. gauging-sag
.. ”suntan?
:9: y

v.» 4.3:? Pa.-
_ . ) Lit/1 v) I

I.
. :13. A v 3.)
vi... iii}:

a .z

,w 3.5, .

1 :ai... I.

.31.?!

no.» ;

: .i: ~l .

v. is»? .
-1: .

r

u. ‘19
{at $1.3!
51..

5

ztaazl .x

.19.»?! II

n. 3. fly“?!
tit! 1

Infililt}:
snlxxi. 1.x}:
3er (El! J.
5 htt...;:r..
Vt. .227.
s .I.

.3155 «I... .35
5': 3).. 1.
V ten

.2. I
. , . . u . Linn

lxhi ‘ir01t\!~u” 11.15;)! lu1k,‘!u?...llhlt

. y .. all? ‘1. (L.|blrllo!tb.;lcifism}il. .l.‘ .5 e7 ‘4.
2.19! (I. 6!. . :5 Y3. 3.... ’1' .0 3.4". .h lvtli4.1|!lll‘. 26...? r... 911.97., 31.!

. 2...,ixva. 1.3:... .1! Quit 3.3.11.9. .. . I . ix 1....»1;.r...lv1.§..t..lv-:)»3!I.r . .
‘ . .. gnud£ti¥i.?§ltii«tlnu.

‘ livii f;i.r.vt7v Irv.l|r|.tlk

 

 

 
 

 

1.‘l‘tESlS

llfliciIi611 SerTEI LIJNINERISI'TI Llfliflflifiml

3 1293 01563 5885

This is to certify that the

dissertation entitled

6 66/40 /W//4§/£///6”I 0‘4
%/ maze/l”? 5/4/////

presented by

.fi Mix/o 5/09 W” /VW%

has been accepted towards fulfillment
of the requirements for

7!)” 9 degree in [0’7 {fl/“(5

fi/JA

Magél/rofessor

  

 

Date {/Z fi/f/l?

MSU is an Affirmative Action/Equal Opportunity Institution 0—12771

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN REI' U RN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

|| DATE DUE DATE DUE DATE DUE

l

l_ l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[——

MSU Is An Affirmative Action/Equal Opportunity Institution
mtclmmpmsva

 

 

 

 

EFFECT OF TRANSLATION ON mRNA STABILITY
By

Ambrosius Theodorus Cornelis van Hoof

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Graduate Program in Genetics

1997

ABSTRACT
EFFECT OF TRANSLATION ON mRNA STABILITY
By

Ambrosius Theodorus Comelis van Hoof

Regulation of any step in gene expression, including regulation at the level of
mRNA stability, can affect the level of the functional gene product. mRNA degradation
rates can vary over several orders of magnitude, and can be regulated in response to both
endogenous and exogenous stimuli. In addition mRNA decay is tightly linked to
translation. A well-characterized example of this is the destabilizing effect of premature
nonsense codons on eukaryotic mRNA stability. This has been named nonsense-
mediated mRNA decay.

Experiments described in this thesis Show that plants have a nonsense-mediated
mRNA decay pathway. Wild type PHA mRNA is degraded with a half-life of about two
hours in tobacco cells. In contrast PHA mRN A containing a premature nonsense codon is
degraded with a half-life of about 30 minutes. Analyses of several different PHA alleles
showed that introduction of a nonsense codon in the first 60% of the normal coding
region destabilizes the mRNA, while a nonsense codon at 80% of the coding region has
no effect. In addition, nonsense codon—containing PHA mRNA was shown to accumulate
to reduced levels in tobacco leaves and Arabidopsis seedlings, indicating that nonsense-
mediated mRNA degradation iS not restricted to tobacco cell lines

Nonsense—mediated mRNA decay has been reported to occur in a several
eukaryotic systems. Although some aspects of nonsense-mediated mRNA decay are

conserved among eukaryotes, some differences have been reported. Nonsense-mediated

mRNA decay appears to be a cytoplasmic event in yeast , but a nuclear event for some
mammalian genes. Subcellular fractionation of tobacco cells expressing PHA mRNA
subject to nonsense-mediated mRN A decay showed that a large fraction of this mRNA
was bound to polyribosomes. This indicates that in tobacco cells nonsense-mediated
mRNA decay of PHA occurs after the mRNA has been assembled into polyribosomes,
and thus occurs in the cytoplasm.

A second link between translation and mRNA degradation that has been proposed
in several reports is a destabilizing effect of rare codons on mRN A. Direct evidence for
this effect has not been reported, and my experiments indicate that rare codons are not

sufficient to destabilize a reporter transcript in tobacco.

ACKNOWLEDGMENTS

Although I am the sole author on this thesis, it will be obvious to anyone who has
ever been in graduate school that I could not have done all the work described here
without help of the following people, nor would my time at the PRL have been as much
fun without them as it has been. First on the list is of course Paml Green, who contributed
in many ways to the research described here. Her direct contribution to the experiments
described here is very much appreciated, but even more important were the many ways in
which she facilitated this research indirectly and the training she gave me in the process.
Thanks for creating a unique atmosphere in the lab, that made work fun, even on days that
the experiments were routine and on days when the experiments didn't work exactly as
planned. Creating a pleasant atmosphere on days that things work is easy, but
unfortunately that does not cover the majority of the days in a research lab. Thanks for
the food and drinks you bought on many occasions, the coffee, the New York Times, the
journals etcetera. I can't even remember in how many ways Pam improved the way I do
my research and write about it. I do know that without her rules of "one thought per
sentence", and "try to maintain the basic paragraph structure" that I posted over my desk
this thesis would be illegible.

I would also like to thank the other members of my advisory committee: Drs.
Frans de Bruijn, Gus de Zoeten, Jon Kaguni, Chris Somerville and Craig Thompson.
They took time out of their busy schedules, contributed significantly by proposing some
of the experiments, and in other cases by confirming that I was on the right path. I think
they made a very good committee and the meetings we had were always enjoyable.

iii

Thanks for the interest you showed and the lively discussions.

I would like to thank all the members of the Green lab. Linda, who helped with
some of the experiments described here, kept everything in stock and well—organized, and
made sure that anything done by undergraduate help was done right. Tom, who helped
me find my way around the lab during my rotation. Past and present post-docs and
students: Crispin, Christie, Jay, Jim, Masaru, Michael, Miguel, Mike, Mike, Nikki,
Pauline, Pedro, Scott and Wan-Ling. Each of them contributed in different ways.
Sometimes they contributed to the actual experiments by suggestions on how to do an
experiment or by asking a question that set wheels in my head in motion, sometimes they
contributed to the writing of this thesis by pointing out that what I wrote is not what I
meant or by pointing out that I had Split an infinitive and sometimes they contributed to
the improvement of my research skills by constructive or blunt criticism. Several rotation
students (Jim, John, Lisa, Melanie, Starch and Terry) also helped with the research.

Numerous people in the PRL gave me advice, especially when I tried doing some
protein work. Others were just great friends (in good times and in worse time) that
helped in ways I can't explain in one or two sentences. I am sure they know who they are
and I am sorry that I have to move on. Keep in touch. A special thanks to you for reading
this thesis. Not very many people will read it. Let me know what you think of it and I
will buy you a beer.

Finally, but most importantly I would like to thank my parents and sisters. My
parents taught me that getting a good education was one of the most important things at
the start of ones life. My education is almost finished now, and I think I am ready for life
to start. My mom and dad and my Sisters, J 036 and Carin, and brothers-in-law provided

iv

strong moral support during my time in graduate school. They didn't always understand
what I was working on, or why I would want to work on that, or why I couldn't work on it

closer to home, but they were there when I needed them.

TABLE OF CONTENTS

LIST OF FIGURES ................................................................................................... viii

CHAPTER 1: Introduction ....................................................................................... 1
Introduction ................................................................................................... 2
Control of mRN A decay in plants ................................................................. 2
Linkage of mRNA stability and translation .................................................. 27
Scope of this thesis ........................................................................................ 41
References ..................................................................................................... 43

CHAPTER 2: Premature nonsense codons decrease the stability of

phytohemagglutinin mRN A in a position-dependent manner .................................. 55
Abstract ......................................................................................................... 56
Introduction ................................................................................................... 57
Results .......................................................................................................... 61
Discussion ..................................................................................................... 72
Materials and methods .................................................................................. 76
Acknowledgments ......................................................................................... 8 1
References ..................................................................................................... 82

CHAPTER 3: Polyribosomal PHA mRNA can be degraded by the nonsense-
mediated mRN A decay pathway ............................................................................... 86

Abstract ......................................................................................................... 87

vi

Introduction ......................................................................................... ‘ .......... 88

Results ........................................................................................................... 93

Conclusions ................................................................................................... 101
Materials and methods .................................................................................. 104
References ..................................................................................................... 107

CHAPTER 4: Rare codons are not sufficient to destabilize a reporter gene

transcript in tobacco .................................................................................................. 110
Abstract ......................................................................................................... 111
Introduction ................................................................................................... 1 12
Results ........................................................................................................... 1 15
Conclusions ................................................................................................... 121
Acknowledgments ......................................................................................... 123
References ..................................................................................................... 124

CHAPTER 5: Conclusions and future prospects ...................................................... 127
Nonsense-mediated mRNA decay ................................................................ 128
mRNA degradation in plants ....................................................................... 134
References ..................................................................................................... 1 39

APPENDIX: GUT 15 cDNAs from tobacco and Arabidopsis correspond to
transcripts with unusual metabolism and a short conserved ORF ............................ 141

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

vii

LIST OF FIGURES
FIGURE 1-1. Sequence of DST elements identified in seven SA UR genes ............. 6
FIGURE 1-2. The threshold model of gene silencing ............................................... 21
FIGURE 1-3. Proposed mechanisms of SRS4 and PHYA mRNA decay in plants... 25
FIGURE 2-1. PHA alleles introduced into tobacco cells and transgenic plants ....... 62

FIGURE 2-2. Decreased expression of the FS and STOP transcripts in tobacco
protoplasts ................................................................................................................. 64

FIGURE 2—3. Destabilization of the PHA mRN A by the FS mutation in stably
transformed tobacco cells .......................................................................................... 66

FIGURE 2-4. Effects of the FS and STOP mutations on PHA mRNA stability
in tobacco .................................................................................................................. 67

FIGURE 2-5. Reduced accumulation of the FS and STOP transcripts in
transgenic plants ........................................................................................................ 69

FIGURE 2-6. Premature nonsense codons in the first 60% of the PHA coding

region cause mRNA instability ................................................................................. 71
FIGURE 3-1. PHA transcripts with premature nonsense codons copurify with
polyribosomes ........................................................................................................... 96
FIGURE 3-2. Absorption profiles of sucrose gradient ............................................. 97
FIGURE 3-3. RNA gel blot analysis of fractionated sucrose gradients .................... 99
FIGURE 3-4. Distribution of PHA mRNA on sucrose gradients ............................. 100
FIGURE 3-5. Absorption profile of sucrose gradients ............................................. 102

FIGURE 4-1. Experimental design used to test the effect of rare codons on
mRNA stability ......................................................................................................... 116

FIGURE 4-2. Insertion of rare codons does not affect PHA mRNA stability
in BY-2 cells ............................................................................................................. 1 18

viii

FIGURE 4—3. Insertion of rare codons does not affect PHA mRN A levels in

transgenic tobacco plants .......................................................................................... 120
FIGURE 6-1. GUT15 probes hybridize to two polyadenylated transcripts .............. 144
FIGURE 6-2. Comparison of the conserved regions in GUT15 genes ..................... 146

ix

CHAPTER 1

INTRODUCTION

Parts of this chapter have been previously published in Modern Cell Biology.
Reference: A van Hoof and PJ Green(1997). Control of mRNA decay in plants. In:
mRNA metabolism and post-transcriptional gene regulation. JB Harford and DR Morris

eds. Modern Cell Biol. 17:201-216

INTRODUCTION

Over the last several years, emphasis on the study of mRNA stability has
increased significantly. This is especially true for mRNA stability in plants. The first
part of this chapter will give an overview of what we have learned about mRNA
degradation in plants, whereas the second part will concentrate on what we know about

the influence of translation on mRNA stability.

CONTROL OF mRN A DECAY IN PLANTS

Plants have many unique features that make Studies of mRN A turnover attractive
from both a conceptual and a technical perspective. In cases where mechanisms are
common to plants and other higher eukaryotes, plants can offer new approaches with the
potential for novel findings. This chapter will not attempt to comprehensively review
recent work, but instead to highlight those findings that are the most definitive and that
reveal the most and the fewest Similarities to work carried out in other systems. I will
also attempt to identify relevant experimental attributes of plant systems that may not be
familiar to those outside the field. For a broader presentation of work on plant mRNA
decay, readers are referred to several recent reviews (Gallie, 1993; Sullivan and Green,

1993; Abler and Green, 1996).

 

3

SEQUENCES CONTROLLING INI-IERENT mRN A STABILITY

Plant mRN A decay rates appear to be similar to those in other higher eukaryotes,
with the average mRN A having a half-life on the order of several hours (Siflow and Key,
1979; Sullivan and Green, 1993; Taylor and Green, 1995). Recent studies have focused
primarily on unstable mRNAs with half-lives of an hour or less because, as in
mammalian cells, such Short-lived mRNAs often encode regulatory proteins or other
proteins of interest that must be induced transiently. The rapid control of gene expression
facilitated by unstable transcripts may be particularly important in plants because, as
sessile organisms, they must respond rapidly to environmental conditions from which
they cannot escape.

Relatively unstable transcripts with half-lives of about an hour or less include the
PH YA mRN A encoding a major form of the plant photoreceptor phytochrome (Seeley et
al., 1992), the Small-Auxin-Up RNAs (SA URS) (McClure and Guilfoyle, 1989) that are
rapidly regulated by the plant hormone auxin, transcripts of the TCH3 and TCH4 (Braam
and Davis, 1990) genes induced transiently by touch, and nonsense mutants of the
phytohemagglutinin gene (PHA) (van Hoof and Green, 1996), and other mRNAs (Taylor
and Green, 1995). It appears that, as in other systems, rapid mRNA decay is an active
process triggered by specific sequences. Recently, several sequences have been shown to
function as mRNA instability determinants in tobacco. Thus far, the most detailed
studies relate to sequences present in the unstable SA UR transcripts mentioned above
(Newman et al., 1993; Gil and Green, 1996; Sullivan and Green, 1996). These transcripts

are unique to plants, and instability determinants derived from them appear to be novel.

4

Plants also recognize AUUUA repeats as signals for rapid mRNA decay, as do
mammalian cells (Chen and Shyu, 1995), and premature nonsense codons, as do yeast,

Caenorhabditis elegans, and mammalian cells (see Maquat, 1995; and Weng et al.,

1997).

Sequences derived from SA UR transcripts

The SA UR transcripts are among the most unstable plant mRNAs reported, with
half-lives of 10-50 minutes depending on the method of measurement (McClure and
Guilfoyle, 1989; Franco et al., 1990). Although the exact function of the SAUR proteins
is not known, their expression patterns indicate that they may be involved in auxin-
induced cell elongation (McClure and Guilfoyle, 1989). All unstable SA UR mRNAs
reported to date contain a hi ghly—conserved sequence about 40 nucleotides in length
referred to as DST, for downstream element (McClure et al., 1989). Although this
element does not resemble prominent mRNA instability determinants in other systems, it
is found in the 3' untranslated region (UTR), a location Similar to that of many sequences
causing rapid decay in mammalian and yeast transcripts. A synthetic dimer of the
soybean SA UR 15A DST element is sufficient to markedly destabilize B-glucuronidase
(GUS) and B-globin reporter transcripts when mRNA half-lives are measured following
actinomycin D (ActD) treatment of stably transformed tobacco (BY-2) suspension
cultures (Newman et al., 1993). It iS likely that DST sequences also target transcripts for
rapid decay in intact plants because the presence of DST sequences causes coordinate

decreases in mRNA abundance in transgenic tobacco plants (Newman et al., 1993;

Sullivan and Green, 1996).

