular eukaryotes. Half-lives range from less than an hour for unstable
messages, to days or more for stable transcripts with the average being on the order of
several hours (Johnson et al., 1998; Taylor and Green, 1995). The decay rates of some
transcripts can be rather dynamic, being modulated by the coordinated integration of
internal and external stimuli. What then, are the molecular determinants that control the
half-life of a particular transcript at any time in the cell? In recent years, research has
mainly focused on the identification and characterization of structural features of the
mRN A molecule, or cis-acting elements, that influence mRNA decay rates. These studies
have shown that general structural elements found at the ends of virtually all mRNAs, as
well as specific sequence elements located within a transcript, can all contribute to the

overall stability.

In addition to their role as translational enhancers, the 7-methyl-G cap at the 5'
end and the polyadenylate (poly(A)) tail at the 3' end have been shown to increase mRN A
stability in transient assays (Gallic, 1998). By electroporating capped or uncapped
mRNAs, and mRNAs with or without poly(A) tails into tobacco protoplasts, it was found
that the 5’ cap stabilizes reporter transcripts by two- to four-fold and the poly(A) tail by
two- to three-fold (Gallic, 1991). Although it is yet unclear how the cap and poly(A) tail
protect a transcript from degradation, an appealing model is that the physical interaction
between the cap and poly(A) tail, via their associated factors (e. g. poly(A) binding
protein, eIF4G, eIF4E, eIF4B) would sequester the ends of the mRNA protecting them
fi'om the action of nucleases (Gallic, 1998).

Within the body of the mRNA, specific sequence motifs that are only present in a
subset of transcripts can act constitutively to establish the inherent instability (or stability)
of a particular transcript or they can modulate the stability of an mRN A in response to
certain physiological, developmental or environmental cues. Major examples of both
classes of stability determinant are discussed in the following sections. Although they are
presented separately, it should be noted that the division is organizational rather than
biological. Some sequences that appear to affect mRNA decay rates constitutively may
be later found to be regulated under special conditions. Conversely, regulatory sequences
may also contribute to inherent stability in the absence of stimuli. In any event, the
characterization of these sequences is already leading to mechanistic insights as to how

they are recognized in the cell and how that recognition may be controlled.

A. SA UR 15A DST clement

(,_Cj_G_AgactgacATAQATngaggagacAt_'l_"l‘t§TAtaata)2

B. SA UR-ACI 3’UTR

AGTACTATACTACAACATTTCCATAI I I I I I I IAGAI IGTIAGCTAAT'I'I‘

 

CCCCTGGAGATAA'ITGTAAATTGTI‘TCAATGAGAGGAATATAC AATA

CA I AGATCGTAATTGATCAATGCGTAT'ITGCATGTI‘

 

Figure 1.2. The DST sequence element. (A) The prototype DST clement derived from the
soybean SA UR 15A gene is shown as DNA sequence. Highly conserved residues across different
species are shaded and invariant residues underlined. Mutational analysis identified important
residues (bold) for DST function (see text for detail). (B) The SA UR-A C] 3’UTR shown as DNA
sequence includes one DST element (shaded) and several ATAGAT-like (underlined) and GTA-
like (upper bar) subdomains that may contribute to its instability function.

Sequence elements that cogtrol ir_rherent wA stabfly

The DST element. The DST or downstream element was originally identified as a
conserved region in the 3 ’ untranslated region (UT R) of the unstable small guxin—gp
RNA (SA UR) transcripts (McClure and Guilfoyle, 1989). It consists of three highly
conserved subdomains separated by two variable regions (Figure 1.2a). When a synthetic
dimer of the soybean SA UR-15A DST sequence was placed in the 3’UTR of a reporter
transcript, its turnover was significantly faster than that of a spacer or no-insert control in
BY-2 cells (Newman et al., 1993). Subsequent mutational analysis indicated that two
conserved subdomains, designated ATAGAT and GTA regions after the invariant

nucleotides they contain, are critical for DST function. Five— and six-base substitutions in

the ATAGAT and GTA regions respectively, resulted in slower turnover rates in BY-2
cells and higher reporter transcript accumulation in transgenic tobacco plants (Sullivan
and Green, 1996). Two base substitution mutations within these two subdomains
indicated that the first four bases of the ATAGAT subdomain are critical for instability
function in tobacco cell culture. Interestingly, a 2-base substitution in the GTA
subdomain inactivated DST function in transgenic tobacco leaves but not in cell culture.
This finding suggests that the DST element may be differentially recognized in different
cell types (Sullivan and Green, 1996).

Detailed studies of SA UR gene expression in Arabidopsis thaliana have been carried
out on the SA UR-A C] gene. By examining the expression of chimeric genes it was shown
that the promoter region of the gene is responsible for auxin induction and that sequences
downstream of the promoter limit mRNA accumulation in an auxin-independent manner
(Gil and Green, 1996). Measurements of the half-lives of the transcripts encoded by
chimeric genes showed that the 3’UTR acts as a potent mRN A instability determinant
(Gil and Green, 1996) (Figure 1.2b). Interestingly, the SA UR-AC] 3’UTR contains one
canonical DST element and several ATAGAT-like and GTA-like subdomains that may
contribute to instability of the mRNA (Figure 1.2b). This is intriguing since in previous
work two copies of the prototype DST element from SA UR-I5A were needed to cause
instability of a reporter transcript (Newman et al., 1993). Further studies will be
necessary to investigate the contribution of particular sequences to the instability of
SA UR-A C] mRN A, and the importance of context for DST clement function.

The novel structure of DST sequences suggests that they may mediate mRNA decay

through a pathway that is unique to plants. In an effort to address this issue, reporter

transcripts with and without DST sequences were expressed in NIH3T3 fibroblasts
(Feldbriigge et al., 2002). The presence of the DST tetramer accelerated deadenylation
and decay rate of the reporter mRNA in the mammalian cells as compared to control
transcripts lacking DST sequences. However a tetramer soybean DST element mutated in
the ATAGAT domain, which is inactive in tobacco cells, was equally active as the wild-
type version in fibroblasts (Fcldbriigge et al., 2002). Also contrary to what expected, two
reporter transcripts containing different versions of the Arabidopsis DST sequence
element decayed no faster than a control transcript lacking DST sequences. Together
these results indicate that recognition of the DST sequence element in mammalian cells
follow different rules as compared to plants cells.

To understand the molecular mechanisms underlying DST firnction, a genetic
strategy was devised to isolate mutants defective in DST-mediated mRN A degradation
(Johnson et al., 2000). Two independent mutants were isolated, dst] and dst2, that
showed elevated mRN A levels of a hygromycin phosphotransferase and a B-
glucuronidase mRN A each containing four copies of the DST element (HPH-DSTx4 and
GUS-DSTx4 respectively) (Johnson et al., 2000). The dst] and dst2 mutants also
exhibited elevated mRN A levels of the endogenous SA UR-A C] gene (Johnson et al.,
2000). In addition, decay of HPH-DSTx4, GUS-DSTx4 and SA UR-A C] mRNAs was
slower in the dst mutants as compared to parental. Because no morphological or
developmental defects were apparent, DNA microarray analysis was used to investigate
the molecular phenotypes of the dst] mutant (Pérez-Amador et al., 2001). Eighteen of the
approximately 7800 genes represented on the Arabidopsis Functional Genomics

Consortium (AF GC) DNA microarray exhibited increased mRNA levels when

comparing dst] mutant and parental plants. In addition, seven genes with decreased levels
in dst] as compared to the parental plants were also identified. Seven of these twenty five
genes contained DST-like sequences in their 3’ UTR indicating they might be primary
targets of the DST-mediated decay pathway (Pércz-Amador et al., 2001). Surprisingly,
eight out of the twenty five genes were regulated by the circadian clock suggesting a
connection between circadian rhythms and the dst] mutation (Pérez-Amador et al., 2001).
This number is higher than what expected based on current estimates of the total number
of clock-controlled genes in Arabidopsis. DNA microarray experiments have shown that
2 to 6% of Arabidopsis mRNAs can oscillate (Harmer etal., 2000; Schaffer et al., 2001).
The relationship between this sequence-specific mRN A decay pathway and circadian
control of gene expression in Arabidopsis is the subject of the Chapter 3 of this
dissertation. Efforts to clone the dst] gene are currently underway (Lidder & Green

unpublished results).

AUUUA-remats. Adenylatc/uridylate-rich elements (ARES) represent a common
determinant of RNA stability in mammalian cells. Transcripts that contain ARES are
selectively targeted for rapid decay (Chen and Shyu, 1995). ARES are approximately 50-
150 nucleotides long, usually contain multiple copies of the AUUUA motif and a high
content of uridine, and are located in the 3’UTR of mRNAs encoding a variety of proto-
oncoproteins, cytokines and transcription factors (Chen and Shyu, 1995). Accordingly
AUUUA sequences play important roles in the post-transcriptional regulation of gene
expression during processes such as cell growth, differentiation, the immune response,

etc. in mammalian systems. Due to the significance of AUUUA elements in mammals, a

synthetic AUUUA repeat was tested for the ability to act as instability determinant in
plants. Reporter transcripts containing 1 l repeats of the AUUUA motif in their 3’UTRS
were degraded much more rapidly in stably transfonned tobacco cells and accumulated to
a lower level in transgenic tobacco plants than those of control constructs (Ohme—Takagi
et al., 1993) . The effect appeared to be AUUUA-Specific because two other sequences
with the same Size and A+U content had no effect in parallel experiments. These results
suggest that the mRN A decay pathway mediated by AUUUA repeats is conserved
between animals and plants. However, the natural targets of the plant AUUUA-mediated
decay pathway remain to be identified. Possible candidates include the PvPRP] transcript
from Phaseolus vulgaris, aAmy3 transcript from Oryzae sativa (discussed below) and
three genes from Arabidopsis (discussed in Chapter 2). Recent experiments have Shown
that ARE-mediated decay is also present in the yeast Saccharomyces cerevisiae

(V asudevan and Pcltz, 2001). Together these data suggest that ARE-mediated decay is

conserved in eukaryotes, from yeast to mammals.

Nonsense codons. Premature nonsense codons decrease mRN A stability by activating
nonsense-mediated decay pathways in several eukaryotic systems. The yeast nonsense-
mediated decay pathway, discussed further below, iS one of the best understood at the
molecular and genetic levels. Nonsense mediated decay is presumably part of a mRN A
surveillance system that rapidly removes abnormal mRNAs to prevent the formation of
truncated or otherwise potentially detrimental polypeptides (Hilleren and Parker, 1999;

Culbertson, 1999).

10

Initial evidence that nonsense-mediated mRNA decay occurred in plants came
from studies conducted on natural alleles of the soybean Kunitz trypsin inhibitor (Jofuku
et al., 1989) and bean phytohernagglutinin A (PHA) genes (Voelker et al., 1986). Cells
containing alleles with early stop codons accumulate low levels of mRNA (Jofuku et al.,
1989; Voelker er al., 1986) even when transcriptional rates were normal (Jofuku et al.,
1989). Nonsense mediated decay was also found responsible for the reduced mRN A
accumulation of two mutant alleles of the WAXY gene in rice (ISShiki et al., 2001).
Generally nonsense codonS affect mRN A abundance in a position-dependent manner, the
closer to the initiation codon the greater the likelihood of affecting mRN A abundance.
The effect of stop codons positioned at variable distance from the translation start codon
of a reporter gene was directly addressed using the initially isolated PHA allele and other
PHA alleles constructed in vitro (van Hoof and Green, 1996). By measuring mRNA
decay rates in stably transformed tobacco cell lines it was demonstrated that transcripts
with nonsense codons positioned 20, 40 and 60% of the way through the normal coding
region yielded highly unstable mRNAs, whereas a transcript with a nonsense codon at
80% was as stable aS wild type. These findings strongly support the idea that plants have

a nonsense—mediated decay pathway Similar to that found in other eukaryotes.

mflereggal cor_rtrol of nLRN A stability

Light modulation. Light regulation at the post-transcriptional level has been well
documented for the pea photosynthetic electron carrier ferredoxin I (Fed-1) gene (for a
review see Dickey et al., 1998). AS with many other photosynthetic genes, Fed-1

expression is induced by light, mRN A levels being five-fold higher in the light than in

11

darkness. When mRN A decay rates were measured, a two-fold higher mRNA half-life
was observed for the transcript in light versus darkness in transgenic tobacco seedlings
demonstrating that light regulation occurs through a change in mRN A stability (Petracek
et al., 1998). A sequence element was identified within the transcribed region that could
confer light responsiveness to a reporter gene under the control of a constitutive promoter.
This internal light regulatory element (iLRE), spans a portion of the 5’ UTR and the first
20 codonS of the coding region (Dickey et al., 1998). The observations that Fed-1 light-
induced mRNA accumulation correlates with its polyribosomal association, prompted a
model in which efficient translation of F ed-I in the light is associated with increased
mRNA stability (Dickey et al., 1998). This model was also initially argued based on the
observation that nonsense mutations, which block ribosomal progression, abrogate Fed-1
mRNA accumulation in response to light (Dickey et al., 1994; Dickey et al., 1998).
However as discussed above, nonsense mutations trigger rapid mRN A degradation
through the nonsense mediated decay pathway in plants (van Hoof and Green, 1996).
More recent studies have indeed found that normal degradation of Fed-1 mRNA in the
dark occurs through a pathway that is different from that mediated by the presence of
nonsense mutations (Petracek et al., 2000). Further mutation analysis of the iLRE
identified two regions that are critical for its function, a CATT repeat in the 5’ UTR and
the translation initiation region (Dickey et al., 1998). Two different substitution
mutations were made within the CA'IT repeat that blocked F ed-I mRN A accumulation,
one of which affected ribosome loading (Dickey et al., 1998). The Simplest explanation
of these studies is that Fed-1 mRNA is stable in illuminated plants when associated with

polyribosomes. In darkness, inefficient translation renders the transcript less stable

12

through a process involving the CATT repeat located in the 5’ portion of the message
(Dickey et al., 1998). In an effort to identify trans-acting factors that mediate the
regulation of Fed-1 mRNA translation and stability, proteins that bind the iLRE were
isolated (Ling et al., 2000). Among several RNA binding activies that were found to
associate with the iLRE, the heat Shock protein HSP101 was identified. HSP101 was
required to achieve high translation activity of iLRE containing reporter transcripts in
yeast (Ling et al., 2000). In addition, reporter transcripts containing iLRE were more
efficiently translated in plant protoplasts expressing HSP101 than control transcripts
(Ling et al., 2000). This data suggest HSP101 might be important for the light-mediated
translational regulation of Fed-1 . Further studies Should help determine whether the
CA’IT region is a stability or instability determinant, as well as the exact mechanistic

relationship between mRNA stability and translation.

Sucrose regulation. or-Amylases are endo-amylolytic enzymes, which catalyze the
hydrolysis of or-l,4 linked glucose polymers and have an important role in degradation of
starch in higher plants. The expression of the rice or-arnylase gene family is coordinately
induced by sucrose starvation and suppressed by sucrose availability, a process that
depends on both transcriptional and post-transcriptional mechanisms (Sheu et al., 1996).
The sucrose-mediated effect on mRNA stability has been analyzed in detail for one of the
most abundant (It-amylase genes, a4my3. When mRNA decay rates were measured, the
half-life of aAmy3 was about 1.5 h in the presence of sucrose and increased to 6 h in
sucrose-starved cells (Sheu et al., 1996). By examining the expression of chimeric genes

in stably transformed rice cells it was Shown that the wimy3 3’UTR was sufficient and

I3

probably the major determinant for controlling the stability of wimy3 mRNA in response
to sucrose availability (Chan and Yu, 1998b; Chan and Yu, 1998a). Further analysis of
the 3’UTR identified two subdomains, called I and III, that could each fimction as a
sugar—dependent stability determinant (Chan and Yu, 1998b; Chan and Yu, 1998a). In
addition, secondary structure analysis predicted extensive duplex formation in the aAmy3
3’UTR, and interestingly, conserved A/U rich regions were found in the loop of
subdomains I and III (Chan and Yu, 1998b). Whether these A/U rich regions or the
structural motifs that contain them are involved in modulation of mRNA stability in
response to sucrose levels remains to be elucidated. Moreover, as in other cases of
regulated mRNA stability, it is not yet clear whether a trans-acting factor slows down the
turnover of the transcript in the presence of sucrose or Speeds up the turnover in its
absence. Treatment with the translation inhibitor cycloheximide enhanced the
accumulation of aAmy3 transcript in the presence or absence of sucrose (Sheu et al.,
1996). In contrast, cycloheximide did not Si gnificantly affect transcriptional rates of
wimy3 regardless of whether or not the cells were provided with sucrose (unpublished
results cited in (Sheu et al., 1996)). These observations might suggest that labile proteins
are involved in aAmy3 mRNA decay. However cycloheximide may interfere with the
normal decay of the message in other ways, e. g. translation of the message may be

required for degradation to take place.

Methionine rggglation. Cysthatione-y-synthase (CGS) catalyzes the first committed step

of the biosynthesis of the amino acid methionine (Met) and is thought to be the major Site

of regulation for the pathway in plants (Ravanel et al., 1998). Regulation of CGS occurs

14

primarily at the level of gene expression and not through metabolic control of the enzyme
activity (reviewed in (Ravanel et al., 1998)). To elucidate the molecular mechanism
underlying regulation of Met biosynthesis, Chiba et a]. (1999) characterized a mutant that
accumulates high levels of soluble Met (Inaba et al., 1994). This mtoI-I mutant also
showed high levels of CGS mRN A, protein and enzyme activity as compared to wild-
type (Chiba et al., 1999). Wild-type plants respond to Met addition by destabilizing CGS
mRNA. However degradation of CGS mRNA in the mic] -1 mutant background was
Slower than in the wild-type and was not affected by the presence of Met (Chiba et al.,
1999). Sequence analysis of the mtoI-I and four other alleles of the mto] locus revealed
single base changes that altered the amino acid sequence in the N-terminus of the CGS
protein (Chiba et al., 1999). The region that contained these mutations (MTOl) was
necessary and sufficient to confer Met responsiveness to reporter transcripts in transient
expression experiments (Chiba et al., 1999). In addition, Silent mutations in the W0]
region did not affect Met response indicating that the nucleotide sequence is not
important for the regulation. Detailed analysis of the first exon of CGS identified the
sequence (A)RRNCSNIGVAQ(I) (with uncertainty in the first and last position) as
required for the feedback regulation mediated by Met (Ominato et al., 2002).
Interestingly, the decrease in CGS mRNA levels following addition of the amino acid
correlated with accumulation of a mRN A decay intermediate truncated at the 5’end
(Chiba et al., 1999). Together, this evidence suggests a mechanism in which translation
of the first exon of CGS in the presence of Met, including the (A)RRNCSNIGVAQ(I)

sequence, would destabilize its own mRNA generating a 3’ end fragment (Chiba et al.,

1999). A similar model has been proposed for autorcgulation of the B-tubulin gene in

15

mammalian systems by unassembled B-tubulin subunits ('Iheodorakis and Cleveland,

1993)

Biotic stress. One of the best examples of modulation of mRN A stability in response to
biotic stress (commonly a result of infection from bacteria, fungi or viruses) has been
characterized in common bean cells. Fungal elicitor treatment of bean cells results in
down-regulation of the PvPRPl gene, which encodes a cell wall proline-rich protein32.
Direct proof that the major control mechanism of this down-regulation is modulation of
mRNA stability was provided by the observation that PvPRPl mRNA half-life in the
presence of the elicitor was shorter than in its absence. Moreover transcriptional rates
remained constant regardless of the presence or absence of the elicitor (Zhang et al.,
1993). Subsequent studies identified a 50 kDa protein (PRP-BP) that can be specifically
crosslinked to the 3’ UTR of the PvPRPI transcript (Zhang and Mehdy, 1994). Using
deletion analysis, the binding Site for PRP-BP was mapped to a 27 nt, U-rich site that
contains one copy of the AUUUA motif. It remains to be demonstrated that this binding
Site iS important for the regulation of transcript stability. Nevertheless the observation
that PvPRP] mRNA degradation, in response to fungal elicitor treatment, was preceded
by increased PRP-BP binding activity in bean cells suggests that this protein and the cis-
element it binds are involved in the regulation (Zhang and Mehdy, 1994). PRP-BP
activity in vitro was increased by the reducing agents DTT or B-mercaptoethanol and
reversibly eliminated with the -SH oxidizing agent diamide or the -SH alkylating agent
N-methylmaleimide (Zhang and Mehdy, 1994). The defense response in bean and many

other Species is accompanied by production of active oxygen Species and other redox

16

perturbations. Hence, these observations suggest that PRP-BP binding activity could be
modulated by the redox changes that take place during the plant defense response

(Mehdy and Brod], 1998; Zhang and Mehdy, 1994).

Other stimuli. Hormones play an indisputable role in the regulation of a multitude of
physiological and developmental processes in plants. Although it is clear that hormones
can influence gene expression at both transcriptional and post-transcriptional levels, a
detailed understanding of the molecular basis of hormone action, especially at the post-
transcriptional level is lacking. One recent example of hormonal regulation of mRN A
stability arose during a study of cytokinin effects on the soybean mRNA Ciml 35. The
predicted Ciml protein product is related to a group of proteins termed B-expansins,
which are involved in cell wall expansion during the vegetative and/or reproductive
phases of plant development. Cim] mRN A abundance increases 20-60-fold upon addition
of cytokinin to cytokinin-starved soybean suspension cultures. When half-life of the
Cim] mRNA was determined following actinomycin D treatment, cytokinin addition to
cytokinin-starved soybean cells increased the mRNA half-life of Cim] about 4-fold
(Downes and Crowell, 1998). Further experiments were aimed to characterize the role of
protein phosphorylation/dephosphorylation in cytokinin-mediated induction of Cim] . It
was observed that accumulation of Cim] message was stimulated by staurosporine
(kinase inhibitor) in the absence of cytokinin and inhibited by okadaic acid (phosphatase
inhibitor) in the presence of cytokinin. These results suggest a role for protein
phosphatases in cytokinin regulation of Cim] mRN A abundance (Downes and Crowell,

1998).

17

In addition to the examples described above, a number of other mRNAs have
been reported to Show modulation of mRN A stability in response to biotic stress, abiotic
stress (e.g. cold, heat, salinity), hormone treatments, etc. In the majority of these cases,
conclusions have been drawn after finding a poor correlation between the rate of
transcription and mRN A accumulation in response to the stimulus. However most of this
research is still at a preliminary stage and the mechanisms through which this modulation
is achieved have not yet been reported (for reviews see (Gallic, 1993; Johnson et al.,

1998; Mehdy and Brodl, 1998; Marcotte, 1998)).

Stable mRNAs in plants

Although it is clear that unstable mRNAs contain instability sequences, no
discrete stabilizing determinant has been demonstrated to be responsible for the long half-
life of an extremely stable transcript in plant systems. The search for mRN A stabilization
sequences has lagged behind those of destabilizing elements in eukaryotes in general, but
at least one example has been well characterized in mammals. Studies aimed at
understanding the mechanism for the selective stabilization of the (it-globin message
during erytlrroid cell development, identified a pyrimidine-rich sequence in the 3'UTR
that was responsible for the long half-life of this transcript (reviewed in (Russell et al.,
1997)). This finding negates a previous idea that all mRNAs are stable by default rather
than by the presence of stabilizing sequences. Further, it makes it likely that stabilizing
sequences exist in other systems as well. Specific mRNA sequences could, for example,
contribute to the stability of seed storage protein mRNAs (Marcotte, 1998), as in the case

of wild cultivars of oat (Johnson et al., 1999), perhaps by influencing their

18

compartrnentalization. Studies conducted in these and other plant systems may provide
insights into the mechanisms of mRNA stabilization. The identification of the sequence
elements, and trans-acting factors they interact with, as well as understanding the
mechanisms of their function could provide tools to improve transgene expression in crop
plants, and would certainly contribute to a more complete understanding of mRN A

metabolism.

Genomic approaches for the study of mRNA decay

While enormous advances have been made in understanding the mRNA decay
process and the determinants of mRNA stability, basic questions remain unanswered. For
example, how many unstable transcripts are there in the cell? What is the physiological
Significance of rapid mRN A turnover? Further, most of what we now know about the cis-
and trans-acting factors involved in mRN A decay derives fiom studies on a relatively
small group of model genes. Therefore the relevance of these studies for the whole cell or
intact organism is unclear.

In an effort to address these questions several groups have developed genome-scale
approaches to study mRN A degradation in E. coli (Bernstein et al., 2002; Selinger et al.,
2002), S. cerevisiae (Wang et al., 2002) and A. thaliana (Gutierrez et al., 2002),
discussed in Chapter 2 of this dissertation). Experimentally all these studies combine time
courses performed after general transcription is arrested with the highly parallel power of
microarray technology. Thus it is possible to monitor genome-wide mRN A
disappearance over time. The experimentally measured decline in mRN A levels can then

be used to estimate the stability of each of the transcripts represented on the array.

19

New insights into mRN A stability derived from micrgrray studies

The first indication that this approach held promise came from a study carried out in
Richard Young’s laboratory in 1998 (Holstege et al., 1998). Holstege et a]. (1998) used
oligonucleotide arrays to analyze global mRNA levels in yeast carrying mutations in
components of the yeast transcription initiation machinery. In one of their experiments,
inactivation of the therrno-sensitive RNA polymerase II in the rpr-] mutant background
(N onet et al., 1987) allowed them to estimate half-lives for more than 5000 yeast mRNAs

(ht_tp://web.wi.mit.edu/young/expression/) (Holstege et al., 1998).

More recently, Bernstein et a1. (2002) used Spotted DNA microarrays to study

 

mRNA degradation in Escherichia coli. Rifampicin was used to prevent initiation of new
transcripts in cells growing at 30 °C in LB and M9 + 0.2% glucose media. Overall, the
mRN A half-life distributions were Similar in the two growth conditions, despite a 3-fold
Slower generation time in the M9 + glucose media. mRN A half-lives ranged fi'om 1-2
min for the most unstable to 10 min or more for the most stable with approximately 80%
of all mRNAs exhibiting half-lives between 3 and 8 min. An exponential decay model
was used to fit the time-series data and only those with a regression coefficient greater
than 0.7 were considered for further studies. This and other filtering criteria allowed half-
life measurements for 3835 transcripts with a mean of 5.7 min in M9 + glucose and 2267
transcripts with a mean of 5.2 min in LB media. Several aspects of the transcripts
measured were analyzed. Interestingly and contrary to what was anticipated, mRN A
stability was found to be a poor predictor of mRNA abundance (Bernstein et al., 2002).