The DST element consists of three highly conserved regions separated by two
variable regions, as Shown in figure 1-1. The ATAGAT and T—-GTA of the second and
third conserved regions, respectively, are invariant in all SA UR DST elements. These
elements were the focus of a recent mutational analysis of the DST element carried out in
stably transformed BY-2 cells and in transgenic tobacco plants (Sullivan and Green,
1996). The mutant DST elements were evaluated within the 3' UTR of a fl-globin
reporter transcript. Five- or Six-base substitution mutations in the ATAGAT or the T--
GTA regions were sufficient to inactivate the element, stabilizing the transcript to the
level of Spacer or no-insert controls. These mutations also inactivated the element in
transgenic plants, based on mRNA accumulation levels. Smaller two-base mutations in
these regions had varying effects, ranging from little or no effect to significant increases
in reporter mRNA half-life and accumulation. Interestingly, a two-base substitution
changing the invariant GTA to CCA inactivated the element in plants but not in BY-2
cells. Taken together, these results indicate that both the ATAGAT and GTA regions are
required for DST to function as an instability determinant in leaves of transgenic plants
or in BY-2 cells, and that the sequence requirements for DST to function in the former are
more stringent (Sullivan and Green, 1996).

In addition to work on isolated DST elements, studies on one natural SA UR gene,
SA UR-ACI (Gil et al., 1994), have been carried out to identify regions of the transcript
that are responsible for its instability (Gil and Green, 1996). Similar to other SA UR
promoters (Li et al., 1991, 1994), the SA UR-A CI promoter region mediates auxin

induction of the gene (Gil and Green, 1997). In contrast, sequences downstream of the

Plant Gene DST Sequence

Soybean 15A GGAG- 5- CATAGATTG- 7- CATTTTGT AT
Soybean X15 GGAG- 5- CATAGATTA- 7- CATTTGGTAC
Soybean SB GGAG- 4- CATAGATTA- 7- CATTTTGT AC
Soybean X10A GGAT- 5- GATAGATTA- 8- AAATTTGTAC
Soybean 10A5 GGAG- 5- GATAGATTA- 8- AAATTTGTAC
Mungbean ARG7 GGTT- 2- CATAGATTA- 8- ATTTTTGTAA
Arabidopsis AC1 GGAA- 9 - CATAGATCG - 8 - CA_AT Gcsfr AT
Consensus fiAg 3:391 cATAGATT a -7/3-°/AAT/A1Tt GT Ac

Figure 1-1. Sequence of DST elements identified in seven SA UR genes. The element
consists of three conserved regions (shown in gray) separated by two variable regions.
Numbers indicate the number of bases between the conserved regions. DST sequences in
five soybean SA UR genes (McClure et al., 1989), one mung bean gene (Yamamoto et al.,
1992), and one Arabidopsis thaliana SA UR gene (Gil et al., 1994) are aligned. For the
consensus, residues are shown as follows: underlined residues are invariant among all
seven genes; capitalized residues are conserved in at least six genes; capitalized residues
separated by a slash are conserved among all seven genes, with each residue present in
three or four genes; lowercase residues are conserved in four or five of the genes.

7

SA UR-ACI promoter limit mRNA accumulation in a manner that is independent of auxin
(Gil and Green, 1996). Effects on mRNA stability were assayed in BY-2 cells using a
tetracycline-repressible promoter system (Gossen and Bujard, 1992; Weinman et al.,
1994) to shut off transcription. The results showed that the SA UR-ACI 3' UTR, but not
the coding region, functions as a potent mRNA instability determinant in tobacco (Gil
and Green, 1996). The SA UR-ACI 3' UTR has a DST element located 10 bases from the
poly(A) addition Site (Gil et al., 1994) which certainly could contribute to this instability.
However, the 3' UTR also contains several other sequences resembling DST subdomains
(i.e., the ATAGAT and T--GTA regions) which may also be involved. Based on work
done with the SA UR-ACI gene (Gil and Green, 1996) and with the DST dimer (Newman
et al., 1993), it appears that the SA UR mRNAs are unstable independent of auxin
treatment, presumably to facilitate rapid changes in mRNA levels following auxin-

mediated transcriptional changes (Franco et al., 1990).

AUUUA sequences

AU-rich elements (ARES) are perhaps the most actively studied mRNA
instability sequences in mammalian cells (Chen and Shyu, 1995). These elements are
found in the 3' UTRS of many unstable mammalian transcripts and often contain multiple
AUUUA motifs (Caput et al., 1986; Shaw and Kamen, 1986). Studies arguing that
AUUUA sequences are important mRNA instability determinants in mammalian cells
include the demonstration that synthetic AUUUA sequences destabilize reporter

. transcripts (Vakalopoulou et al., 1991) and that mutation of AUUUA sequences in natural

8
ARES (e.g., in c-fos) can inactivate instability function (Shyu et al., 1991). AU-rich
elements may contribute to mRNA instability in yeast, but AUUUA sequences do not
appear to be critical determinants (Muhlrad and Parker, 1992).

To evaluate whether AUUUA sequences can function in plant cells, an AUUUA
repeat consisting of 11 copies of the element, was inserted into the 3' UTR of GUS or B-
globin reporter transcripts and mRNA half-lives were measured in stably transformed
BY-2 cells following ActD treatment. Reporter transcripts containing the AUUUA repeat
degraded much more rapidly than did control transcripts containing no insert or a spacer
with interspersed G's and C's. Because a similar element containing an AUUAA repeat
did not cause rapid decay, it appears that plants Specifically recognize the AUUUA
sequence, not simply sequences rich in AS and Us. As in mammalian cells, not all
AUUUA sequences act as instability determinants (Greenberg and Belasco, 1993; Walker
et al., 1995), in part because the minimal domain appears to be longer in some cases
(Lagnado et al., 1994; Zubiaga et al., 1995).

The effect of the AUUUA repeat on mRNA accumulation in BY-2 cells and
transgenic plants was even greater than its effect on mRNA stability, as measured in the
latter system (Ohme-Tagaki et al., 1993; Sullivan and Green, personal communication).
A similar discrepancy between mRN A accumulation and mRNA stability was observed
for an ARE from granulocyte-monocyte colony-stimulating factor (GM-CSF) in
mammalian cells (Savante-Bhonsale and Cleveland, 1992). In the case of GM-CSF, it
was suggested that mRNA half-lives measured in the presence of ActD may be
overestimates if recognition of the ARE is dependent on transcription. A similar model

could explain the data obtained with the AUUUA repeat in plants. Alternatively the

9

element may be recognized by more than one mechanism.

The effect of AUUUA sequences on mRNA stability and abundance in plants may
have important relevance to biotechnology because some AU-rich foreign transcripts fail
to accumulate in transgenic plants. The best example of this problem comes from
attempts to express insecticidal proteins called B. t. toxins in plants using genes derived
from the bacterium Bacillus thuringiensis (Diehn et al., 1996). Even when B. t. toxin
genes are introduced into plants under the control of strong plant promoters, very little
B. t. toxin mRNA can be detected. In a practical sense, this problem has been overcome
by creating synthetic B. t. toxin genes that more closely resemble plant genes and are
highly expressed (Perlak et al., 1990, 1991). The major modifications have involved
removal of several AUUUA sequences and other potential RNA-processing signals,
increasing the GC content, and improving the codon usage. It has been suggested that
these changes stabilize the mRNA. In one case this hypothesis was examined directly by
comparing the stability of wild-type and synthetic B. t. toxin transcripts. The wild-type
B. t. toxin transcript (with a GC content of 34%) was much less stable than the synthetic
version (with a GC content of 64%), but the contribution of AUUUA elements to this
effect remains to be evaluated (EJ De Rocher and PJ Green, personal communication;

Diehn et al., 1996).

10

NOVEL STIMULI AFFECTING mRN A STABILITY IN PLANTS

Fungal elicitors

Some of the most striking examples of differentially-regulated mRNA stability in
plants are related to a plant's need to adapt rapidly to an ever-changing environment. One
case that illustrates this point occurs during the response of common bean to pathogen
attack. Plants respond to pathogen attack by changing the expression level of many genes
(Alexander et al., 1994). Often these changes can be re-created by treating cell cultures
with compounds of plant or pathogen origin, referred to as elicitors, that the cell
recognizes as signals of pathogen attack. In common bean, the PvPRPI gene, which
encodes a cell wall protein, is among those regulated in response to a fungal elicitor
(extracellular polysaccharides from fungi) (Sheng et al., 1991). Presumably this gene iS
down-regulated by elicitor treatment to alter the composition of the cell wall, which may
contribute to plant protection.

Several lines of evidence indicate that the decay rate of the PvPRPI transcript is
regulated by elicitor treatment (Zhang et al., 1993). Most important, decay in the
presence of elicitor was faster than in its absence when mRN A decay rates were measured
following treatment of cultured bean cells with ActD. Further supporting evidence comes
from nuclear run-on transcription experiments which showed a constant transcription
rate, regardless of elicitor treatment. The magnitude of the elicitor effect on mRNA
stability cannot be calculated directly because the decay of the PvPRPI transcript appears

to be dampened by ActD. In the presence of ActD, the half-life of the PvPRPI transcript

11

is 18 h and 60 h with and without elicitor respectively, but the mRNA disappears with a
half-life of only 45 min if elicitor is added in the absence of ActD. Decay of the PvPRPI
transcript is also sensitive to the protein synthesis inhibitors emetine and anisomycin,
leading to the proposal that the mechanism requires ongoing transcription and translation
(Zhang et al., 1993).

Ultraviolet (UV) crosslinking experiments carried out with protein extracts from
unelicited cells have identified a 50 kD polypeptide, called PRP-BP, that can be
crosslinked to the 3' half of the PvPRPI transcript (Zhang and Mehdy, 1994). No
proteins have been found to cross-link to the 5' half. Using a series of 3' and 5' deleted
transcripts, the binding site for PRP—BP was mapped to a 27 nucleotide U-rich site that
contains one copy of the AUUUA motif. Crosslinking of PRP-BP could be competed
specifically by poly U and by poly AU but not by several other competitors (Zhang and
Mehdy 1994). It remains to be shown that this binding site is necessary for regulated
instability of the transcript, but interestingly, the binding activity responds to elicitor
treatment. The binding activity increases about fivefold one hour after elicitor treatment,
concurrent with PvPRPI destabilization. This increase in binding activity is not caused
by its de novo synthesis, but seems to be regulated post-translationally. Binding activity
in extracts from unelicited cells could be increased by treatment with dithiotreitol (DTT)
or [I-mercaptoethanol, and could be reversibly eliminated by treatment with N-
ethylmaleimide (NEM) or diamide (Zhang and Mehdy 1994). One explanation of these
results is that binding of PRP-BP triggers rapid decay of the PvPRPI transcript in a
manner that is redox regulated (Zhang and Mehdy 1994).

Another legume transcript for a putative cell wall protein (MsPRPZ) also appears

12

to be regulated post-transcriptionally. In this case, the regulatory signal is osmotic stress.
This transcript was found to be induced in salt-treated cells without a change in
transcription rate (Deutch and Winicov, 1995). Interestingly the 3' UTR of MsPRPZ
contains a sequence that is similar to the putative cis-acting element in PvPRPI . It will
be interesting to see whether the two genes, which share only 26% amino acid identity,

are using similar mechanisms to regulate mRNA stability in response to different signals.

Light

Light is perhaps the most important external Signal to which plants must respond.
Therefore, the effects of light on gene expression have been studied extensively, and post—
transcriptional events have been found or suggested in a number of cases (Thompson and
White, 1991). The best illustration of a gene that is post-transcriptionally regulated in
response to light is the pea fed-1 gene, which encodes the photosynthetic electron carrier
ferredoxin I (Elliott et al., 1989). As with many other photosynthetic genes, fed-I
expression is induced by light, with mRNA accumulating to about fivefold higher levels
in the light than in the dark. Although no direct measurements of mRNA half-lives have
been made due to technical limitations, strong circumstantial evidence argues that light
regulation of fed-1 occurs at the level of mRNA stability. When run-on transcription
experiments were performed using nuclei from light-grown and dark-adapted plants, no
differences in transcription rates were observed (unpublished data cited in Dickey et al.,
1992). Moreover, the cis-acting sequences responsible for light-induction were localized

to the transcribed region of the gene (Elliott et al., 1989). When different regions of fed-1

13

were fused to a reporter gene, the internal light regulatory element (iLRE) was found to
include sequences from both the 5' UTR and the 5' end of the coding region. In
particular, a fusion of the 5' UTR and the first 20 codons of fed-1 to the chloramphenicol
acetyl transferase coding region rendered the chimeric transcript light-responsive (Dickey
et al., 1992, 1994). It has also been shown that an open reading frame is required for
differential accumulation of fed-I in response to light (Dickey et al., 1994). Light-
regulation of the fed-I gene is diminished or abolished by a missense mutation in the start
codon and by nonsense, but not missense, mutations in the 5' portion of the coding region.
This is the strongest evidence that light acts at the level of RNA stability because
nonsense codons would not be expected to influence light-regulated transcription. It is
possible that these data can be explained by nonsense-mediated decay overriding the
normal stability of fed-I mRNA in the light. However, perhaps it is more likely that for

the iLRE to function correctly, it must interact with a translating ribosome.

Other stimuli

A number of plant genes are regulated by the availability of nutrients such as
carbohydrates. Recently, it has been Shown that sucrose availability can influence oc-
amylase gene expression at the level of mRNA stability in rice suspension cultures. The
addition of sucrose decreases the abundance of oc-amylase transcripts (Yu etal., 1991)
due to both transcriptional and post—transcriptional regulation (Sheu et al., 1994). When
mRN A decay rates were measured following ActD treatment, the half-life of the pool of

oc-amylase mRNA was about 12 h in sucrose-starved cells but decreased to less than an

14

hour when sucrose was provided (Sheu et al., 1994). oc-Amylase in rice is encoded by at
least eight different genes that produce very little mRNA in the presence of sucrose.
However, under sucrose starvation conditions, two transcripts give rise to 90% of the oc-
amylase pool. Although the individual half-lives of the two highly expressed transcripts
and of a poorly expressed transcript differ, each is about fourfold more stable in the
absence of sucrose (S.-M.Yu, personal communication).

A large number of other genes in plants have been proposed to be differentially
regulated at the level of mRNA stability, but in most cases the work is at a relatively early
stage. For example, it has been proposed that histone mRNAs can be post-
transcriptionally regulated, possibly in response to the cell cycle in the case of histone H3
(Kapros et al., 1995). If this is the case, then the sequences and mechanisms responsible
could be novel since plant histone H3 transcripts end with a poly(A) tail (Chaboute et al.,
1988; Ohtsubo and Iwabuchi, 1994) rather than with the regulatory 3' inverted repeat
well-characterized in mammalian cells (Marzluff and Hanson, 1993). In any event, it is
clear from half-life measurements in nonsynchronized cells that the putative cell cycle-
dependent forms of histone H3 mRNA degrade with half-lives of 1 h in tobacco (Taylor
and Green, 1995) or 2 h in alfalfa cells (Kapros et al., 1995), while the mRNA for a
putative replacement-type histone had a half-life of 6 h in alfalfa cells (Kapros et al.,
1995). Aside from these examples, there are many other cases for which differential
regulation of mRNA stability has been proposed (reviewed in Gallie et al., 1993; Abler
and Green, 1996), the majority of which are based on observed discrepancies between

nuclear run-on transcription rates and RNA accumulation levels.

15

CONTRIBUTION OF mRNA DECAY TO GENE SILENCING AND ANT ISENSE

MECHANISMS

The nature of gene silencing in plants

Perhaps the most intriguing phenomenon associated with the control of mRN A
stability in plants is a form of gene silencing also known as cosuppression or sense
suppression. This type of gene silencing was discovered during reverse genetic studies
aimed at elucidating the function of cloned plant genes via overexpression approaches.
Most often, these experiments involved introduction of the cloned plant gene into plants
under the control of a strong (viral) promoter and the subsequent regeneration of
transgenic plants. One would expect the expression of the transgene to vary among these
plants due to chromosome position effects, so that high levels of overexpression would
be achieved in some plants and lower to no overexpression in others. Indeed this occurs,
but the surprising observation is that in some transgenic plants, accumulation of the
mRNA from both the transgene and the endogenous gene is markedly suppressed. That is,
in these cases, the transgene reduces its own expression and that of the endogenous gene
(reviewed in Chasan, 1994; Flavell, 1994; Matzke and Matzke, 1995). The process is
homology dependent, but the transgene and the endogenous gene do not need to be
identical. Either one gene or several closely related genes can be Silenced.

There are at least two distinct mechanisms of gene Silencing in plants,
transcriptional and post-transcriptional. Transcriptional gene Silencing may be similar to

other epigenetic phenomena in plants, animals, and fungi (Matzke and Matzke, 1995).

l6

Post-transcriptional gene silencing has been studied in the most detail in plants, but
recently has also been described in Neurospora crassa (Cogoni et a]. 1996). However, as
discussed below, this type of gene silencing is also associated with some forms of plant
viral resistance and may account for some of the position effects on transgene expression
previously thought to affect only transcription. With respect to the latter, it has been
observed that some transgenic plants that exhibit low transgene mRNA levels actually
have high rates of nuclear run-on transcription and must therefore be silenced post-
transcriptionally (e.g., Lindbo et al., 1993). There is a general assumption that post-
transcriptional gene silencing involves accelerated decay of the mRNAs that are
diminished, although the supporting data are somewhat indirect, as illustrated in the
examples that follow. Within the confines of this chapter, it will only be possible to
present a snapshot of post-transcriptional gene silencing and the major models proposed
to explain it. For a more thorough discussion, readers are referred to several reviews on
this topic (Chasan, 1994; Flavell 1994; de Carvalho Niebel et al., 1995b; Matzke and

Matzke, 1995; Baulcombe and English 1996).

Examples of post-transcriptional gene silencing

One gene whose post-transcriptional gene silencing has been studied in detail is a
gene for the basic isoform of [31,3-glucanase of Nicotiana plumbaginifolia. When this
gene was introduced into tobacco plants (Nicotiana tabacum), the plant with the highest
transgene mRNA level was selected for further study and self—fertilized (de Carvalho et

al., 1992). Progeny of the original transfonnant that were homozygous for the transgene

17

did not express the transgene, while heterozygous progeny did. The endogenous basic
Bl,3-glucanase isoform, but not the acidic isoforrn, was also silenced in the homozygotes
(de Carvalho N iebel et al., 1995a). The effect was observed at the RNA level, but nuclear
run-on transcription experiments showed that there was no difference in transcription rate
between heterozygotic unsilenced plants and homozygotic silenced plants (de Carvalho et
al., 1992; de Carvalho-Niebel et al., 1995a). This was confirmed by analyzing the level
of RNA copurifying with nuclei using a reverse transcriptase polymerase chain reaction
(RT-PCR) assay. Both spliced and unspliced, polyadenylated RNA were more abundant
in homozygous silenced nuclei than in heterozygous unsilenced nuclei (de Carvalho

N iebel et al., 1995a). These data provide strong evidence that gene Silencing occurs post-
transcriptionally and presumably involves degradation of the mature form of the mRN A,
either just before export from the nucleus or in the cytoplasm.

Other prominent examples of post-transcriptional gene silencing relate to the
chalcone synthase (CHS) and dihydroflavonol-4-reductase (DFR) genes in petunia.
Again, a percentage of the plants transformed with a CHS or DFR transgene showed gene
silencing. Both genes are involved in flower pigmentation, so that in these cases, the
silenced state (white flowers) can easily be distinguished from the normal state (purple
flowers). The white flowers have low CHS or DFR RNA levels (Napoli et al., 1990; van
der Krol et al., 1990), and, at least in the case of CHS, transcription rates and nuclear
RNA levels are normal compared to those of purple flowers (van Blokland et al., 1994).

There are also instances of engineered virus resistance that are representative of
post-transcriptional gene silencing and provide further insight into the process.

Transgenic plants transformed with a virus-derived gene sometimes show resistance to

18

viral infection. In one particularly interesting example, some of the plants transgenic for
the Tobacco Etch Virus (TEV) coat protein transgene were initially successfully infected
by TEV (a cytoplasmically replicating RNA virus), but after some time they "recovered"
(Lindbo et al., 1993). This recovery resulted in the production of new leaves that had no
detectable virus. Protoplasts isolated from recovered leaves did not support virus
replication, but protoplasts from untransformed and transformed uninfected plants did
(Lindbo et al., 1993). The most interesting finding was that the transgene was silenced in
the recovered tissue. mRNA accumulated to reduced levels in these recovered leaves
even though the transcription rates of the transgene were the same as those in uninfected
plants (Lindbo et al., 1993). Apparently, the viral infection had triggered the silencing of
the transgene and viral resistance, most likely by causing rapid decay of transgene and
viral transcripts (Lindbo et al., 1993).

English et a1. (1996) recently provided additional evidence that transgene-
mediated virus resistance can be due to post-transcriptional gene silencing. An important
prediction of this hypothesis was that a silenced transgene should be able to confer virus
resistance to viruses that have sequence Similarity, even if this gene was not originally
part of the virus. The cytoplasmically replicating Potato virus X (PVX) tolerates insertion
of a GUS gene that is not normally part of the viral genome. This PVX.GUS virus was
used to infect plants that contained either a post-transcriptionally silenced GUS gene, an
expressed GUS gene, or no GUS gene. As expected, infection of nontransformed plants
or plants that had not silenced the GUS gene showed numerous GUS-positive lesions,
indicating a successful infection by PVX.GUS. These plants also accumulated PVX.GUS

RNA to high levels. In contrast, plants carrying the silenced GUS gene showed very few

l9

GUS-positive lesions after inoculation with PVX.GUS and accumulated little if any
PVX.GUS RNA. Additional experiments showed that this outcome was critically
dependent on sequence identity between the silenced transgene and the infecting virus. A
silenced version of the PG or nptII gene conferred resistance to PVX viruses carrying the
same gene, while none of these silenced genes affected infection by a PVX virus carrying
the gene for green fluorescent protein (English et al., 1996). In this study and in previous
analyses of these same transgenic lines (Hobbs et al., 1990), a correlation was observed
between the post-transcriptional gene Silencing phenomenon, the number of GUS genes
inserted in the genome, and methylation levels of specific restriction sites in the GUS
gene.

Similar observations have been reported for post-transcriptional gene silencing of
an nptII gene. In this case, one particular transforrnant containing a silenced nptII gene
was self-fertilized and the progeny were analyzed. All progeny transcribed their nptII
genes at approximately the same rate, but some plants accumulated 20- to 40-fold more
RNA than other progeny and the original transforrnant. The progeny that had silenced
genes (and the original transforrnant) had two independently segregating loci of nptII and
methylation of several restriction sites in the transgenes. In contrast, progeny that
expressed nptll at high levels had only one locus (although multiple copies at that locus)
of nptII and less or no methylation on these same restriction Sites (Ingelbrecht et al.,
1994). It is not clear whether methylation is a cause and/or an effect of gene silencing in

these cases.

20

Models for post-transcriptional gene Silencing

Two favored models that have been proposed to explain how post-transcriptional
gene silencing is triggered are the threshold model and the ectopic pairing model. Neither
of these models has been proven valid, and arguments for and against each can be made.
Here we will briefly present both models, without attempting to comprehensively discuss
all the supporting data. A proposed explanation for how these models lead to increased
mRNA decay is also included. For a more detailed discussion the reader is referred to
several recent reviews (Flavell 1994; de Carvalho Niebel et al., 1995b; Baulcombe and
English, 1996)

In the threshold model (Lindbo et al., 1993), gene silencing is triggered when
homologous mRNA accumulates above a certain threshold (Figure 1-2B). This excess
mRNA is sensed by one or more trans-acting factors, which in turn trigger increased rates
of mRNA decay. In the absence of a gene-silencing mechanism, this excess RNA would
be expected to accumulate to a high level, as in Figure 1-2A. The trans-acting factors can
be RNA-binding proteins, aberrant RNAS, or antisense RNAS (see below). An
observation supporting this model is that those transgenes with the‘highest transcriptional
activity are generally (but not always) the ones silenced. A possible weakness is that it is
difficult to accommodate the observation that a promoterless transgene can cause
silencing (van Blokland et al., 1994) without proposing technical problems with the
experiments.

In the ectopic pairing model (reviewed by Baulcombe and English 1996), gene

silencing is triggered by direct interaction of two DNA molecules, resulting in an altered

21

(DC) GD

GE 00)

B g 53.9 RNAthreshold
Um ........................ OLE) .........................
C: C

30) 39

(U: (ti-C

<2 <3:

2:: Z;

(I; {I

 

 

 

 

 

Figure 1-2. The histograms represent the mRNA abundance from a given gene (black
bar) in the absence (1) or presence of an additional copy of that gene (open bar)
transcribed at relatively low levels (2) or at high levels (3). A depicts the expected
situation in the absence of a gene-silencing mechanism, and B depicts what has been
found in several cases of gene silencing. The RNA threshold model proposes that gene
silencing is triggered when the total amount of homologous RNA is higher than some
threshold level. This trigger leads to a stable state in which RNA is rapidly degraded and
therefore accumulates to very low levels.

22

state of the DNA (or chromatin). The exact nature of this altered state is unknown, but
recent results showing a correlation between post-transcriptional gene Silencing and
methylation may indicate that methylation is involved. This altered state of the DNA
leads to the production of a small amount of aberrant RNA in addition to the normal
RNA. This aberrant RNA is then transported to the cytoplasm, where it triggers
degradation of the mRNA. This model explains why a promoterless transgene can induce
silencing (van Blokland et al., 1994), and it is easy to imagine how methylation could
correlate with a post-transcriptional event that involves DNA-DNA interactions.
However, a rather complicated hypothesis (Baulcombe and English 1996) is required to
accommodate the observation that a cytoplasmically replicating RNA virus can induce

silencing, as in the TEV example.

Role of antisense RNA

In both models, triggering gene silencing leads to an increased degradation rate of
the homologous mRNA. It has been proposed that an RNA-dependent RNA polymerase
(RdRP) may be involved in this process. RdRP activity has been characterized in several
plant species and has been purified from tomato (Schiebel et al., 1993a; 1993b). The
substrate for RdRP in the threshold model could be the excess RNA, while in the ect0pic
pairing model it could be the aberrant RNA. The antisense RNA produced in this manner
is envisioned to form a hybrid with the sense RNA and thus be targeted for degradation
by a double-strand specific RNase.

Antisense RNA has been widely used to reduce expression of homologous genes

23

in a variety of plants (reviewed in Bourque, 1995). In many cases, antisense effects in
plants are associated with diminished levels of the sense RNA. The exact mechanism by
which this occurs has not been elucidated, but there is precedent for the involvement of
RNA degradation. This was directly demonstrated in transgenic tobacco expressing an
antisense RBCS gene by measuring decay rates of the sense mRNA after addition of
cordycepin (J iang et al., 1994). In untransformed control plants RBCS mRNA decayed
with a half-life of about 5 hours, versus about 1 hour in plants expressing antisense RBCS
RNA.

Clearly, more detailed study of the mechanism(s) of gene silencing and antisense
RNA effects in plants is required. To date, the power of genetic analysis has not been
fully utilized, although one report has shown that at least two loci of Arabidopsis are
required for silencing of the rolB gene (Dehio and Schell, 1994). Identification of
additional loci and cloning of some of the genes involved should greatly advance our
understanding of this intriguing phenomenon. A more detailed study of the involvement

Of methylation also seems warranted.

ANALYSIS OF mRNA DECAY PATHWAYS

The analysis of mRNA degradation intermediates is one of the most effective
Strategies to uncover steps in particular mRNA decay pathways. As in mammalian cells
and yeast, prominent mRN A decay intermediates have not been detected for most

transcripts in plants. This indicates that once degradation has been initiated, subsequent

Steps are very rapid. However, two plant transcripts that appear to be exceptions to this

 

24

rule are the oat PH YA mRNA and the soybean SRS4 mRNA. Accordingly, detailed
analysis of these transcripts has provided most of our insight into possible mRNA decay
pathways in plants.

When the SRS4 gene is expressed in soybean and transgenic petunia, a series of
short SRS4 transcripts can be observed in addition to the full-length mRNA on RNA gel
blots (Thompson et al., 1992). These short transcripts have been proposed to be bona
fide decay intermediates of SRS4 mRNA for several reasons: (1) they accumulate when
the gene is expressed from the Cauliflower Mosaic Virus 358 promoter, as well as from
the SRS4 promoter (Thompson et al., 1992), (2) the fragments are polysome associated
(Thompson et al., 1992), (3) radioactive tracers added to homogenized samples are not
degraded in vitro during subsequent RNA purification steps (Thompson et al., 1992); and,
most importantly, (4) the same short transcripts can be produced by adding in vitro-
synthesized SRS4 RNA to a cell-free mRN A decay system containing polysomes (Tanzer
and Meagher, 1994) or to an 8150 extract (Tanzer and Meagher, 1995). Mapping the
ends of the short transcripts showed that most proximal SRS4 RNAS with intact 5' ends
Could be matched with a distal SRS4 RNA with an intact 3' end, indicating that each pair
0f products probably arose from endonuclease cleavage of the intact SRS4 mRNA,
perhaps directed by local secondary structure (Tanzer and Meagher, 1995). This
Cndonuclease cleavage appeared to be independent of decapping or of deadenylation of
the transcript. On the basis of these and other data, the model of SRS4 RNA decay shown
in figure 1-3A was proposed (Tanzer and Meagher, 1995). According to this model,
Cleavage by a stochastic endonuclease initiates decay, followed by 3' to 5' or 5' to 3'

exonuclease digestion of the proximal and distal products.

25

 

—‘M(A) n

' exonuclease
. (75%)
stochastic endonucle e deadenylation \
(25%)

8 - MAM) n
$0” n V ’
.— ; ) MAM) n

:-

 

 

 

I

 

exonucleases
exonucleases

@- -3

 

 

 

 

=3
@=-'A
@___._-..W .
A

Figure 1-3. Proposed mechanisms of SRS4 and PH YA mRNA decay in plants. A depicts
Stochastic endonucleolytic decay without prior deadenylation, as proposed for SRS4
ITIRNA (Tanzer and Meagher, 1995). B depicts exonucleolytic decay pathways, as
Proposed for PH YA mRNA (Higgs and Colbert, 1994). The major degradation pathways
for each transcript are indicated by heavy arrows. The 5' cap, the 3' and 5' untranslated
rfbgions, and the coding region of the transcripts are indicated by solid dots, thin lines, and
Open bars, respectively. AAA(A)n represents the poly(A) tail of the mRN A. White
enzymes are stochastic endonucleases, gray and spotted enzymes are exonucleases.

1)mentheses around the 5' cap indicate that the presence or absence of the cap has not
been determined.

26

RNA preparations from plants expressing PH YA genes also yield short RNA
fragments in addition to the full-length transcript. Similar to the case of SRS4, three
observations (Higgs and Colbert, 1994) argue that the short PH YA mRNAs are true
decay intermediates: (1) the PH YA RNA fragments are present in RNA preparations
isolated according to different procedures while other endogenous or added RNAS remain
intact, (2) RNA isolated from a polysome-based in vitro system also contains the PH YA
fragments, and (3) these fragments are associated with polysomes in vivo as well.
Interestingly, the PH YA fragments are continuously distributed over a size range of about
200 nt to 4.2 kb (Higgs and Colbert, 1994), a pattern differing from that of the discrete
sized fragments of the SRS4 transcript discussed above. The structure of the PHYA
fragments was analyzed using different probes. Based on the results, two pathways that
primarily involve exonuclease activities were proposed (Higgs and Colbert, 1994) as
illustrated in figure 1-3B. In this model, the majority (about 75%) of the PHYA mRNA is
degraded by a 5' to 3' exoribonuclease before removal of the poly(A) tail. The remainder
of the mRNA is deadenylated prior to degradation by 5' to 3' and 3' to 5'
exoribonucleases (Higgs and Colbert, 1994). An alternative model involving a stochastic
endoribonuclease cannot be ruled out; however, considering the continuous distribution
0f fragments, this seems less likely (Higgs and Colbert, 1994).