Although there were cases in which stable mRNAs did accumulate to high levels, no

20

positive correlation between transcript stability and abundance was observed.’In an effort
to understand the determinants of mRNA stability in E. coli, structural features were also
inspected. The enzyme RNascE catalyzes the rate-limiting endonucleolytic cleavage that
initiates decay of many mRNAs in E. coli (Coburn and Mackic, 1999). However the
density of putative Sites of cleavage by RN aseE was not predictive of mRN A stability.
Secondary structures are known to Slow the action of mRNases in E. coli (Bouvet and
Belasco, 1992; Grunberg-Manago, 1999), but the fiee energy of folding, G/C content, or
length of 5’ or 3’ UTR sequences could not be correlated with mRNA half-lives. Further,
no correlation could be detected between half-lives and open reading frame (ORF) length,
operon length or codon usage. Interestingly, despite the fact that stability did not correlate
with obvious molecular characteristics it seemed to be Similar in transcripts that encode
related metabolic functions. For example, mRNAs encoding amino acid synthesis or
macromolecule synthesis/modification functions Showed Shorter half-lives than average,
whereas those belonging to the cell-envelope maintenance or recycling of small
molecules category showed half-lives longer than average.

mRN A decay rates appear to be related to physiological function in a second study
in E. coli (Selinger et al., 2002). Oligonucleotide arrays containing on average one 25-
mer probe every 30 bp throughout the entire bacterial genome were used to measure
stability of transcripts corresponding to 1036 ORF S and 329 operons. The drug rifampicin
was used to shut-off transcription in bacteria cultures grown in LB at 37 °C. In
accordance to the assumed first order kinetic of mRNA degradation, the intensity of most
mRNAs studied decreased exponentially over time with an average mRNA half-life of

6.8 min. In this study, translation and post-translational modification functions were

21

under-represented in the group of labile mRNAs. In contrast, transcripts that encode
putative enzymes were Si grrificantly over-represented among Short lived mRNAs.
Furthermore, genes involved in energy metabolism were over-represented among
transcripts with half-lives between 10 and 20 min (Selinger etal., 2002). The high
resolution of the oligonucleotide microarrays used in this study allowed mechanistic
aspects of mRNA degradation in E. coli to be examined. Disappearance of the S’UTR,
3’UTR and three equally Spaced internal regions was monitored over time in operons
with 2 ORFS or more. Consistent with current models of mRNA degradation in bacteria
(Coburn and Mackie, 1999), 5’ ends of operons degraded on average more quickly than
the rest of the transcript with stability increasing in the 3’ direction. Interestingly,
hierarchical clustering of the decay patterns for 149 Operons indicated that this pattern
predominates but it is not the only one. Alternatives to the 5’ to 3 ’ model of bacterial
mRNA degradation have been postulated (Coburn and Mackie, 1999). This approach
Should greatly aid in evaluating the Significance of the different pathways of mRNA

degradation and identifying the cellular targets in bacteria.

The past year was also fruitful for global mRNA stability studies in eukaryotic
systems. Wang et a]. (2002) reported the use of spotted DNA microarrays to investigate
mRNA decay rates in the yeast Saccharomyces cerevisiae. In this study, mRNA synthesis
was halted by thermal inactivation of the temperature sensitive RNA polymerase II
(rpr-I) (Nonet et al., 1987). Similar to bacteria, exponential decay was a good model to
explain mRNA disappearance in yeast and allowed half-lives of 4687 yeast transcripts to
be determined. The half-life measurements ranged from ~3 to 90 min and Showed a

global mean value of 23 min. Several features of the mRNAs measured were investigated,

22

but no simple correlation was observed between mRNA half-lives and ORF Size, codon
bias, ribosome density or mRN A abundance. However consistent with the data in bacteria,
coordination of mRN A decay rates and gene fimction was apparent. The mRNAs
encoding components of cellular complexes such as the nucleosome core, the 208
proteasome core, the ribosome or the trehalose phosphate synthase complex, showed
strikingly similar turnover rates. This observation was extended to 95 other complexes
analyzed emphasizing the remarkable coordination of mRNA decay in yeast.
Coordination of the decay of yeast mRNAs was also observed in more broadly related
physiological functions. For example, transcripts encoding enzymes that participate in the
central systems of energy metabolism (glycolySiS / gluconeogenesis, the tricarboxylic
acid cycle and the glyoxylate cycle) were among those that live the longest. In contrast,
transcripts encoding the proteins of the mating pheromone signal transduction pathway
turned over relatively rapidly.

Lam et al. (2001) used a specialized lymphocyte array (spotted DNA microarray) to
estimate mRNA stabilities in lymphoid cell cultures. The initial objective of this study
was to determine the mode of action of the anti-cancer drug flavopiridol. It was found
that flavopiridol inhibited general transcription in the cell cultures, presumably by
inhibiting the transcription elongation factor P-TEFb (Price, 2000) in a similar fashion as
5,6-dichloro-1-[i-D-ribofi1ranosyl-benzimidazole (DRB). The effect of flavopiridol on
mRNA levels was comparable to that obtained with DRB or the DNA interchalating
agent actinomycin D. Thus Lam et a]. (2001) used flavopiridol to stop transcription and
determine the mRNA half-lives of 2794 mammalian mRNAs. Although the gene sample

analyzed was biased towards lymphocyte-related functions, some of the conclusions

23

reached in this study echoed those of other systems. The great majority of the well-
measured transcripts decreased in abundance with first order kinetics after transcription
inhibition. In addition, association between mRN A stability and gene function was also
observed in the mammalian cell cultures studied. Transcripts encoding apoptosis
regulators often decayed rapidly, as did mRNAs coding for several key cell-cycle
regulators. ARES are among the best characterized instability sequences in eukaryotes
and are often found in labile mRNAs encoding proto-oncoproteins, cytokines and
transcription factors (Chen and Shyu, 1995). The relationship between the number of
transcripts containing ARES and mRNA stability was investigated. The number of ARE-
containing mRNAs increased as the mRN A stability decreased. However, the great
majority of unstable transcripts lack ARES and 10% of stable mRNAs contained ARE
sequences. Hence ARES are not predictive of rapid mRN A turnover. AS in other systems,
additional unknown transcript features Should contribute to the individual mRN A
stabilities observed.

mRNA decay studies with microarray technology have also been conducted in
Arabidopsis thaliana (Gutierrez et al., 2002). What we have learned in Arabidopsis, to
date the only effort in an intact multicellular eukaryote, is the foundation of this doctoral

dissertation and will be discussed in depth in Chapter 2.

Common themes anflew trends iLmRNA decay

The coupling of global transcriptional Shut-off assays with DNA microarray analysis

has been a successful approach for monitoring mRNA decay on a global basis. The few

24

studies reported to date support basic notions of mRNA degradation. It is generally
assumed that mRNA degradation, like radioactive decay, is a stochastic process.
Therefore the change in mRN A concentration at any time is a first-order process, that
depends only on the amount of mRNA present at the time (Brawerrnan, 1993; Ross,
1995; Caponigro and Parker, 1996). Based on the data obtained in bacteria, yeast and
mammalian cell cultures, it now appears that first order decay kinetics is indeed a good
approximation to model mRNA disappearance of most cellular transcripts after
transcription ceases. Detailed analysis of individual time-series should help identify the
exceptions to this rule. Characterization of the features of transcripts with unusual decay

kinetics Should provide additional insight into the mRN A degradation process.

The studies in yeast and bacteria are in agreement with the main pathways of mRNA
decay. By comparing global decay rates of oligo(dT) and random primer labeled mRNA
samples, Wang et al. (2002) Showed that 3’ ends are more labile on average than the
body of mRNAs. This evidence supports the two major pathways of mRNA degradation
in yeast (Caponigro and Parker, 1996). In addition, Selinger et al. (2002) Showed that
disappearance of most operons with 2 ORF S or more proceeded in a 5’ to 3’ direction
consistent with the major model of mRNA degradation in bacteria (Coburn and Mackie,
1999). These results indicate that microarray studies Should be useful to address
mechanistic aspects of mRNA degradation. Such an application would be particularly
attractive for less characterized organisms. Perhaps sub-genie resolution affymetrix
experiments as those described by Selinger et al. (2002) could be used to identify
transcripts with stable decay intermediates. Analysis of the decay intermediates Should be

instrumental in dissecting the steps in the corresponding decay pathways.

25

The genome-wide studies described have also challenged previous notions about
mRNA stability. The view that mRN A abundance is directly correlated with mRN A
stability has not been supported in the studies in bacteria, yeast or plants. Direct
measurements of mRNA abundance in bacteria and indirect estimates based on
fluorescent intensity in yeast and plants (Chapter 2) indicate that mRNA stability is a
poor predictor of mRN A abundance. Because mRNA levels are determined primarily by
the balance between the rate of synthesis and degradation (Hargrove et al., 1991), this
data implies that transcription plays a predominant role in determining mRNA steady-
statc levels. This data also suggest that the regulation of mRNA half-life may have an
alternative biological significance. For example as illustrated in Chapter 2 for touch-
controlled genes, a role for rapid mRN A turnover might be to allow rapid and transient
changes in transcript abundance in response to environmental cues. Alternatively, and as
illustrated in Chapter 3, regulation of mRN A stability might be essential to achieve
precise expression patterns that are not possible through transcriptional regulation alone.

An interesting genome-wide property that emerged from these studies is the
coordination of transcript stability based on functional and physiological associations.
Overall, long-lived mRNAs seem to be involved in central metabolic functions whereas
those involved in regulatory systems turn over relatively rapidly. At least in yeast, this
association goes beyond broad physiological relationships and seems important in
ensuring proper expression of the components of stoichiometric complexes such as the

ribosome.

26

Future prospects

Microarray technology has already proved to be a valuable tool to study post-
transcriptional regulation of gene expression in various model organisms. But the use of
this approach for the study of mRN A degradation and post-transcriptional processes in
general is in its infancy. New applications for this approach can be easily foreseen. For
example, determining mRN A turnover rates over different developmental, environmental
or other treatments should help us evaluate the extent and Significance of this mechanism
of regulation. In addition, the role of new candidate regulators of the mRN A degradation
process as well as components of the decay machinery could be readily addressed by
comparing mRNA turnover rates in KO mutants and wild-type. Because mRNAs will
decay with Slower rates in the relevant KO mutants, this analysis Should help categorize
transcripts on the basis of their decay strategy.

The exciting corollary of these studies is that much remains to be learned about
mechanisms of mRNA degradation in biological systems, for example regarding the
structural features that determine the stability of individual transcripts. It is likely that the
next years will see more applications of this tool to address the questions posed. Detailed
knowledge of this level of regulation is necessary to better comprehend the complex gene

expression program in plants and other systems.

27

References

Abler,M.L. and Green,P.J. (1996). Control of mRNA stability in higher plants. Plant Mol.
Biol. 32, 63-78.

Bernstein,J.A., Khodursky,A.B., Lin,P.H., Lin-Chao,S., and Cohen,S.N. (2002). Global
analysis of mRNA decay and abundance in Escherichia coli at single-gene resolution
using two-color fluorescent DNA microarrays. Proc. Natl. Acad. Sci. USA 99, 9697-9702.

Bouvet,P. and Belasco,].G. (1992). Control of RN ase E-mediated RNA degradation by
5'-terminal base pairing in E. coli. Nature 360, 488-491.

Brawerman,G. (1993). mRNA degradation in eukaryotic cellszan overview. In Control of
Messenger RNA Stability, J .Belasco and G.Brawerman, eds. Academic Press, Inc. San
Diego, CA), pp. 149-160.

Caponigro,G. and Parker,R. (1996). Mechanisms and control of mRNA turnover in
Saccharomyces cerevisiae. Microbiol. Rev. 60, 233-249.

Chan,M.T. and Yu,S.M. (1998a). The 3 ' untranslated region of a rice alpha-amylase gene
fimctions as a sugar-dependent mRNA stability determinant. Proc. Natl. Acad. Sci. USA
95, 6543-6547.

Chan,M.T. and Yu,S.M. (1998b). The 3 ' untranslated region of a rice alpha-amylase
gene mediates sugar-dependent abundance of mRNA. Plant J. 15, 685-695.

Chen,C.Y.A. and Shyu,A.B. (1995). AU—rich elements - Characterization and importance
in messenger-RNA degradation. T. Biochem. Sci. 20, 465-470.

Chiba,Y., Ishikawa,M., Kijima,F., Tyson,R.H., Kim,J., Yamamoto,A., Nambara,E.,
Leustek,T., Wallsgrove,R.M., and Naito,S. (1999). Evidence for autorcgulation of
cystathionine gamma-synthase mRNA stability in Arabidopsis. Science 286, 1371-1374.

Coburn,G.A. and Mackie,G.A. (1999). Degradation of mRNA in Escherichia coli: An old
problem with some new twists. Prog. Nucleic Acid Res. Mol. Biol.Vol. 62 62, 55-108.

Culbertson,M.R. (1999). RNA surveillance - unforeseen consequences for gene
expression, inherited genetic disorders and cancer. Trends Genet. 15, 74-80.

Dickey,L.F., Nguyen,T.T., Allen, GO and Thompson,W.F. (1994). Light modulation of
Federroxin mRNA abundance requires an open reading fi'ame. Plant Cell 6, 1171-1176.

Dickey,L.F., Petracek,M.E., Nguyen,T.T., Hansen,E.R., and Thompson,W.F. (1998).
Light regulation of Fed-l mRNA requires an element in the 5 ‘ untranslated region and
correlates with differential polyribosome association. Plant Cell 10, 475-484.

28

Downes,B.P. and Crowell,D.N. (1998). Cytokinin regulates the expression of a soybean
beta-expansin gene by a post-transcriptional mechanism. Plant Mol. Biol. 37, 437-444.

F eldbriigge M., Arizti, P., Sullivan, M.L., Zamore, P.D., Belasco, J.G. and Green, P.J.
(2002). Comparative analysis of the plant mRNA-destabilizing clement, DST, in
mammalian and tobacco cells. Plant Mol. Biol.49, 215-223.

Gallie,D.R. (1991). The cap and poly(A) tail function synergistically to regulate
messenger-RNA translational efficiency. Genes Dev. 5, 2108-2116.

Gallie,D.R. (1993). Posttranscriptional regulation of gene-expression in plants. Annu.
Rev. Plant Physiol. Plant Mol. Biol. 44, 77-105.

Gallie,D.R. (1998). A tale of two termini: A functional interaction between the termini of
an mRNA is a prerequisite for efficient translation initiation. Gene 216, 1-11.

Gil,P. and Green,P.J. (1996). Multiple regions of the Arabidopsis SA UR-A C] gene
control trascript abundance: the 3' untranslated region functions as an mRN A instability
determinant. EMBO J. 15, 1678-1686.

Grunberg-Manago,M. (1999). Messenger RNA stability and its role in control of gene
expression in bacteria and phages. Annu. Rev. Genet. 33, 193-227.

Gutiérrez,R.A., Ewing,R.M., Cherry,J.M., and Green,P.J. (2002). Identification of
unstable transcripts in Arabidopsis by cDNA microarray analysis: rapid decay is
associated with a group of touch- and Specific clock-controlled genes. Proc. Natl. Acad.
Sci. U. S. A 99,11513-11518.

Hargrove,J.L., Hulsey,M.G., and Bealc,E.G. (1991). The kinetics of mammalian gene-
expression. Bioessays 13, 667-674.

Harmer,S.L., Hogenesch,J.B., Straume,M., Chang,H.S., Han,B., Zhu,T., Wang,X.,
Kreps,J.A., and Kay,S.A. (2000). Orchestrated transcription of key pathways in
Arabidopsis by the circadian clock. Science 290 , 2110-2113.

Hilleren,P. and Parker,R. (1999). Mechanisms of mRNA surveillance in eukaryotes.
Annu. Rev. Genet. 33, 229-260.

Holstege,F.C.P., JenningS,E.G., Wyrick,J.J., Lee,T.I., Hengartrrer,C.J., Green,M.R.,
Golub,T.R., Lander,E.S., and Young,R.A. (1998). Dissecting the regulatory circuitry of a
eukaryotic genome. Cell 95, 717-728.

Inaba,K., Fujiwara,T., Hayashi,H., Chino,M., Komeda,Y., and Naito,S. (1994). Isolation
of an Arabidopsis thaliana Mutant, mtol , That Overaccumulates Soluble Methionine
(Temporal and Spatial Patterns of Soluble Methionine Accumulation). Plant Physiol. 104,
881-887.

29

Isshiki,M., Yamamoto,K., Satoh,H., and Shimamoto,K. (2001). Nonsense-mediated
decay of mutant waxy mRNA in rice. Plant Physiol. 125, 1388-1395.

Jofuku,K.D., Schipper,R.D., and Goldberg,R.B. (1989). A frameshifi mutation prevents
kunitz trypsin-inhibitor messenger-RNA accumulation in soybean embryos. Plant Cell 1,
427-435.

Johnson,M.A., Baker,E.J., Colbert,J.T., and Green,P.J. (1998). Determinants of mRNA
stability in plants. In A look beyond transcription: Mechanisms determining mRN A
stability and translation in plants, J .Bailey-Serres and D.R.Gallie, eds. (Rockville:
American Society of Plant Physiologists Press), pp. 40-53.

Johnson,M.A., Pérez-Amador,M.A., Lidder,P., and Green,P.J. (2000). Mutants of
Arabidopsis defective in a sequence-Specific mRN A degradation pathway. Proc. Natl.
Acad. Sci. U. S. A 97, 13991-13996.

Johnson,R.R., Chaverra,M.E., Cranston,H.J., Pleban,T., and Dyer,W.E. (1999).
Degradation of oat mRNAs during seed development. Plant Mol. Biol. 39, 823-833.

Lam,L.T., Pickeral,O.K., Peng,A.C., Rosenwald,A., Hurt,E.M., Giltnane,J.M.,
Averett,L.M., Zhao,H., Davis,R.E., Sathyamoorthy,M., Wahl,L.M., HarriS,E.D.,
MikovitS,J.A., Monks,A.P., Hollingshead,M.G., Sausville,E.A., and Staudt,L.M. (2001).
Genomic-scale measurement of mRN A turnover and the mechanisms of action of the
anti-cancer drug flavopiridol. Genome Biol. 2, 1-11.

Ling,J., WellS,D.R., Tanguay,R.L., Dickey,L.F., Thompson,W.F., and Gallie,D.R. (2000).
Heat Shock protein HSP101 binds to the Fed-1 internal light regulatory element and
mediates its high translational activity. Plant Cell 12, 1213-1227.

Marcotte,W.R. (1998). Developmental regulation of translation and mRNA Stability. In A
look beyond transcription: Mechanisms determining mRNA stability and translation in
plants, J.Bailey-Serres and D.R.Gallie, eds. American Society of Plant Physiologists), pp.
64-67.

McCarthy,J.E.G. (1998). Posttranscriptional control of gene expression in yeast.
Microbiol. Rev. 62, 1492-1553.

McClure,B.A. and Guilfoyle,T. (1989). Rapid redistribution of auxin-regulated RNAS
during gravitropism. Science 243, 91-93.

Mehdy,M.C. and Brodl,M.R. (1998). The role of stress in regulating mRNA stability. In
A look beyond transcription: Mechanisms determining mRN A stability and translation in
plants, J .Bailey-Serres and D.R.Gallie, eds. American Society of Plant Physiologists), pp.
54-63.

Mitchell,P. and Tollervey,D. (2000). mRNA stability in eukaryotes. Curr. Opin. Genet.
Dev. 10, 193-198.

30

Newman,T.C., Ohme-Takagi,M., Taylor,C.B., and Green,P.J. (1993). DST sequences,
highly conserved among plant SA UR genes, target reporter transcripts for rapid decay in
tobacco. Plant Cell 5, 701-714.

Nonet,M., Scafe,C., Sexton,J., and Young,R. (1987). Eukaryotic RNA-polymerase
conditional mutant that rapidly ceases messenger-RNA synthesis. Mol. Cell. Biol. 7,
1602-1611.

Ohme-Takagi,M., Taylor,C.B., Newman,T.C., and Green,P.J. (1993). The effect of
sequences with high AU content on mRNA stability in tobacco. Proc. Natl. Acad. Sci.
USA 90,11811-11815.

Ominato,K., Akita,H., Suzuki,A., Kijima,F., Yoshino,T., Yoshino,M., Chiba,Y.,

Onouchi,H., and Naito,S. (2002). Identification of a Short highly conserved amino acid
sequence as the functional region required for Posttranscriptional autorcgulation of the
cystathionine gamma-synthase gene in Arabidopsis. J. Biol. Chem. 277, 36380-36386.

Pérez-Amador,M.A., Lidder,P., Johnson,M.A., Landgraf,J., Wisman,E., and Green,P.J.
(2001). New molecular phenotypes in the dst mutants of Arabidopsis revealed by DNA
microarray analysis. Plant Cell 13, 2703-2717.

Petracek,M.E., Dickey,L.F., Nguyen,T.T., Gatz,C., Sowinski,D.A., Allen,G.C., and
Thompson,W.F. (1998). Ferredoxin-l mRNA is destabilized by changes in
photosynthetic electron transport. Proc. Natl. Acad. Sci. USA 95, 9009-9013.

Price,D.H. (2000). P-TEFb, a Cyclin-Dependent Kinase Controlling Elongation by RNA
Polymerase 11. Mol. Cell. Biol. 20, 2629-2634.

Ravanel,S., Gakiere,B., Job,D., and Douce,R. (1998). The Specific features of methionine
biosynthesis and metabolism in plants. Proc. Natl. Acad. Sci. U. S. A 95, 7805-7812.

Ross,J. (1995). Messenger-RNA stability in mammalian-cells. Microbiol. Rev. 59, 423-
450.

Russell,J.E., Morales,J., and Liebhaber,S.A. (1997). The role of mRNA stability in the
control of globin gene expression. Prog. Nucleic Acid Res. Mol. Biol. Vol 57 57, 249-
287.

Schaffer,R., Landgraf,J., Accerbi,M., Simon,V., V, Larson,M., and Wisman,E. (2001).
Microarray Analysis of Diurnal and Circadian-Regulated Genes in ArabidOpSiS. Plant
Cell 13, 113-123.

Selinger, D. W., Saxena, R. M., Cheung, K. J ., Church, G. M., and Roscnow, C. Global
RNA half-life analysis in Escherichia coli reveals positional patterns of transcript
degradation. Genome Res. 2002. In Press

31

Sheu,J.J., Yu,T.S., Tong,W.F., and Yu,S.M. (1996). Carbohydrate starvation stimulates
differential expression of rice alpha-amylase genes that is modulated through complicated
transcriptional and posttranscriptional processes. J. Biol. Chem. 2 7] , 26998-27004.

Sullivan,M.L. and Green,P.J. (1996). Mutational analysis of the DST element in tobacco
cells and transgenic plants: Identification of residues critical for mRNA instability. RNA
2, 308-315.

Taylor,C.B. and Green,P.J. (1995). Identification and characterization of genes with
unstable transcripts (GUTS) in tobacco. Plant Mol. Biol. 28, 27-38.

Theodorakis,N.G. and Cleveland,D.W. (1993). Translationally coupled degradation of
tubulin mRNA. In Control of Messenger RNA Stablilty, J .Belasco and G.Brawerman, eds.
Academic Press, Inc. San Diego, CA), pp. 219-238.

Tucker,M. and Parker,R. (2000). Mechanisms and control of mRNA decapping in
Saccharomyces cerevisiae. Annu. Rev.Biochem. 69, 571-595.

van Hoof,A. and Green,P.J. (1996). Premature nonsense codons decrease the stability of
phytohemagglutinin mRN A in a position-dependent manner. Plant J. 10, 415-424.

Vasudevan, S. and Pcltz, SW. (2001). Regulated ARE-mediated mRNA decay in
Saccharomyces cerevisiae. Mol. Cell 7, 1191-1200.

Voelker,T.A., Staswick,P., and ChrispeelS,M.J. (1986). Molecular analysis of 2
phytohemagglutinin genes and their expression in Phaseolus-vulgaris cv pinto, a lectin-
deficient cultivar of the bean. EMBO J. 5, 3075-3082.

Wang,Y.L., Liu,C.L., Storey,J.D., Tibshirani,R.J., Herschlag,D., and Brown,P.O. (2002).
Precision and functional specificity in mRNA decay. Proc. Natl. Acad. Sci. USA 99,
5860-5865.

Wilusz,C.J., Wormington,M., and Peltz,S.W. (2001). The cap-to-tail guide to mRNA
turnover. Nature Rev. Mol. Cell Biol. 2, 237-246.

Zhang,S.Q. and Mehdy,M.C. (1994). Binding of a 50-kD protein to a U-rich sequence in
an messenger-RNA encoding a proline-rich protein that is destabilized by fungal elicitor.
Plant Cell 6, 135-145.

Zhang,S.Q., Sheng,J.S., Liu,Y.D., and Mehdy,M.C. (1993). Fugal elicitor-induced bean
proline-rich protein messenger-RNA down-regulation is due to destabilization that iS
transcription and translation dependent. Plant Cell 5, 1089-1099.

32

CHAPTER 2

Identification of unstable transcripts in Arabidopsis by cDNA
microarray analysis: Rapid decay is associated with a group of

touch- and specific clock-controlled genesII

 

“ The original form of this chapter was published in “Gutierrez , R.A., Ewing R.M., Cherry, J.M., and
Green, P.J. (2002). Identification of unstable transcripts in Arabidopsis by cDNA microarray analysis:
rapid decay is associated with a group of touch- and specific clock-controlled genes. Proc. Natl. Acad. Sci.
USA. 99: 11513-11518”.

33

Introduction

Regulation of the stability of mRNAs is an important process in the control of
gene expression. This point is perhaps most evident in the wide range of half-lives that is
typically observed for nuclear encoded transcripts. In plants, Similar to what reported in
mammalian systems, half-lives of mRNAs Span several orders of magnitude. Unstable
messages have half-lives of less than 60 rrrinutcs, very stable ones on the order of days,
with the average being on the order of several hours (Siflow and Key, 1979; Hargrove et
aL,l991)

Most research has emphasized the study of unstable mRNAs. These transcripts
have attracted attention because they often code for regulatory functions that are
important for growth and development. For example, mRNAs known to be highly
unstable include those for the transcription factors c-myc and c-fos in mammalian cells
(Greenberg and Belasco, 1993) and the mating-type transcripts in yeast (Peltz and
Jacobson, 1992). In plants, transcripts that fall into this category include the mRNAs for
photo-labile phytochrome (Seeley et al., 1992) and several auxin-inducible transcripts
(McClure and Guilfoyle, 1989; Koshiba et al., 1995). The instability of these mRNAs
facilitates fast changes in mRN A levels that result in transient and tightly controlled gene
expression (Treisrnan, 1985).

Previous work on unstable transcripts has concentrated on the identification of
sequence elements and trans-acting factors that regulate the stability of individual or
small groups of transcripts. In eukaryotic cells, transcripts destabilized by multiple

overlapping of AUUUA sequences or other AU-rich elements (ARES) located in 3’

34

untranslated regions (UTRS), have been a major focus (Shaw and Kamen, 1986; Ohme-
Takagi et al., 1993; Chen and Shyu, 1995; Vasudevan and Pcltz, 2001). Several proto-
oncogene, cytokine, and transcription factor mRNAs involved in growth and
differentiation are recognized for rapid decay via ARES (Shaw and Kamen, 1986; Ohme-
Takagi et al., 1993; Chen and Shyu, 1995; Vasudevan and Pcltz, 2001). In plants, one of
the best characterized instability sequences is the DST or downstream element (McClure
et al., 1989; Newman et al., 1993). This instability determinant is found in the 3’UTR of
the small auxin pp RNA (SA UR) genes. DST elements have a complex structure (Sullivan
and Green, 1996) and the recognition requirements appear to be unique to plants
(F eldbriigge et al., 2002). Other sequences that cause instability have also been described
(Ross, 1995; Caponigro and Parker, 1996; Gutierrez er al., 1999). Nevertheless, the
number of structural features identified to date that target transcripts for rapid decay is
relatively modest, and many more are likely to be discovered (Taylor and Green, 1995).
Although study of individual transcripts is a viable avenue to address this
problem, genomic-scale analysis is necessary to evaluate the nature of unstable
transcripts within an organism and the regulatory associations they Share. Genomic
approaches using DNA microarrays have emphasized the study of mRN A levels and how
are those levels affected under different conditions (Brown and Botstein, 1999; Schaffer
et al., 2000), but they have been rarely applied to the study of post-transcriptional
processes. The first indication that this approach held promise came from data presented
in a web Site (http://wcb.wi.mit.edu/young/expression/) referred to in Holstege et al.
(Holstege et al., 1998) that estimated stabilities of yeast mRNAs. More recently, Lam et

al. (Lam et al., 2001) used a Specialized lymphocyte array to estimate mRNA stabilities

35

in lymphoid cell cultures. Both investigations suggested that global analysis of mRN A
stability in intact multi-ccllular organisms should be feasible and more revealing.