Only the deadenylated fraction of the PH YA mRNA (25%) appears to be a good
Candidate for decay by the major deadenylation-dependent decappin g pathway elucidated
in yeast by Parker and coworkers (reviewed by Tharun and Parker, 1997). The alternative
deadenylation-independent decapping pathway also described in yeast (Muhlrad and

Parker, 1994) could explain the decay of the remainder of PH YA mRNA, but at present it

27

is unknown whether or not decapping is required for degradation of the PH YA transcript.
In contrast, a novel mechanism appears to be responsible for SRS4 mRNA decay. It will
be interesting to see if other plant mRNAs that do not give rise to visible mRNA decay

intermediates degrade by the same pathways.

LINKAGE OF mRN A STABILITY AND TRANSLATION

There are numerous indications that translation and mRNA degradation are
intimately linked processes in viva. This is true for both prokaryotes and eukaryotes, but
prokaryotic and eukaryotic mRNA structure and translation are sufficiently different from
each other to suggest that one can not expect to be able to extrapolate findings from one
to the other. In this section I will highlight some of the more interesting examples of
links between the two processes in eukaryotes. Most of these links are not yet fully

understood, and clearly more work is needed.

The poly(A) tail

Both ends of eukaryotic mRNAs play rate-determining and regulated roles in
mRNA degradation and translation. A major pathway of mRNA decay in yeast is
initiated by shortening of the poly(A) tail, followed by removal of the cap structure from
the 5' end of the transcript (the deadenylation-dependent decapping pathway) (Decker and
Parker, 1994). Although little is known about the pathways of mRNA degradation in

higher eukaryotes, given the conservative nature of evolution (as exemplified in the

28
similarities in transcription, splicing, and polyadenylation mechanisms among all
eukaryotes), it seems plausible that some of the basic mechanisms of mRNA decay are
similar in higher eukaryotes. The limited amount of available data indicate that
shortening of the poly(A) tail is also a critical step in the degradation of some higher
eukaryotic mRNAs. However, it remains to be shown that this is indeed the case for the
bulk of the mRNA. Removal of the poly(A) tail is the first step in the degradation of the
unstable c-fos (Wilson and Treisman, 1988; Shyu et al., 1991) and c-myc (Swartwout and
Kinniburgh, 1989; Laird-Ofringa et al., 1990) mRNAs in mammalian cells. Furthermore,
determinants of poly(A) tail-shortening rates are identical to, or at least overlap with, the
determinants of mRNA instability. Specifically, B-globin reporter transcripts that
normally are stable and slowly deadenylated can be made to deadenylate and degrade
quickly by insertion of the c-fos 3' UTR, a fragment of the c-fos coding region, or the
sequence UUAUUUAUU (Shyu et al., 1991; Zubiaga et al., 1995), suggesting that
removal of the poly(A) tail is a required initial step in the degradation of the c-fos
message and possibly of other mRNAs. As discussed above removal of the poly(A) tail
does not appear to be a critical step in the degradation of SRS4 and PH YA transcripts in
plants.

The poly(A) tail not only is important for regulating mRNA decay but also has an
important role in translation. The presence of a poly(A) tail increases the translation rate
of an mRNA both in rabbit reticulocyte lysates (Doel and Carey, 1976) and in vivo. For
example, RNA synthesized in vitro and introduced into protoplasts by electroporation
was more efficiently translated if a poly(A) tail was added to the 3' end (Gallie et al.,

1989; Gallic, 1991).

29

The exact mechanism by which the poly(A) tail affects translation and mRNA
degradation is not known, but the available evidence suggests that the poly(A) binding
protein (PAB) is important (Caponigro and Parker, 1995; Jacobson, 1996). These data
suggest that the first step in the degradation of an mRNA may inhibit further translation
of the mRNA. It seems likely that more links exist, probably acting through PAB. Plants
express what appear to be multiple PABS in a tissue-specific manner (Belostolsky and
Maegher, 1993). It remains to be determined whether all these proteins function in
regulating both translation and mRN A turnover. It is tempting to speculate that there may
be differences in some of their functions, and that the repertoire of expressed PABS

regulate translation or turnover in a cell type-Specific manner.

The cap structure

The 5' end of eukaryotic mRNAs consists of a modified G-nucleotide linked to the
rest of the transcript by a unique linkage. Recently it has become evident that the second
step of the major mRNA decay pathway in yeast is the removal of this cap structure
(Decker and Parker, 1994). However, the role of decapping in the decay of mRNA in
higher organisms is completely unknown. The cap structure also has a well-documented
role in translation initiation (Sonenberg, 1996). The first step in translation is the binding
of the cap by eIF4F. This is followed by assembly of a large complex, including the 408
ribosomal subunit, onto the 5' end of the mRNA (Merrick and Hershey, 1996). Assembly
of such a large complex may influence the rate of decapping of the transcript by sterically

hindering access to the cap by the decapping enzyme. Conversely, decapping would be

30

predicted to influence translation by preventing assembly of the preinitiation complex.
Clearly more work needs to be done to investigate the importance of decapping in mRN A

decay in plants and how this process interacts with translation.

Inhibition of translation stabilizes many unstable mRNAs

Inhibition of translation can have large effects on the degradation rates of specific
mRNAs. Usually this is studied by inhibiting global translation with chemical inhibitors,
but more refined analysis has been performed in certain cases (see below). The drug most
widely used to inhibit translation is cycloheximide (chx). A large number of unstable
transcripts are stabilized by addition of chx (Jacobson and Peltz, 1996), although there are
also cases where chx has no effect on mRNA stability. This has also been found to be the
case in plants. Taylor and Green (1995) identified unstable mRNAs by differential
hybridization and studied the effect of chx on the disappearance of these transcripts
during actinomycin D treatment. For all eight mRNAs analyzed the disappearance of the
transcripts during actinomycin D treatment was inhibited in the presence of chx.

However it is clear that induction by chx can also be mediated by promoter sequences
(e.g. Gil and Green, 1996), thus it is important to measure mRNA half lives with and
without chx, before concluding that chx influences the stability of a particular transcript.

Besides chx treatment, there are additional ways to inhibit global translation.
There are other drugs that inhibit translation. Some act at the same step as chx, while
others act at different steps in translation. Inhibition of global translation without the use

of drugs can be achieved by using a (conditional) mutation in a factor involved in

31

translation. One example of this is a temperature sensitive mutation in the CCAl gene of
yeast. The CCAl gene encodes tRNA-nucleotidyltransferase, the enzyme that adds the
trinucleotide CCA to the 3' end of tRNAs. The 3' terminal CCA sequence is required for
tRNA function, but it is not encoded by eukaryotic tRNA genes (Deutscher, 1990). In
strains that have a defect in CCAl, the cell is depleted of functional tRNAs and
translation is inhibited. Unstable mRNAs are stabilized in this strain at non-permissive
temperature, presumably because of the defect in translation (Peltz et al., 1992).

It is unclear by what mechanism inhibition of global translation affects mRNA
stability, and it seems likely that there is more than one way in which this may occur.
The two simplest explanations are that either translation of the mRNA in cis is required
for its normal degradation, or that inhibition of translation of some other mRNA (in
trans) depletes the cell of some unstable protein involved in mRNA degradation. In the
trans mechanism, The required protein could be an RNase or other factor important for
degrading RNA. This model would predict that all drugs that inhibit translation would
have the same effect, regardless of what step in translation they affect. Alternatively,
under the cis model, association of the mRNA of interest with ribosomes is important for
its normal degradation. One clear, but possibly unique, example of this mechanism is the
degradation of B-tubulin mRNA in mammalian cells (Cleveland, 1988; Theodorakis and
Cleveland, 1996). The degradation rate of fl-tubulin mRNA is increased when free
tubulin monomers accumulate. Degradation of the B-tubulin mRNA is initiated when the
nascent peptide emerges from the ribosome. The N-terminus of the protein is recognized
specifically, triggering the degradation of the mRNA. If translation of the tubulin mRN A

is inhibited, it is stabilized because it is the mRNA-ribosome complex that is required for

32

recognition of B-tubulin mRNA.

In most other cases it is not clear which of these two models best explains
translation-dependent degradation of a particular transcript. Addressing this question
requires inhibition of translation of only the mRNA of interest. This can be achieved in
several different ways. Translation of a particular mRN A can be inhibited by
introduction of a stable secondary structure in the 5' UTR, by mutation of the start codon,
or by introduction of a nonsense codon. AS discussed below, each of these possibilities
has been used in different studies. However introduction of a nonsense codon often has a
destabilizing effect independent of the normal decay pathway for the specific mRNA (see
the section on nonsense-mediated mRNA decay), so inhibiting translation by insertion of
a nonsense codon may not be informative of the normal pathway of mRNA degradation.
One disadvantage of all of three possible strategies is that they require mutations in the
gene of interest, and thus the gene of interest has to be introduced as a transgene.

A stable secondary structure in the 5' UTR inhibits translation because during the
normal translation initiation process, initiation factors and the 40S ribosomal subunit bind
at the cap structure at the 5' end of the mRNA and subsequently scan the mRNA to find
the start codon. Introduction of a sufficiently stable secondary structure in the 5' UTR
inhibits the scanning of the ribosome and thus prevents it from reaching the start codon
(Kozak, 1989). This strategy has been used in several cases. For example Aharon and
Schneider (1993) introduced a hairpin structure into the 5' UTR of a reporter gene that
also contained an AU-rich destabilizing sequence. This mRNA was inefficiently
translated and was not targeted for rapid degradation. In an elegant control experiment

translation of the mRNA was restored by introducing an internal ribosome entry site of

33

viral origin (IRES) downstream of the hairpin. This restores translation and instability of
the mRNA, which shows that the hairpin in the 5' UTR functioned by inhibiting
translation and not by some other means, for example inhibiting an exonuclease. This is
an important consideration, because secondary structures are known to inhibit a wide
variety of exonucleases in yeast (Vreken and Raue, 1992), Hela cell extracts (Ford et al.,
1997), chloroplasts (Stern and Gruissem, 1987), and E. coli (Mott et al., 1985)

A variation on this scheme is the introduction of an iron responsive element (IRE)
in the 5' UTR, similar to what is naturally present in the human ferritin message. The IRE
is a stem loop structure capable of binding a trans-acting factor, the iron responsive
protein (IRP). The IRE structure by itself is not sufficiently stable to inhibit ribosomal
scanning, but when the IRP is bound translation of the mRNA is inhibited (Raoult et al.,
1996). In mammalian cells, the binding activity of IRP is regulated by the availability of
iron. Thus translation of specific mRNAs can be regulated by iron, and its effect on
mRNA stability can be examined. The IRE/IRP system has been used to investigate the
link between translation and stability of reporter genes destabilized by insertion of AU
rich sequences from c-fos (Koeller et al., 1991; Winstall et al., 1995) and GM-CSF
(Winstall et al., 1995) or to study translation-dependent stability of a modified transferrin
mRNA (Koeller et al., 1991). Koeller et al. (1991) found that IRP binding inhibited
translation of the reporter transcripts as effectively as chx, but that the stability of the
reporter mRNAs was not affected. This suggested that a labile trans-acting factor is
necessary for the rapid degradation of these mRNAs. Interestingly, Winstall et al. (1995)
using the same experimental design, but slightly different reporter genes and a different

mammalian cell line, reached the opposite conclusion. In this case IRP binding led to

34

decreased translation of the reporter transcripts and decreased RNA turnover. This effect
was critically dependent on a functional IRE in the reporter genes; control transcripts that
had a defective IRE because of a one-base deletion were not affected. It is not clear why
these two experiments resulted in opposing conclusions.

Yeast and plants do not have an IRE-binding activity (Rothenberger et al., 1990;
Oliveira et al., 1993). Instead plants appear mainly to use regulation of mRNA
abundance as the major control of ferritin expression (van der Mark, 1983; Lescure et al.,
1991; Gaymard et al., 1996), while yeast does not appear to use ferritin as the major iron
storage form (Raguzzi et al., 1988). However, all the essential components of the
IRE/IRP system have been cloned from mammals and thus can be introduced into yeast
or plants. This was successfully accomplished in yeast by expressing the IRP from a
strong, but tightly regulated promoter and introducing an IRE in a luciferase reporter gene
(Oliveira et al., 1993). This resulted in regulated expression of luciferase, but the system
has not been widely used to address questions related to mRNA stability. A similar
strategy was followed to introduce this system into plants, but luciferase was expressed
approximately equally in the presence or absence of an IRE and/or IRP (van Hoof and
Green, unpublished). It is not clear whether this reflects a difference between the two
organisms, or whether technical problems, for example a failure to produce sufficient
levels of IRP, are responsible for this.

Translation can also be inhibited by mutation of the start codon. To inhibit
translation completely one has to mutate all AUG triplets in all three reading frames,
which for most transcripts means introducing multiple mutations along the transcript.

According to the universally accepted scanning mode] of translation initiation, this would

35

lead to scanning of the 40S ribosome until it reaches the 3' end of the transcript, which is
an unnatural situation. This strategy is not often used because it requires a lot of effort to
generate an unnatural situation. More often only the normal start codon is mutated, while
downstream AUGs are left unchanged. For example, mutating the normal start codon of
the pea fed] gene led to a reduction in the magnitude of regulated stability, indicating that
translation of the internal light regulatory element is important for its function (Dickey et

aL,1994)

Coding region instability determinants

Further indications that translation and mRNA degradation are linked are that
some determinants of mRNA half-life have been found in the coding region, and that
their effects are dependent on their translation. One well-characterized example of a
coding region determinant that links translation and mRNA degradation can be found in
the yeast matal gene. A 65 bp region of the matal gene can destabilize a reporter gene
when it is inserted into the coding region (Parker and Jacobson, 1990; Caponigro et al.,
1993). This region consists of two parts. The first part is rich in rare codons, while the
second part is AU-rich (Parker and Jacobson, 1990). That translation is important for the
function of this determinant was indicated by the fact that the first part can be replaced by
other rare codons (Caponigro et al., 1993). This part of the coding region by itself does
not have destabilizing activity but can stimulate the destabilizing activity of the AU rich
region, probably by pausing the translating ribosomes so that they are in the correct

position to interact with the second part of the coding region (Caponigro et al., 1993;

36

Hennigan and Jacobson, 1996).

Other coding region determinants of mRNA stability have been found in the
mammalian c-myc and c-fos transcripts and in the yeast STE3 and SPOl3 transcripts
(Shyu et al., 1991; Wisdom and Lee, 1991; Heaton et al., 1992; Surosky and Esposito,
1992) but are not as well-defined. Whether any of these coding region determinants
interact directly with ribosomes is unknown, but at least in some cases their translation is
required. Normally stable reporter genes can be destabilized by insertion of c-myc or c-

fos coding region determinants (of 250- 320 bp), but this effect is overcome if translation
of the mRN A is prevented by inserting a stable secondary structure in the 5' UTR
(Schiavi et al., 1994) or by mutating the normal start codon (Wisdom and Lee, 1991).
However, recognition of the c-fos coding region determinant is dependent on the RNA
sequence itself (Wellington et al., 1993), not on the amino acid sequence or the codon
usage (as described above for fl-tubulin and mam] respectively).

In contrast to the coding region determinants that require translation for their
activity, some determinants normally found in the 3' UTR are inactive when inserted into
the coding region. This was demonstrated for the AU-rich region of the 3' UTR of GM-
CSF. This element can destabilize a B—globin reporter transcript when inserted into the 3'
UTR of the reporter gene but does not confer instability when it is inserted in the coding
region of the same reporter gene (Savant-Bhonsale and Cleveland, 1992).

One explanation for the different effects of inserting instability determinants in the
coding region versus the 3' UTR is that to function correctly these elements need to bind
specific proteins, and ribosomes may interfere with the correct assembly of this RNA-

protein complex. Alternatively, passing ribosomes may alter the secondary structure of

37

the mRN A, which may affect the function of instability determinants.

The fed] internal light regulatory element mentioned above (which includes the 5'
UTR and part of the coding region) is the only characterized coding region determinant
from a plant gene. The SA UR-ACI coding region also contains sequences that result in
decreased mRNA abundance. However, transcripts containing the SA UR-A CI coding
region were found to be stable when mRN A half-lives were measured directly (Gil and
Green, 1996). It is not clear what step in gene expression is affected by the SA UR-A CI

coding region.

Nonsense-mediated mRNA decay

In most if not all eukaryotes some mRNAs that contain nonsense mutations are
rapidly degraded. This process is the best-characterized example of a link between
translation and mRNA degradation and has been named nonsense-mediated mRN A decay
or mRNA surveillance. Nonsense-mediated mRNA decay has been studied in yeast
(Peltz et al., 1994), mammals (Maquat, 1995), Xenopus (Whitfield et al., 1994), C.
elegans (Pulak and Anderson, 1993) and plants (see chapters two and three of this thesis).
Similar effects described in E. coli (e.g. Nilson et al., 1987) may be functionally alike to

nonsense-mediated decay, but may operate by a different mechanism.

38

Conserved characteristics of Nonsense-mediated mRNA decay

A number of the main characteristics of nonsense-mediated mRN A decay are
conserved among all eukaryotes. In yeast (Losson and Lacroute, 1979; Peltz et al., 1993),
mammals (Cheng et al., 1990), and plants (see chapter two), nonsense codons in the 5'
part of an mRNA are effective in triggering the decay of the transcript, while nonsense
codons in the 3' part of the same transcript have no effect. In both yeast and mammals
this is dependent at least in part on certain sequences downstream of the premature
nonsense codon, although the sequence requirements are very different in these two
systems (see below).

The available evidence indicates that recognition of nonsense codon-containing
transcripts occurs by the translational machinery. In both yeast and mammals nonsense
codon-containing mRNAs are stabilized by expression of suppressor tRNAs capable of
translating the appropriate nonsense codon, but not by other tRNAs (Losson and
Lacroute, 1979; T akeshita et al., 1984; Gozalbo and Hohmann, 1990; Belgrader et al.,
1993). In addition inhibition of translation by insertion of a strong secondary structure in
the 5' UTR stabilizes mammalian mRNAs containing premature nonsense codons
(Belgrader et a1, 1993).

In both yeast (Leeds et al., 1991; 1992) and nematodes (Pulak and Anderson,
1993), there are Specific trans-acting factors that are required for nonsense-mediated
mRNA decay, and some of these factors are homologous to each other. For example, the
best-characterized factor involved in nonsense-mediated mRN A decay is UPF 1p from

yeast. The smg2 gene product from C. elegans is also involved in nonsense-mediated

39

mRNA decay, and has extensive sequence similarity to UPFlp. Recently cDNAs with
sequence similarity to UPFlp were isolated from mouse and human, and a fusion gene
encoding most of the human protein and N- and C-terrninal parts of the yeast UPF]
protein was able to complement a defect in the yeast gene (Perlick et al., 1996), which
may indicate that this protein functions in nonsense-mediated mRNA decay in mammals.
However, the UPFl gene product has two genetically separable functions. In addition to
its role in nonsense-mediated mRNA decay UPFlp also functions in translation (W eng et
al., 1996a, b). It remains to be shown whether the mammalian homologs also function in
both roles.

As has been discussed above, poly(A) shortening is an important first step in the
degradation of most mRNAs in yeast and of at least some mRNAs in mammalian cells.
The decay of nonsense codon-containing transcripts in these same systems does not
require Shortening of the poly(A) tail. In yeast, nonsense-mediated mRNA decay is
achieved by decapping the fully polyadenylated transcript, followed by a 5' to 3'
exonucleolytic decay of the body of the transcript (Muhlrad and Parker; 1994). The
requirement of poly(A) shortening for nonsense-mediated mRN A decay in mammals was
tested by fusing a nonsense codon containing ,B-globin coding region to the c-fos
promotor. Transient induction of the c-fos promotor by serum addition enabled the
measurement of both mRNA degradation and poly(A) shortening rates. While this
mRNA was very unstable, the poly(A) tail was shortened slowly (Shyu et al., 1991). In
addition, expression of fl-globin genes with inserted nonsense codons in transgenic mice
leads to accumulation of shorter-than-expected transcripts (as well as full length mRNA).

These shorter transcripts appear to be decay intermediates, and are missing varying

40

amounts of the 5' end of the transcript, yet copurify with poly(A)+ RNA (Lim et al., 1989;
1992). This suggests that nonsense-mediated mRNA decay mechanisms may be
conserved between yeast and mammals, and that it involves 5' to 3' exonucleolytic decay

without previous shortening of the poly(A) tail.

Differences is nonsense-mediated mRNA decay in eukaryotes

In contrast to the conserved features of nonsense-mediated mRNA decay
mentioned above, there are certain variations among the process in different eukaryotes.
Two distinctions have been described in some detail. The first difference concerns the
mechanism by which cells discriminate between normal nonsense codons and premature
nonsense codons. To distinguish between the two types of nonsense codons, additional
sequence elements are needed. One of the elements from yeast (TGYYGATGYYYYY)
is fairly well-characterized (Zhang et al., 1995) and appears to function by destabilizing
transcripts if present downstream of the nonsense codon, while having no effect when
located upstream of the nonsense codon. It is clear that yeast has additional cis-acting
sequences, but these have not been well-characterized (Peltz et al., 1993). Mammalian
cells discriminate between premature and wild-type nonsense codons by their position
relative to introns: Nonsense codons upstream of the last intron of the coding region of
TPI, aprt and MUP genes are recognized as being premature nonsense codons and the
transcripts are targeted for rapid degradation (Urlaub et al., 1989; Cheng et al., 1990;
1994; Kessler and Chasin, 1996; Belgrader and Maquat, 1994). In contrast nonsense

codons in the last exon of the coding region are not recognized as being premature, and

41

hence the mRNA is stable. In neither yeast nor mammalian cells is it clear exactly how
these additional sequences function.

The other well-characterized difference in nonsense-mediated decay is the
subcellular location of the pathway. In yeast nonsense-mediated mRN A decay is a
cytoplasmic process, while in mammalian cells at least some nonsense codon-containing
transcripts are degraded while they are still associated with the nucleus. The evidence
supporting these conclusions is discussed in detail in the introduction of chapter three of

this thesis.

SCOPE OF THIS THESIS

From the data described above it seems clear that translation and mRNA
degradation are intimately linked processes. A wide variety of links have been described
in mammalian cells and yeast, but comparatively little is known about these links in
plants. The work described in this thesis is aimed at providing us with a better
understanding of the effects of translation on mRNA stability in plants.

Chapter 2 of this thesis describes the first convincing evidence that plants, like
other eukaryotes, contain a nonsense-mediated mRNA decay pathway. This was
achieved by directly measuring the stability of transcripts with premature nonsense
codons and comparing that with the stability of the wild—type transcript. In addition
chapter 2 describes data that show that nonsense codonS in the 5' part of an mRNA are
effective in triggering mRNA degradation, while nonsense codons further downstream

have no such effect, and that this does not require the presence of introns. Chapter 3

42

describes my attempts to determine the subcellular location of the nonsense-mediated
mRNA decay pathway in plants. The results indicate that this pathway appears to be
cytoplasmic, making plants more Similar to yeast than mammals in this respect.

In contrast to the well documented effects of premature nonsense codons on
mRNA stability, there has been Speculation but very little data on the effects of rare
codons on mRNA stability. Chapter 4 describes an attempt to Show directly that rare
codons can destabilize PHA reporter transcripts, but the resulting data indicate that rare
codons are not sufficient to destabilize this reporter mRN A.

An appendix to this thesis describes the characterization of an unusual set of
transcripts from Arabidopsis and tobacco. The first report of this group of transcripts
came from a screen for genes with unstable transcripts (GUTs) by Taylor and Green
(1995), but subsequently, similar transcripts were found in Arabidopsis, cucumber, and
potato. These transcripts are not typical protein-encoding mRNAs and are unlikely to be
translated. This may be related to the extremely rapid degradation of the transcript in

tobacco.

43

REFERENCES

Abler ML, Green P] (1996): Control of mRNA stability in higher plants. Plant Mol. Biol.
32: 63-78.

Aharon T and Schneider RJ (1993): Selective destabilization of short-lived mRNAs with
the granulocyte-macrophagc colony-stimulating factor AU-rich 3' noncodin g region is
mediated by a cotranslational mechanism. Mol. Cell. Biol. 13: 1971-1980.

Alexander D, Lawton K, Uknes S, Ward E, Ryals J (1994): Defense—related gene
induction in plants. Genetic Engineering 16: 195-212.

Apostel I, Heinstein PF, Low PS (1989): Rapid stimulation of an oxidativc burst during
elicitation of cultured plant cells. Plant Physiol. 90: 109-116.

Baulcombe DC, English JJ (1996): Ectopic pairing of homologous DNA and post
transcriptional gene silencing in transgenic plants. Curr. Opinion Biotechnology 7: 173-
180

Belgrader P, Cheng J, and Maquat LE (1993): Evidence to implicate translation by
ribosomes in the mechanism by which nonsense codons reduce the nuclear level of
human triosephosphate isomcrase mRNA. Proc. Natl. Acad. Sci. USA 90: 482-486.

Belgrader P, Cheng J, Zhou X, Stephenson LS, and Maquat LE (1994): Mammalian
nonsense codons can be cis effectors of nuclear mRNA half-life. Mol. Cell. Biol. 14:
8219-8228.

Belostotsky DA and Meagher RB (1993): Differential organ-specific expression of three
poly(A)-binding protein genes from Arabidopsis thaliana. Proc. Natl. Acad. Sci. USA 90:
6686-6690.

Bourque JE (1995): Antisense strategies for genetic manipulations in plants. Plant Sci
105: 125-149.

Braam J, Davis RW (1990): Rain-, wind-, and touch-induced expression of calmodulin
and calmodulin-related genes in Arabidopsis. Cell 60: 357-364.

Caponigro G, Muhlrad D, and Parker R (1993): A small segment of the MATa'I transcript
promotes mRN A decay in Saccharomyces cerevisiae: A stimulatory role for rare codons.
Mol. Cell. Biol. 13: 5141-5148.

Caponigro G and Parker R (1995): Multiple functions for poly(A)-binding protein in
mRNA decapping and deadenylation in yeast. Genes Dev. 9: 2421-2432.

44

Caput D, Beutlcr B, Hartog K, Thayer R, Brown-Shimer S, Cerami A (1986):
Identification of a common nucleotide sequence in the 3'-untranslated region of mRN A

molecules specifying inflammatory mediators. Proc. Natl. Acad. Sci. USA 83:
1670-1674.

Chaboutc M-E, Chaubct N, Clement B, Gigot C, Philipps G (1988): Polyadcnylation of
histone H3 and H4 mRNAs in dicotyledonous plants. Gene 71: 217-223.

Chasan R (1994): Making sense (suppression) of viral RNA-mediated resistance. Plant
Cell 6: 1329-1331.

Chen CYA, Shyu AB (1995): AU-rich elements: Characterization and importance in
mRNA degradation. Trends Biochem. Sci. 20: 465-470.

Cheng J, Fogel-Petrovic M, and Maquat LE (1990): Translation to near the distal end of

the penultimate exon is required for normal levels of spliced triosephosphate isomerase
mRNA. Mol. Cell. Biol. 10: 5215-5225.

Cheng J, Belgrader P, Zhou X, and Maquat LE (1994): Introns are cis effectors of the
nonsense-codon-mediated reduction in nuclear mRNA abundance. Mol. Cell. Biol. 14:
63 17-6325.

Cleveland DW (1988): Autorcgulated instability of tubulin mRNAs: a novel eukaryotic
regulatory mechanism. Trends Biochem. Sci. 13: 339-343.

Cogoni C, Irelan JT, Schumacher M, Schmidhauser TJ, Sclker EU, Macino G (1996):
Transgene silencing of the al-I gene in vegetative cells of Neurospora is mediated by a

cytoplasmic effector and does not depend on DNA-DNA interactions or DNA
methylation. EMBO J. 15: 3153-3163.

dc Carvalho F, Gheysen G, Kushnir S, Van Montagu M, Inzé D, Castrcsana C (1992):
Suppression of B-l,3-glucanase transgene expression in homozygous plants. EMBO J.
1 1: 2595-2602.

dc Carvalho Niebel F, Frendo P, Inzé D, Cornelissen M, Van Montagu M (1995b):
Co-suppression of B-1,3-glucanase genes in Nicotiana tabacum. Curr. Top. Microbiol.
Irnmunol. 197: 91-103.

de Carvalho Niebel F, Frendo P, Van Montagu M, Cornelissen M (1995a):
Post-transcriptional cosuppression of B-l,3-glucanase genes does not affect accumulation
of transgene nuclear mRNA. The Plant Cell 7: 347-358.

Decker CJ and Parker R (1994): Mechanisms of mRNA degradation in eukaryotes.
Trends Biochem. Sci. 19: 336-340.

45

Deutscher MP (1990): Ribonuclcases, tRN A nucleotidyltransfcrase, and the 3' processing
of tRNA. Prog. Nucleic Acid Res. Mol. Biol. 39: 209-240.

Dehio C, Schell J (1994): Identification of plant genetic loci involved in a
posttranscriptional mechanism for meiotically reversible transgene silencing. Proc. Natl.
Acad. Sci. USA 91: 5538-5542.

Deutch CE, Winicov I (1995): Post-transcriptional regulation of a salt-inducible alfalfa

gene encoding a putative chimeric proline-rich cell wall protein. Plant Mol. Biol. 27:
41 1-418.

Dickey LF, Gallo-Meagher M, Thompson WF (1992): Light regulatory sequences are
located within the 5' portion of the Fed-I message sequence. EMBO J. 11: 2311-2317.

Dickey LF, Nguyen T-T, Allen GC, Thompson WF (1994): Light modulation of
ferredoxin mRNA abundance requires an open reading frame. Plant Cell 6: 1171-1176.

Diehn SH, De Rocher EJ, Green PJ (1996): Problems that can limit the expression of
foreign genes in plants: Lessons to be learned from B. t. toxin genes. Genetic Engineering
18. 83-99.

Doel MT and Carey NH (1976): The translational capacity of deadenylated ovalbumin
messenger RNA. Cell 8: 51-58.

Elliott RC, Dickey LF, White MJ, Thompson WF (1989): Cis-acting elements for light
regulation of pea ferredoxin I gene expression are located within transcribed sequences.
Plant Cell 1: 691-698.

English JJ, Mueller E, Baulcombe DC (1996): Suppression of virus accumulation in
transgenic plants exhibiting silencing of nuclear genes. Plant Cell 8: 179-188.

Flavell RB (1994): Inactivation of gene expression in plants as a consequence of specific
sequence duplication. Proc. Natl. Acad. Sci. USA 91: 3490-3496.

Ford LP, Bagga PS, and Wilusz J (1997): The poly(A) tail inhibits the assembly of a
3'-to-5' exonuclease in an in vitro RNA stability system. Mol. Cell. Biol. 17: 398-406.

Franco AR, Gee MA, Guilfoyle TJ (1990): Induction and superinduction of
auxin-responsive mRNAs with auxin and protein synthesis inhibitors. J. Biol. Chem. 265:
15845-15849.

Gallic DR, Lucas WJ, and Walbot V (1989): Visualizing mRNA expression in plant
protoplasts: factors influencing efficient mRNA uptake and translation. Plant Cell 1:
301-311.

46

Gallic DR (1991): The cap and poly(A) tail function synergistically to regulate mRN A
translational efficiency. Genes Dev. 5: 2108-2116.

Gallic DR (1993): Posttranscriptional regulation of gene expression in plants. Annu. Rev.
Plant Physiol. Plant Mol. Biol. 44: 77-105.

Gaymard F, Boucherez J, Briat JF (1996): Characterization of a ferritin mRNA from
Arabidopsis thaliana accumulated in response to iron through an oxidative pathway
independent of abscisic acid. Biochem. J. 318: 67-73

Gil P, Green PJ (1996): Multiple regions of the Arabidopsis SA UR-ACI gene control
transcript abundance: The 3' untranslated region functions as an mRNA instability
determinant. EMBO J. 15: 1678-1686.

Gil P, Green PJ (1997): Regulatory activity exerted by the SAUR-AC1 promoter region
in transgenic plants. Plant Mol. Biol. in press.