In this study, we examined mRNA degradation in intact Arabidopsis plants using
cDNA arrays containing more than 11,000 clones. Similar to the situation in other
organisms, the identity and percentage of unstable transcripts in plants had not been
evaluated on this scale. Our analysis indicated that at least 1% of the transcripts
represented on our arrays decayed with half-lives of less than 60 minutes. Further, we
identified specific functional and regulatory associations among groups of unstable
mRNAs that provide insight into the biological Significance of rapid mRNA decay

mechanisms.

36

Materials and Methods

Tiny-life measurements and preparation of RNA samples.

 

Half-lives were determined as described by Seeley et al. (Seeley et al., 1992) with
the following modifications. Arabidopsis thaliana ecotype Columbia were grown on
plates containing 1x Murashige and Skoog salts, 1x Gamborg's vitamins and 1% sucrose
for two weeks at 22 °C and 16/8h light/dark cycles. The plants were then transferred to a
flask with incubation buffer (Seeley et al., 1992). After a 30 min incubation, 3'-
deoxyadenosine (cordycepin) was added to a final concentration of 0.6 mM (time 0).
Tissue samples were harvested at regular intervals thereafter and quickly frozen in liquid
nitrogen. Total RNA was isolated and analyzed by northern blot using standard
techniques. Cordycepin was used to inhibit transcription in these studies because its use is
routine in plants (Seeley etal., 1992; Johnson et al., 2000) in contrast to other inhibitors
such as alpha-amanitin, and is more effective in leaf tissue than Actinomycin D (Johnson

et al., 2000) presumably due to poor penetration.

Hybridization of cDNA Microarray;

The 11,521 element cDNA microarray, print name MSU-2_03-00, prepared by the
AF GC was used in all experiments (Schaffer etal., 2001). 100 pg of total RNA
corresponding to time 0 and 120 min after cordycepin treatment was labeled during first
strand cDNA synthesis with Cy3- and Cy5-labeled dUTP, respectively, as previously

described (Schaffer et al., 2001). Three independent cordycepin treatments (biological

37

replicas) were performed and RNA samples were isolated. Each pair of samples from the
0 and 120 time points was used in two microarray hybridizations, the second with reverse
labeling relative to the first (technical replicas). Measurement of the fluorescence
corresponding to hybridization intensities was performed with the ScanArray 4000
Microarray Acquisition System (Packard BioChip Technologies. Billerica, MA). We
used the ScanAlyze v2.44 software (http://rana.1bl.gov/EisenSoftware.htm) to extract the
information of the images generated. The raw data for these experiments is available
hour the Stanford Microarray Database (http://genome-www.stanford.edu/microarray/)

(Sherlock et al., 2001), ExptID: 11374, 11333, 11339, 11323, 11375 and 11342.

Microprray data analysis.

Stringent quality control measures were applied to define the working data set.
Spots with abnormal Shapes or high local background were discarded manually. Spots
with channel intensity values smaller than the mean plus two standard deviations of each
Slide background or with GTB2 values (GTB2 indicates the fraction of pixels in the spot
that have intensity values 1.5 times the background) smaller than 0.65 in more than two
channels were discarded because of low Signal. The quality of the hybridization was also
evaluated by visual inspection of the gradients, using the most sensitive setting of the
"Array Color Plot" tool implemented in the Stanford Microarray Database (SMD). Slides
that Showed gradients in more than 25% of the array surface and/or that had R-Squared
values > 0.15 (indicating a strong dependence on spatial location) were not used for data
analysis. The percent of the array surface that exhibited gradients was also used to order

the Slides from worst to best or best to worst in Figure 2. 1 b.

38

The z-score method in log space with a 90% trimmed data set was used for global
normalization of the data (Schaffer et al., 2001). The difference in mRNA levels between
the time points considered (0 and 120 min) can be used to estimate the rates of decay

using the equation: In (Normalized Ratio) = -kdecayt, with the half-life being: i”; = 0.693 /

kdecay, because mRN A degradation generally obeys first-order kinetics (Lam et al., 2001).

Statistical analysis of the ratios was performed using the t-test as described in the text.

Sequence and gene expression analysis

Sequences of 59 clones representing Arabidopsis thaliana genes for pnstable

transcripts (A tG UT s) were determined and found to be consistent with sequences
deposited in Genbank. For EST identification, the BLASTN program was used to search

the completed Arabidopsis genome sequence downloaded from The Institute for

Genomic Research (TIGR). Functional categories were obtained from the Munich

Information Center for Protein Sequences. For analysis of gene expression data across

multiple experiments, the Cluster and T reeview Software were used (Eisen et al., 1998)

(httpz/lranalbl.gov/EisenSoftwarehtm). Visual images were generated with the Treeview

software using the output generated by the hieraIChical clustering program of the Cluster

software. The uncentered correlation similarity metric was used to perform average

linkage clustering.

Sequences of AtG UTs (S’UTR, coding Sequence, 3’UTR) were obtained from the

TIGR Arabidopsis genome. Because UTRS are not an annotated feature of the

Arabidopsis genome, the 3'UTR sequences of AtGUTS were assembled by extracting the

average length of an Arabidopsis 3'UTR, that is 150 bp downstream of the annotated stop

39

codon. Similarly, 5'UTR sequences were assembled by extracting 75 bp upstream of the
annotated start codon. Average 3UTR and 5'UTR sizes were estimated from the 5,000
full lenghthNAs released by CERES (also available from TIGR ).

The oligomer counting method was used in an effort to identify candidate
instability determinants in AtGUTs (van Helden et al., 1998). This is a rigorous and
exhaustive method that is based on the detection of over-represented oligonucleotides in
the input set of sequences as compared to a control set. Frequencies of overlapping (1 bp
window) oligonucleotides (up to 6 nt in length) were determined in the 3’UTR sequences
of AtGUTs and also in the 3’UTR sequences of a control set of genes as described by van
Helden et al., (van Helden et al., 1998). The control set was derived fi‘om 4064 clones
corresponding to stable transcripts on the array. To assess Significance, 1000 random
samples of the same Size as the test set were taken from the control sequences and
oligonucleotide frequencies were determined in these samples. The criteria for
significance were as follow: (1) Oligonucleotide was at least 2-fold more abundant in
unstable than in whole control set; (2) Oligonucleotide had frequencies > 2 stdev above
mean frequency in the 1000 random samples from control set; (3) Oligonucleotide was
present in >10% of test sequences.

Additional MEME (Multiple Expectation Maximization for Motif Elicitation)
searches (Bailey and Elkan, 1994) were canied out as described at http://meme.sdsc.edu.
MEME is a computational tool for discovering Short sequence patterns (motif) that occur
repeatedly in a group of related DNA or protein sequences. MEME output describes each
motif it finds by the probability of each possible nucleotide (if using DNA or RNA

sequences) at each position in the motif. MEME indicates the location in the input

40

sequences and provides a p-value to evaluate the significance of each motif. Various
combinations of the MEME parameters were tested: motif distribution, 1-3; number of
motifs, 3-5; motif width, 5-50. Programs written in the Practical Extraction and Report

Language (Perl) were used for sequence extraction and manipulation.

' ' 0- w m slid r best slide
Add transcription inhibitor (harvest time 0) aa- 3; slide t: :0": slide

Incubate for 120 min (harvest time 120)

Extract RNA from each time point and

70
so
label with Cy3 or Cy5 50

   
 

Hybridiae 11521 spot AFGC microarray slide

Percentage of
Non-reproducible spots

20
10 _"—“'

Normalize data and
determine ratio

0
4
Estimate half-life Number of slides

Figure 2. 1. Strategy for monitoring mRNA stability using cDNA microarrays.

(A) RNA samples corresponding to 0 and 120 rrrin after the addition of the transcriptional
inhibitor cordycepin were labeled with Cy3 and Cy5 respectively and used to hybridize 11K
microarray slides. Each pair of RNA samples were reverse labeled for a separate microarray
hybridization. These hybridizations were performed with samples from three independent
cordycepin treatments for a final data set of Six Slides. Half-life values were then estimated hour
the normalized ratios. (B) Non-reproducible spots decrease as a function of the number of slides,
nearly leveling out when the data from four slides is combined. The quality of the Slides, best to
worst or worst to best based on the extent of visible gradients (see Materials and Methods), does
not significantly affect the reproducibility of the data when two or more Slides are considered,
although the curves are slightly steeper with better slides.

41

Results and Analysis

Monitoring mRNA stability using cDNA microarrays.

mRNA decay rates, expressed as half-life values, are typically measured by
monitoring the disappearance of a transcript by northern blot after transcription of the
corresponding gene has been halted. We combined this Simple experimental strategy with
the highly parallel power of DNA microarray analysis (Schena et al., 1995) as outlined in
Figure 2.1a. Total RNA samples corresponding to 0 and 120 min time points after
transcriptional inhibition with cordycepin were isolated. 100 pg of total RNA from each
of these samples was used to synthesize cDNA probes by incorporating Cy3- or Cy5-
labeled dUTP during oligo(dT)-primed reverse transcription. The probes were combined
and used for hybridization of the 11,521 elements cDNA microarray (11K microarray)
prepared on glass Slides by the AF GC. We performed three biological replica
experiments, each with a reverse-labeling technical replicate. The purpose of these
repetitions was to increase the likelihood of detecting significant differences in mRNA
levels, while decreasing the likelihood of false positives, which might be common on
microarray studies with one or two Slides (Lee et al., 2000). Quite reasonably, the number
of non-reproducible normalized intensity ratios 22 decreased as a function of the number
of Slides, nearly leveling out below 5% when 4 slides were considered (Figure 2.1b).
Based on these data, and to be rigorous, we defined our working data set as all those

clones with reproducible normalized intensity ratios 22 in 5 of the 6 Slides.

42

At legs; 1% of slopes on the 11K Arabidopsis microarrays corresmnd
to unstable message;

 

To identify and characterize the most unstable transcripts from our working data
set, we focused our attention on the transcripts that were most diminished after treatment
with cordycepin for 120 minutes. Clones whose median normalized intensity ratios were
2 4 (0 versus 120 min) and that met several quality control criteria (see Materials and
Methods) were used for further analysis. In this study, expressed sequence tags (EST)
that overlap with the same annotated open reading frame were considered as representing
the same gene. This is a reasonable assmnption because the expression patterns of groups
of ESTS that match the same open reading fi'ame were well correlated across multiple
experiments (data not shown). Based on this criterion, the selected clones corresponded
to 100 genes that were termed Arabidopsis thaliana genes with pnstable transcripts
(AtGUTs), because the calculated half-lives of the encoded transcripts were 60 min or
leSS (Table 2.1). Sirrrilarly, we identified 225 genes with moderately unstable mRNAs,
whose estimated half-lives ranged from 60 to 120 min. The great majority of the
transcripts in Arabidopsis appeared to decay with rates greater than two hours, consistent
with the idea that most messages in plants are relatively stable (Siflow and Key, 1979).

The 11K microarray used in these studies represents an estimated 7800 unique
genes (Schaffer et al., 2001) so the 100 AtGUTS we identified correspond to about 1%.
This number likely represents an underestimate of unstable Arabidopsis transcripts
especially when extrapolated to the whole genome for several reasons. First, unstable
mRNAs with low steady state levels may be underrepresented in the EST collections

used for the 11K microarray. Second, some unstable transcripts likely fall below the

43

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F 529a .5655 08892 clean. 87o 8m armour. can W :«SR, 82$ Ego
i can dew—>50 m
: €3.52 .3 88393218 SASS: Niche; Swan, mom Ema; 3a 633% m8~3<<h topmmm
ole»; 9:203 =mEm _a._o>ow 2 5:55. , ,
v 9:: 8858588350 amnion: 8128; BE 5m sewn; n8 Scam; c582 vacuum
v seen 9:93 85885. 8:95: 8-w8._ 2: arm; :n 8:8. F88: Evmevfl
F owocogxoo: 8:35. Find”. 97o 9... Emma QB 8°83 «Name tonne
_ 5065 2:98 83°82 zoflhmmmmo imam 8 amen Ea 83.3“ 82932 firmed
8 529a gash Emcee”: Sinai RES.» 5; Swan «.3 SSSN @883 $508—
v 990985329220 25:55 Seaman: oEIoEwmu Shoo; k: new; won @888 among $825:
8 Coca 585:: 0828: 21392.2 83 mom Swen Zn Sacco 883 C82:
v 5an 9:23“. 948mm: manages. onto «9. «mm; 98 $88— 882 EVER
P Sosa gash 2834.2 836.8 97o com new: o8 834R momooz >5me
2 852838582 cans“. 883%.. 9..ng «to , m8 sumo; com 888 38m; Canaan
1°.th Sim 28. Quicken.
_ is: 2:288 _ 2.8.2:. _ 38.. Thain—5minTish—5.823; .e Tag—9:888;

 

SEQ: 32.85.. 232: 5.3 8:8 2.2.2.. 22.83.92 4.“ 2...:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 8.9.. 3.7.-.8 88882 2:28: 8.m8.~ 88 2-8... 8.8.. 82.8 3sz E88

8 .853 2-23 883-2982.". 8.82:. 8:88. 8.... 8e 9-8.8 ~84 8:8 882 E08.

8 8.2.. :35.an 88%.... 81:8“. 8-m88W 88 2.8.... «.8 .8888 88.2.. .888

v 8 ox_.-..o.8.W 88.82 W 82% m.-m8.eW 8 2.8.“ cu. Wet-ma «8%. 58..
W 5:5 Ego—o 38:88. 2.0ng W W W W

8 W 8.9.. :35an 828.2 3.8.“. W o M 8: 3-8.. 8.... W888 §§W axe-.3.
_.

8 529a 885A 8888:. 81.88 W 8.... 8.. SE8 8.... W 9848 8..8<<W oxen...
. r.

8 . 98.925 388W 8.88% 2.888". W-.W8mooc 8. Won-m: a... Waconefi 88on E8“.

. 1 8.8 555:1. SEREW 3:85.22 W-.8m8n 8..” W288 8.... .58....MW. 895W tax...

A 8.388: 38:28.5 Sash“ 2888:. 818.88“. .3... W «8 W88... 3.. 8888 88th E828.

.. W as vans-W 98.85. .130. W 8.... m8 HEd... 3.. W 8.8L. 88tW #88...
W .25 .888 28882 83.8”- i. W W W

o. 8.9.. 565.5 8.822 W clan-8.... W 8.1... W 2c 8.88 .8 «SEW Ream-W £88.
‘ W .

.. area 3855 8.882 W «no.9. ~...o 3.. Wino-o 9? W838 .882: hxouuu.

8 as... 88 SSE-=82 8888:. oisoxs. c. .a c... 8.8.. 98 98.8 .882... $98.

.. 38.8885285851. 88882 . 2.9.23“. Wow-mg: So 2d... c8 588W .83. P88

45

 

W 885.8".

4

. .
W .888.
W .

 

 

 

 

 

 

 

 

 

 

 

 

 

.mewé

..
8 5.22.. 8.52:8 52.8 82. 88.92 m..-w8.m . 8. . .52... 3.. 8.88: E88
. . .
8 529.. .5555 88382 2.88". 8.6 8a 6.8... «.9. . .858. 8882. R888
8 8.9.. 2,85“. 0888.2 o..n.zo.n. 8.8 8.. Sm... 3... . 888, .88... Era:
8 529.. 985.. 8888.2 9-8.8. 8.m_8.. 8 2mm... 8... 888 88m: #588.
8 229.. 2.85.. 8885. 28.9.8. m. ..o 8.. 3m? 8.9. . 28.38. 8.8822. £338
. e...- 5..-oc..oa 2.88.5.2 88.92 28.0... 8.8 8.. Sm... .8 88.88, 888322 .588
v .><m 599.. 8.2.925 883.2 3.8. o 8. 8m...” «.8 8.8.. 8882 t..08~
o. m 82... 522.. 8.3.8-885... 8888:. 88.5.8 98.8.8 3. www.8- 99. .888 888. E88...
o. 529.. .8858... .8892 8.8”. 8.... .8 2-8.. 98 8°88 8882 .59.“.
£29.. 3 . .m - . .
v 58.. 88 82.500 3.98 88 «2 n .5... 8.. n8 .. mm m a «v 838.8 88: D88.
.m N 8882 29.2 8.822 “-8.8“. o 8.. 8.8... m8 . .88.. 8.82 £8.88
8 599.. 556.5 888.2 888 8.-.. 8m .78... 8.. . 8.88 8.8322. E22:
m 3...» 8.2.. 9...-5628. 8.88:. 88.88“. 8.... «8 83.8 .8 . 88.8. 888.. .808
89.8.8 8.80 2&8. «.3838... 8.8... :22. oz 2.8 W .889 .88.. p.828.

 

46

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

888:8 9.58 Eggs m§< is»? :29: oz 67mg «.8 8.85 89.0! #868.
o 33595.08” 23.92 flung”. onto mum W89; 23. W 8:8, 882 Emig—
8 52.3 355k 88%: r Splflmau. imam“. 8. «32 m9. “853% 8v~3<< Kasai
9 6mm“: 8 cmmafioofi seen 233“”: olefin o 30 2m: «.3 Emma, 8301‘ EVEN:
: SE 2.9585 8mm§h< .8329: 958a. m8 9mm; 98 .8553 :53 D88:
v 9:25 <zo Sa->xm>> 2:93 23%? «Exam; mtb «ma 07m; v.3 88mm 83%. £9on
v 5205 5:269»: 8:292 925% o 8o 936a «.8 $28 8%; Ema<8
8 5055 32653»: 83%.2 mnnuth 37.. v8 Sm? Ev «88: 0852 #535
v asscoitomceEEsEm; 935$? 835$ a?» 34 9m? 9% $88 882 Elm?
v 59.5% mama Sign. 033$: 2-82mi :3 5m 3mg 0:. .8m8mm 22E PSESF
8 seem E355 85%? 2.29» 23 m3 «wmwa 2v ‘ F893 nm$8<< $853
8 5205 $855 03332 @383 o 3» 9mm; «.2 Smear 83oz homunum
n .ocoawcezgaoa Sash. 8&3»: Edit» 658.8 mom Lemma .5. 6835 «3.85” “cams:
mm 529383533 82an 53%”. 8a hoe $918

2mg. ,

Wm—-wm.m

m 28mg“ 8892

47

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ON—IF 5..th inood _,

,, , a, m ,
9. cx__-c_a§3a82-o>>, 863m: M 8.83: Smog. 8m 8m: «.8 .8kfiu, 822 Comma
M W ,
, , , M
F 5205 2.55;, Scam: 96mm: E-m8.f 8“ Emma 98 .8239 5%, mxo<$F
8 Eocxcascasa BEBE, 08892 _ «35¢ 8; 8m Emma. mdm “8839 Swap the:
F 52% 9.55.. 8.392 8633 538+ 92 imam 9% 92.3. was: Emmommw
8 seen 2.55“. ovmmmomi ompuuzmfi «smog. §~ 2-th 9% _ 838V 28D gear
o 5205 9:25 883m: ”150m: 8-m8.~ «mm 2mm; mam. 58% 5.892, E483
_ 9:52 6522923686 8.59:. VIENE «2.» m8 2.28 «mm 3&8 88$ Fflog
v 529a ox__-.os£co=%um§: 088mm: Flomxms. 870 m8 :mmd 3m 8089 $52.. tvEmmP
2 5905 9:93. ommwmug‘ omwufiomfi RES; mun «$3 #5 $33“ 2882 5.5an
E 85.83985883 o:m__e_w 28%? SEN“. 2.6 2o .33 Em .3318 8083 axtmfi
3-3222 mom I. .o - . .
v 385,8an wzmcoameocoacm 8m? 2 o: 3; mm. mm Cu? mom ans—em mEmm<< E185
P 5203 365%: SK??? 889: 3.3 :m 2%.» QB ,8839 Reoozfi axomxoo
P 2.393%.33 5.8.82 age: ammuoNKE 83 «8 9mm; 98 8:85 8—om\.<<m #555
3 5203 365m 883»: m9 8mm; m3 _ .885 FEE:

8835..

48

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 5205 565.5. 88532 niece; 9.3 a, 2m 8.88 :m 358.va 8383“ EEO?“
3 «Sosa 8.29.538539 oEoumE 2:888. 97o n6 8.8.. 8.8 ":88; 8.353,”H E28
2 n 62.83 .5205 2:.5858fi 88%;. ~32“. , 8.8 . 9:. "8.8.. 88 w .889 848:, E35:
v 850922739858 88—8: . 88% 8-m8.m a, 9-8.8 “.8 £885 388,2. FREE
8 seen 5555 888.2 A clans—u. «:8 o; 8.98 8.8. 4.388, 8852 E82
8 seen 555.5 888% $5.; «:8 m8 2.88 v.8 83.83 «8892, CE?
8 285g .5065 835828: 88%»: 3.83.2 27¢ 8m :m: 98 38.3 899? P58.
8 529a E855 882.92 3&8”. 83 one ovum; m8 883... 8:85 D828,
2 885388.: SH. 98585 8309.: 2.558 87o 8m 8min 0.8 3.88 883 C88
8 £205 565:: 8882 mofimmmm» mcé Ev 8mg 53 L888. 8N8m<< $349:
9 seen 2,85". H.388: 3J8; 8-m8.m 5: 9w; 0.3 88% o8w8<< $80me

5:89.88“ 868 Sgasmauougfmsmmm :28. oz 8mm.” «.3. 8E8. “.88.“. #5988
v 3885 .2 3°98: SWSBE , ~98 as 9-8.8 8.8 48839 888? $338—

5.. EZN 529a 59.: 8n 2 5.28 . ,

8 8:82... onto 3:. 69.8.. «.8 .0885 SEE, #382

£205 5655, 88°85. N

49

 

$6 833mb? calm—om. 83

Nomad

«

 

 

 

 

 

 

 

 

 

 

 

 

 

v 5205 uczoflm.c_-—n_oo 2 E=E_w_ 5m mdm _ ovumvnmm 505053 $6369

mm E205 9:55; 85.35. W 86.9.9. n96 03. homo; 9mm 5».an Emma;W PEER:
5 F ox__.a_.-u8:uc_.Eu:9o‘ 8989.5. «Flora ‘wmmoo; 8w mom—ed 1mm _ 255me 8585* mxowfi:
v £205 59.: oEN 02E 026:5 owno 595. $6402 0 Em .8-wo.— «.mm _ 8330 885 <3»<I._.<

:5. 805303 9.68 SEES mxmqoané‘ 6:6? :29: 02 E... oTwmd 9mm _ moonvom «.3392 1x38—

F F 590.5 9:32. ommowm :< lem<mwu 8 To man womvd Adm NEQKN NM Fomn<< thug:
op 9:52:00 flange 5:650:04 omMmNmmh< omlmrwm E 0 0mm 0 Tune de ganmmv mmmmmm Em Emm—
o_. 508.5 0265.". owmnwmva omINN_v_>_ hwmood «mm o TmNm v.8 mouanN 280th C. 5(th
:... $0838 9.68 .3ng m.80EoE< 6:89.. 529.: 02 z... 8-mw.m odm nmmmomw Nomwmz tbmmu.

_. mo_oEor_ £203 COEDBE ommfimmmh< 0N FIoPIm... o wpw momwd on Nmoomhm owomwng nFmImx

m mwm>=ow ooflnam Snmnmwz IQDmh on To 8v momwd mfim 8m 5m 58 5:. P528
9 £295 9:.-533EEO 93¢qu YoPIm—u ow-woo.w in 5mm; v.5 vwmmmnu noon E<< twin".
v 5205 14:9 ”03:: .65 9:23.“. oeowmm :< vuwpzou Emood ., 3N domwd «Km .,, @58an 89822. paws—mo—
— omflmficgé 9.25220, omvamuwi 2.29”. , 87w nun .,oTwm.F :m 388 58 F. P503.

50

 

 

 

 

: 5205 5655 883E 5&8: RE Ev Smog 98 8939 38¢. £93?

2 529a 9:25 88mg: 8322; 33 Se 8mm.“ m8 338 .83; #552:

F 3.93 632223556 SCNEE WEmfi o 08 .8-mv.~ m8 .. $53 8.3% #88,
272%.: cum I _ . . , ._ H

v 26535503 26:88. 95:5 82 E 8F 3; 83 8m 9m: N 08 _. 38% F85! £303;

51

detection limits of the microarray technique or might not meet our stringent
reproducibility requirements. However, the channel intensity distribution for AtG UTs
resembled that of the whole array, suggesting the AtGUTs identified were not strongly
biased to either high or low expression levels on the 1 1K microarray. The identification
of highly expressed AtG UTs was an added bonus from this analysis because these
transcripts should greatly facilitate future studies of steps in their degradation. Finally,
multiple members of closely related gene families may be missed because the 11K
microarray was not designed to resolve gene family members. Thus, on the basis of our
work, it seems valid to estimate, that at least 1% of the genes of Arabidopsis correspond
to unstable transcripts.

We used conventional northern blot analysis of cordycepin time courses with
several time points to confirm the 11K microarray data. Four randomly selected
transcripts with half-lives of less than 60 min showed comparable turnover rates in full
cordycepin time courses (Figure 2.2a). Two representative examples are shown in Figure
2.2b-c. In addition, we assessed the statistical significance of the ratio values for the
genes of interest. Using the t-test and the conservative Bonferroni method to adjust p
values (Samuels, 1989) all selected AtGUTs showed significantly different ratios from the
mean of the population at a < 0.0001 (Table 2.1). Several genes with moderately unstable
messages according to the microarray studies were at least moderately unstable in
northern blots (data not shown). Four stable transcripts according to microarray data were
also found stable by northern blot analysis further validating our results. Two

representative examples of these stable transcripts are shown in Figure 2.2b—c.

52

.>
9°
.0

  

 

 

 

 

... m
I momma-33m
5 mo ” ‘5 3° 5° 9° 120 ("I“) . zronzsmz-«m
‘9' , O rzzcrsn/z-nooun
: 4 108017 . A .
g 8° (At-WU) ... 3‘ 5 g 1
so man i 7, a
g 40 W”) . 8% 0.1
‘ ”197' "”5"? {“5 TILL"; “ ‘4"???
rzzcrs .. ~~ ~ < n
a 1" (ASNI) i”. ‘2: "a g;
o -— ' v
(é, 40 WA Danny 04"., 5.. m 150
(QB “0%; Time after addition
‘bs ‘ ‘ ‘x ‘ ‘ ofeordyeepin (min)

Clones

Figure 2.2. Confirmation of the instability of transcripts identified by microarray
analysis. (A) Half-life values determined by northern blot are comparable to estimates from
microarray analysis for four randomly selected unstable transcripts. Northern blot values are
representative of at least three independent cordycepin time courses. (B) Representative northern
blot analysis of cordycepin time courses for two randomly selected unstable and two stable
transcripts. Samples consisted of 10 pg of total RNA isolated from the indicated time points. (C)
Quantitation of the decrease in mRNA abundance and half-life estimation. The signal for eIF 4A
does not change significantly during the time courses and was used as a reference for equal
loading.

General structural features of genes with unstable and stable
transcripts are similar.

The identification of the AtGUTs allowed us to evaluate them for structural
properties that might play a role in determining their instability. We compared the
sequences of the 100 AtG UTs against genes that encode stable transcripts under our
conditions. AtG UTs were evenly distributed throughout the Arabidopsis genome and
showed no significant differences in nucleotide composition, number of introns, size of
the coding sequence and codon usage as compared to genes with stable mRNAs.