Gil P, Liu Y, Orbovic V, Verkamp E, Poff KL, Green PJ (1994): Characterization of the
auxin-inducible SA UR-ACI gene for use as a molecular genetic tool in Arabidopsis. Plant
Physiol. 104: 777-784.

Gossen M, Bujard H (1992): Tight control of gene expression in mammalian cells by
tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89: 5547-5551.

Gozalbo D and Hohmann S (1990): Nonsense suppressors partially revert the decrease of
the mRNA level of a nonsense mutant allele in yeast. Curr. Genet. 17: 77-79.

Greenberg ME, Belasco JG (1993): Control of the decay of labile protooncogcne and
cytokinc mRNAs. In Belasco JG, Brawerman G (eds): Control of Messenger RNA
Stability. San Diego: Academic Press, pp. 199-218.

Heaton B, Decker C, Muhlrad D, Donahue J, Jacobson A, and Parker R (1992): Analysis
of chimeric mRNAs derived from the STE3 mRNA identifies multiple regions within
yeast mRNAs that modulate mRNA decay. Nucleic Acids Res. 20: 5365-5373.

Hennigan AN and Jacobson A (1996): Functional mapping of the translation-dependent
instability element of yeast MATtxl mRNA. Mol. Cell. Biol. 16: 3833-3843.

Higgs DC, Colbert JT (1994): Oat phytochrome A mRN A degradation appears to occur
via two distinct pathways. Plant Cell 6: 1007-1019.

Hobbs SLA, Kpodar P, DeLong CMO (1990): The effect of T-DNA copy number,
position and methylation on reporter gene expression in tobacco transformants. Plant
Mol. Biol. 15:851-864.

47

Ingelbrecht 1, Van Houdt H, Van Montagu M, Depicker A (1994): Posttranscriptional

silencing of reporter transgenes in tobacco correlates with DNA methylation. Proc. Natl.
Acad. Sci. USA 91: 10502-10506.

Jacobson A (1996): Poly(A) metabolism and translation: The closed loop model. In
Translational Control. M. Hershey, M. Mathews, and N. Sonenberg, eds. (Cold Spring
Harbor Press), Plainview, NY, USA.

Jacobson A and Peltz SW (1996): Interrelationships of the pathways of mRNA decay and
translation in eukaryotic cells. Annu. Rev. Biochem. 65:

Jiang C-Z, Kliebenstcin D, Kc N, Rodermel S (1994): Destabilization of rch sense
transcripts by antisense RNA. Plant Mol. Biol. 25: 569-576.

Kapros T, Robertson AJ, Waterborg JH (1995): Histone H3 transcript stability in alfalfa.
Plant Mol. Biol. 28: 901-914.

Kessler O and Chasin LA (1996): Effect of nonsense mutations on nuclear and
cytoplasmic adenine phosphoribosyltransferase RNA. Mol. Cell. Biol. 16: 4426-4435.

Koeller DM, Horowitz J A, Casey JL, Klausner RD, and Harford JB (1991): Translation
and the stability of mRNAs encoding the transferrin receptor and c-fos. Proc. Natl. Acad.
Sci. USA 88: 7778-7782.

Lagnado CA, Brown CY, Goodall GJ (1994): AUUUA is not sufficient to promote
poly(A) shortening and degradation of an mRNA: The functional sequence within
AU-rich elements may be UUAUUUA(U/A)(U/A). Mol. Cell. Biol. 14: 7984-7995.

Laird-Offringa IA, De Wit CL, Elffcrich P, and Van dcr Eb AJ (1990): Poly(A) tail
shortening is the translation-dependent step in c-myc mRN A degradation. Mol. Cell. Biol.
10: 6132-6140.

Leeds P, Peltz SW, Jacobson A, and Culbertson MR (1991): The product of the yeast
UPF] gene is required for rapid turnover of mRNAs containing a premature translational
termination codon. Genes Dev. 5: 2303-2314.

Leeds P, Wood JM, Lee B-S, and Culbertson MR (1992): Gene products that promote
mRNA turnover in Saccharomyces cerevisiae.. Mol. Cell. Biol. 12: 2165-2177.

Li Y, Hagen G, Guilfoyle TJ (1991): An auxin-responsive promoter is differentially
induced by auxin gradients during tropisms. Plant Cell 3: 1167-1175.

Li Y, Strabala TJ, Hagen G, Guilfoyle TJ (1994): The soybean SAUR open reading frame
contains a cis element responsible for cycloheximide-induced mRNA accumulation. Plant
Mol. Biol. 24: 715-723.

48

Lim S-K, Mullins JJ, Chen C, Gross KW, and Maquat LE (1989): Novel metabolism of
several BO-thalesemic B-globin mRN As in the erythroid tissues of transgenic mice.
EMBO J. 8: 2613-2619.

Lim S-K, Sigmund CD, Gross KW, and Maquat LE (1992): Nonsense codons in human
B-globin mRNA result in the production of mRN A degradation products. Mol. Cell. Biol.
12: 1149-1161.

Lindbo JA, Silva-Rosales L, Proebsting WM, Dougherty WG (1993): Induction of a
highly specific antiviral state in transgenic plants: implications for regulation of gene
expression and virus resistance. Plant Cell 5: 1749-1759.

Losson R and Lacroute F (1979): Interference of nonsense mutations with eukaryotic
messenger RNA stability. Proc. Natl. Acad. Sci. USA 76: 5134-5137.

Maquat LE (1995): When cells stop making sense: Effects of nonsense codons on RNA
metabolism in vertebrate cells. RNA 1: 453-465.

Marzluff WF, Hanson RJ (1993): Degradation of a nonpolyadenylated messenger:
Histone mRNA decay. In Belasco JG, Brawerman G (eds): Control of Messenger RNA
Stability. San Diego: Academic Press, pp. 267-290.

Matzke MA, Matzke AJM (1995): Homology-dependent gene silencing in transgenic
plants: What does it really tell us? TIG 11: 1—3.

McClure BA, Guilfoyle T J (1989): Rapid redistribution of auxin-regulated RNAs during
gravitropism. Science 243: 91-93.

McClure BA, Hagen G, Brown CS, Gee MA, Guilfoyle TJ (1989):Transcription,
organization and sequence of an auxin-regulated gene cluster in soybean. Plant Cell 1:
229-239.

Merrick WC and Hershey JWB (1996): The pathway and mechanism of eukaryotic
protein synthesis. In Translational control. JWB Hershey, MB Mathews, and N.
Sonenberg, eds. (Cold Spring Harbor: Cold Spring Harbor Laboratory Press), pp. 31-70.

Mott JE, Galloway JL, and Platt T (1985): Maturation of escherichia coli tryptophan
operon mRN A: evidence for 3' exonucleolytic processing after rho-dependent
termination. EMBO J. 4: 1887-1891.

Muhlrad D, Parker R (1992): Mutations affecting stability and deadenylation of the yeast
MFA2 transcript. Genes Dev. 6: 2100-2111.

Muhlrad D, Parker R (1994): Premature translational termination triggers mRN A
decapping. Nature 370: 578-581.

49

Napoli C, Lemieux C, J orgensen R (1990): Introduction of a chimeric chalcone synthase

gene into petunia results in reversible co-suppression of homologous genes in trans..
Plant Cell 2: 279-289.

Newman TC, Ohme-Takagi M, Taylor CB, Green P] (1993): DST sequences, highly
conserved among plant SA UR genes, target reporter transcripts for rapid decay in tobacco.
Plant Cell 5: 701-714.

Nilsson G, Belasco JG, Cohen SN, and Von Gabain A (1987): Effect of premature
termination of translation on mRN A stability depends on the site of ribosome release.
Proc. Natl. Acad. Sci. USA 84: 4890-4894.

Ohme-Takagi M, Taylor CB, Newman TC, Green PJ (1993): The effect of sequences
with high AU content on mRNA stability in tobacco. Proc. Natl. Acad. Sci. USA 90:
11811-11815.

Ohtsubo N, Iwabuchi M (1994): The conserved 3'-flanking sequence, AATGGAAATG,
of the wheat histone H3 gene is necessary for the accurate 3'-end formation of mRNA.
Nucleic Acids Res. 22: 1052-1058.

Oliveira CC, Goossen B, Zanchin NI, McCarthy JEG, Hentze MW, and Stripecke R
(1993): Translational repression by the human iron-responsive factor (IRF) in
saccharomyces cerevisiae.. Nucleic Acids Res. 21: 5316-5322.

Parker R and Jacobson A (1990): Translation and a 42-nucleotide segment within the
coding region of the mRNA encoded by the MATa'I gene are involved in promoting rapid
mRNA decay in yeast. Proc. Natl. Acad. Sci. USA 87: 2780-2784.

Peltz SW, Donahue JL, and Jacobson A (1992): A mutation in the tRNA
nucleotidyltransferase gene promotes stabilization of mRNAs in Saccharomyces
cerevisiae. Mol. Cell. Biol. 12: 5778-5784.

Peltz SW, Brown AH, and Jacobson A (1993): mRNA destabilization triggered by
premature translational termination depends on at least three cis-acting sequence elements
and one trans—acting factor. Genes Dev. 7: 1737-1754.

Peltz SW, He F, Welch E, and Jacobson A (1994): Nonsense-mediated mRNA decay in
yeast. Prog. Nucleic Acid Res. Mol. Biol. 47: 271-298.

Perlak FJ, Deaton RW, Armstrong TA, Fuchs RL, Sims SR, Greenplate JT, Fischhoff DA
(1990): Insect resistant cotton plants. Bio/Technology 8: 939-943.

Perlak FJ, Fuchs RL, Dean DA, McPherson SL, Fischhoff DA (1991): Modification of
the coding sequence enhances plant expression of insect control protein genes. Proc. N at].
Acad. Sci. USA 88: 3324-3328.

50

Perlick HA, Medghalchi SM, Spencer FA, Kendzior R], Jr, and Dietz HC (1996):
Mammalian orthologues of a yeast regulator of nonsense transcript stability. Proc. Natl.
Acad. Sci. USA 93: 10928-10932.

Pulak R and Anderson P (1993): mRNA surveillance by the Caenorhabditis elegans smg
genes. Genes Dev. 7: 1885-1897.

Raguzzi F, Lesuisse E, and Crichton RR (1988): Iron storage in Sacharomyces cerevisiae.
FEBS Lett. 231: 253-258.

Rothenberger S, Mullner EW, and Kuhn LC (1990): The mRNA-binding protein which
controls ferritin and transferrin receptor expression is conserved during evolution.
Nucleic Acids Res. 18: 1175-1179.

Rouault TA, Klausner RD, and Harford JB (1996): Translational control of ferritin. In
Translational control. JWB Hershey, MB Mathews, and N Sonenberg, eds. (Cold Spring
Harbor: Cold Spring Harbor Laboratory Press), pp. 335-362.

Savant-Bhonsale S, Cleveland DW (1992): Evidence for instability of mRN As containing
AUUUA motifs mediated through translation-dependent assembly of a >208 degradation
complex. Genes Dev. 6: 1927-1939.

Schiavi SC, Wellington CL, Shyu A-B, Chen C-YA, Greenberg ME, and Belasco JG
(1994): Multiple elements in the c-fos protein-coding region facilitate mRNA

deadenylation and decay by a mechanism coupled to translation. J. Biol. Chem. 269:
3441-3448.

Schiebel W, Haas B, Marinkovic S, Klanner A, Sanger HL (1993a): RNA-directed RNA
Polymerase from tomato leaves: I Purification and physical properties. J. Biol. Chem.
268: 11851-11857.

Schiebel W, Haas B, Marinkovic S, Klanner A, Sanger HL (1993b): RNA-directed RNA
polymerase from tomato, leaves: 11 Catalytic in vitro properties. J. Biol. Chem. 263:
11858-11867.

Shyu A—B, Belasco JG, and Greenberg ME ( 1991): Two distinct destabilizing elements in

the c-fos message trigger deadenylation as a first step in rapid mRNA decay. Genes Dev.
5: 221-231.

Seeley KA, Byme DH, Colbert JT (1992): Red light-independent instability of oat
phytochrome mRNA in vivo. Plant Cell 4: 29-38.

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

51

Sheng J, D'Ovidio R, Mehdy MC (1991): Negative and positive regulation of a novel
proline-rich protein mRNA by fungal elicitor and wounding. Plant J. 1: 345-354.

Sheu J -J , Jan S-P, Lee H-T, Yu S-M (1994): Control of transcription and mRNA turnover
as mechanisms of metabolic repression of a-amylase gene expression. Plant J. 5:
655-664.

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

Sonenberg N (1996): mRNA 5' cap-binding protein eIF4E and control of cell growth. In
Translational control. JWB Hershey, MB Mathews, and N Sonenberg, eds. (Cold Spring
Harbor: Cold Spring Harbor Laboratory Press), pp. 245-270.

Stern DB and Gruissem W (1987): Control of plastid gene expression: 3' inverted repeats
act as mRN A processing and stabilizing elements, but do not terminate transcription. Cell
51: 1145-1157.

Sullivan ML, Green P] (1993): Post-transcriptional regulation of nuclear-encoded genes
in higher plants: The roles of mRNA stability and translation. Plant Mol. Biol. 23:
1091-1104.

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

Surosky RT and Esposito RE (1992): Early meiotic transcripts are highly unstable in
saccharomyces cerevisiae.. Mol. Cell. Biol. 12: 3948-3958.

Swartwout SG and Kinniburgh AJ (1989): c-myc RNA degradation in growing and
differentiating cells: Possible alternate pathways. Mol. Cell. Biol. 9: 288-295.

Takeshita K, Forget BG, Scarpa A, and Benz E] Jr. (1984): Intranuclear defect in
beta-globin mRN A accumulation due to a premature termination codon. Blood 64: 13—22.

Tanzer MM, Meagher RB (1994): Faithful degradation of soybean rch mRNA in vitro.
Mol. Cell. Biol. 14: 2640-2650.

Tanzer MM, Meagher RB (1995): Degradation of the soybean ribulose-l,5-bisphosphate

carboxylase small-subunit mRN A, SRS4, initiates with endonucleolytic cleavage. Mol.
Cell. Biol. 15: 6641-6652.

Taylor CB, Green P] (1995): Identification and characterization of genes with unstable
transcripts (GUTs) in tobacco. Plant Mol. Biol. 28: 27-38.

52

Tharun S, Parker R (1997): Mechanisms of mRN A turnover in eukaryotic cells. In
mRNA metabolism and post-transcriptional gene regulation. Modern Cell Biol. 17: 181-
199 Morris DR, Harford JB , eds. (New York: Wiley and Sons),

Theodorakis NG and Cleveland DW (1996): Translationally coupled degradation of
mRNA in eukaryotes. In Translational control. JWB Hershey, MB Mathews, and N.

Sonenberg, eds. (Cold Spring Harbor: Cold Spring Harbor Laboratory Press), pp.
63 1 -65 2.

Thompson DM, Tanzer MM, Meagher RB (1992): Degradation products of the mRN A
encoding the small subunit of ribulose-1,5-bisphosphate carboxylase in soybean and
transgenic petunia. Plant Cell 4: 47-58.

Thompson WF, White MJ (1991): Physiological and molecular studies of light-regulated
nuclear genes in higher plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 42: 423-466.

van Blokland R, Van der Geest N, Mol JNM, Kooter J M (1994): Transgene-mediated
suppression of chalcone synthase expression in Petunia hybrida results from an increase
in RNA turnover. Plant J. 6: 861-877.

van der Krol AR, Mur LA, Beld M, Mol J NM, Stuitje AR (1990): Flavonoid genes in
petunia: Addition of a limited number of gene copies may lead to a suppression of gene
expression. Plant Cell 2: 291-299.

van der Mark F, Bienfait F, and van den Ende H (1983): Variable amounts of translatable
ferritin mRNA in bean leaves with various iron content. Biochem. Biophys. Res.

Commun. 1 15: 463—469.

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

Vakalopoulou E, Schaack J, Shenk T (1991): A 32-kilodalton protein binds to AU-rich
domains in the 3' untranslated regions of rapidly degraded mRNAs. Mol. Cell. Biol. 11:
3355-3364.