We did not expect to find a simple sequence that would be present in the 3’UTR

of all or most AtGUTs because previous observations suggest that many instability

S3

sequences exist (Ross, 1995; Caponigro and Parker, 1996; Gutierrez et al., 1999).
Consistent with this prediction, neither an oligonucleotide frequency approach (van
Helden et al., 1998) nor a probabilistic approach using the MEME software (Bailey and
Elkan, 1994) was indicative of a simple sequence common to all or most AtGUTs
compared to controls (see Materials and Methods for more details) . However a few
AtGUTs have potential ARE-like instability sequences (Greenberg and Belasco, 1993;
Ohme-Takagi et al., 1993; Chen and Shyu, 1995) typified by repeats of the AUUUA
motif: a putative nematode-resistance gene (At2g4000) which encodes the most unstable
transcript in our conditions and two genes of unknown function (At1g72450 and

At2g4l 640). Though the functional significance of these sequences remains to be
determined, these transcripts are potential targets for the AUUUA-mediated decay
pathway in Arabidopsis. Similarly two AtGUTs, the senescence-associated gene SENI
(At4g35770) and a putative light regulated gene similar to the ccr gene from Citrus
paradisi (At3g26740), have DST-like elements in their 3’UTRs. Interestingly, the
expression of these two transcripts is altered in dst], a mutant deficient in DST-mediated
decay (Pérez-Amador et al., 2001). Therefore, these transcripts are potential primary

targets of the DST-mediated decay pathway in Arabidopsis (Pérez-Amador et al., 2001).

AtGUTs are predicted to play a role in a broad range of cellplg
processes but most prominently in transcription.

 

To explore the potential cellular roles of AtGUTs, we analyzed how they were

distributed among the functional categories assigned by the Munich Information Center

for Erotein Sequences (MIPS; Figure 2.3). More than half of the AtGUTs could be

54

IAnnotated genome (Dec 2000)
[1 Genes with unstable transcripts (Dec 2000;
DGenes with unstabletranpscripts May 200 )

  

CELLULAR ORGANIZATION

IONIC HOMEOSTASIS

CELL RESCUE, DEFENSE. CELL DEATH AND AGEING
CELLULAR COMMUNICATION/SIGNAL TRANSDUCTION
CELLULAR BIOGENESIS

INTRACELLULAR TRANSPORT

TRANSPORT FACILITATION

PROTEIN DESTINATION

PROTEIN SYNTHESIS

TRANSCRIPTION

CELL GROWTH, CELL DIVISION AND DNA SYNTHESIS
ENERGY

METABOLISM

 

 

 

 

 

 

 

0 510152025303540
Percentage per category

Figure 2. 3. Instability is associated with a broad range of plant processes. Genes with
unstable transcripts were classified according to the scheme of MIPS. AtGUTs are predicted to
participate in a broad range of cellular processes with transcriptional functions over-represented
compared to what is expected based on the whole genome annotation. To allow comparison,
AtGUTs were classified based on the information released for the annotation of the whole A.
thaliana genome sequence in Dec. 2000. The most updated annotation for the AtGUTs is also
included (May 2002), although a whole genome annotation based on this updated information is
not yet available.

assigned to a MIPS category (Table 2.1) with the remainder lacking similarity to known
proteins. The distribution of predicted functions for AtGUTs suggests they participate in a
broad range of plant processes and in roughly the same proportion as the whole
complement of Arabidopsis genes. Interestingly, enrichment was observed for
transcriptional functions. AtGUTs encode transcriptional functions more than twice the
expected frequency based on the whole Arabidopsis genome annotation (The Arabidopsis
Genome Initiative, 2000). BLAST search analysis (Altschul et al., 1990) indicated that 14
of the 21 AtGUT s that belong to this transcriptional class were not found in the sequenced
genome of H. sapiens, Mus musculus, Rattus norvegicus, C. elegans, D. melanogaster, S.
cerevisiae, Synechocystis, Eubacteria and Archebacteria (BLASTX program, p-value <

0.01) (Table 2.2). This is in line with the detailed analysis of Arabidopsis transcription

factors performed by Riechmann et al. (3 7) that indicated that 45% of those annotated on

55

Table 2.2. Arabidopsis genes with unstable messages that belong to the MIPS
transcriptional category (04) as of May 2002. It should be noted that this
category includes transcription as well as other aspects of RNA metabolism.

Locus

 

genes as
2. Plant specific genes according to the classification by Reichman et al. (2001) (37).

the genome are fi'om families specific to plants, reflecting the independent evolution of
many plant transcription factors. It is possible that plants might also have evolved
mechanisms for regulating the stability of these transcripts that are distinct from those of
other eukaryotes. Plant specific mechanisms might not be exclusive to transcriptional

functions but could also extend to other AtGUTs which are unique to plants.

56

Rapid mRNA dggradation is associated witl_r Arabidopsis responses to
mechanical stimulation and circadian rhythms.

 

To begin understanding the physiological implications of instability in
Arabidopsis, we examined expression of the identified AtGUTs for patterns of regulation.
Hierarchical cluster analysis was performed (see Materials and Methods) using the
publicly available microarray data for the AtGUTs, deposited in the Stanford Microarray
Database (Sherlock et al., 2001) by the AFGC. At the time of this study, 112 slides
corresponding to 47 different experiments carried out under various treatments,
environmental conditions, developmental stages or in different genotypes were available.

Two main clusters of genes were observed. The largest contained 32 genes, the
majority of which were induced by mechanical stimulation (touch; see Table 2.3 for
SMD experiment identifiers) (Figure 2.4). Several of the genes in this cluster also
appeared repressed in an auxin treatment and induced in the histone deacetylase mutant
axe1-4 relative to wild type Arabidopsis plants (Murfett et al., 2001). In addition, most
showed organ-preferential expression with low levels in flowers and high levels in roots
compared to a reference sample prepared from a mixture of plant organs (Figure 2.4 and
Table 2.3). The identification of a touch-induced cluster of AtGUTs is consistent with
touch responses being fast (Braarn and Davis, 1990), and instability being critical when
rapid changes in mRN A steady state levels are to be achieved. In fact, the touch gene
transcripts initially characterized were detected within minutes of treatment and
disappeared very rapidly thereafter, consistent with rapid turnover (Braarn and Davis,
1990). Interestingly, 32 (34%) of the 95 genes induced by the touch treatment in SMD

encode unstable transcripts (Table 2.3). In contrast, only 0.8% of the genes repressed

57

 

Ills! 5.253.883. 1:531:88!
53:62.25!st insist-lei}

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

[cl-8.3.. .2
I L tr .
iig=g§LEE§ol¢I .BIIIH..I .II. .. I .11 1‘ I! t t .r
g.§!§i¥1¢§§h£.§f§ougg algg.lismo~!g&go§ .Bmatm.§pio§es.ua=€.-B§.
. I. It». 1 . (E14. I111.» in . .1411. I.|P..J. 4L»
1:. «$111. littlllhl. . n .. Ii... - 4 .. . con—(1.532553%;
I 1
was down 8:922 .2. 38:
cows down 8:22.: .9 925E
K8 .onmm $52 .9, Eaton.
8; .38 .958 323:5 .9 Eu: 0.: E,
a: +5636
88— .88 .5326
.uvun .88 .68 Etc—
.38 .mmmm New“ .88—
.mto— .38 .mtop .aan
33 .360 on». as; .2 E32... 193
$5 .98 35:8 33:5 .2. 5253: €5<
min .18 .868 033:5 61:253.: ‘88.,
sepia: >23... one... 5* .2252
AN 8 3:9. 3 5.23 09:20 5 608925 9 Sum D3 315......

.59..
8:3 31.62 35:22....

85.52.3938... u§<§§25xcoau<aaslso.Iiaflcsgoasoscoatfissodaifl

58

Touch induced High In ml-4

—| Auxln re reused mutant Low In flowers,
I I Hi In In roots

Clones

  

Experiments

 

Low in flowers, FR induced
PM expressed Low in roots
— —

Clones

 

Experiments
Figure 2. 4. Cluster analysis indicates that a set of AtGUT s is induced by mechanical
stimulation (touch) and another is controlled in a diurnal fashion. Hierarchical cluster
analysis of expression data for the AtGUTs across multiple microarray experiments was
performed with Cluster and Treeview software (28). Expression characteristics shared by genes in
each cluster are indicated (PM = post meridiem; FR= far red light). Each row represents a gene
and each column represents an experiment. Small labels on top of the clusters are SMD
experiment identifiers. Small labels on right side of clusters indicate Arabidopsis thaliana loci or
EST clone identifiers.
under high C02 conditions (experiments 7561 and 7562) and 5% of the genes induced
after 1.5h and 3h H202/N02 co-treatment (experiments 7523 and 9371) did the same. The
number of genes differentially expressed in the C02 and H202/N02 experiments is in the
same range as the overall number of AtGUTs identified. These results indicate that not all
physiological responses have an equal instability component. Moreover, our data suggest
that mRNA instability is an important regulatory component of the touch response.

It is relevant to note that we were unable to confirm the auxin regulation shown

by genes inside the touch cluster. Indole acetic acid (1AA) treatment of two weeks old

Arabidopsis plants did not affect the expression of three different genes in the touch

59

cluster over a 24 hours time course (data not shown). These three genes showed a
transient induction irrespective of the presence or absence of IAA in the solution sprayed.
However, IAA treatment strongly induced the expression of the small auxin up gNA
gene SA UR—A C1 (McClure et al., 1989) compared to the control (data not shown). Our
data suggest that, at least in the conditions tested, the genes that belong to the touch
cluster are not regulated by the phytohorrnone auxin but are induced by spraying alone.
Presumably the genes in the touch cluster are more sensitive to experimental variation
especially if there is mechanical stimulation involved.

A second smaller cluster contained 6 genes whose expression was regulated
diumally (Figure 2.4). In addition, some of these genes were induced by a far-red light
treatment and showed organ-preferential expression with low levels in flowers and roots
(Figure 2.4). Interestingly, some of these genes (At5g67480, At3g62550, At3g26740,
Atl g80920) have been also shown to be controlled by the circadian clock with a peak in
mRNA abundance in the afternoon (Schaffer et al., 2001). Further, 12 other AtGUTs
(At3g15450, At1g37130, At1g75900, At4g31500, At2g39730, At2g32150, Atl gl3260,
At3g55240, At2g35260, Atl 349500, At2g29450, At4g32060) have been reported as
clock-controlled genes (Table 2.3) (Schaffer et al., 2001; Harmer et al., 2000). At similar
transcriptional rates, different mRNA stabilities would translate into circadian mRNA
profiles with distinct phases and amplitudes (So and Rosbash, 1997). Therefore,
instability might be essential for the mRN A oscillatory patterns observed for specific
AtGUTs that are regulated by the clock. Post-transcriptional regulation of mRN A stability
has been shown to play a role in the expression of the Drosophila clock gene per (So and

Rosbash, 1997). There is also evidence that transcription makes a small contribution to

60

the observed circadian expression pattern of the Arabidopsis NIA2 gene (Pilgrim et al.,
1993). Hence, modulation of mRN A stability could also contribute to the clock-regulated
expression of specific AtGUTs. An interesting common feature of the genes in the touch
and diurnal cluster was their low expression level in flowers (Figure 2.4). In fact 40% of
the AtGUTs identified showed similar diminished expression suggesting rapid mRNA
turnover might not be as prominent in flowers as in other organs. Perhaps reproductive
tissues do not require rapid response control as much as vegetative tissues do, but instead
favor the more economical long-lived mRN As which are common in specialized cells
(Weiss and Liebhaber, 1994).

Because genes that showed similar patterns of expression might share regulatory
mechanisms, the 3’UTR sequences of AtGUTs that belong to the touch and light
regulated clusters were analyzed for the presence of common sequence elements as
described previously. Neither the oligonucleotide frequency approach nor the MEME
software provided candidate sequence motifs overrepresented among these subsets of
AtGUTs. Signals that control mRNA metabolism are often composite sequence elements,
e.g. mRNA localization signals (Bashirullah et al., 1998), whose key motifs have been
difficult to identify. Alternatively, a sequence conserved in a small number of AtGUTs
might not be detected by our approach.

Perhaps the most intriguing question for future studies prompted by our work
relates to the AtGUTs regulated by touch, and light. Are these AtGUTs "constitutively"
unstable, or is their mRN A stability regulated in response to light, touch, or other signals?
The approach we describe has the potential to address this and other new exciting

questions about post-transcriptional processes in plants. By comparing global turnover

61

rates under different conditions, post-transcriptional regulatory networks might be
unraveled. This large-scale approach for analysis of post-transcriptional control
mechanisms should help uncover the extent and significance of this level of regulation in

plants in response to a variety of stimuli.

62

References

Altschul,S.F., Gish,W., Miller,W., Myers,E.W., and Lipman,D.J. (1990). Basic local
alignment search tool. J. Mol. Biol. 215, 403-410.

Bailey, T. L. and Elkan, C. Fitting a mixture model by expectation maximization to
discover motifs in biopolymers. Proceedings of the Second International Conference on
Intelligent Systems for Molecular Biology , 28-36. 1994. AAAI Press.

Bashirullah,A., Cooperstock,R.L., and Lipshitz,H.D. (1998). RNA localization in
development. Annu. Rev. Biochem. 67, 335-394.

Braarn,J. and Davis,R.W. (1990). Rain-induced, wound-induced, and touch-induced
expression of calmodulin and calmodulin-related genes in Arabidopsis. Cell 60, 357-364.

Brown,P.O. and Botstein,D. (1999). Exploring the new world of the genome with DNA
microarrays. Nature Genet. 21, 33-37.

Caponigro,G. and Parker,R. (1996). Mechanisms and control of mRNA turnover in
Saccharomyces cerevisiae. Microbiol. Rev. 60, 233-249.

Chen,C.Y.A. and Shyu,A.B. (1995). AU-rich elements - Characterization and importance
in messenger-RNA degradation. Trends Biochem. Sci. 20, 465-470.

Eisen,M.B., Spellman,P.T., Brown,P.O., and Botstein,D. (1998). Cluster analysis and
display of genome-wide expression patterns. Proc. Natl. Acad. Sci. USA 95, 14863-
14868.

F eldbriigge,M., Arizti,P., Sullivan,M.L., Zamore,P.D., Belasco,J.G., and Green,P.J.
(2002). Comparative analysis of the plant mRNA-destabilizing element, DST, in
mammalian and tobacco cells. Plant Mol. Biol. 49, 215-223.

Greenberg,M.E. and Belasco,J.G. (1993). Control of the Decay of Labile Protooncogene
and Cytokine mRNAs. In Control of Messenger RNA Stability, J.Belasco and
G.Brawerman, eds. Academic Press, Inc. San Diego, CA), pp. 199-218.

Gutiérrez,R.A., MacIntosh,G.C., and Green,P.J. (1999). Current perspectives on mRNA
stability in plants: multiple levels and mechanisms of control. Trends Plant Sci. 4, 429-

438.

Hargrove,J.L., Hulsey,M.G., and Beale,E.G. (1991). The kinetics of mammalian gene-
expression. Bioessays 13, 667-674.

63

Harmer,S.L., Hogenesch,J.B., Straume,M., Chang,H.S., Han,B., Zhu,T., Wang,X.,
Kreps,J.A., and Kay,S.A. (2000). Orchestrated transcription of key pathways in
Arabidopsis by the circadian clock. Science 290 , 21 10-21 13.

Holstege,F.C.P., Jennings,E.G., Wyrick,J.J., Lee,T.I., Hengartrrer,C.J., Green,M.R.,
Golub,T.R., Lander,E.S., and Young,R.A. (1998). Dissecting the regulatory circuitry of a
eukaryotic genome. Cell 95, 717-728.

Johnson,M.A., Pérez-Amador,M.A., Lidder,P., and Green,P.J. (2000). Mutants of
Arabidopsis defective in a sequence-specific mRNA degradation pathway. Proc. Natl.
Acad. Sci. U. S. A 97, 13991-13996.

Koshiba,T., Ballas,N., Wong,L.M., and Theologis,A. (1995). Transcriptional regulation
of Ps-IAA 4/5 and Ps-IAA6 early gene-expression by indoleacetic-acid and protein-
synthesis inhibitors in pea (Pisum-sativum). J. Mol. Biol. 253, 396-413.

Lam,L.T., Pickeral,O.K., Peng,A.C., Rosenwald,A., Hurt,E.M., Gilmane,J.M.,
Averett,L.M., Zhao,H., Davis,R.E., Sathyamoorthy,M., Wahl,L.M., Harris,E.D.,
Mikovits,J.A., Monks,A.P., Hollingshead,M.G., Sausville,E.A., and Staudt,L.M. (2001).
Genomic-scale measurement of mRNA turnover and the mechanisms of action of the
anti-cancer drug flavopiridol. Genome Biol. 2, 1-1 1.

Lee,M.L.T., Kuo,F.C., Whitmore,G.A., and Sklar,J. (2000). Importance of replication in
microarray gene expression studies: Statistical methods and evidence fi'om repetitive
cDNA hybridizations. Proc. Natl. Acad. Sci. USA 97, 9834-9839.

McClure,B.A. and Guilfoyle,T. (1989). Rapid redistribution of auxin-regulated RNAS
during gravitropism. Science 243, 91-93.

McClure,B.A., Hagen,G., Brown,C.S., Gee,M.A., and Guilfoyle,T.J. (1989).
Transcription, organization, and sequence of an auxin-regulated gene cluster in soybean.
Plant Cell 1, 229-239.

MurfetLJ., Wang,X.J., Hagen,G., and Guilfoyle,T.J. (2001). Identification of arabidopsis
histone deacetylase HDA6 mutants that affect transgene expression. Plant Cell 13, 1047-
1061.

Newman,T.C., Ohme-Takagi,M., Taylor,C.B., and Green,P.J. (1993). DST sequences,
highly conserved among plant SA UR genes, target reporter transcripts for rapid decay in
tobacco. Plant Cell 5, 701-714.

Ohme-Takagi,M., Taylor,C.B., Newman,T.C., and Green,P.J. (1993). The effect of
sequences with high AU content on mRNA stability in tobacco. Proc. Natl. Acad. Sci.
USA 90,11811-11815.

Peltz,S.W. and Jacobson,A. (1992). mRNA stability: in trans-it. Curr. Opin. Cell Biol. 4,
979-983.

Pérez-Amador,M.A., Lidder,P., Johnson,M.A., Landgraf,J., Wisman,E., and Green,P.J.
(2001). New molecular phenotypes in the dst mutants of Arabidopsis revealed by DNA
microarray analysis. Plant Cell 13, 2703-2717.

Pilgrim,M.L., Caspar,T., Quail,P.H., and McClung,C.R. (1993). Circadian and light-
regulated expression of nitrate reductase in Arabidopsis. Plant Mol. Biol. 23, 349-364.

Ross,J. (1995). Messenger-RNA stability in mammalian-cells. Microbiol. Rev. 59, 423-
450.

Samuels,M.L. (1989). Statistics for Life Science, M.L.Samuels, ed. (San Francisco:
Dellen), pp. 504-505.

Schaffer,R., Landgraf,J., Accerbi,M., Simon,V., V, Larson,M., and Wisman,E. (2001).
Microarray Analysis of Diurnal and Circadian-Regulated Genes in Arabidopsis. Plant
Cell 13, 113-123.

Schaffer,R., Landgraf,J., Pérez-Amador,M., and Wisman,E. (2000). Monitoring genome-
wide expression in plants. Curr. Opin. Biotech. 11, 162-167.

Schena,M., Shalon,D., Davis,R.W., and Brown,P.O. (1995). Quantitative monitoring of
gene-expression patterns with a complementary-DNA microarray. Science 270, 467-470.

Seeley,K.A., Byrne,D.H., and Colbert,J.T. (1992). Red Light-Independent Instability of
Oat Phytochrome mRNA in Vivo. Plant Cell 4, 29-38.

Shaw,G. and Kamen,R. (1986). A conserved AU sequence from the 3' untranslated
region of the GM-CSF messenger-RNA mediates selective messenger-RNA degradation.
Cell 46, 659-667.

Sherlock,G., Hemandez-Boussard,T., Kasarskis,A., Binkley,G., Matese,J.C., Dwight,S.S.,
Kaloper,M., Weng,S., Jin,H., Ball,C.A., Eisen,M.B., Spellman,P.T., Brown,P.O.,
Botstein,D., and Cherry,J.M. (2001). The Stanford Microarray Database. Nucl. Acids Res.
29, 152-155.

Siflow,C.D. and Key,J.L. (1979). Stability of polysome-associated polyadenylated RNA
from soybean suspension culture cells. Biochemistry 18, 1013-1018.

So,W.V. and Rosbash,M. (1997). Post-transcriptional regulation contributes to
Drosophila clock gene mRNA cycling. EMBO J. 16 , 7146-7155.

Sullivan,M.L. and Green,P.J. (1996). Mutational analysis of the DST element in tobacco
cells and transgenic plants: Identification of residues critical for mRNA instability. RNA
2, 308-315.

Taylor,C.B. and Green,P.J. (1995). Identification and characterization of genes with
unstable transcripts (GUT s) in tobacco. Plant Mol. Biol. 28, 27-38.

65

The Arabidopsis Genome Initiative (2000). Analysis of the genome sequence of the
flowering plant Arabidopsis thaliana. Nature 408, 796-815.

Treisman,R. (1985). Transient accumulation of c-fos RNA following serum stimulation
requires a conserved 5' element and c-fos 3' sequences. Cell 42, 889-902.

van Helden,.l., Andre,B., and Collado-Vides,J. (1998). Extracting regulatory sites from
the upstream region of yeast genes by computational analysis of oligonucleotide
fi'equencies. J. Mol. Biol. 281, 827-842.

Vasudevan,S. and Peltz,S.W. (2001). Regulated ARE-mediated mRNA decay in
Saccharomyces cerevisiae. Mol. Cell 7, 1191-1200.

Weiss,I.M. and Liebhaber,S.A. (1994). Erythroid cell-specific determinants of alpha-
globin messenger-RNA stability. Mol. Cell. Biol. 14, 8123-8132.

66

CHAPTER 3

Circadian rhythms and control of mRNA decay: oscillation of the
Arabidopsis Cor-like and SENI transcripts is dependent on normal

DST-mediated mRNA degradationm.

 

m This chapter is part of a bigger project developed in collaboration with Ms. Preet Lidder (PhD candidate.
MSU-DOE Plant Research Laboratory). For completeness, in this Chapter, I discuss experiments carried
out by Lidder P. hereinafier cited as “Lidder, P. and Green, P.J. unpublished results”.

67

Introduction

Plants, like many other organisms, have internal clocks that command biological
rhythms with a period close to 24 hr. These rhythms provide selective advantages
because they allow anticipation of the daily changes in environmental conditions (Green
et al., 2002; Ouyang et al., 1998). Examples of processes that can exhibit circadian
rhythms in plants are leaf movement, hypocotyl elongation, stomatal opening and floral
induction (reviewed in (McClung, 2001)). At the molecular level, DNA microarray
experiments have shown that 2 to 6% of Arabidopsis mRNAs can oscillate (Schaffer et
al., 2001; Harmer etal., 2000). The underlying “master” oscillator that controls these
rhythms is believed to include the late elongated hypocotil or LHY (Schaffer et al., 1998),
the circadian clock-associated protein 1 or CCAI (Wang and Tobin, 1998) and the timing
of CAB expression 1 or TOCI (Millar et al., 1995; Strayer et al., 2000) genes. Similar to
what is observed in other systems (Glossop et al., 1999), LHY, CCAl and TOCl proteins
would form a regulatory loop in which TOCl positively regulates LHY and CCAI gene
expression, and LHY and CCAl proteins in turn repress T 0C1 gene expression (Alabadi
et al., 2001). This loop would output cyclic positive or negative expression commands to
the downstream targets.

Transcriptional regulation is a well established mechanism for the circadian
expression of LHY, CCAI, TOCI and other clock-controlled genes (CCGs). Recently, a
novel sequence element termed the evening element was identified in the promoter region
of 31 Arabidopsis circadian genes, 30 of which peaked at the end of the subjective day
(Harmer et al., 2000). This element was shown to be important for cycling of the CCR2

and TOCI promoter activities (Harmer et al., 2000; Alabadi et al., 2001) and has been

68

proposed as the binding site for LHY and CCAl transcription factors (Alabadi et al.,
2001).

In addition to the cycling in transcriptional activity, it is also clear that post-
transcriptional mechanisms that affect mRNA accumulation can play a role in CCG
expression. The best characterized example corresponds to the per gene (reviewed in
(Stanewsky, 2002)), one of the components of the pacemaker in Drosophila (Panda et al.,
2002). The mRNA of the Drosophila clock gene per can oscillate in the absence of its
natural promoter region (So and Rosbash, 1997). Furthermore, the 5’ upstream portion of
the gene is not sufficient to confer wild—type like oscillation of a luciferase reporter
transcript (Stanewsky et al., 1997). Regulatory elements located in both promoter and
transcribed regions of per gene are necessary to replicate wild-type cycling (So and
Rosbash, 1997; Stanewsky et al., 1997). The functional significance of this post-
transcriptional level of regulation has been well documented (Stanewsky, 2002). For
example, a per-transgene completely devoid of its promoter sequences but including parts
of the first intron is capable of restoring rhythmic behavior in per” mutant flies solely
through post-transcriptional cycling of its mRNA (F risch et al., 1994; So and Rosbash,
1997). In addition, disruption of the 3’ untranslated region of the per gene affects
circadian behavioral rhythms in Drosophila (Chen et al., 1998). Hence post-
transcriptional mechanisms are not only important for normal circadian gene expression
but also for proper gene function.

Post-transcriptional mechanisms have also been invoked to explain circadian
oscillation of a plant gene. Run-on experiments have shown that transcription makes a

small contribution to the observed circadian expression pattern of the Arabidopsis NIA2

69

gene (Pilgrim et al., 1993). Interestingly, overexpression of the Arabidopsis RNA binding
protein AtGRP7/CCR2 can negatively affect accumulation of its own mRNA and that of
the CCG AtGRP8 (I-leintzen et al., 1997). Effectors of mRNA cycling at the post-
transcriptional level are likely to include clock-controlled RNA binding proteins.

Although no examples of changes in mRNA stability regulated by the circadian
clock have been reported, recent evidence suggests a more prominent role of control of
mRNA stability in CCG expression in Arabidopsis. Genes with altered expression in dst],
a mutant deficient in DST-mediated mRNA decay, were biased towards CCGs (Pérez-
Amador et al., 2001). In addition, 16 out of 100 genes with unstable transcripts recently
identified are known clock-controlled genes (Gutierrez et al., 2002). This is a higher than
expected proportion of CCGs with labile mRNAs based on current estimates of total
CCGs in Arabidopsis (Schaffer et al., 2001; Harmer et al., 2000). Together these data
suggest rapid mRNA turnover, perhaps at specific times of the day, and the DST-
mediated rapid mRNA decay, might be important for the specific circadian oscillation of
CCGs in Arabidopsis.

To explore the relationship between mRNA stability and circadian gene expression
we analyzed the stability of selected CCGs with unstable transcripts in Arabidopsis plants
at two times during the day. We found two CCGs that exhibit circadian regulation of
mRNA stability. In addition, analysis of mRNA levels throughout the day in wild-type
and dst] mutant lines for these genes uncovered a previously unknown connection
between diurnal mRN A oscillations and the sequence specific mRNA degradation

mediated by the DST element.

70

Materials and Methods

Arabidopsis strains and ggowth conditions

Arabidopsis thaliana ecotype Columbia, and dst] mutant and 1519 parental lines
(Johnson et al., 2000) were grown on agar plates containing 1x Murashige and Skoog
salts, 1x Gamborg's vitamins and 1% sucrose for two weeks at 22 °C and 16h light and
8h dark cycles unless indicated otherwise. Lighting was provided by fluorescent light

bulbs.

Half-life measurements and preparatiop of RNA samples.

Half-lives were determined as described by Seeley et al. (Seeley et al., 1992) with
the following modifications. Two-week old Arabidopsis plants were transferred to a flask
with incubation buffer (Seeley et al., 1992). After a 30 min incubation, 3'-
deoxyadenosine (cordycepin) was added to a final concentration of 0.6 mM (time 0).
Tissue samples were harvested at regular intervals thereafter and quickly frozen in liquid
nitrogen. Total RNA was isolated and analyzed by northern blot using standard

techniques.

71

Results

Stability of Cor-like and SEN] r_r_rRNAs changes during the day.

 

Previously, it was reported that some Arabidopsis genes that encode highly unstable
mRNAs are also regulated by the circadian clock (Gutierrez et al., 2002). To determine
whether regulation of mRNA stability contributes to the expression of these genes we
measured mRN A decay rates at two times during the day. Two-week old plants grown on
16 h light / 8 h darkness cycles (16/8 LD) were used for cordycepin time courses one
hour after dawn (zeitgeber time 1 or ZTl) as detailed in the Materials and Methods.
Similar experiments were performed 8 h after dawn (ZT8). As shown in Figure 3.la-c,
the putative light regulated gene similar to the cor gene from Citrus paradisi (Ccr-like;
At3g26740), exhibited mRNA stability changes during the day. The transcript for this
gene was significantly more stable in the morning, ZTl , as compared to the afiemoon
(ZT8) (Figure 3.1c). Similarly, turnover of the senescence associated gene 1 (SEN! ;
At4g35770) mRNA was also differentially regulated during the day (Figure 3.2a-c). The
transcript was also more stable in the morning as compared to the afternoon. These data
indicate regulation of mRNA stability is important for the diurnal oscillation of Cor-like
and SEN], two Arabidopsis genes that are controlled by the circadian clock (Schaffer et
al., 2001; Harmer et al., 2000). In contrast, no difference in mRNA turnover rates was
observed for the clock controlled NIA2 gene that encodes nitrate reductase 2 (Atlg37130,
Figure 3.3a-c) and for the light regulated ATHB6 gene which encodes a homeodomain
transcription factor (At2g22430, Figure 3.3c). The mRNA for these genes decayed at

similar rates in the morning and in the aflemoon in the conditions tested.

72

A' Mornlng (Z'l‘l) Alternoon (ZT8)
0153060901200153060901200nln)

Cor-like .‘ in s...

9”“ flue-onus “““ww

 

 

 

 

13- A zrr (rm = 175 a 55 min) C-
o zrs (1,0 = 3s a: 5.0 min)

E 10
a 150
l.’ 200
< - a
a 1 El 150
u a: roo
‘3 50
3 0.1 ' * fl

0 so 100 150 °

ZT-l ZT-O
Time (min)

Figure 3.1. Cor-like mRNA stability is regulated during the day. (A) Representative
northern blot analysis of cordycepin time courses performed in the morning, 1 hour after dawn
(zeitgeber timel or ZTl), and in the afternoon, 8 hours after dawn (ZT 8) for Cor-like and the
loading control eIF4A mRNAs. Samples consisted of 10 pg of total RNA isolated from the
indicated time points. (B) Quantitation of the decrease in mRNA abundance and half-life
estimation. The signal for eIF4A does not change significantly during the time courses and was
used as a reference for equal loading. (C) Half-life values determined by northern blot indicate
Cor-like mRNA is more stable in the morning than in the afiemoon. Values are representative of
three independent cordycepin time courses.

73

Morning (2T1) Afternoon (ZT8)

 

0l5306090120015306090120(nlln)

"-1-.. f.
‘

SEN] as; r = - ~:- , -. . . i as.
. L
.1 - fw ’ .3;

eF4A M M «an a...» {M a; , mm

 

 

 

 

 

 

 

 

 

 

 

 

B. A ZTl (t,,2= 99 :1: 13 min) C,
o zrs (z,,,= 37 a 3.0 min)
5 10
a 250
0
'- 9 200
< 1 a.
g a
In
0
5
g 0.. . .
0 50 100 150
zr-r zr-s
Time (min)

Figure 3. 2. SEN] mRNA stability is regulated during the day. (A) Representative northern
blot analysis of cordycepin time courses performed in the morning, 1 hour after dawn (ZTl), and
in the afternoon, 8 hours after dawn (ZT8), for SEN] and eIF 4A mRNAs. Samples consisted of
10 pg of total RNA isolated from the indicated time points. (B) Quantitation of the decrease in
mRNA abundance and half-life estimation. The signal for the stable eIF4A mRNA was used as a
reference for equal loading. (C) Half-life values indicate SEN] mRNA is more stable in the
morning than in the afternoon. Values are representative of three independent cordycepin time

courses.

74

A.
Morning (ZTl) Afternoon (ZT8)

 

015306090120015306090120(mln)

NIAZ ”-~~ ...... -----.—..

am «ovum... flfi.-“‘

P”

NIA2 C-

A zrr (t,,2=49iS.O min)
0 zrs(:,,,=43is.o min)

 

 

 

 

 

 

'0 16 l 250
3 0

3 E, 150
< :1

1 a
g a...
i ° so
a
g 0.1 - - o

0 50 100 150 zr-r zr-s zr-r zr-a

Time (min) ——

NIA2 AT H86

Figure 3. 3. NIAZ mRNA stability is comparable in the morning and in the afternoon. (A)
Representative northern blot analysis of cordycepin time courses performed in the morning, 1
hour afier dawn (ZTl), and in the afiemoon, 8 hours after dawn (ZT8), for NIA2 and eIF 4A
mRNAs. Samples consisted of 10 pg of total RNA isolated from the indicated time points in a
cordycepin time course. (B) Quantitation of the decrease in mRNA abundance and half-life
estimation. The signal for eIF 4A was used as a reference for equal loading. (C) Half-life values
determined by northern blot indicate NIA2 and ATHB6 mRNAs decay at similar rates in the
morning and in the afternoon. Values are representative of three independent cordycepin time
courses.

75

The well characterized clock gene LHY was used to control for timing of the experiments.
LHY mRNA has a circadian expression pattern with a peak around dawn and with very
low levels throughout most of the day (Schaffer et al., 1998). As expected, LHY was
easily detectable in the morning and was near background levels in the afternoon (Figure
3.4a). LHY mRN A was relatively stable in the morning (ti/2 > 130 min. Figure 3.4b), but
low levels precluded determination of its half-life in the afiemoon experiments. Together
these experiments indicate the half-life changes observed are not the result of differences
in the global cellular mRNA turnover rates in the morning and afternoon. They also
indicate that regulation at the level of mRNA stability is not a general property but rather

specific to some clock-controlled genes.

Cor-like app SEN] mRNA stability changes are dictated by the
circadian clock.

Two possible mechanisms could explain the changes in mRN A stability observed
during the day for Cor-like and SEN] genes. Signaling pathways activated by the changes
in light patterns during the normal day cycle could be responsible as previously observed
for the pea FED] (Elliott et al., 1989) and other genes (Silverthome and Tobin, 1990;
Vorst et al., 1993). Alternatively, the circadian clock could promote the change. To
discriminate between these two possibilities, mRNA half-lives were determined under
free running conditions. Arabidopsis plants were grown for 12 days in 16/8 LD cycles. In
the morning of the 12th day they were transferred to continuous light conditions. Half-
lives were then measured 1 (circadian time 1 or CTl) and 8 (CT8) hours after the

subjective dawn of the 14m day. As shown in Figures 3.5 and 3.6, both Ccr-like and

76

A. Morning (zrr) Afternoon (zrs)

 

 

015306090120 015306090120(min)

MYCCOCD.

elm DOOOCU m as Isl! .. w 4' *

B. a zrr (t,,,> 130 min)

 

.s
O

 

 

 

 

Relative mRNA remaining

 

P
.5

so 1130 150
Time (min)

0

Figure 3. 4. LHY mRNA is highly expressed in the morning and decreases to background
levels in the afternoon. (A) Representative northern blot analysis of cordycepin time courses
performed in the morning, 1 hour afier dawn (ZTl), and in the aftemoon, 8 hours after dawn
(ZT 8), for LHY and eIF4A mRNAs. Samples consisted of 10 pg of total RNA isolated fi'om the
indicated time points. (B) Quantitation of the decrease in mRNA abundance and half-life
estimation. The signal for eIF 4A does not change significantly during the time courses and was
used as a reference for equal loading. Values are representative of three independent cordycepin
time courses.

77

Subjective Dawn (CTl) Subjective Afternoon (CT8)

 

 

0 15 30} 60 90 120 0 15 30 60 90 120 (mm)

eIF4A n.0,; 1.. z”... ham-'7 w m ...4.,; {W m: g; ; rig-,5 w

A CT] (t,,2=4lOi 150 min) C.
0 CT8(t,/2=76:l: 18min)

9‘

 

an
6

Relative mRNA remaining
fl
Half-life (min)

e § § § § § § §

 

 

 

 

.9
in:

 

50 100 150
Time (min)

9

Figure 3. 5. Cor-like mRNA stability is regulated by the Arabidopsis circadian clock. (A)
Representative northern blot analysis of cordycepin time courses performed in the subjective
morning, 1 hour after subjective dawn (circadian time 1 or CT 1), and in the subjective afternoon,
8 hours after subjective dawn (CT 8) for Cor-like and eIF4A mRNAs. Samples consisted of 10 pg
of total RNA isolated from the indicated time points in a cordycepin time course. (B) Quantitation
of the decrease in mRNA abundance and half-life estimation. The stable eIF4A transcript was
used as a reference for equal loading. (C) Half-life values indicate Ccr-like mRNA is more stable
in the subjective morning than in the subjective afternoon. Values are representative of three
independent cordycepin time courses.

78

Subjective Dawn (CTl) Subjective Afternoon (CT8)
015306090120 0153060901207(nnn)

SEN] use was... ufihhuu'

 

eIF4A my a...- M w m w w...» has In! 4.1.1» we in“

 

 

 

 

 

B. A CTl (tm=380:l: 110 min) C.
a 0 CT8(t,,2=953: 12 min)
a 10 g 700
l
A son
a 3 50°
‘3 ' 3' 400
g . 2. ..
a
s =1 1°”
2:: g roo
% o
M 0.1 7 7 ‘
0 50 100 150

Figure 3. 6. SEN] mRNA stability is regulated by the Arabidopsis circadian clock. (A)
Representative northern blot analysis of cordycepin time courses performed in the subjective
morning, 1 hour after subjective dawn (CT 1), and in the subjective afiernoon, 8 hours after
subjective dawn (CT 8), for SEN! and eIF4A mRNAs. Samples consisted of 10 pg of total RNA
isolated from the indicated time points. (B) Quantitation of the decrease in mRNA abundance and
half-life estimation. The signal for eIF4A does not change significantly during the time courses
and was used as a reference for equal loading. (C) Half-life values indicate SEN] mRNA is more
stable in the subjective morning than in the subjective afternoon. Values are representative of
three independent cordycepin time courses.

79

SEN] mRNA stability was regulated under continuous light conditions as seen previously
in the day/night cycles. The transcripts were significantly more stable in the subjective
morning (CTl) as compared to the subjective afternoon (CT 8). In addition, Ccr-like and
SEN] mRNAs were more stable in continuous light as compared to the equivalent times
of the day under regular 16/8 LD cycles. In contrast to SEN] and Ccr-like mRNAs, the
stability of the NIAZ transcript was not affected by the time of the day and was
comparable at CT 1 and CT8 (Figure 3.7). LHY expression was readily detectable at CT]
and close to background levels at CT8, consistent with the planned timing of the
experiments (Figure 3.7a). As before, LHY mRNA was relatively stable in the subjective
morning (ti/2 > 180 min, Figure 3.7b). This data indicates regulation of Cor-like and

SEN] mRNA stability is controlled by the Arabidopsis circadian clock.

DST] fu__nction is involved in the normal oscillatogy expression of Ccr-
like and SEN] genes.

 

Car-like and SEN] genes contain DST-like sequences in the 3’UTR and showed
altered expression levels in dst] mutant as compared to wild-type (Pérez-Amador et al.,
2001). In addition both genes encode unstable mRNAs in the afternoon (Figure 3.1 and
3.2). These features suggest they are targets of the DST-mediated decay pathway. To test
whether DST] function is necessary for the normal diurnal expression of Cor-like and
SEN], mRN A levels were examined throughout the day in dst] mutant and 1519 parental
lines. Two-week old Arabidopsis plants grown on 16/8 LD cycles were harvested every

two hours after dawn. Total RNA was isolated and mRNA levels were examined by

80

A.

Subjective Dawn (Cl‘l) Subjective Afternoon (CT8)

 

 

015306090120 015 30609012001111!)

eIF 4A

”W “W"? ‘9’" the! but it...» ».-- out u w 5"“ a.“
B NIAZ LHY
' A CTl (tm=6l 21:4.4 min)

0 CT8 (ti/2:55i3‘0 min) ‘ CT] (tl/2> 180111111)

 

 

an
6

Relative mRNA remaining
fl
/
Relative mRN A remaining
p-s
D
15

 

 

 

 

 

 

 

 

0.1 r 0.1 fl '
0 50 100 150 0 50 100 150
Time (Mill) Time (min)
6 NIA2

700

600

0 500

E

I 400

E 300

200

100

o

 

CT] CT8

Figure 3. 7. LHY mRNA is highly expressed in the subjective morning and decreases to
background levels in the subjective aftemoon. NIA2 mRNA stability is comparable in the
two conditions. (A) Representative northern blot analysis of cordycepin time courses performed
in the morning, 1 hour after dawn (ZTl), and in the afternoon, 8 hours after dawn (ZT8), for NIA2,
LHY and eIF4A mRNA. Samples consisted of 10 pg of total RNA isolated from the indicated
time points. (B) Quantitation of the decrease in mRNA abundance and half-life estimation. The
stable eIF 4A mRNA was used as a reference for equal loading. (C) Half-life values indicate NIA2
mRNA decays at similar rates in the subjective morning and in the subjective afiemoon. Values
are representative of three independent cordycepin time courses.

81

northern blotting. As shown in Figure 3.8a-b, Cor-like peaked late during the day in the
parental 1519 line. In contrast, in the dst] mutant Cor-like mRNA started to accumulate
later and peaked at least 2 hours later than in the 1519 line (Figure 3.8a-b). In addition to
the delay in the phase, the amplitude of Ccr-like mRNA oscillation was also reduced as
compared to the parental (Figure 3.8b).

The impact of the dst] mutation on SEN] mRNA oscillation was less dramatic but
nevertheless significant. We considered the curves of mRN A levels throughout the day to
be significantly different between the two lines when the error bars of three consecutive
time points or more did not overlap. SEN] gene was greatly induced during the dark
period as reported previously (Oh et al., 1996; Schaffer et al., 2001) (Figure 3.9a-b)..
Transcriptional control is thought to be the main mechanism responsible for this dark
induction (Chung et al., 1997) and as shown in Figure 3.9a-b was mostly unaffected by
the dst] mutation. In contrast to the dark-induced mRN A levels, the clock-controlled
accumulation of SEN] mRNA in the afternoon (Harmer et al., 2000) was abolished in the
dst] mutant (Figure 3.9a-b). These data indicate DST] function is required for the normal

oscillation of Cor-like and SEN] mRNAs.

The effect of the dst] mutation is specific to a subset (@STI targets

To further understand the impact of the dst] mutation on the diurnal oscillation of
targets of the DST-mediated decay pathway, we examined mRN A levels throughout the
day for the SA UR-A CI gene (Johnson et al., 2000; Pérez-Amador et al., 2001). This gene
was first shown to be controlled by the circadian clock in the microarray experiments of

Harmer et al. (2000). SA UR-A C] transcript levels were monitored throughout the day by

82

0 2 4 6 8 10 12 l4 I6 18 20 22 (ZT(IIrI))

Ccr-like s22 . ‘. ‘ ’ ‘1 i aa-
.1... as e QQ QQQQ Qa- - ..

 

 

 

dst]
Ccr-like I" .. ‘.‘. fi Q
"M Q QQ Q Q Q Q Q Q Q
B.
*0— Parental (1519) +11er mutant
[ —
Cor-like
%
E
i
U
'3
.9.
é
o s 10 Is 20 25
ZT (hrs)

Figure 3. 8. Diurnal oscillation of Ccr-like mRNA is altered in the dst] mutant. (A)
Representative northern blot analysis of time courses performed throughout an entire day for Ccr-
like mRNA and eIF4A mRNA in dst] mutant and parental 1519 plants. Samples consisted of 10
pg of total RNA isolated from the indicated times of the day after dawn (ZT=0). (B) Quantitation
of mRNA levels. Data fi'om three independent experiments was used for the time points ZT=0 to
ZT =16. Data from two independent experiments was used for the latest time points (ZT =18,

=20 and ZT =22). All values are relative to the highest mRNA accumulation in either of the
two genetic backgrounds. The signal for eIF 4A was used as a reference for equal loading. The
rectangle above the graph illustrates the 16h day (white segment) and 8h night (grey segment)
period used in these experiments.

83

'— 0 2 4 6 8 10 12 14 16 18 20 22 (ZT(hra))

L—J s151v1 .QQOQQ. m

eIF4A Qmfimmmfindmwn

 

 

 

 

 

 

 

 

 

 

an .
SEN! . " 2“ "" "' 0..
eIF4A Q..O.Q O O,” at as an:

B.

—<>— Parental (1519) + dst] mutant
E _
Sen]
'3
E.
i
0
i
0 5 10 15 20 25
zr (hrs)

Figure 3. 9. Diurnal oscillation of SEN] mRNA is altered in the dst] mutant. (A)
Representative northern blot analysis of time courses performed throughout an entire day for
SEN] mRNA and eIF4A mRNA in dst] mutant and parental 1519 plants. Samples consisted of 10
pg of total RNA isolated firm the indicated times of the day after dawn (ZT=0). (B) Quantitation
of mRNA levels. Data from three independent experiments was used for the time points ZT=O to
ZT=16. Data from two independent experiments was used for the latest time points (ZT=18,

=20 and ZT=22). All values are relative to the highest mRNA accumulation in either of the
two genetic backgrounds. The signal for eIF 4A was used as a reference for equal loading. Inset
shows a magnification of SEN] mRNA levels between ZT =2 and ZT =16. Y- and X-axis in the
inset are “Relative mRNA levels” and “ZT (hrs)” respectively. The rectangle above the graph
illustrates the 16h day (white segment) and 8h night (grey segment) period used in these

experiments.

northern blotting as described before. As shown in Figure 3.10a, SA UR-A C1 oscillation
was similar in mutant and parental plants. These data suggest the dst] mutation does not
disturb the diurnal expression patterns of all DSTl targets, but is associated with a subset

of them.

Diurnal expression of other CCGs is not compromised in the dst]
mutant.

 

As an initial effort to evaluate the impact of the dst] mutation on general CCG
expression, the diurnal oscillation of additional CCGs was investigated in mutant and
parental lines. Two-week old dst] and 1519 parental plants grown on 16/8 LD cycles
were harvested every two hours after dawn. Total RNA was isolated and transcript levels
were examined for AtGRP7/CCR2, a well characterized CCG that oscillates with opposite
phase to LHY and CCAI and that is thought to be a slave oscillator downstream of the
master clock (Heintzen et al., 1997). As shown in Figure 3. 1 Ob, AtGRP7/CCR2 mRNA
oscillation was nearly indistinguishable in dst] mutant and parental lines. Oscillation of
LHY and CCAI expression, two genes thought to be components of the central clock, was
also examined. As shown in Figure 3.10c-d, LHY and CCAI mRN A oscillation was
similar in the dst] mutant and in the parental 1519 line and was consistent with what
previously described (Wang and Tobin, 1998; Schaffer et al., 1998). Interestingly, albeit
similar in amplitude and phase, both LHY and CCAI mRNAs appeared to start
accumulating later in the dst) mutant as compared to the parental. These results indicate

that dst] mutant affect diurnal expression of selected Arabidopsis CCGs.

85

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

-0- Parental (1519) +dafl mutant . —0— Parental (1519) +113" mutant
1 — l -
SA UR-ACI CCAI
1.1
1 * ___.._ ._
éu- __ 3
g .5 4 ._~.._ _._- -. “m. .__,_..._. a
s “ l ’ ‘1 5
“H We! ..
I . . . T A m
0 5 10 I5 20 25 0 5 lo 15 2O 25
z'ram) Z'l'thn)
B. D.
-<>- Punt-H1519) +dst1 mutant -<>- Parental (1519) +4." mutant
1 _ _
LHY

AtGRP‘I/ C CR2

 

 

 

 

 

 

 

 

 

 

 

Relative-MAM
. r. f t 8
L
fl.
J

_. I
+ . .

i l
RelativemRNAlevela
.. a : gréfi.-.A— z .—
{\ifi

. l

 

 

 

 

0 5 u 15 a 15 0 5 10 15 20 25

ZT (hrs) 21' (hrs)

Figure 3.10. Diurnal oscillation of SA UR-A C1 , AtGRP7/CCR2, CCAI and LHY in dst]
mutant and 1519 parental plants. Transcript levels throughout an entire day for (A) SAUR—
ACI, data represented corresponds to two independent experiments. (B) AtGRP7/CCR2, (C)
CCAI, and (D) LHY. mRNA levels were determined by northern blot analysis of time course
experiments as described earlier (e.g. legend to Figure 3.9).

86

Regulation of SEN] mRNA stabilitv is defective in the dst] mutant

To determine whether altered mRNA degradation in the dst] mutant plays a role in
the abnormal diurnal expression of SEN] mRNA, half-life measurements were carried
out at ZTl and 2T8 in both dst] and 1519 plants. Preliminary results are presented in
Figure 3.11 with permission from the author (Preet, L. and Green, P.J. unpublished
results). Consistent with previous experiments (Figure 3.2), SEN] mRN A was rapidly
degraded in the afternoon (ZT8) and stabilized in the morning (ZTl) in parental plants
(Figure 3.11a). Interestingly, the opposite was true for the dst] mutant (Figure 3.1 lb).
SEN] mRNA decayed faster in the morning (ZTl) than in the afternoon (ZT8). Moreover,
SEN] mRNA decay measured in the morning in dst] plants was comparable to the rate of
decay measured in the afternoon in 1519 plants (Figure 3.11c). This data suggests normal
DST] fimction is required for proper timing of the degradation of SENI mRNA. This data
further suggests that the aberrant oscillation of SEN] mRNA in dst] might be caused at

least partly by a defective regulation of its mRNA stability.

87

 

1519

 

 

 

 

Moral-“211) Abram-(1T!)
015306090120 015 306090120

323,1. , .7, _ . , '_l . . '~. .f. r r. L . .
SEN! ;DIQ.~~ 9..“ It *r
1. 'f‘ N {Jr-1'. ‘1'. . fr“ ‘3» . ‘

'1‘" W‘ in: I... a?! ~35?

eIF4A A... a. .. ...

 

 

C° Morning (ZTl)

 

 

0 1519 (In - as Ida)
I 4er a”, - as u.)

 

 

0.01 f . - 1
10 45 78 100 135
Time (mln)

SEN] ... ‘3'! m

elFlA

 

dst]

 

 

 

 

Moral-[(211) Aha-noon (2T!)

015306090120 015 306090120(fll|)
: ~ ~ ' .- U .1. I.."_L.'(a.fi-'l"u .
quaad“
‘ , 31:31:;‘fil. ."1. . 1

as.) no a-' ~.<r-"

 

Afternoon (ZT8)

 

 

 

0 1519 (rm-40 u.)
I an (:m- 10 an)

 

 

Figure 3.11. Regulation of SENI mRNA stability is altered in the dst] mutant. (A)
Northern blot analysis of cordycepin time course performed in 1519 plants, 1 hour after dawn
(2'11) and 8 hours after dawn (ZT8), for SEN] and eIF 4A mRNAs. (B) Northern blot analysis of
cordycepin time course performed in dst! plants, 1 hour afier dawn (ZTl) and 8 hours afier dawn
(ZT8), for SEN] and eIF 4A mRNAs. Samples consisted of 10 ug of total RNA isolated from the
indicated time points. (C) Quantitation of the decrease in mRNA abundance and half-life
estimation. The signal for eIF 4A does not change significantly during the time courses and was

used as a reference for equal loading.

88

Discussion

It is clear that transcriptional control plays an important role in the circadian
expression of many genes. Indeed, a transcriptional regulatory loop lays at the heart of
the master oscillator in plants and other systems ((Alabadi et al., 2001); reviewed in
(Panda et al., 2002)). However, transcription alone is not sufficient to explain the
oscillation in mRNA levels of every clock gene (Pilgrim et al., 1993; So and Rosbash,
1997). In fact, both transcriptional and post-transcriptional control of mRNA
accumulation are necessary for normal per function and circadian behavior in flies (So
and Rosbash, 1997; Stanewsky et al., 1997; Chen et al., 1998). Although it is now clear
that post-transcriptional mechanisms can contribute significantly to the expression of
clock genes, to date there is no direct evidence that the stability of per transcript, or that
of any other gene, can be controlled by the circadian clock.

We have shown that Ccr-like and SEN! mRNA decay is regulated as a function of
the time of the day. Both Ccr-like and SEN] mRN As were stable in the morning at 2'1“]
and highly unstable in the afternoon at ZT8 in Arabidopsis plants grown in 16/8 LD
cycles. This change in mRNA turnover rates prevailed in free running conditions
indicating that it is controlled by the circadian clock. Cor-like and SEN] mRNAs were
more stable in the subjective morning (CT 1) as compared to the subjective afternoon
(CT8) two days after plants entrained in 16/8 LD cycles were transferred to continuous
light. In contrast to Cor-like and SEN], the NIA2 and ATHB6 mRNAs decayed at similar
rates at 2T] and ZT8. Curiously, previous run-on experiments in 5 weeks plants grown
on soil indicated that NIA2 transcription was relatively constant throughout the day

(Pilgrim et al., 1993). These results suggested that mRNA stability changes are

89

responsible for the approximately 2-fold circadian oscillation of the NIA2 mRNA
(Pilgrim et al., 1993). However, our results suggest that under the conditions studied,
regulation of mRNA stability does not contribute to diurnal oscillation of the NIAZ
transcript. Developmental regulation could account for the discrepancy. Alternatively, the
presence of sucrose in our experiments could override the circadian regulation. Sucrose
in the growth media have been shown to induce nitrate reductase activity irrespective of
the light conditions (Crawford et al., 1992). Together these results indicate that the
changes observed are not the result of a global change in mRNA turnover rates. Instead
regulation of mRNA stability is a prOperty of selected clock-controlled genes. This data
further suggests that post-transcriptional control of mRNA stability is important for the
circadian expression of Ccr-like and SEN] genes. Circadian rhythms allow anticipation
of the daily changes in environmental conditions, thus providing an adaptive advantage
(Ouyang et al., 1998). Consistent with this anticipatory nature of clock-controlled
processes, increased mRNA stability preceded the circadian mRNA accumulation phase
and rapid turnover was followed by the decrease in transcript levels.

Intriguingly, Cor-like and SEN] mRN As were more stable in continuous light as
compared to the equivalent times of the day under regular 16/8 LD cycles. An appealing
hypothesis to explain this moderate stabilization would be that the circadian clock
promotes mRN A degradation. In the absence of the zeitgeber (time-giver), the clock-
induced degradation could dampen or shift phase over time leading to the observed
partial stabilization. Dampening and shift in the phase of circadian phenotypes is
commonly observed under free running conditions. In this scenario, regulation of mRNA

stability could still be achieved if the decay machinery but not the circadian mechanism

90

that triggers the decay is altered. Alternatively, continuous light or the absence of a dark
period might repress or otherwise impair the normal function of the mRN A decay
machinery involved in Cor-like and SEN] transcript degradation. Further research is
required to distinguish between these possibilities.