Vreken P and Raué HA (1992): The rate-limiting step in yeast PGKI mRNA degradation
is an endonucleolytic cleavage in the 3'-termina1 part of the coding region. Mol. Cell.
Biol. 12: 2986-2996.

Walker EL, Weeden NF, Taylor CB, Green PJ, Coruzzi GM (1995): Molecular evolution
of duplicate copies of genes encoding cytosolic glutamine synthetase in Pisum sativum.
Plant Mol. Biol. 29: 1111-1125.

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

53

Wellington CL, Greenberg ME, and Belasco JG (1993): The destabilizing elements in the
coding region of c-fos mRNA are recognized as RNA. Mol. Cell. Biol. 13: 5034-5042.

Weng YM, Czaplinski K, and Peltz SW (1996a): Genetic and biochemical
characterization of mutations in the ATPase and helicase regions of the Upfl protein.
Mol. Cell. Biol. 16: 5477-5490.

Weng YM, Czaplinski K, and Peltz SW (1996b): Identification and characterization of
mutations in the UPF] gene that affect nonsense suppression and the formation of the
Upf protein complex but not mRNA turnover. Mol. Cell. Biol. 16: 5491-5506.

Weng Y, Ruiz-Echevarria MJ, Zhang S, Cui Y, Czaplinski K, Dinman JD, and Peltz SW
(1997): Characterization of the nonsense-mediated mRN A decay pathway and its effect
on modulating translation termination and programmed frameshifting. In mRN A
metabolism and post-transcriptional gene regulation. Modern Cell Biol. 17: 241-263
Morris DR, Harford JB , eds. (New York: Wiley and Sons),

Whitfield TT, Sharpe CR, and Wylie CC (1994): Nonsense-mediated mRNA decay in
Xenopus oocytes and embryos. Dev. Biol. 165: 731-734.

Wilson T and Treisman R (1988): Removal of poly(A) and consequent degradation of
c-fos mRNA facilitated by 3' AU-rich sequences. Nature 336: 396-399.

Winstall E, Gamache M, and Raymond V (1995): Rapid mRNA degradation mediated by
the c-fos 3' AU-rich element and that mediated by the granulocyte-macrophage

colony-stimulating factor 3' AU-rich element occur through similar polysome-associated
mechanisms. Mol. Cell. Biol. 15: 3796-3804.

Wisdom R and Lee W (1991): The protein-coding region of c-myc mRN A contains a
sequence that specifies rapid mRNA turnover and induction by protein synthesis
inhibitors. Genes Dev. 5: 232-243.

Yamamoto KT, Mori H, Irnaseki H (1992): cDNA cloning of indole-3-acetic
acid-regulated genes: Aux22 and SAUR from mung bean (Vigna radiata) hypocotyl
tissue. Plant Cell Physiol. 33: 93-97.

Yu S-M, Kuo Y-H, Sheu G, Sheu J -J , Liu LF (1991): Metabolic derepression of
a-amylase gene expression in suspension-cultured cells of rice. J. Biol. Chem. 266:
21131-21137.

Zhang S, Sheng J, Liu Y, Mehdy MC (1993): Fungal elicitor-induced bean proline-rich
protein mRN A down-regulation is due to destabilization that is transcription and
translation dependent. Plant Cell 5: 1089-1099.

54

Zhang S, Mehdy MC (1994): Binding of a 50-kD protein to a U-rich sequence in an

mRNA encoding a proline-rich protein that is destabilized by fungal elicitor. Plant Cell 6:
135-145.

Zhang S, Ruiz-Echevarria MJ, Quan Y, and Peltz SW (1995): Identification and

characterization of a sequence motif involved in nonsense-mediated mRN A decay. Mol.
Cell. Biol. 15: 2231-2244.

Zubiaga AM, Belasco JG, Greenberg ME (1995): The nonamer UUAUUUAUU is the
key AU-rich sequence motif that mediates mRNA degradation. Mol. Cell. Biol. 15 :
2219-2230.

CHAPTER 2

PREMATURE NON SENSE CODONS DECREASE THE STABILITY OF

PHYTOI—IEMAGGLUTININ mRN A IN A POSITION-DEPENDENT MANNER

This Chapter was originally published in The Plant Journal
Reference: A van Hoof and PJ Green (1996) Premature nonsense codons decrease the
stability of phytohemagglutinin mRNA in a position-dependent manner. Plant J. 10:415-

424.

55

56

ABSTRACT

Premature termination of translation has often been associated with decreased
mRN A accumulation in plants, but the affected step in gene expression has not been
identified. To investigate this problem, I have examined the expression of wild-type and
mutant alleles of the bean phytohemagglutinin (PHA) gene in tobacco cells and transgenic
plants. Measurement of mRN A decay rates in stably transformed cell lines demonstrated
that premature nonsense codons markedly destabilized the mRNA. This decreased
stability was also reflected by decreased accumulation of transcripts containing premature
nonsense codons in transgenic plants. The positional dependence of the nonsense codon
effect was evaluated by introducing premature nonsense codons at different distances
from the PHA AUG start codon. Transcripts with nonsense codons about 20%, 40% or
60% of the way through the normal PHA coding region yielded highly unstable mRNAs,
whereas a transcript with a nonsense codon at 80% was as stable as wild-type. The ability
to recognize and rapidly degrade certain transcripts with early nonsense codons could
provide plant cells with a means to minimize the production of wasteful and possibly

deleterious truncated proteins.

57

INTRODUCTION

Eukaryotic gene expression is a highly regulated process that involves many steps.
A majority of the studies on gene expression have focused on transcription and the
regulation thereof. More recently it has become clear that control of gene expression can
also result from regulation exerted at later steps, such as splicing (Smith et al., 1989;
Luehrsen et a1 ., 1994), mRN A degradation (Sullivan and Green, 1993; Surdej et al.,
1994), or translation (Altmann and Trachsel, 1993; Hershey, 1991). Considerable
evidence suggests that these steps are not independent, but are influenced by each other
(e.g. Urlaub etal., 1989; Dickey et al., 1994). In particular, it appears that translation and
mRNA degradation can be tightly coupled (Peltz et al., 1991; Sullivan and Green, 1993).
For some transcripts, translation is necessary for normal rates of mRN A degradation
(Cleveland, 1988; Edwards and Mahadevan, 1992; Aharon and Schneider, 1993),
whereas for others interruption of translation leads to increased degradation rates (Peltz et
al., 1992; Losson and Lacroute, 1979).

In plants, most of the data indicating a linkage between translation and mRNA
stability derive from the effects of mutations creating premature nonsense codons. Several
studies indicate that these mutations lead to decreased transcript accumulation. In one
case it was found that a natural allele of a Kunitz trypsin inhibitor gene (Kti3) of soybean
caused low accumulation of mRNA. This allele contained a point mutation about 70%
into the coding region that introduced a premature nonsense codon (J ofuku et al., 1989).
Transcription rates of the wild-type and mutant genes, as measured by nuclear run-on

analysis, were not significantly different. This suggested that the effect was post-

58

transcriptional, possibly at the level of mRNA stability (J ofuku et al., 1989). In an other
case it was found that a natural allele of the phytohemagglutinin (PHA) gene from
common bean (Phaseolus vulgaris) led to greatly reduced levels of PHA mRN A (Voelker
et al., 1986). Sequencing showed that this allele contained a frame-shift mutation that led
to a premature nonsense codon in the new reading frame about 20% of the way through
the PHA coding region. Replacement of the wild-type coding region with that from the
frame—shift allele decreased expression in transgenic plants, and correction of the
mutation restored wild-type expression levels (Voelker et al., 1990). From these studies it
can be concluded that the frame-shift mutation decreased mRN A levels, but the affected
step in gene expression was not identified. It appears that a premature nonsense codon
may also decrease mRN A accumulation for a chimeric patatin gene in transgenic tobacco
(Vancanneyt et al., 1990). An additional example of premature nonsense codons
interfering with normal transcript accumulation has been found for the pea ferredoxin
gene (FEDI) which is post-transcriptionally regulated by light (Dickey et a1. 1994). The
cis-acting sequences regulating the response to light partially overlap with the coding
region. When nonsense codons were introduced into the FED] coding region, the
accumulation of the mRN A failed to respond to light (Dickey et al., 1994). If light
stabilizes the FED] mRNA as has been proposed (Dickey et al. 1992), then the
introduction of nonsense codons could interfere with this effect. Alternatively, nonsense
codons may destabilize or otherwise affect the abundance of the FEDI message by a
pathway that is independent of light. Measurements of mRNA decay rates will be
required to resolve among these possibilities. Finally, recent advances in map-based

cloning have led to the molecular characterization of a number of mutant alleles from

59

Arabidopsis thaliana and other plants. In several cases, a correlation between the '
presence of nonsense or frame-shift mutations and reduced mRNA levels can be seen
(e. g. Dehesh et al., 1993; Reed et al., 1993). This suggests that premature nonsense
codons may decrease mRNA accumulation when present in a wide variety of plant genes.

Premature termination of translation also has a pronounced effect on post-
transcriptional mRNA metabolism in Saccharomyces cerevisiae (yeast). The effect was
shown to occur at the level of mRNA stability by measuring mRNA decay rates after
transcriptional inhibition, or by estimating mRN A decay rates by approach to steady state
labeling (for a discussion of this method, see Ross, 1995). Prominent examples include
transcripts of the yeast genes URA3 (Losson and Lacroute, 1979), URAl (Pelsy and
Lacroute, 1984), LEU2 and HIS4 (Leeds etal., 1991), and PGKl (Peltz et al., 1993), all
of which were shown to be destabilized by premature nonsense codons. Molecular
genetic studies in yeast and Caenorhabditis elegans indicate that several gene products
participate in nonsense-mediated mRN A decay (Leeds et al., 1991; Pulak and Anderson,
1993).

Examples of reduced mRNA abundance caused by premature termination of
translation in mammalian systems have also been reported. Similar to the situation in
yeast, premature nonsense codons decreased the stability of the human (Maquat et al.,
1981) and rabbit B-globin mRNAs (Shyu et al., 1991), and the Rous sarcoma virus gag
pre-mRNA (Barker and Beeman, 1991), when mRNA half-lives were monitored in cells
following transcriptional inhibition. However, premature nonsense codons that decrease
mRNA levels are not always associated with decreased mRNA stability in similar

experiments. Messenger RNA decay measurements made following transcriptional

60

inhibition, were not indicative of accelerated mRNA decay for hamster dihydrofolate
reductase (U rlaub et al., 1989), human triosephosphate isomerase (Cheng et al., 1990), v-
src of avian sarcoma virus (Simpson and Stoltzfus, 1994), or minute virus of mice NS2
(N aeger et al., 1992) mRNAs that contained premature nonsense codons. The exact
mechanism(s) by which nonsense codons decrease mRNA abundance in these cases,
(some of which depend on the presence of introns), is not known (Maquat, 1995).

Our knowledge of the effects of premature nonsense codons in plants is even more
limited because, prior to this report, rates of mRNA degradation were never measured for
any of the plant examples. The goal of this study was to use the PHA gene as a model to
better understand the decreased mRNA accumulation caused by the presence of
premature nonsense codons in certain plant transcripts. To this end, I investigated the
expression of a set of natural and in vitro generated PHA alleles containing premature
nonsense codons in cultured tobacco cells and transgenic plants. These studies
demonstrate that premature nonsense codons decrease PHA mRNA accumulation at the
level of mRNA stability, and therefore indicate that a nonsense-mediated mRNA decay
pathway exists in higher plants. I also examined how far into the PHA coding region a

nonsense codon can be positioned before it no longer causes rapid mRNA decay.

61

RESULTS

Premature nonsense codons lead to decreased mRNA accumulation in protoplasts.

The PHA gene system from common bean has several features that make it
attractive for the study of premature nonsense codons. First, the wild-type (WT) allele
and a natural frame-shift (FS) allele of PHA have been well characterized. As illustrated
in figure 2-1, a frame-shift mutation in codon 11, which creates a nonsense codon at
codon 53, is the only difference between the FS and WT alleles (Voelker et al., 1990).
Second, the decreased mRNA accumulation attributable to the FS allele in bean can be
recapitulated in tobacco (Voelker et al., 1990), which is easy to transform and regenerate.
Finally The PHA structure is relatively simple. It contains a 5'UTR of 10-15 bp (Hoffman
and Donaldson, 1985), a coding region of 825 bp that has a typical codon usage and is not
interrupted by introns (Hoffman and Donaldson, 1985 and data not shown) and a 3'UTR
of 126-132 bp (as determined by RT-PCR of the 3'UTR, data not shown).

It seemed likely that the decreased abundance of the FS PHA mRNA was caused
by the premature nonsense codon, but alternative effects of the FS mutation could not be
excluded. To resolve this issue, I constructed a third allele, designated STOP, by
introducing a nonsense codon via site-directed mutagenesis. The STOP mutation was
derived from WT by making a two base substitution in codon 53 to create a nonsense
codon (Figure 2-1). Thus, the only common characteristic differentiating FS and STOP
from WT is a premature nonsense codon at codon 53. Previous studies comparing PHA

alleles in tobacco were carried out in seeds where the PHA promoter is preferentially

62

#—
LCaMV 353 I PHA I PHA 3' I

 

 

 

 

 

 

 

STOP I 52 aa

 

FIGURE 2-1. PHA alleles introduced into tobacco cells and transgenic plants. Initially
three different PHA alleles (WT, FS and STOP) were constructed. Each allele contained
the CaMV 35$ promoter (CaMV 35S), a PHA coding region (PHA) and 1.4 kb of 3'
flanking sequence (PHA3'). The transcribed region is indicated by an arrow. A 1-bp
deletion in the frame-shift allele (FS) is indicated by AC. This deletion leads to a frame-
shifted reading frame, indicated by the shaded box and a premature nonsense codon in the
53rd codon. A control allele (STOP) was constructed that contains a premature nonsense
codon in the same position as the FS, without the frame-shift mutation.

63

expressed (Voelker et al., 1990). To facilitate analyses using tobacco leaf tissue and
cultured tobacco cells, the PHA promoter was replaced by the 358 promoter in alleles
used for this study.

Previous work has shown that the established tobacco cell line BY-2 (also called
NTl) provides a useful system for mRNA stability measurements (Newman et al., 1993;
Ohme—Takagi et al., 1993). Therefore, in preparation for such analyses each of the PHA
alleles was initially examined for expression levels following electroporation into BY—2
protoplasts. Total RNA was isolated 14 hr after electroporation and analyzed by RNA gel
blotting. Figure 2-2a shows that the mRNA from WT PHA accumulated to a substantial
level, while the mRNA from FS or STOP PHA was barely detectable. Quantitation of the
transcript levels for the different PHA alleles relative to that of a GUS gene, co-
electroporated as an internal standard is represented in Figure 2—2b. These data show that
the abundance of the PHA FS and STOP transcripts was reduced approximately fivefold
as compared to WT PHA mRN A. The faint extra band that can be seen above the PHA
mRNA in Figure 2-2a was observed whenever PHA was expressed, but is more
prominent after the long exposures needed to visualize destabilized versions of the PHA
mRNA (e.g. Figure 2-3b). The band was enriched in poly A‘” fractions, was DNase I
resistant, hybridized to a strand-specific probe, was not observed with other 358 promoter
constructs, but was observed when another promoter (from the wheat cabl gene) was
used to control PHA (data not shown). Therefore, this band most likely represents
polyadenylation at a site far downstream of the major polyadenylation site. Because the
FS and STOP alleles cause decreased mRNA accumulation in the absence of the PHA

promoter, it is likely that the effect is post-transcriptional. The decreased transcript

64

'WTFSSTOP'

PHAF

 

 

PHA Allele

FIGURE 2-2. Decreased expression of the FS and STOP transcripts in tobacco
protoplasts. (a). PHA mRNA accumulation. Tobacco BY-2 protoplasts were
electroporated with Bluescript derivatives containing the WT, FS, and STOP PHA alleles
described in Figure 2-1a. A Bluescript derivative containing a 35S-GUS gene was co-
electroporated with each PHA allele to serve as an internal standard. RNA was isolated 14
hr after electroporation and 20ug was analyzed on RNA gel blots. (-) represents control
electroporations without PHA DNA.(b). Relative abundance of PHA transcripts. RNA
levels of three independent experiments were quantified using a PhosphorImager. The
PHA transcript level for the different alleles was standardized to that of the GUS
transcript and the relative transcript accumulation for WT was set to 1. Shown is the
mean relative abundance of PHA. Error bars indicate the standard error (SE).