Cor-like and SEN] mRN As are thought to be targets of the DST-mediated mRNA
degradation pathway (Pérez-Amador et al., 2001; Gutierrez et al., 2002). To evaluate the
relevance of this sequence-specific mRNA decay pathway for the expression of Cor-like
and SEN] genes, the diurnal oscillation of their transcripts was studied in dst], a mutant
defective in DST-mediated mRN A degradation (Johnson et al., 2000). We showed that
Cor-like and SEN] mRN A oscillations were altered in dst] mutant as compared to
parental plants. In contrast, oscillation of SA UR-A C1 transcript, another target of this
mRNA decay pathway, was comparable in dst] mutant and 1519 parental line. These
results indicate that DST] function is necessary for the normal diurnal oscillation of
selected targets of the DST-mediated decay pathway. The reason for this selectivity is not
clear at this point. No obvious bias in the 3’UTR sequences could be detected. In addition,
two copies of the DST sequence element failed to confer diurnal oscillation to a reporter
mRNA (data not shown). Therefore sequence requirements can not be predicted.
However previous studies have suggested different modes of recognition of the DST
sequence in different cell types (Feldbriigge et al., 2002; Sullivan and Green, 1996).
There is also precedent for differences in DST-element function depending upon the
sequence context in which is present (Newman et al., 1993). Further experiments are
needed to understand the function of the DST-element and the molecular basis of this

selectivity.

91

Based on the characteristics of the dst] mutant, a plausible explanation for the
molecular phenotypes discussed above is defective mRN A decay. In fact, preliminary
experiments argue that DST] function is important for proper timing of mRN A
degradation. SEN] mRN A was stable in the afternoon and very unstable in the morning
in the dst] mutant, a pattern of regulation that was completely opposite to what is seen in
parental plants (Preet, P. and Green, P.J. unpublished results). This alteration in the
pattern of regulation of SEN] mRNA decay can explain the difference in mRNA levels
throughout the day in dst] mutant as compared to parental. In wild type plants,
accumulation of mRN A during the morning is preceded by stabilization of the message
and decrease in the transcript levels during the afternoon is preceded by SEN] mRN A
rapid turnover. In contrast, in the dst] mutant, rapid decay of SEN] transcript in the
morning would prevent the message from accumulating early during the day. The
increase in SEN] mRN A stability in the afternoon in the dst] mutant should lead to
accumulation of the transcript with a peak later during the day. However no accumulation
of SEN] is evident in dst] mutant during the 16 h light period. A possible explanation for
this result is that SEN] mRNA accumulates towards the end of the day and it is masked
by the strong accumulation induced during the dark period in normal day/night cycles.
Alternatively, the stabilization of SEN] mRN A could be temporally shorter in the dst]
mutant as compared to parental and insufficiently long for detectable increase in mRN A
levels. Additional experiments conducted in continuous light conditions should help
distinguish between these two possibilities.

With regard to the relationship between DST] and the circadian clock, one

interpretation of the data is that DST] fimctions downstream of the master circadian

92

oscillator, receives signals from the clock and affects the expression of specific CCGs at
the post-transcriptional level. This model would predict that dst] plants do not have a
severe clock defect. Normal oscillation of SA UR-A C] and AtGRP7/CCR2 are consistent
with this model. In addition, preliminary mapping experiments of dst] are also in
accordance with such explanation. DST] would be located at the bottom of Chromosome
1, far fiom the genes lmown to be important for clock function (Lidder, P. and Green, P.J.
unpublished results). However, mutants in clock-associated genes can affect diurnal
oscillation of CCGs (Somers et al., 2000; Alabadi et al., 2002; Mizoguchi et al., 2002).
The clock phenotypes of such mutants, are often more evident under free-running
conditions (Harmer et al., 2001). Hence, some circadian alterations in the dst] mutant
might not be readily detected in the presence of the zeitgeber, due to resetting of the
phase every morning. This could explain that despite reaching similar levels and peaking
at the same time of the day, LHY and CCA] mRNAs appeared to start the accumulation
phase later in dst] than in 1519 plants. In this model, dst] would be associated with or
affect the function of the central oscillator. This in turn would affect downstream
circadian output pathways including oscillation of CCGs. To better understand the
relationship between the dst] mutation and circadian rhythms, classical circadian
phenotypes are currently being analyzed in collaboration with Dr. C Robertson McClung
(Dept. of Biological Sciences, Dartmouth College).

Regardless of the mechanistic relationship between DST] and the circadian clock,
we have uncovered a previously unknown connection between a sequence-specific
mRNA decay pathway and expression of CCGs. Our results also strengthen the idea that

DST sequences are more functionally versatile than previously anticipated. Recognition

93

and mode of action of the DST-mediated decay pathway might go beyond simple rapid
turnover. Precise regulation of decay might be essential for proper expression of selected
CCGs in Arabidopsis. Although DST-sequence recognition appears to be unique to plants,
the relationship between post-transcriptional control and circadian rhythms might be of

significance to other eukaryotes.

94

References

Alabadi,D., Oyama,T., Yanovsky,M.J., Harmon,F.G., Mas,P., and Kay,S.A. (2001).
Reciprocal regulation between TOCl and LHY/CCAl within the Arabidopsis circadian
clock. Science 293, 880-883.

Alabadi,D., Yanovsky,M.J., Mas,P., Harmer,S.L., and Kay,S.A. (2002). Critical role for
CCAl and LHY in maintaining circadian rhythmicity in Arabidopsis. Curr. Biol. 12, 757-
761.

Chen,Y., Hunter-Ensor,M., Schotland,P., and Sehgal,A. (1998). Alterations of per RNA
in noncoding regions affect periodicity of circadian behavioral rhythms. J. Biol. Rhythms
13, 364-379.

Chung,B.C., Lee,S.Y., Oh,S.A., Rhew,T.H., Nam,H.G., and Lee,C.H. (1997). The
promoter activity of senl , a senescence-associated gene of Arabidopsis, is repressed by
sugars. J. Plant Physiol. 15] , 339-345.

Crawford,N.M., Wilkinson,J.Q., and Labrie,S.T. (1992). Control of nitrate reduction in
plants. Aust. J Plant Physiol. 19, 377-385.

Elliott,R.C., Dickey,L.F., White,M.J., and Thompson,W.F. (1989). cis-acting elements
for light regulation of pea ferredoxin I gene expression are located within transcribed
sequences. Plant Cell I, 691-698.

Feldbriigge,M., Arizti,P., Sullivan,M.L., Zamore,P.D., Belasco,J.G., and Green,P.J.
(2002). Comparative analysis of the plant mRNA-destabilizing element, DST, in
mammalian and tobacco cells. Plant Mol. Biol. 49, 215-223.

Frisch,B., Hardin,P.E., Hamblen-Coyle,M.J., Rosbash,M., and Hall,J.C. (1994). A
promoterless period gene mediates behavioral rhythmicity and cyclical per expression in
a restricted subset of the Drosophila nervous system. Neuron 12, 555-570.

Glossop,N.R., Lyons,L.C., and Hardin,P.E. (1999). Interlocked feedback loops within the
Drosophila circadian oscillator. Science 286, 766-768.

Green,R.M., Tingay,S., Wang,Z.Y., and Tobin,E.M. (2002). Circadian rhythms confer a
higher level of fitness to Arabidopsis plants. Plant Physiol 129, 576-584.

Gutiérrez,R.A., Ewing,R.M., Cherry,J.M., and Green,P.J. (2002). Identification of
unstable transcripts in Arabidopsis by cDNA microarray analysis: Rapid decay is
associated with a group of touch- and specific clock-controlled genes. Proc. Natl. Acad.
Sci. USA 99,11513-11518.

Gutiérrez,R.A., Ewing,R.M., Cherry,J.M., and Green,P.J. (2002). Identification of
unstable transcripts in Arabidopsis by cDNA microarray analysis: rapid decay is

95

associated with a group of touch- and specific clock-controlled genes. Proc. Natl. Acad.
Sci. U. S. A 99,11513-11518.

Harmer,S.L., Hogenesch,J.B., Straume,M., Chang,H.S., Han,B., Zhu,T., Wang,X.,
Kreps,J.A., and Kay,S.A. (2000). Orchestrated transcription of key pathways in
Arabidopsis by the circadian clock. Science 290 , 21 10-21 13.

Harmer,S.L., Panda,S., and Kay,S.A. (2001). Molecular bases of circadian rhythms.
Annu. Rev. Cell Dev. Biol. 1 7, 215-253.

Heintzen,C., Nater,M., Apel,K., and Staiger,D. (1997). AtGRP7, a nuclear RNA-binding
protein as a component of a circadian- regulated negative feedback loop in Arabidopsis
thaliana. Proc. Natl. Acad. Sci. U. S. A 94, 8515-8520.

Johnson,M.A., Pérez-Amador,M.A., Lidder,P., and Green,P.J. (2000). Mutants of
Arabidopsis defective in a sequence-specific mRNA degradation pathway. Proc. Natl.
Acad. Sci. U. S. A 97, 13991-13996.

McClung,C.R. (2001). Circadian rythms in plants. Annu. Rev. Plant Physiol Plant Mol.
Biol. 52, 139-162.

Millar,A.J., Carre,I.A., Strayer,C.A., Chua,N.H., and Kay,S.A. (1995). Circadian clock
mutants in Arabidopsis identified by luciferase imaging. Science 267, 1 161-1163.

Mizoguchi,T., Wheatley,K., Hanzawa,Y., Wright,L., Mizoguchi,M., Song,H.R.,
Carre,I.A., and Coupland,G. (2002). LHY and CCA] are partially redundant genes
required to maintain circadian rhythms in Arabidopsis. Dev. Cell 2, 629-641.

Newman,T.C., Ohme-Takagi,M., Taylor,C.B., and Green,P.J. (1993). DST sequences,
highly conserved among plant SA UR genes, target reporter transcripts for rapid decay in
tobacco. Plant Cell 5, 701-714.

Oh, S.A., Lee, S.Y., Chung, I.K., Lee, CH. and Nam, H.G. (1996). A senescence-
associated gene of Arabidopsis thaliana is distinctively regulated during natural and
artificially induced leaf senescence. Plant Mol. Biol. 30, 739-754.

Ouyang,Y., Andersson,C.R., Kondo,T., Golden,S.S., and Johnson,C.H. (1998).
Resonating circadian clocks enhance fitness in cyanobacteria. Proc. Natl. Acad. Sci. U. S.
A 95, 8660-8664.

Panda,S., Hogenesch,J.B., and Kay,S.A. (2002). Circadian rhythms fi'om flies to human.
Nature 417, 329-335.

Pérez-Amador,M.A., Lidder,P., Johnson,M.A., Landgraf,J., Wisman,E., and Green,P.J.

(2001). New molecular phenotypes in the dst mutants of Arabidopsis revealed by DNA
microarray analysis. Plant Cell 13, 2703-2717.

96

Pilgrim,M.L., Caspar,T., Quail,P.H., and McClung,C.R. (1993). Circadian and light-
regulated expression of nitrate reductase in Arabidopsis. Plant Mol. Biol. 23, 349-364.

Schaffer,R., Landgraf,J., Accerbi,M., Simon,V., V, Larson,M., and Wisman,E. (2001).
Microarray Analysis of Diurnal and Circadian-Regulated Genes in Arabidopsis. Plant
Cell 13, 113-123.

Schaffer,R., Ramsay,N., Samach,A., Corden,S., Putterill,J., Carre,I.A., and Coupland,G.
(1998). The late elongated hypocotyl mutation of Arabidopsis disrupts circadian rhythms
and the photoperiodic control of flowering. Cell 93, 1219-1229.

Schaffer,R., Landgraf,J., Accerbi,M., Simon,V., V, Larson,M., and Wisman,E. (2001).
Microarray Analysis of Diurnal and Circadian-Regulated Genes in Arabidopsis. Plant
Cell 13, 113-123.

Seeley,K.A., Byrne,D.H., and Colbert,J.T. (1992). Red Light-Independent Instability of
Oat Phytochrome mRNA in Vivo. Plant Cell 4, 29-38.

Silverthome,J. and Tobin,E.M. (1990). Posttranscriptional regulation of organ-specific
expression of individual rbc-S messenger-RNAS in Lemna-gibba. Plant Cell 2, 1181-
1190.

So,W.V. and Rosbash,M. (1997). Post-transcriptional regulation contributes to
Drosophila clock gene mRNA cycling. EMBO J. 16 , 7146-7155.

Somers,D.E., Schultz,T.F., Milnamow,M., and Kay,S.A. (2000). ZEITLUPE encodes a
novel clock-associated PAS protein from Arabidopsis. Cell 10], 319-329.

Stanewsky,R. (2002). Clock mechanisms in Drosophila. Cell Tissue Res. 309, 11-26.

Stanewsky,R., Jamison,C.F., Plautz,J.D., Kay,S.A., and Hall,J.C. (1997). Multiple
circadian-regulated elements contribute to cycling period gene expression in Drosophila.
EMBO J. 16, 5006-5018.

Strayer,C., Oyama,T., Schultz,T.F., Raman,R., Somers,D.E., Mas,P., Panda,S.,
Kreps,J.A., and Kay,S.A. (2000). Cloning of the Arabidopsis clock gene TOCl , an
autorcgulatory response regulator homolog. Science 289, 768-771.

Sullivan,M.L. and Green,P.J. (1996). Mutational analysis of the DST element in tobacco
cells and transgenic plants: Identification of residues critical for mRN A instability. RNA
2, 308-315.

Vorst,O., Vandam,F., Weisbeek,P., and Smeekens,S. (1993). Light-regulated expression

of the Arabidopsis-thaliana ferrodoxin-A gene involves both transcriptional and
posttranscriptional processes. Plant J. 3, 793-803.

97

Wang,Z.Y. and Tobin,E.M. (1998). Constitutive expression of the CIRCADIAN CLOCK
ASSOCIATED l (CCAl) gene disrupts circadian rhythms and suppresses its own
expression. Cell 93, 1207-1217.

98

CHAPTER 4

Mammalian determinants of long mRNA half-life and their

significance to plants

99

Introduction

Plant transcripts decay with rates similar to those in other multicellular eukaryotes.
The spectrum of mRNA half-lives extends from an hour or less for unstable messages to
days or more for very stable transcripts. Most mRN As however, appear to decay with
half-lives in the order of hours (Siflow and Key, 1979; Hargrove et al., 1991). The rate
with which individual mRN A species are degraded is determined by a combination of
general and specific structural features (Sachs, 1993; Ross, 1995; Chen and Shyu, 1995;
Caponigro and Parker, 1996; Gutierrez et al., 1999). Structures that are common to
virtually all eukaryotic mRNAs are the 5’ cap and the 3’ polyadenylate (poly(A)) tail.
Although it is not yet clear how the cap and poly(A) tail removal are regulated in plants,
these structures are thought to play a general role in the overall stability of plant
transcripts. Transcript-specific sequence elements that can affect mRNA stability in
plants and other eukaryotes have been identified (reviewed in Ross, 1995; Caponigro and
Parker, 1996; Gutierrez et al., 1999). Adenylate/uridylate-rich elements (ARES) a
common determinant of mRNA stability in mammalian cells are one of the best
characterized. ARES are approximately 50-150 nucleotides long, usually contain multiple
copies of the AUUUA motif and a high content of uridine. ARES are located in the
3’UTR of mRNAs encoding a variety of proto-oncoproteins, cytokines and transcription
factors (Chen and Shyu, 1995). Transcripts that contain them are selectively targeted for
rapid decay (Chen and Shyu, 1995). Due to the significance of AUUUA elements in
mammals, a synthetic AUUUA repeat was tested for the ability to act as instability
determinant in plants. Reporter transcripts containing 1 1 repeats of the AUUUA motif in

their 3’UTRs were degraded much more rapidly in stably transformed tobacco cells and

100

accumulated to a lower level in transgenic tobacco plants than those of control constructs
(Ohme—Takagi et al., 1993). These results suggest that the mRNA decay pathway
mediated by AUUUA repeats is conserved between animals and plants. In plants, the
DST or downstream element has been best characterized. Originally identified as a
conserved sequence in the 3’ untranslated region (UTR) of the unstable small auxin-up
RNA (SA UR) transcripts (McClure et al., 1989) DST sequences were first showed to
target reporter transcripts for rapid decay in BY—2 cells (Newman et al., 1993).

It is now clear that unstable mRNAs in plants contain instability sequences, however
it is still unclear whether stable ones contain discrete stabilizing determinants, conditional
stabilization elements or are stable by default. In fact comparatively little attention has
been devoted to the long-lived end of the mRNA stability spectra in plants or other
systems. However the relevance of stable mRNAs is well justified from a cellular
economy perspective (Russell et al., 1997) (Mehdy and Brodl, 1998) and is perhaps most
evident in cases where it is lost (below) (Russell et al., 1997). The best understood case
of mRN A stabilization occurs during red blood cell differentiation. The (Jr-globin mRNA
is selectively stabilized during erythroid cell development and at the late reticulocyte
stage accounts for 95-98% of the total mRNA (Russell et al., 1997). A single sense
mutation in the translation termination codon of the or-globin mRN A severely
destabilizes the transcript (Weiss and Liebhaber, 1994). This destabilization results in
greater than 95% reduction in (It-globin expression and accounts for one of the most

common non-deletional or-thalassernias throughout the world (Russell et al., 1997).

Efforts to understand the molecular basis of or-globin mRNA longevity identified a

pyrimidine-rich sequence in the 3’ untranslated region as essential for the transcript’s

101

stability (Weiss and Liebhaber, 1995). Several proteins can bind to this region and form a
ribonucleoprotein complex in vitro (Wang et al., 1995). Mutations that affect the
formation of this complex correlated with destabilization effect on the a-globin transcript,
suggesting that the complex is a major determinant of a-globin mRNA stability (Wang et
al., 1995). Formation of the a-complex was also observed in mouse erythroleukernia
cells (MEL), human erythroleukernia cells (K562), fibroblasts (C127), human epithelioid
carcinoma cells (HeLa) and rat adrenal pheochromocytoma cells (PC-12) (Wang et al.,
1995; Holcik and Liebhaber, 1997). Furthermore, other stable transcripts (e.g. B-globin,
lS-lipoxygenase, tyrosine hydroxylase and arm-collagen) possess similar sequences in
their 3’ UTR. These sequences are capable of forming related ribonucleoprotein
complexes suggesting that the mechanism of a-globin stabilization might have a broader
significance in mammalian cells (Holcik and Liebhaber, 1997).

Besides their crucial role in the regulation of endogenous genes, post-transcriptional
mechanisms that promote long-lived mRNAs are also becoming important considerations
in plant biotechnology. Attaching a strong promoter to a gene will not always guarantee
high expression in plants (Diehn et al., 1996). Expression of a foreign gene can be limited
at the mRNA level through mechanisms such as aberrant splicing, aberrant
polyadenylation, and rapid mRNA degradation (Diehn et al., 1998; De Rocher et al.,
1998).

As a first step to gain insight into the molecular principles that govern slow mRNA
decay rates in plants we investigated the effect of the pyrimidine-rich sequence derived
fi'om the mammalian (at-globin transcript and a related synthetic sequence on mRN A

stability in three different plant expression systems.

102

Materials and methods

Plant materials and culture

Arabidopsis thaliana ecotype Columbia plants were grown in soil under a 12-h
light/12-h dark cycle at 20°C. Nicotiana tabacum cv Bright Yellow 2 (BY-2) cells (An,
1985; NAGATA et al., 1992) were cultured as described by Newman et al. (1993). Maize
(Zea mays cv BMS) cells (a gift from Pioneer Hi-Bred International, Inc., Des Moines,

IA) were grown as described previously (DeRocher et al., 1998).

Gene constructions

In order to test the activity of the putative stability sequences in plants, expression
cassettes consisting of a double enhanced 35S promoter followed by the dihydrofolate
reductase coding region (DHF R), and the 3’UTR of the pea rubisco small subunit E9
gene (E9 tail) were made (pl 801; Figure 4.1). A duplication of the 42 nt long
stabilization sequence from the (at-globin 3’UTR was then cloned downstream of the
DHFR coding region and upstream of the E9 tail to generate construct p1802 (DHFR-
gSE). In a separate construct, a synthetic sequence (84 nt long) derived from the
consensus for the stability element (Holcik and Liebhaber, 1997) was inserted instead to
generate construct p1803 (DHFR-cSE). The plant expression cassettes from pl 801,
p1802 and pl 803 were then moved into the Hind 111 site of plasmid pBllZl (Jefferson et

al., 1987) to generate constructs p1804, p1805, p1814 respectively. pB1121 posses a

103

2xCaMV 35$ DHFR rch E9 3’

 

 

 

 

 

 

 

p1804 : H I 1 1 .-
2xCaMV 358
p1805 : f ’
a—globln
”nonsense
«(D-coma"!
Tyrodnollydroxylaae
Consensus = (CCCU)2,
2xCaMV 358 DHFR : [was E9 3’
p1814 : I ' F l f l :

Figure 4. 1. Reporter system for testing putative stabilization elements derived fiom
mammalian transcripts. Test sequences were introduced into a plant expression cassette as
illustrated, downstream of the dihydrofolate reductase coding region, and upstream of the 3’UTR
of the pea rubisco small subunit E9 gene. All chimeric genes were under the control of a double
enhanced CaMV 35S promoter. p1804: DHFR control (DHFR), p1805: DHFR and (at-globin
stability determinant (DHFR-gSE). p18 14: DHFR and (CCCU)21, a synthetic sequence based on
the sequence alignment of (Jr-globin, Lipoxygenase, 0t(I)-collagen and Tyrosine Hydroxylase
3’UTRs (see text for details) (DHFR-cSE).

nopaline synthase driven neomycin phosphotransferase II gene that confers kanamicin

resistance to transgenic plants.

Protoplast preparation and transformation

Protoplasts from BY-2 or BMS cells were prepared and transformed as described
previously (Diehn, 1998). Briefly, the cells from a 3-4 days subculture were incubated in
2 % cellulysin, l % cytolase and 0.2 % pectolyase in buffer KMC 700 (8.65g KCl,

16.47g MgClz.6H20, 12.5g CaClz.2HzO, 5g MES in 900ml ddeo, pH 6 with KOH;

104

osmolarity 700mOsm) for 3-5 h at 28°C with gentle agitation. Hydrolytic enzymes were
prepared as described previously (van Hoof and Green, 1996). The cells were monitored
under microscope periodically to evaluate protoplast formation. The protoplasts were
then sieved, washed in KMC 650 (like KMC 700 except osmolarity was adjusted to
650mOsm), washed in KMC 600 (like KMC 700 except osmolarity was adjusted to
600mOsm) and then counted. For transformation, 0.5 mL of the final protoplast
suspension (8x106 cells/mL) were incubated with 100 uL of plasmid DNA (1 pg/uL) in
the presence of 0.5 mL PEG solution (40% PEG 8000 (w/v), 0.4M mannitol, 0.1M
Ca(N03)2.4HzO, pH 9.0). After 30 min at room temperature, W6 solution (0.38g KCl, 9g
CaC12.2HzO, 9g NaCl, 9g glucose, lg MES (2[N-morpholino]ethanesulfonic acid) in
600ml ddeO, pH 6.0, osmolarity 550-600mOsm) was added stepwise in 1, 2, and 5 mL
volumes 5 min apart. After centrifugation, protoplasts were resuspended in F W media
(4.3g 1'1 MS salts (Gibco BRL, Gaithersburg, MD), 5m11'l 200x 135 vitamins, 30g 1"
sucrose, 1.5g 1'1 proline, 54g 1'1 mannitol, 3mg 1'1 2,4-D, pH 5.7 with KOH) and
transferred to 100 mm Petri plates, where they were incubated at 28°C in the dark for 12

or 16h before harvesting.

Plant transformation

Arabidopsis plants were transformed using the infiltration method described in
Chapter 5. Trangenic plants were selected by germinating the seeds of infiltrated plants
on agar plates containing 1x Murashige and Skoog salts, 1x Gamborg's vitamins, 1%

sucrose and 50 ug/mL kanamycin as selectable marker.

105

RNA isolation and northern blot analysis

RNA was extracted from plant tissue as described previously (Newman et al., 1993).
RNA was isolated fiom BY-2 or BMS protoplasts using the same method except that the
protoplasts were not grounded under liquid N2, but thawed directly in 2.5 mL of
guanidiniurn thiocyanate solution while vortexing. Aliquots of total RNA from
transiently transformed protoplasts were treated with DNaseI for 15 min at 37°C to
remove residual plasmid DNA previous to northern blot analysis. Equal amounts of total
RNA (20 rig/lane) were used for RNA gel blots and hybridization using standard

techniques.

106

Results

OL-Globin stabilization element and a related sequence do pot 1pcrea_s_e
abundance of a reporter transcript in mai_ze or in tobaccurotoplasts

 

To test if the (It-globin pyrimidine-rich element is recognized in plants as an
stability determinant, two copies of the 42 nt sequence fi'om the a-globin transcript
3’UTR (Holcik and Liebhaber, 1997) was cloned between the DHFR coding region and
the 3’ UTR of the pea rubisco small subunit E9 (p1805, DHFR-gSE. Figure 4.1).
Sequences similar to the (at-globin stability element have been identified in three other
stable mammalian mRNAs (Figure 4.1; (Holcik and Liebhaber, 1997)). A synthetic
“stabilization” element was then synthesized based on the consensus in the sequence
comparison and introduced in a separate plant expression cassette (pl 814, DHFR-cSE.
Figure 4.1). The relative mRNA abundance for the different constructs was studied in
maize and tobacco cells. Protoplasts fi'om maize and tobacco cells were prepared and
transiently transformed with plasmids pl 804 (DHFR), p1805 (DHFR-gSE) and p1814
(DHFR-cSE). Total RNA was isolated from samples harvested 12 and 16 h after
transformation and analyzed by northern blot hybridization using 32P-labeled DHFR
coding region as probe. As shown in Figure 4.2a-b, the reporter transcripts DHFR-gSE
and DHFR-cSE did not accumulate to higher levels than the control DHFR mRNA in
maize or tobacco protoplasts. In fact, the sequences appeared to decrease mRNA
accumulation of the reporter mRN As. These results suggest that the sequences tested are

not recognized as stability determinants in maize and tobacco cell cultures.

107

A. B.

 

DHFR (plain) DHFR @1804)

161:

burn-gs: @1805) DHFR-38E (p1805)

DHFR-cSE (p1814) 12 h

DHFR-cSE (plan)

No DNA control 12 h

 

No DNA control

 

 

00.20.40.603112 00120:40:601811.2

Relative mRN A abundance Relative mRN A abundance

Figure 4. 2. The alt-globin stabilization element and the synthetic poly(CC C (0 sequence
do not increase the abundance of a reporter transcript in maize or in tobacco cells. (A)
Maize protoplasts were transiently transformed with plasmids p1804 (DHFR control), p1805
(DHFR-gSE) and p1814 (DHFR-cSE) as described in the text. Cells were harvested 12 and 16h
after transformation and 20 pg of total RNA was used for northern blot analysis using the DHFR
coding region as a probe. The relative abundance of the reporter transcripts is expressed as a
fraction of the control construct (p1804), which lacks the putative stabilization sequence. (B)
Similar experiments were performed in tobacco protoplasts.

oc-Globin stabilization element and a related sequence do not increase
abundance of a reporter mRNA in transgenic Arabidopsis seedlings

As an alternative to the cell culture expression system, the same constructs were
stably introduced into Arabidopsis thaliana via Agrobacterium-mediated transformation.
In order to have a representative population of transgenic plants, approximately 100
seedlings were randomly selected from the first generation of transformed plants and

pooled for total RNA isolation and northern blot analysis as before. As shown in Figure

4.3, steady state levels of the DHFR-gSE and DHFR-cSE did not accumulate to higher

108

levels than the control DHFR transcript. Similar to what observed in protoplasts, DHFR-
gSE and DHFR-cSE reporters appeared to accumulate to lower levels than the control
reporter transcript. These results are consistent with that observed in maize and tobacco
cells and suggest that in the conditions tested the or-globin stability determinant and the

synthetic poly(CCCU) sequence do not confer higher stability to the reporter transcript.

 

 

 

D
‘er‘-.Z" ‘.‘ 5:4-
ourn (p1804) .
. I .1 .
-

 

 

DHFR-gSE (p1805) .' ‘_

 

No DNA control [

 

 

 

o 0.2 014 0:0 013 1 1.2
Relative mRNA abundance

Figure 4. 3. The at-globin stabilization element and the synthetic poly(CCClD sequence
do not increase the abundance of a reporter transcript in transgenic Arabidopsis plants.
Arabidopsis plants were transformed with the constructs pl 804 (control), pl 805 (DHFR-gSE)
and p1814 (DHFR-cSE). Kanamycin resistant seedlings from independent transformation events
were pooled for total RNA isolation. 20 pg of total RNA was used for northern blot analysis
using the DHFR coding region as a probe. The relative abundance of the reporter transcript is
expressed as a fraction of the control construct (pl 804), which lacks a putative stabilization
sequence.

109

Discussion

Plant transcripts exhibit a wide spectrum of cytoplasmic half-lives, from as short as a
few minutes to as long as several days. While a significant amount of information is
available regarding mechanisms underlying rapid mRN A turnover (Ross, 1995;
McCarthy, 1998; Gutierrez et al., 1999), the mechanisms responsible for the long half-
life of transcripts in plants and other systems are largely unknown. An important first step
to understand the mechanisms of regulation of mRNA stability is the identification of cis-
acting sequences.

In an effort to gain insight into mRNA stabilization in plants, we tested the best
understood stabilization sequence element, the pyrimidine-rich element found in the
human a-globin transcript, in three plant expression systems. A similar strategy was used
to demonstrate that AUUUA repeats, a common instability determinant in mammalian
transcripts, can target mRNA for rapid degradation in plants (Ohme-Takagi et al., 1993).
Because sequence motifs can function better when multimerized (Kuhlemeier et al.,
1987; Lam and Chua, 1990; Newman et al., 1993), two copies of the or-globin
stabilization element were placed between a reporter DHF R coding region and the E9
polyadenylation signal. Sequences similar to the or-globin stability element have been
found in other stable mammalian transcripts (Holcik and Liebhaber, 1997). Therefore a
synthetic poly(CCCU) element derived item the consensus in the sequence comparison
(Holcik and Liebhaber, 1997) was also tested.

The results obtained in the three systems consistently indicated that the (Jr-globin and

the synthetic sequence do not increase the abundance of reporter transcripts in plants. In

110

fact, the sequences tested appear to diminish reporter transcript accumulation. These
results strongly argue that at least in the conditions tested, these sequence elements are
not recognized as mRN A stability determinants. It is not clear what caused reduced
mRNA accumulation. It is possible that these sequences are recognized as determinants
for rapid mRN A degradation, but influence over other post-transcriptional processes (e. g.
polyadenylation) can not be ruled out with the available data. It is also possible that under
different conditions the elements could manifest their mRN A stabilization function.
However, because no obvious plant condition appears to mirror the red blood cell
environment, testing would have to be conducted by trial and error. Such approach might
be challenging and has no guaranteed success because these sequences might not have a
functional significance in plants and stability elements might be different. Because of
these considerations, this strategy was not pursued any further.

As an alternative for the future, a directed evolution approach as that described by
Chrzanowska-Lightowlers and Lightowlers (2001) to find sequences that increase mRNA
stability in mammalian cells could be adapted for plants. Briefly, in such approach a
randomized sequence cassette is placed in the 3’UTR of a reporter transcript and a library
is made with the heterogeneous population of transcripts. Several rounds of expression
and selection for high levels of the reporter are then performed, for example, by
extracting RNA several hours after transient transformation experiments as described in
this Chapter. The putative stability sequences are then rescued using PCR and fused to
the same original reporter to prepare new “enriched” libraries. This process is iterated
until individual sequences that confer long life to the reporter mRNA are identified.

Though this strategy proved successful in mammalian cells (Chrzanowska-Lightowlers

111

and Lightowlers, 2001), one drawback is that the biological significance of the sequences
identified is not readily determined. This might not be of concern for applied purposes.
However the most attractive strategy to find stabilization sequences that are relevant to
plant mRNAs might be to combine deletion analysis of long-lived transcripts with half-
1ife measurements using the promoter system described in the next Chapter. Candidate
stable transcripts for analysis can be obtained from the global study of mRN A stability
described in Chapter 2. Table 4.1 contains examples of very stable transcripts detected in
the microarray experiments and that have been confirmed by northern blot analysis in at
least one extended cordycepin time course. In addition, sequences located downstream of
the coding region of plant genes have been shown to increase the abundance of reporter
transcripts via an unknown mechanism (summarized in Table 4.2). Fusion of these
sequences to unstable reporter transcripts and mRN A half-life measurements in
transcriptional pulse type experiments could also lead to the identification of signals that
protect transcripts from degradation, the first step towards understanding mechanisms of

long mRNA life in plants.

112

Table 4.1.

Examples of Arabidopsis genes with stable transcripts. The genes indicated
showed no detectable difference when comparing the 0 and 120 min time
point of cordycepin time courses in a microarray experiment (described in
Chapter 2) and were also stable by northern blot in at least one extended
cordycepin time course (data not shown).

 

Locus Description

 

Atl g08360 Putative ribosomal protein L10

 

Putative 2,3-bisphosphoglycerate-independent

“1309780 phosphoglycerate mutase

 

At3 g13920 Eukaryotic protein synthesis initiation factor 4A

 

At3g47340 Asparagine synthetase l

 

 

At5g60390 Translation elongation factor eEF-l alpha chain

 

 

 

113

Table 4.2. Sequence elements located downstream of the translation stop codon can
increase mRN A abundance of endogenous as well as reporter genes in
different plant systems.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Organism Gene Expression pattern Studies Effect on Reference
expression
Flaveria Mel Leaves GUS Increase in (Ali and
bidentis chimeras, leaves Taylor,
deletion 2001)
analysis
Sesbania SrEnodZ Nodule parenchyma GUS Increase in (Chen et
rostrata chimeras, nodule al., 1998)
deletion parenchyrna.
analysis
Potato SUS4 Developing tubers; basal GUS increase in (Fu et al.,
tissues of axillary buds and chimeras, tuber, root, 1995)
shoots; root cap and deletion leaves and
meristem.. analysis stems.
Arabidopsis Glabrous] Developing trichomes, leaf GUS Increase in leaf (Iarkin et
thaliana primordia and stipules. chimeras, primordia, al., 1993)
deletion trichomes and
analysis stipules.
Bmssica AX92 Root cortex of embryos and GUS Increase in root (Dietrich
napus seedlings. chimeras cortex of et al.,
embryos and 1992)
seedling
Petunia SSU301 Leaves Chimeras Increase in (Dean et
in tobacco leaves al., 1989)
plants
Potato Pl-II Leaves, wound inducible CAT Increase in (An et al.,
chimeras, leaves 1989)
deletion
analysis in
tobacco
plants.

 

114

 

References

Ali,S. and Taylor,W.C. (2001). The 3 ' non-coding region of a C-4 photosynthesis gene
increases transgene expression when combined with heterologous promoters. Plant Mol.
Biol. 46, 325-333.

An,G.H. (1985). High-efficiency transformation of cultured tobacco cells. Plant Physiol.
79, 568-570.

An,G.H., Mitra,A., Choi,H.K., Costa,M.A., An,K.S., Thomburg,R.W., and Ryan,C.A.
(1989). Functional-analysis of the 3' control region of the potato wound-inducible
proteinase inhibitor-II gene. Plant Cell 1, 115-122.

Caponigro,G. and Parker,R. (1996). Mechanisms and control of mRNA turnover in
Saccharomyces cerevisiae. Microbiol. Rev. 60, 233-249.

Chen,C.Y.A. and Shyu,A.B. (1995). AU-rich elements - Characterization and importance
in messenger-RNA degradation. Trends Biochem. Sci. 20, 465-470.

Chen,R., Silver,D.L., and de Bruijn,F.J. (1998). Nodule parenchyma-specific expression
of the Sesbania rostrata early nodulin gene SrEnod2 is mediated by its 3 ' untranslated
region. Plant Cell 10, 1585-1602.

Chrzanowska-Lightowlersz. and Lightowlers,R.N. (2001). Fending off decay: a
combinatorial approach in intact cells for identifying mRNA stability elements. RNA 7,
435-444.

De Rocher,E.J., Vargo-Gogola,T.C., Diehn,S.H., and Green,P.J. (1998). Direct evidence
for rapid degradation of Bacillus thuringiensis toxin mRNA as a cause of poor expression
in plants. Plant Physiol. 1] 7, 1445-1461.

Dean,C., Favreau,M., Bondnutter,D., Bedbrook,J., and Dunsmuir,P. (1989). Sequences
downstream of translation start regulate quantitative expression of 2 petunia rch genes.
Plant Cell I, 201-208.

DeRocher,E.J., Vargo-Gogola,E.C., Diehn,S.H., and Green,P.J. (1998). Direct evidence
for rapid degradation of Bacillus thuringiensis toxin mRN A as a cause of poor expression
in plants. Plant Physiol. 1 I 7, 1445-1461.

Diehn, S. H. (1998). B. t. toxin gene expression and differential utilization of
polyadenylation signals in grarnineous and dicotyledonous plants. PhD Dissertation.
Michigan State University.

115

Diehn,S.H., Chiu,W.L., De Rocher,E.J., and Green,P.J. (1998). Premature
polyadenylation at multiple sites within a Bacillus thuringiensis toxin gene-coding region.
Plant Physiol. 1] 7, 1433-1443.

Diehn,S.H., De Rocher,E.J., and Green,P.J. (1996). Problems that can limit the
expression of foreign genes in plants: lessons to be learned from B. t. toxin genes. In
Genetic Engineering: Principles and Methods, J .K.Setlow, ed. (New York: Plenum
Press), pp. 83-99.

Dietrich,R.A., Radke,S.E., and Harada,J.J. (1992). Downstream DNA-sequences are
required to activate a gene expressed in the root cortex of embryos and seedlings. Plant
Cell 4, 1371-1382.

Fu,H.Y., Kim,S.Y., and Park,W.D. (1995). High-level tuber expression and sucrose
inducibility of a potato SUS4 sucrose synthase gene requires 5'-flanking and 3'-flanking
sequences and the leader intron. Plant Cell 7, 1387-1394.

Gutiérrez,R.A., MacIntosh,G.C., and Green,P.J. (1999). Current perspectives on mRNA
stability in plants: multiple levels and mechanisms of control. Trends Plant Sci. 4, 429-
438.

Hargrove,J.L., Hulsey,M.G., and Beale,E.G. (1991). The kinetics of mammalian gene-
expression. Bioessays 13, 667-674.

Holcik,M. and Liebhaber,S.A. (1997). Four highly stable eukaryotic mRNAs assemble 3'
untranslated region RNA-protein complexes sharing cis and trans components. Proc. Natl.
Acad. Sci. USA 94, 2410-2414.

Jefferson,R.A., Kavanagh,T.A., and Bevan,M.W. (1987). GUS fusions-beta-
glucoronidase as a sensitive and versatile gene fusion marker in higher-plants. EMBO J.
6, 3901-3907.

Kuhlemeier,C., F luhr,R., Green,P.J., and Chua,N.-H. (1987). Sequences in the pea rbsS-
3A gene have homology to constitutive mammalian enhancers but function as negative
regulatory elements. Genes Dev. 1, 247-255.

Lam,E. and Chua,N.H. (1990). GT-l binding-site confers light responsive expression in
transgenic tobacco. Science 248, 471-474.

Larkin,J.C., Oppenheimer,D.G., Pollock,S., and Marks,M.D. (1993). Arabidopsis
GLABROUSl gene requires downstream sequence for function. Plant Cell 5, 1739-1748.

McCarthy,J.E.G. (1998). Posttranscriptional control of gene expression in yeast.
Microbiol. Rev. 62, 1492-1553.

McClure,B.A., Hagen,G., Brown,C.S., Gee,M.A., and Guilfoyle,T.J. (1989).
Transcription, organization, and sequence of an auxin-regulated gene cluster in soybean.
Plant Cell I, 229-239.

116

Mehdy,M.C. and Brodl,M.R. (1998). The role of stress in regulating mRNA stability. In
A look beyond transcription: Mechanisms determining mRNA stability and translation in
plants, J .Bailey-Serres and D.R.Gallie, eds. American Society of Plant Physiologists), pp.
54-63.

Nagata,T., Nemoto,Y., and Hasezawa,S. (1992). Tobacco BY-2 cell-line as the Hela-cell
in the cell biology of higher-plants. Int. Rev. Cytol. 132, 1-30.

Newman,T.C., Ohme-Takagi,M., Taylor,C.B., and Green,P.J. (1993). DST sequences,
highly conserved among plant SA UR genes, target reporter transcripts for rapid decay in
tobacco. Plant Cell 5, 701-714.

Ohme-Takagi,M., Taylor,C.B., Newman,T.C., and Green,P.J. (1993). The effect of
sequences with high AU content on mRNA stability in tobacco. Proc. Natl. Acad. Sci.
USA 90,11811-11815.

Ross,J. (1995). Messenger-RNA stability in mammalian-cells. Microbiol. Rev. 59, 423-
450.

Russe11,J .E., Morales,J., and Liebhaber,S.A. (1997). The role of mRNA stability in the
control of globin gene expression. Prog. Nucleic Acid Res. Mol. Biol. Vol 57 5 7, 249-
287.

Sachs,A.B. (1993). Messenger-RNA degradation in eukaryotes. Cell 74, 413-421.

Siflow,C.D. and Key,J.L. (1979). Stability of polysome-associated polyadenylated RNA
from soybean suspension culture cells. Biochemistry 18, 1013-1018.

van Hoof,A. and Green,P.J. (1996). Premature nonsense codons decrease the stability of
phytohemagglutinin mRNA in a position-dependent manner. Plant J. 10, 415-424.

Wang,X.M., Kiledjian,M., Weiss,I.M., and Liebhaber,S.A. (1995). Detection and
characterization of a 3' untranslated region ribonucleoprotein complex-associated with
human alpha-globin messenger-RNA stability. Mol. Cell. Biol. 15, 2331.

Weiss,I.M. and Liebhaber,S.A. (1994). Erythroid cell-specific determinants of alpha-
globin messenger-RNA stability. Mol. Cell. Biol. 14, 8123-8132.

Weiss,I.M. and Liebhaber,S.A. (1995). Erythroid cell-specific messenger-RNA stability
elements in the alpha-2-globin 3' nontranslated region. Mol. Cell. Biol. 15, 2457-2465.

117

CHAPTER 5

Promoter region of the At-EXPLI gene drives transient expression

of reporter transcripts: A new regulated promoter system to study

mRNA degradation in Arabidopsisw.

 

W Part of the initial cloning of the At-EM’L] promoter and part of the analysis of reproducibility of At-
EH’L] gene expression (Figure 5.2) were carried out by the undergraduate student Mr. Chris Behrens as
pm of his professorial assistantship.

118

Introduction

The process of mRNA degradation is an important regulatory point for gene
expression. In multicellular eukaryotes, turnover rates can vary from minutes, for the
most rapidly degraded mRN As, to several days for those transcripts that are most stable
(Gutierrez et al., 1999). In addition, modulation of these rates can play a significant role
during the transition to new mRNA steady state levels triggered by developmental,
environmental, or physiological cues. The faster the mRNA is degraded the faster it
reaches a new steady state level (Hargrove et al., 1991).

Determining the turnover rate of a transcript and how that rate changes in response
to internal or external signals is key in evaluating the contribution of this post-
transcriptional control mechanism to the expression of any gene. Moreover, accurate
knowledge of the kinetics of mRN A degradation is essential in order to understand the
molecular mechanism that underlies this level of control of gene expression.

Methods that monitor the rate of loss of mRN A after transcription is halted are the
most popular to investigate the process of mRNA degradation (Parker et al., 1991).
Primarily because they provide information on mRNA integrity, they can be used to
study high- and low-abundance mRNAs, require little amounts of radioactivity, and are
technically simple (Parker et al., 1991). Typically, a chemical that inhibits global
transcription is utilized and time courses are generated by extracting tissue samples at
regular intervals after the tissue has been exposed to the inhibitor. Different chemicals
have been used in different model organisms to inhibit the RNA polymerase. For
example, the DNA interchalating agent actinomycin D has been widely applied in studies

in mammalian systems (Brawerrnan, 1993) while the chain terminator cordycepin has

119

been common in plants (Seeley et al., 1992; Johnson et al., 2000). The main disadvantage
of using drugs to stop global transcription is that prolonged incubation with inhibitors can
have a severe impact on the cellular physiology (Brawerman, 1993). As a consequence,
usually only short incubation periods are possible, which limits the application of this
approach to the study of unstable mRNAs. However even with short incubation times,
normal decay of cellular mRNAs can be altered in the presence of the inhibitor (e.g. Shyu
et al., 1989; Gil and Green, 1996; Petracek et al., 1998). In S. cerevisiae another option to
stop global transcription is to thermally inactivate the temperature sensitive RNA
polymerase II (tpbl-I) (Nonet et al., 1987). Although this approach has yielded
comparable results with other methods for selected mRNAs (Herrick et al., 1990) and has
been used in global studies of mRNA turnover (Wang et al., 2002), it is not yet clear to
what extent normal degradation is affected by the elevated temperature.

As an alternative to global transcriptional repression, a regulated promoter system
can be used to stop mRN A synthesis. Although turnover rate of only individual
transcripts can be measured with this approach, it has the main advantage that cell
physiology is not severely compromised. Examples of regulated promoters systems
include the GAL] promoter in yeast (Parker et al., 1991) and the c-fos promoter in
mammalian systems (Shyu et al., 1989)

In plants, promoter systems that are positively regulated by chemical ligands have
been developed (e.g. Aoyama and Chua, 1997; Salter et al., 1998; Bohner et al., 1999;
Martinez et al., 1999; Granger and Cyr, 2001). Although these are good inducible
systems, they are not well suited to determine mRN A stability, primarily because

transcriptional repression is not fast or easily achieved.

120

Efforts to develop negatively regulated promoters have yielded far fewer options in
plant systems. In fact the only alternatives are variants of the tetracycline(Tc)-repressible
system initially developed in tobacco plants (W einmann et al., 1994). This system has
been usefirl for mRNA degradation studies in tobacco cell cultures (Gil and Green, 1996)
and tobacco plants (Petracek et al., 1998). It has also been adapted to Arabidopsis (Love
et al., 2000), but it is not yet clear whether this “ToplO promoter system” would be
usefirl to study mRN A degradation in this plant. Transcriptional down-regulation is
achieved in the presence of To in Arabidopsis, but repression kinetic studies were carried
out at the protein level and detectable changes were observed only after several hours
(Love et al., 2000). Nevertheless, the ToplO promoter system is not a strong promoter
(approximately lO-fold lower than the CaMV 35S promoter. Gil & Green, unpublished
data) which makes difficult the study of labile mRNAs. Furthermore, due to the
heterogeneity in the mRN A population that results from Top10 constitutive transcription,
this approach will not readily allow the study of critical mechanistic aspects of mRNA
degradation such as deadenylation and decay kinetics and precursor/product relationship
for mRNAs with stable decay intermediates.

For successful mRN A degradation studies an ideal regulated promoter system should
exhibit at least the following properties: (i) Extremely low or no basal levels of
expression in the absence of inducing conditions (ii) A high level of expression in the
induced stage (iii) Rapid transcriptional induction and repression to generate a
synchronized population of mRNAs (iv) Highly reproducible expression patterns (v) No
external chemical signal necessary to activate and/or repress transcription. Recently,

Shimizu-Sato et al. (2002) described a light controlled promoter system that appears to

121

comply with these rules. Although promising, its potential for mRN A turnover
measurements remains to be investigated.

Here we describe a new regulated promoter system that meets all the ideal
requirements for mRNA stability studies. We have successfully used this system to
recapitulate the instability of a well studied unstable reporter mRN A. The system makes
use of the natural transcriptional regulatory properties of the Arabidopsis thaliana WLI
gene (At-EXPLI). At-EXPLI is member of a small group of genes known as expansin-
like which belong to the expansin superfamiliy (http://www.bio.psu.edu/expansins/).

In the conditions described, the promoter region of the Arabidopsis At-EXPL] gene
behaves as a strong promoter system that is transiently activated to produce a

synchronized population of transcripts.

122

Materials and Methods

Arabidopsis strain and growth comlitiw

Arabidopsis thaliana ecotype Columbia was used for all experiments. Plants were
germinated on agar plates containing 1x Murashige and Skoog salts, 1x Gamborg's
vitamins and 1% sucrose at 22 °C and 16h light and 8h dark cycles. After two weeks they
were either used for the half-life experiments described below or transferred to soil to

complete their life cycle.

Plasmid constructs

The polymerase chain reaction (PCR) was used to amplify a 2650 bp region of
Chromosome HI upstream of the expansin-like gene At-EXPL] (locus number
At3g45970). The DNA fragment was amplified with primers pg956
(TCGTCACAATGGAGT'I‘ACATGAGTAGAAG) and pg957
(CT'I'TAAGATCTAATAAGAGAGAGAGATATGTG) and cloned into the pGEM-T
Easy Vector (Promega. Madison, WI) according to instructions provided by the
manufacturer. Sequence of the PCR fragment was carried out to discard any mistake
introduced during the amplification. The sequence verified upstream region of At-EXPL]
(pEXPLl) was excised from the pGEM-T Easy Vector as a fiagment Sac I / pEXPLl I
Bgl I]. This DNA fragment replaced the CaMV 35S promoter present in the plant
expression cassette in plasmids p1088 and p1060 at corresponding restrictions sites.

p1088 originally contained an expression cassette consisting of Rm 1] / Sac I / CaMV

123

35$ promoter / Bgl II / human B-globin coding sequence / 2 DST instability sequences /
3’UTR from the pea Rubisco small subunit rch-E9 / Pvu II (Newman et al., 1993).
Plasmid p1060 was identical to p1088 except that there were two copies of a control
spacer instead of the DST instability sequence (Newman et al., 1993). The new derivative
of p1088 was termed p2102 and the corresponding derivative of p1060 was termed p2103.
The complete expression cassette from plasmids p2102 and p2103 was then removed
with Pvu II and inserted into the Sma I site of the vector pCAMBIA3301 (CAMBIA.
Canberra, Australia) to generate plasmids p2112 and p2113 respectively. Orientation of
plant expression cassettes in the CAMBIA vectors was verified to be the same. p21 12
and p2113 were then introduced into A. thaliana via Agrobacterium mediated
transformation as described below. To ensure integrity of the plasmid constructions,
restriction mapping with at least 3 restriction enzymes and sequence verification of the

junctions was carried out at each cloning step.

Arabidopsis transfornptipp

Transformation of Arabidopsis plants was carried out with the Agrobacterium
tumefaciens infiltration method as detailed previously (Bariola et al., 1999). Briefly,
individual plasmids were electroporated into A. tumefaciens strain GV3101 pMP90
(Koncz and Schell 1986) using a Gene-Pulsar (BioRad). The aerial portion of soil grown
4 to 5-week old Arabidopsis plants was then submerged in a solution of the transformed
Agrobacterium. Vacuum of 18 in Hg was applied for 5 min and the plants were then
allowed to set seed under normal growth conditions. The transformed seeds were selected

by germination on agar plates containing 1x Murashige and Skoog salts, 1x Gamborg's

124

vitamins, 1% sucrose and 50 ug/mL of gluphosinate ammonium (C14030000, Crescent
Chemical Co.). The herbicide resistant plants were transferred after two weeks to soil or

used for further studies.

Half-life measurements witl_r a trapscriptional inhibitor

Half-lives were determined as described by Seeley et al. (Seeley et al., 1992) with
the following modifications. Two-week old Arabidopsis plants grown on plates were
transferred to a flask with incubation buffer (Seeley et al., 1992). After a 30 min
incubation, 3'-deoxyadenosine (cordycepin) was added to a final concentration of 0.6 mM
(time 0). Tissue samples were harvested at regular intervals thereafter and quickly fi'ozen
in liquid nitrogen. Total RNA was isolated and analyzed by northern blot using standard

techniques.

 

flplf-life measurements using the At-EXPL] promoter:

To determine half-lives using the At-EXPL] promoter system we followed the same
protocol described above for half-life measurement with a transcription inhibitor except
that the drug was never added. Plants were transferred to the incubation buffer and tissue
samples were harvested at regular time intervals thereafter and quickly frozen in liquid
nitrogen. Total RNA was isolated and analyzed by northern blot using standard

techniques.

125

Results

At-EXPLI gene is transiently induced

We investigated the expression of the expansin-like gene At-EIG’L] (At3g45970),
which encodes one of the most unstable Arabidopsis mRN As recently identified through
microarray analysis (Gutierrez et al., 2002). At-EXPLI exhibited an unusual pattern of
gene expression. Initially undetectable in two-week old Arabidopsis plants grown on agar
plates, it was rapidly and transiently up-regulated following submersion of the plants in
incubation buffer during our routine mRNA half-life experiments (Figure 5.1a).
Interestingly, this pattern of gene expression occurred irrespective of the presence or
absence of cordycepin (Compare Figure 5.1a to 5.1b) and was unique among 14 genes
with unstable and moderately unstable transcripts (data not shown). Similar transient

induction of the c-fos promoter following the addition of serum to serum-starved

Incubation (min) 0 30 180
Cordycepin - - ..

At-EM’LI .

A. B.
Incubation (min) 0 30 180
Cordycepin - + +
At-EXPL] .
efla hit . 5"

his
efla g Q Q
l . .7 3.57:1"

Figure 5.]. Transient induction of the At-EIG’LI gene. (A) Northern blot analysis of At-
WLI mRNA at the indicated time points after transferring two-week old plants to an incubation
buffer. Cordycepin was added after 30 min of incubation. Samples consisted of 10 pg of total
RNA isolated from the indicated time points. (B) As (A) but in the absence of cordycepin. The
stable eIF4A mRNA was used as a loading control.

126

NIH-3T3 cells has been successfully exploited to determine mRN A degradation rates in
mammalian cells (Shyu et al., 1989). In fact, such transient-induction system it is now a
mainstay of mammalian mRN A decay research. As noted above, such a system is
unavailable for plants. If the observed At-EIGDL] transient induction is transcriptionally
controlled, it could be used as a regulated promoter system to study the mRN A

degradation process in Arabidopsis thaliana.

Transient induction of At-QXPL] is highly reprodpciblg

As a prerequisite for developing At-EXTLI as a regulated promoter system, we
evaluated the reproducibility of its expression pattern in biological replicate experiments.
Two-week old Arabidopsis plants were used to perform extended time course
experiments in the absence of cordycepin as described in the Materials and Methods. A
representative time course and summary data are shown in Figure 5.2. In the three
independent experiments performed, At-EXPL] mRN A levels were undetectable at the
beginning of the time course, accumulated very rapidly to easily detectable levels and
reached maximum at 45 min. At-EXPL] transcript levels dropped very rapidly afterwards
to nearly undetectable levels by 150 min (Figure 5.2). The transient induction pattern of
At-EH’L] was highly reproducible as evidenced by the small standard error bars shown

in Figure 5.2.

127

. 0 rs 30 .45 so so rzorso 2_ro_ (min)
Ar-EJG’L] is “ O I w; - ‘

eIF4AwwWI-fiewwwww

 

 

 

 

 

 

 

 

 

B.
n
37] \\i\
\N

Time (min)

Figure 5.2. At-EXPL] gene expression is highly reproducible. (A) Representative northern
blot analysis of At-EMDL] mRNA at the indicated time points after transferring two-week old
Arabidopsis plants to an incubation buffer (see text for details). Samples consisted of 10 pg of
total RNA isolated from the indicated time points. (B) Average and standard error of mRNA
levels for three independent experiments are shown. All values were normalized to the
corresponding eIF4A mRNA levels. The signal for eIF 4A does not change significantly during
the time courses and is used as a reference for equal loading.

At-EXPLI gene transcription is severely dowp-r ated 45 min after its
induction

To evaluate the extent of transcriptional down-regulation of At-EXPL] gene during
the declining phase we monitored At-EXPL] mRN A levels in the presence of the general
transcription inhibitor cordycepin and compared them to those in the absence of the drug.

Two-week old Arabidopsis plants were used to perform time course experiments in the

presence and in the absence of cordycepin as described in the Materials and Methods. As

128

A.

Incubation (min) 30 45 so 90 120180 45 so 90 120 150210
Cordycepin + + + + + +

(J

At-EJG’LI . i as . D . . as ;.
“one ~1- '15-?- - - a u

'03" 1' .12

 

B' O + Cordycepin ’1/2 = 32 :t: 1.8
O - Cordycepin tm= 35 :1: 1.7

 

 

2

 

Relative mRNA remaining

 

 

 

0.01

Time (ruin)

Figure 5. 3. At-EIG’L] mRNA disappears with similar speed in the presence or absence of
cordycepin. (A) Representative northern blot analysis of time course experiments in the presence
and in the absence of cordycepin for At-EJGDL] and efl amRNA. Samples consisted of 10 pg of
total RNA isolated from the indicated time points. (B) Quantitation of the decrease in mRNA
abundance and half-life estimation. The stable ef] atranscript was used as a reference for equal
loading. Values are representative of three independent cordycepin time courses.

shown in Figure 5.3 the rate of disappearance of At-EMDL] mRNA was indistinguishable
in the two conditions. The only detectable difference was that decline in the presence of
cordycepin began upon addition of the drug, 15 min earlier than the natural decline of At-
EJG’L] mRNA in the absence of the inhibitor (Figure 5.3b). We conclude from these

results that At-EXPL] gene expression is down-regulated after 45 min, reaching levels

129

comparable to those in the presence of the general transcription inhibitor cordycepin.
This data suggest mRN A transcription from the At-EXPLI promoter ceases naturally
approximately 45 min afier the plants have been submerged in the incubation buffer. This
data also indicates that the RNA chain terminator cordycepin does not interfere with the
mRNA decay machinery that degrades the A t-EXPLI mRN A. These results encourage
the use of this chemical as a global transcriptional inhibitor for mRN A decay studies in

plants.

Progner region of At-EXPLI gene drives transient expression of
morter trapscripg

To demonstrate that the DNA regulatory elements that control the transient induction
of the At-EXPL] gene lay within its promoter region, the upstream sequence of At-EXPLI
was firsed to reporter transcripts of known stability. The polymerase chain reaction (PCR)
was used to amplify a 2650 bp region of Chromosome III upstream of the predicted At-
E20’L1 coding sequence. The PCR amplification product was directly cloned into a
commercial vector and sequence verified. A Bgl I] site was engineered via the
downstream primer pg957 to facilitate cloning into the plant expression cassettes. The
promoter region was then excised as a 2584 fragment that included 37 bp of the
5’untranslated region of At-EXPL] using a natural Sac I restriction site and the new Bgl I]
site. This fragment was fused to a chimeric reporter previously built by Newman et al.
(1993) consisting of the human B-globin coding sequence, two copies of the DST

instability sequence and the E9 polyadenylation signal (globin-DSTxZ; Figure 5.4a).

130

Globin-DSTx2 (p2113): DST x2
r—fi

| At3g45970 5’ rgglon ] ,B-hGlobin [1 ] rchE9 3’ ]

Globln-spacerx2 (p2112): Spacer x2
r—m

| At3g45970 5’ rggion f fi-hGlobin 1 I l rch E9 3' j

2113-1 2113-2 2113-3

 

 

0 45 150 0 45 150 0 45 150 (mln)

, .--¢—--,
. o

globin-DSsz m" ' as fl ..

eIF4A H g m fioqu w aha-L “ 2‘.“ ,
fist-n. a ; . . ._ ‘ , .

2112-1 2112-2 2112-3

 

o 45 150 0 45150 0 45150 (mln)

globin-spacerxz I . . . W ~~ iii

eIF4A aflHHHUHUfl

Figure 5. 4. At-EXPL] promoter region drives transient expression of two globin reporter
mRNAs in transgenic Arabidopsis plants. (A) A 2584 bp DNA fiagment containing the
promoter region of the At-EH’LI gene (At3g45970) was fused to a reporter consisting of the
human B-globin coding region, two copies of the DST instability sequence and the E9
polyadenylation signal (globin-DSTXZ. p2113). A second construction was made in which a
control spacer sequence replaced the DST element (globin-spacerx2. p2112) as illustrated. (B)
Northern blot analysis of time course experiments for three independent transgenic lines
expressing the globin-DSTx2 or the globin-spacerx2 reporter transcript. eIF 4A was used as a
reference for equal loading. Samples consisted of 10 pg of total RNA isolated from the indicated
time points in the absence of cordycepin.

131

A second construction was made by fusing the At-EXPL] promoter fiagment to a control
chimeric reporter in which a spacer sequence between the B-globin coding sequence and
the E9 polyadenylation signal replaced the DST element (globm-spaceer; Figure 5.4a)
(Newman et al., 1993). Transgenic Arabidopsis plants carrying these two constructs were
made via Agrobacterium mediated transformation. Seeds from independent Tl plants
were germinated on agar plates in the presence of gluphosinate ammonium. Two-week
old herbicide-resistant plants were then transferred to leaf incubation buffer and time
courses in the absence of cordycepin were performed. As shown in Figure 5.4b the three
lines tested for each construct exhibited reporter mRN A profiles comparable to those of
the endogenous At-EJG’L] gene. Reporter transcripts in all lines tested were near
background levels at the beginning of the experiment, accumulated to easily detectable
levels by 45 nrin and dropped afterwards to levels proportional to their stability (Figure
5.4b). These results demonstrate that the promoter region of At-EXPLI is sufficient to

drive transient expression of sequences fused downstream.

At-EXPLI promoter system allows measurement of mRNA stability.

In order to test the utility of the At-EXPLI promoter system to study mRNA degradation
in plants, lines carrying the At-EXPL] promoter globin-DSTxZ and globin-spaceer
fusions were used for half-life experiments in the absence of transcription inhibitor.
Seeds from three independent Tl plants per construct were used for these experiments.
As expected, reporter mRNA levels were near background levels at the beginning of the

experiment, accumulated rapidly to detectable levels and peaked at 30 min for globin-

132

DST2 and 45 min for the globin-spacerx2 mRN A (Figure 5.5a,c). Reporter transcript
levels dropped afterwards with a rate proportional to their mRN A
stability. Figure 5.5a-b shows a representative time course experiment. Time course
experiments extended for up to 5 hrs failed to detect resumption of At-EJG’L] promoter
activity (data not shown). Summary data for the three lines studied per construct are
presented in Table 5.1. The stability of the globin-DSTx2 transcript (35 :1: 1.5 min) agreed
well with that previously reported in tobacco cells for the same mRNA (33 :h 2.0 min)
(Newman et al., 1993). Surprisingly, we found a 1.6-fold difference in mRNA stability
between the two reporters in Arabidopsis plants (Table 5.1), in contrast to the 4.8-fold
difference previously reported in studies in tobacco cell cultures (Newman et al., 1993).
Assuming zero order mRNA synthesis and first order mRN A degradation kinetics,
theoretical calculations predict that given the same transcriptional activity, unstable
mRNAs will reach a new steady state level faster than stable mRNAs (Hargrove et al.,
1991). Our experimental data support this mathematical model in plants. On average the
more unstable globin-DSTx2 reporter mRN A reached maximum accumulation 15 rrrin
earlier than the globin-spaceer mRNA did under the same conditions (Figure 5.5c). In
contrast endogenous At-EXPL] mRNA levels peaked at 45 min regardless of the genetic
background (data not shown). Given the enormous range of mRNA stabilities observed in
cellular mRN As in comparison with the less than 2-fold difference in the transcripts
studied, this data emphasizes the impact that mRN A decay rates can have on plant gene

expression.

133

Figure 5. 5. At-EM’LI promoter system can be used to study mRNA degradation in
Arabidopsis. (A) Representative northern blot analysis of time courses in the absence of
cordycepin. Transgenic Arabidopsis plants expressing the globin-DSTxZ or globin-spaceer
reporter mRN As under the control of the At-WL] promoter were used for these experiments.
EndogenousAt-EMDL] was used as internal reference and eIF 4A mRNA as a loading control in
each time course experiment. Samples consisted of 10 pg of total RNA isolated hour the
indicated time points. (B) Quantitation of mRNA abundance and half-life estimation in a
representative experiment. The signal for eIF 4A was used for normalization and the values were
made relative to the maximum mRNA accumulation in the time course. Half-lives were
calculated using the 45 min time point and thereafter as illustrated. (C) Average mRNA levels
and standard error for three transgenic lines for each construct.

134

 

A' o 15 so 45 so so no 150 m (min)
globin-DSTxZ “fig g... as
At-EXPL] 4. - “5.1““
eIF4A assess so assess”
0 rs 30 45 so so no 150 210 (min)
globin-spacerxz

At—EXPL]

 

eIF4A wmfli‘flfi UW

B‘ l globin-spaceer (rm - 50 i 2.4) Q globin-D8132 (rm - 35 :t: 15)
o At-EXPL] (rm - 32 s 2.3)

10 .11. 1 _._..._._,,,.1, "H, .1:

i

o At-EXPL] (1m '37 i 1.7)
10.... ..,...., .. ”in... .. a... M.,

 

.-
0
1
p

 

 

7'

 

 

Relative mRNA remaining
0

 

Relative mRNA remaining
9

 

O
.
e
u—
a

s:
e:

G
a
5
E
E
N
:9
°-‘

50 100 150 200 250
Time (min) Time (min)

—I— globin-spacerxz
+ globin-DSTxZ

 

0 so 100 rso zoo 250
Time (min)

135

 

Table 5.1. Summary of mRNA stability measurements with regulated promoter system.

 

Construct Mean Absolute Half-Life (min)

 

 

Test gene Ref. Gene Test mRNA Ref. mRN A Mean Test/Ref.“ Relative Half-Life“

 

Globla-spacerxz At-EXPLI so a 2.4 32 a: 2.3 1.60 e on 1.00 s 0.07

GlobiI-DSTX2 At—EM’L] 35 t 1.5 37 :1: 1.7 0.94 :t 0.01 0.59 :l: 0.“)6

 

a. Calculations were made as previously described (Newman et al., 1993)

136

Discussion

It is now acknowledged that regulation of the stability of cytoplasmic mRNAs is
an important step in the control of gene expression. To understand the molecular
mechanism underlying this level of modulation, a detailed analysis of the steps involved
in mRNA degradation is essential. Methods to study the decay of mRN As have been
devised. Most commonly, direct measurements of mRNA stability are performed by
stopping transcription with chemical inhibitors and monitoring mRN A disappearance
thereafter. However transcriptional pulsing approaches are perhaps best suited for
uncovering critical mechanistic aspects of the mRN A degradation process such as
deadenylation and decay kinetics and precursor/product relationship. Such approaches are
not available for plant systems.

Here we describe a new regulated promoter system that allows for transcriptional
pulse experiments in Arabidopsis. This system takes advantage of the natural regulatory
properties of the Arabidopsis thaliana EXPLI promoter region. We showed that At-
EH’LI gene expression was transiently activated after submerging two-week old
Arabidopsis plants in incubation buffer. At-EXPL] expression pattern was highly
reproducible in our experimental conditions and the main regulatory elements that control
its transient induction are contained within the promoter region of the gene. We took
advantage of these characteristics to study the stability of the globin-DSTx2 and globin-
spacerx2 reporter mRN As. The At—EXPL] regulated promoter system recapitulated the
instability of the globin-DSTx2 transcript (35 :l: 1.5) previously studied in tobacco cell
culture systems (33 :I: 2 min) (Newman et al., 1993). Unexpectedly, the globin-spacerx2

mRNA decayed faster in Arabidopsis (50 :1: 2.4) than anticipated based on the previous

137

 

measurements performed in tobacco (113 d: 6 min) with actinomycin D (Newman et al.,
1993). Although sequences resembling known instability determinants can not be
identified in the control spacer used in these experiments, we can not rule out the
possibility that the Arabidopsis degradation machinery recognizes some features of this
artificial transcript as instability determinants. Other sequences in the chimeric transcript,
e. g. within the globin coding sequence, might also contribute to its increased turnover in
Arabidopsis plants. In addition, it is known that mRNA stability studies with actinomycin
D can lead to underestimation of decay rates in plants and other systems (Belasco and
Brawerman, 1993; Gil and Green, 1996; Petracek et al., 1998). Furthermore, the
instability function of distinct sequences within the same transcript can be affected to
different degrees by actinomycin D treatment (Shyu et al., 1989). Therefore the function
of additional instability elements in the globin-spaceer reporter could have been masked
in the tobacco experiments by the use of actinomycin D, leading to an overestimation of
the transcript stability.

It is important to mention that the turnover rate of the At-EXPL] mRNA was
indistinguishable in time course experiments performed in the presence or in the absence
of cordycepin. This data indicates that this general transcription inhibitor does not
interfere with the mRNA decay machinery that normally degrades the At-EXPL] mRNA.
This is in contrast to what obtained with other chemicals (Gil and Green, 1996; Petracek
et al., 1998) and encourage the use of cordycepin for mRNA decay studies in plants.

A great advantage of the system described is that the endogenous At-EXPL]
mRNA can be used as a control for the transient induction of At-EJGDL] promoter activity

as well as an mRN A stability internal reference to compare across multiple experiments.

138

Previous studies have shown that the use of a reference transcript as an internal standard
is a very effective strategy to normalize for experimental variation that may occur due to
small differences in growth conditions, developmental stage, physiological status, etc
(Newman et al., 1993).

It is not yet known what stimulus or stimuli control the transient transcriptional
activation of At-Efl’LI. General knowledge about the expansin-like gene family is very
limited. However, semi-quantitative RT-PCR experiments indicated At-WLI is
expressed in most Arabidopsis organs (Dan Cosgrove, personal communication). In
addition, promoter B-glucuronidase fusion experiments suggest At-EXPLI is induced by
wounding (Dan Cosgrove, personal communication). Preliminary studies in our
laboratory are consistent with At-EH’L] being induced in response to mechanical
damage (data not shown). Based on our experimental design, we can propose mechanical
stimulation/damage and perhaps submersion are the most likely signals involved in
regulating At-EJG’LI promoter activity. Previous studies indicate that mechanical
stimulation (touch) can induce transient gene expression. In fact the touch gene
transcripts initially characterized were all detected within minutes of treatment and
disappeared very rapidly afterwards, similar to what observed for the At-EXPLI mRNA
(Braam and Davis, 1990). Transient gene expression in response to wounding has also
been reported (e.g. (Rojo et al., 1999)). Submergence could also be implicated in At-
EXPLI expression as it is known that submersion alone can effectively induce expression
of two or-expansin genes in Oryza sativa (Cho and Kende, 1997). Additional experiments
will be necessary to determine the signals(s) that regulate At-EJODLI promoter activity.

Nevertheless, under the conditions described for our assays it reproducibly behaves as a

139

strong promoter system that is transiently activated. This transient induction results in a
synchronized population of transcripts that can be monitored overtime. Hence, this tool
will be most useful to study mRN A decay kinetics and specific aspects of the mRNA
degradation process such as deadenylation. The strength of this promoter will greatly
facilitate the study of labile mRNAs. In addition, because no toxic chemicals are needed
to activate or represses the At-EXPL] promoter, time courses can be extended allowing

the study of stable plant transcripts.

140

References

Aoyarna,T. and Chua,N.H. (1997). A glucocorticoid-mediated transcriptional induction
system in transgenic plants. Plant J. I], 605-612.

Belasco,J.G. and Brawerrnan,G. (1993). Experimental Approaches to the Study of mRNA
Decay. In Control of Messenger RNA Stability, J .Belasco and G.Brawerman, eds.
Academic Press, Inc. San Diego, CA), pp. 475-493.

Bohner,S., Lenk,l., Rieping,M., Herold,M., and Gatz,C. (1999). Transcriptional activator
TGV mediates dexamethasone-inducible and tetracycline-inactivatable gene expression.
Plant J. 19, 87-95.

Braarn,J. and Davis,R.W. (1990). Rain-induced, wound-induced, and touch-induced
expression of calmodulin and calmodulin-related genes in Arabidopsis. Cell 60, 357-364.

Brawerrnan,G. (1993). mRNA degradation in eukaryotic cellszan overview. In Control of
Messenger RNA Stability, J .Belasco and G.Brawerman, eds. Academic Press, Inc. San
Diego, CA), pp. 149-160.

Cho,H.T. and Kende,H. (1997). Expression of expansin genes is correlated with growth
in deepwater rice. Plant Cell 9, 1661-1671.

Gil,P. and Green,P.J. (1996). Multiple regions of the Arabidopsis SA UR-A C] gene
control trascript abundance: the 3' untranslated region functions as an mRN A instability
determinant. EMBO J. 15, 1678-1686.

Granger,C.L. and Cyr,R.J. (2001). Characterization of the yeast copper-inducible
promoter system in Arabidopsis thaliana. Plant Cell Reports 20, 227-234.

Gutiérrez,R.A., Ewing,R.M., Cherry,J.M., and Green,P.J. (2002). Identification of
unstable transcripts in Arabidopsis by cDNA microarray analysis: rapid decay is
associated with a group of touch- and specific clock-controlled genes. Proc. Natl. Acad.
Sci. U. S. A 99,11513-11518.

Gutiérrez,R.A., MacIntosh,G.C., and Green,P.J. (1999). Current perspectives on mRNA
stability in plants: multiple levels and mechanisms of control. Trends Plant Sci. 4, 429-
438.

Hargrove,J.L., Hulsey,M.G., and Beale,E.G. (1991). The kinetics of mammalian gene
expression. Bioessays 13, 667-674.

Herrick,D., Parker,R., and Jacobson,A. (1990). Identification and comparison of stable
and unstable messenger-RNAS in Saccharomyces-cerevisiae. Mol. Cell. Biol. 10, 2269-
2284.

141

Johnson,M.A., Pérez-Amador,M.A., Lidder,P., and Green,P.J. (2000). Mutants of
Arabidopsis defective in a sequence-specific mRN A degradation pathway. Proc. Natl.
Acad. Sci. U. S. A 97, 13991-13996.

Love,J., Scott,A.C., and Thompson,W.F. (2000). Stringent control of transgene
expression in Arabidopsis thaliana using the Topl 0 promoter system. Plant J. 21 , 579-
588.

Martinez,A., Sparks,C., Hart,C.A., Thompson,J., and Jepson,l. (1999). Ecdysone agonist
inducible transcription in transgenic tobacco plants. Plant J. 19, 97-106.

Newman,T.C., Ohme-Takagi,M., Taylor,C.B., and Green,P.J. (1993). DST sequences,
highly conserved among plant SA UR genes, target reporter transcripts for rapid decay in
tobacco. Plant Cell 5, 701-714.

Nonet,M., Scafe,C., Sexton,J., and Young,R. (1987). Eukaryotic RNA polymerase
conditional mutant that rapidly ceases messenger RNA synthesis. Mol. Cell. Biol. 7,
1602-1611.

Parker,R., Herrick,D., Peltz,S.W., and Jacobson,A. (1991). Measurement of messenger
RNA decay rates in Saccharomyces cervisiae. Methods Enzymol. 194, 415-423.

Petracek,M.E., Dickey,L.F., Nguyen,T.T., Gatz,C., Sowinski,D.A., Allen,G.C., and
Thompson,W.F. (1998). F erredoxin-l mRNA is destabilized by changes in
photosynthetic electron transport. Proc. Natl. Acad. Sci. USA 95, 9009-9013.

Rojo,E., Leon,J., and Sanchez-Serrano,J.J. (1999). Cross-talk between wound signalling
pathways determines local versus systemic gene expression in Arabidopsis thaliana. Plant
J. 20, 135-142.

Salter,M.G., Paine,J.A., Riddell,K.V., Jepson,l., Greenland,A.J., Caddick,M.X., and
Tomsett,A.B. (1998). Characterisation of the ethanol-inducible alc gene expression
system for transgenic plants. Plant J. 16, 127-132.

Seeley,K.A., Byme,D.H., and Colbert,J.T. (1992). Red light-independent instability of
oat phytochrome mRNA in vivo. Plant Cell 4, 29-38.

Shimizu-Sato,S., Huq,E., Tepperman,J.M., and Quail,P.H. (2002). A light-switchable
gene promoter system. Nature Biotechnol. 20, 1041-1044.

Shyu,A.B., Greenberg,M.E., and Belasco,J.G. (1989). The c-fos transcript is targeted for
rapid decay by 2 distinct messenger-RNA degradation pathways. Genes Dev. 3, 60-72.

Wang,Y.L., Liu,C.L., Storey,J.D., Tibshirani,R.J., Herschlag,D., and Brown,P.O. (2002).
Precision and functional specificity in mRNA decay. Proc. Natl. Acad. Sci. USA 99,
$860-$865.

142

Weinmann,P., Gossen,M., Hillen,W., Bujard,H., and Gatz,C. (1994). A chimeric
transactivator allows tretracycline-responsive gene expression in whole plants. Plant J. 5,
559-569. ~

143

CHAPTER 6

Final remarks and challenges ahead

144

At the onset of this work, genomic approaches using DNA microarrays were
common only for comparing mRN A steady-state levels between two conditions. They
have been rarely applied to the study of mRN A decay or other post-transcriptional
processes. In this dissertation we used the Arabidopsis Functional Genonrics Consortium
cDNA microarrays to identify the genes with the most unstable mRN As in Arabidopsis
plants. Our study provided basic information regarding rapid mRNA turnover in
Arabidopsis and constituted the foundation for further research.

This study, together with genomic studies in other systems, underscores the
relevance of a fundamental and yet not fully answered question in the field: What are the
cis-acting molecular determinants of rapid turnover? It is clear that discrete sequences
within can act as signals for accelerated mRNA degradation. However those stability
determinants that are known can at best explain a small fraction of the labile mRNA
population. Deletion analysis experiments of new unstable transcripts coupled with
mRN A half-life measurements should help us identify novel molecular determinants of
mRNA stability. The identification of a high number of transcripts with similar turnover
rates prompts the analysis of structural features. However although computational
approaches have been useful for detecting DNA sequences involved in transcriptional
regulation (e.g. (Harmer et al., 2000)), they are not the best to detect sequences that
control gene expression at the mRNA stability level. Perhaps due to the complexity of
these sequence motifs and/or the presence of secondary structures, stability determinants
are difficult to detect above background noise with the current sequence analysis tools. In

the future, and as more stability determinants are known, optimized algorithms might

145

better predict mRNA stability determinants and thus guide experimental research. We
should also consider that perhaps different biological principles play a role in governing
mRNA decay rates as well. Recently, a class of micro-RNAS (miRNAs) was shown to
direct cleavage of Scarecrow-like mRN As in Arabidopsis (Llave et al., 2002). miRNAs
are a numerous class of short ~21 to 22 RNAs present in plants, animals and
microorganisms most of which have unknown function. Perhaps miRNAs play a more
general role in controlling mRN A degradation. Comparing genome-wide patterns of
mRNA decay between Arabidopsis WT and KO mutants impaired in miRN A synthesis or
in the post-transcriptional gene silencing pathway should help evaluate the significance
of this mechanism for global mRNA degradation in plants. Similar studies can be also
carried out to investigate the role of new candidate regulators of the mRNA degradation
process as well as known components of the decay machinery. Because mRNAs will
decay with slower rates in the relevant KO mutants, this analysis should help categorize
transcripts on the basis of their decay strategy.

Another intriguing question for firture studies that emerges from this dissertation
research relates to the AtGUTs regulated by touch, light and other signals. Are these
AtGUTs "constitutively" unstable, or is their mRNA stability regulated in response to
light, touch, or other signals? It is clear that at least in some cases regulation of mRNA
stability occurs. However by comparing global turnover rates under different conditions,
post-transcriptional regulatory networks might be unraveled. This large-scale approach
for analysis of post-transcriptional control mechanisms should help uncover the extent

and significance of this level of regulation in plants in response to a variety of stimuli.

146

The studies of mRNA decay and circadian rhythms opened an exciting avenue for
future research. We observed a new relationship between expression of clock-controlled
genes and the sequence specific mRN A decay pathway mediated by the DST element.
Determining the mechanistic relationship between DST] and the circadian clock should
further our understanding of the importance of this level of control for circadian rhythms
in plants. Isolation of the DST] gene should be a priority. But further physiological
studies, e.g. circadian phenotypes, should help us narrow the context for DST‘l function
and its relationship with circadian gene expression. Such knowledge should aid current
mapping efforts. Additional mutants impaired in DST-mediated decay have been isolated.
The study of circadian phenotypes in these mutants, e. g. dst2, and the relationship to the
dst] mutant in this context should provide further insight into the role of DST-mediated
mRN A degradation for circadian rhythms.

Similarities between the mRNA decay process in plants and other eukaryotic
systems have been shown before. However it is also clear that differences exist. This may
be especially true with respect to determinants of long mRNA life, which are typical of
transcripts expressed in terminally differentiated and specialized cells. Deletion analysis
of stable transcripts and mRNA half-life measurements should help identify the signals
that protect transcripts fi'om degradation in plants. In addition, the availability of a
regulated promoter system in plants should facilitate the study of specific aspects of the
mRNA degradation process in stable and unstable transcripts, such as deadenylation.
Deadenylation is an important control point for transcript decay. This step often initiates
the degradation process in yeast (Caponigro and Parker, 1996) and mammals (Wilusz, et

al., 2001) and its rate can impact mRNA stability. Comparing deadenylation rates in

147

stable and unstable transcripts should help evaluate the significance of this step in the

degradation of plant mRNAs.

148

References

Caponigro,G. and Parker,R. (1996). Mechanisms and control of mRNA turnover in
Saccharomyces cerevisiae. Microbiol. Rev. 60, 233-249.

Harmer,S.L., Hogenesch,J.B., Straume,M., Chang,H.S., Han,B., Zhu,T., Wang,X.,
Kreps,J.A., and Kay,S.A. (2000). Orchestrated transcription of key pathways in
Arabidopsis by the circadian clock. Science 290 , 2110-21 13.

L1ave,C., Xie,Z.X., Kasschau,K.D., and Carrington,J.C. (2002). Cleavage of Scarecrow-
like mRNA targets directed by a class of Arabidopsis miRNA. Science 297, 2053-2056.

Wilusz,C.J., Wormington,M., and Peltz,S.W. (2001). The cap-to-tail guide to mRNA
turnover. Nature Rev. Mol. Cell Biol. 2, 237-246.

149

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

11111le11111111111711le