65

accumulation associated with the FS mutation in BY-2 cells is analogous to the effects of
the FS mutation originally observed in seeds of bean and tobacco (Voelker et al., 1986;
1990). This indicates that the BY-2 system will be a valid model for further study of

premature IlOl'lSEIlSC codons.

Premature nonsense codons cause rapid degradation of mRN A

To test whether the premature nonsense codons exert their effect at the level of
mRNA stability, the WT, FS and STOP alleles were stably introduced into BY-2 cells
using Agrobacterium-mediated transformation. Messenger RNA decay rates Were
measured in individual transformed cell lines that were grown in liquid culture.
Following treatment with actinomycin D to stop transcription, total RNA was isolated
from cells harvested at regular time intervals and analyzed by RNA gel blotting.
Representative experiments measuring the decay of the WT and FS mRNAs are shown in
figure 2-3. WT PHA mRNA was quite stable (Figure 2-3a and 2-3c), whereas the FS
mRNA disappeared rapidly (Figure 2-3b and 2-3c; note the different time scale between
Figures 2-3a and 2-3b). These mRNA half-life measurements were repeated at least four
times on at least two independent transformants for each of the three PHA alleles. The
mean half-lives of the PHA transcripts as well as the standard error are represented in
Figure 2-4. The half-lives of both the FS and STOP transcripts were reduced about 3-fold
relative to the WT (36:6 and 4314 min respectively as compared to 129:14 min),
showing that premature nonsense codons can cause rapid degradation of the PHA mRNA

in tobacco.

66

a d—>12rr

 

mRNA

Remaining

 

 

 

   

FIGURE 2-3. Destabilization of the PHA mRNA by the FS mutation in stably
transformed tobacco cells. (a). RNA gel blot monitoring the decay of WT PHA mRNA.
Tobacco BY-2 cells were stably transformed with a pMONSOS derivative containing the
WT PHA allele. A cell line that expressed the PHA gene was treated with actinomycin D
(ActD) to stop transcription and RNA was isolated at 30 min intervals for 2 hr thereafter.
The gel blot contained 20p g total RNA per lane and was probed for the PHA transcript.
(b). RNA gel blot monitoring the decay of the FS PHA mRNA. As in a, except that a cell
line expressing the FS allele of PHA was used and RNA was isolated from samples
collected at 15 min intervals after ActD treatment for a period of 1 hr. (c). Graphic
representation of mRN A decay rates. Signals from the blots shown in a and b were
quantified using a PhosphorImager, normalized to the zero time point, and subjected to
linear regression analysis to calculate mRNA half-lives. Boxes represent the WT
transcript from a. Circles represent the FS transcript from b.

67

   

100-

PHA mRNA
Half-life
(min)

50-

 

WT FS STOP

FIGURE 2-4. Effects of the FS and STOP mutations on PHA mRNA stability in tobacco.
Messenger RNA half-life measurements were made as shown in Figure 2-3 from at least
four independent experiments carried out using at least two independent tobacco BY-2
cell lines per construct. Mean PHA mRNA half-lives for the individual constructs were
calculated. Error bars represent the SE.

68

Premature nonsense codons lead to decreased mRN A accumulation in transgenic plants

To study the effect of premature nonsense codons in transgenic plants, the WT,
FS, and STOP alleles were introduced into tobacco. Nine to ten transgenic tobacco plants
expressing PHA were regenerated for each construct. RNA was isolated from an
individual leaf from each plant and analyzed by RNA gel blotting. PHA mRNA levels
were quantitated and are shown in figure 2-5a. Although there is considerable variation in
expression levels between individual transformants (presumably due to position effects),
statistical analysis shows that premature nonsense codons significantly decrease the
accumulation of PHA mRNA. Compared to WT, the average mRN A level of FS and
STOP was reduced 5 to 10-fold and this difference is highly significant (P>0.975),
according to the test of Wilcoxon-Mann—Whitney (Nap et al., 1993). As expected there
was no significant difference between mRN A levels for the FS and STOP alleles.
Decreased mRNA accumulation was also clearly evident when equal amounts of RNA
from F8 and STOP plants were pooled and analyzed in the RNA gel blot in figure 2-5c.
As a further confirmation the WT and STOP constructs were introduced into Arabidopsis
and mRNA levels were again compared in individual transgenic plants and pooled RNA
samples. The results were similar to those obtained with tobacco as shown in Figure 2-5b
and d. The average mRNA accumulation for the FS construct was reduced 5-fold, and
this difference was statistically significant (P>0.999). Based on these results, it is likely
that premature termination of PHA mRNA translation leads to increased mRNA

degradation rates in transgenic tobacco and Arabidopsis plants, as it does in tobacco cells.

69

 

 

 

 

 

 

 

 

a 3 C E
I I 8
_ - o
8 a
1 06 I '2- 1 06 I O o
PHA mRNA : o 8 PHA mRNA 2
Abundance — Abundance - 0 9
(arbitrary unlts) — o (arbitrary units) - '8
_ _ o
O 8 'O' O
105 E 2 g 106 E
E (g, o E
I o I
o _ o
4
1O 1
WT FS STOP 04 wr STOP

 

|:§|
§
§
§
3'.
3c
S.
a

| NA lo.975 | 0.990 I P NA 0.999

 

 

- wr FS STOP - WT STOP

PHA >

   

FIGURE 2-5. Reduced accumulation of the FS and STOP transcripts in transgenic plants.
(a). Relative abundance of different PHA transcripts in transgenic plants. RNA levels of
at least nine independent transgenic plants per construct were quantified using a
PhosphorImager. Points represent the data from individual plants. The average for each
construct is indicated by a black bar. Statistical analyses using the Wilcoxon-Mann-
Whitney test are presented in the table below the plot. NA indicates not applicable and P
indicates the probability that the expression level is different from that of WT. (b). Same
as a except that RNA from at least 1 1 independent Arabidopsis transformants was used.
(c). Equal amounts of the RNA from each plant were pooled and 10p g was used for RNA
gel blot analysis. Hybridization was as in Figure 2-3. (-) represents control RNA from an
untransformed tobacco plant. The band above the PHA transcripts is the same as that in
Figures 2-2a and 2-3b and probably represents polyadenylation at a site downstream of
the normal site, as explained in the text. (d). Same as c except that Arabidopsis RNA was
used.

‘70

The effect of premature nonsense codons is position-dependent

In all of the aforementioned experiments, the premature nonsense codon was
located at the same position, 20% of the way through the PHA coding region. However it
is unclear whether premature termination of PHA translation further downstream could
also decrease mRNA stability. To address this question three additional alleles were
made, each containing a premature nonsense codon about 40%, 60%, or 80% of the way
through the normal PHA coding region as shown in Figure 2-6a. Each of these PHA
alleles was stably transformed into BY-2 cells and mRN A decay was monitored as
described above for the FS, STOP (20%), and WT constructs. Multiple half-life
measurements were made on two to four independent cell lines for the 60 and 80%
constructs and one cell line for the 40% construct. The half-lives of transcripts with a
nonsense codon at 40% (35:4 min) or 60% (38:3 min) of the way through the coding
region were indistinguishable from those of the unstable FS and STOP transcripts (Figure
2—6b). In contrast, the transcript with a nonsense codon at 80% was at least as stable
(161:18 min) as WT PHA mRNA. These data indicate that nonsense codons up to 60%
of the way through the coding region can trigger the nonsense-mediated mRNA decay of
PHA mRNA in tobacco, but the machinery apparently fails to discriminate between the

80% nonsense codon and the normal nonsense codon.

71

 

 

 

 

 

a we Stop b
l 1
W I 27500 I
“is ”3" 15°
$109 on
AUG 5109
l
40% m
we SEEP 100
m
PHA RNA
MIG snip HEN-lite
l
30% I7 21703 I— (m)
50 .

 

 

0—
STOP 40% 60% 00%
(20%)

Figure 2-6. Premature nonsense codons in the first 60% of the PHA coding region cause
mRNA instability. a. Introduction of nonsense codons at different positions within the
PHA coding region. PHA alleles containing premature nonsense codons about 40%, 60%,
and 80% of the way through the coding region are compared diagrammatically to the WT
and STOP (20%) alleles. Open bars represent the open reading frame and indicate the
number of amino acids (aa) before the nonsense codons that are indicated by stop arrows.
Untranslated regions are represented by lines. All alleles were cloned into pMON505
derivatives for stable introduction into BY-2 cells. b. Effects of premature nonsense
codons on PHA mRNA stability. Multiple mRNA half—life measurements were made for
the cell lines stably expressing each PHA allele in a. Mean half-lives, calculated as in
Figure 2-4, are represented in the histogram. Error bars represent the SE.

72

DISCUSSION

In this report I examined the effect of premature nonsense codons on PHA mRNA
metabolism. The half-life of PHA transcripts containing a frame-shift (FS) or nonsense
(STOP) mutation in tobacco BY-2 cells was significantly reduced as compared to the
wild-type (WT) PHA transcript, showing that premature termination of translation (20%
of the way through the coding region) can lead to increased rates of mRN A decay. These
results were confirmed by demonstrating that PHA transcripts with nonsense codons 40%
or 60% of the way through the coding region were also rapidly degraded. Further support
for a destabilizing effect of premature nonsense codons was obtained by analyzing the
effect of the FS and STOP mutations on accumulation of PHA mRNA in BY-2
protoplasts and transgenic tobacco and Arabidopsis plants. FS and STOP mRNA
accumulated to barely detectable levels in all three systems.

It was reported previously that the FS mutation reduced the accumulation of PHA
mRNA in bean and transgenic tobacco seeds (Voelker et al., 1986; 1990). Based on these
reports, it was not clear whether this effect was caused by premature termination of
translation, a change in the structure of the mRNA, or a change in codon usage in the
alternative reading frame. Another consideration was that the FS mutation, located in
codon 11, disrupted the PHA signal sequence, which could potentially change its site of
translation from membrane-bound to soluble polyribosomes. To differentiate among these
alternative explanations, in this study I created the STOP allele, which had the same early
nonsense codon as the FS, but no alteration in the reading frame. The PS mutation (a one-

base deletion) and the STOP mutation (a two-base substitution) were separated by more

73

than 100 nucleotides, which makes it very unlikely that they both disturb a stabilizing
sequence in the transcript or have a similar effect on secondary structure. In addition, the
codon usage of the STOP allele is identical to that of the WT PHA. Therefore, premature
termination of translation is the only plausible cause of the rapid degradation of the
mutant transcripts. Moreover, a premature nonsense codon is the common denominator
among the four PHA transcripts that were found to be unstable (FS, STOP, 40%, and
60%) in this report. This work extends previous studies both by showing that it is
premature termination of translation that causes low accumulation of PHA mRNA, and by
demonstrating that the degradation of the transcripts is markedly accelerated.

The nonsense-mediated mRNA decay pathway that facilitates the rapid decay of
certain PHA transcripts may extend to other transcripts in higher plants. As discussed in
the introduction, a number of other transcripts with premature nonsense codons also fail
to accumulate to wild-type levels in plant cells. In addition, E]. De Rocher and P.J.Green
(personal communication) have identified a derivative of a synthetic Bacillus
thuringiensis toxin gene containing a premature nonsense codon. This mutant gene gives
rise to much lower transcript levels than does the wild-type in maize and tobacco
protoplasts and stably transformed BY-2 cells . The work in this report demonstrates that
early nonsense codons in PHA can also be recognized in different plants, organs, and cells
in which PHA is not normally expressed. On the basis of all of these observations it
seems likely that other mRNAs in addition to PHA transcripts may be subject to
nonsense-mediated mRNA decay in plants and that this mechanism Operates in at least
three different plant species.

The effect of premature nonsense codons on PHA mRNA can be greatly

74

influenced by their position in the transcript because termination of translation 80% of the
way through the PHA coding region did not trigger rapid decay. The information that
allows plants to recognize certain nonsense codons as being premature is not contained in
the nonsense codon itself. Therefore, there must either be an additional cis-acting
sequence contained in the transcript, or simply the length of the translated or untranslated
region may render the mRN A unstable. At present, there is no evidence for a minimal
length of a coding region or a maximal length of a 3' untranslated region for efficient gene
expression. However, there is precedence for the presence of cis-acting sequences for
nonsense-mediated mRN A decay in yeast (Peltz et al., 1993). This has been studied in the
most detail for the PGKl gene. Premature nonsense codons in the first two-thirds of the
PGKl coding region cause rapid degradation of the mRNA, whereas nonsense codons in
the last quarter do not (Peltz et al., 1993). Nonsense-mediated mRNA decay of the PGKl
mRNA appears to be mediated by two types of cis-acting sequences, one of which has
been delineated fairly precisely and must be located downstream of a nonsense codon to
have an effect on mRNA abundance (Peltz et al., 1993; Zhang et al., 1995). The second
more loosely defined sequence, located between 67 and 92% of the PGKI coding region,
has been hypothesized to prevent downstream nonsense codons from causing rapid
degradation when it is translated (Peltz et al., 1993).

These results point to the sequences located between 60 and 80% of the way
through the PHA coding region as probable candidates for cis-acting elements controlling
nonsense-mediated decay. An mRNA instability sequence that acts when it is downstream
of a premature stop codon could be present in this region. This scenario would explain

why the PHA transcripts with nonsense codons located in the first 60% of the PHA

75

coding region are unstable, whereas those which terminate translation in the last 20% are
not. If such an element exists, it is likely to differ from the cis element that must be
located downstream of a premature stop codon in PGK 1, discussed above. The PGKl
sequence is not found in PHA and the closest match is found upstream of the 60%
position. Another possibility is that the region between 60% and 80% contains a sequence
that stabilizes the transcript if it is translated, similar to. that of the second element
hypothesized to act in PGK 1. In any event, the requirement for additional sequences
might indicate that not all messages can be destabilized by premature nonsense codons. In
addition, the exact location of these cis-elements in different transcripts could easily
differ and thus determine the positions in the transcript where nonsense codons would
have an effect. Clearly, more experiments will be required to examine the nature of
putative cis-elements in the PHA transcript and to explain how these elements mediate
recognition of premature nonsense codons in a position-dependent manner.

The location in the cell where nonsense-mediated mRNA decay takes place is
somewhat controversial. There is considerable evidence for recognition and degradation
of these transcripts during export from mammalian nuclei and for a likely cytoplasmic
recognition and degradation event in yeast (Maquat, 1995). It is not clear whether these
are fundamentally different mechanisms or variations on the same theme. With the tools
and knowledge resulting from the present study, it should be possible to address the
question of where PHA transcripts with premature nonsense codons are degraded in plant
cells and to investigate other aspects of the mechanism.

Plants, yeast (Leeds et al., 1991), mammals (Maquat, 1995) and C. elegans (Pulak

and Anderson, 1993) all appear to have evolved mechanisms to accelerate the decay of

76

certain transcripts containing early nonsense codons. A major advantage associated with
having a nonsense-mediated mRN A decay pathway would be that it allows organisms to
minimize the production of truncated proteins from defective mRNAs. RNA processing
reactions are not completely accurate, giving rise to abnormal messages by incorporating
an incorrect nucleotide during transcription or by aberrant or incomplete splicing.
Premature nonsense codons could also arise by somatic or germline mutation. Not only is
the production of truncated proteins wasteful, but it could be detrimental to the cell. In C.
elegans, some myosin nonsense mutations that are recessive in a wild-type background
become dominant in mutants that lack the nonsense-mediated mRNA decay pathway
(Pulak and Anderson, 1993). Presumably, the overproduction of truncated myosin
polypeptides interferes with the assembly and/or function of normal myosin. Truncation
of many other proteins could also cause ill effects. Thus, the ability to degrade even a

subset of transcripts with premature nonsense codons may increase evolutionary fitness.

MATERIALS AND METHODS

Plasmid construction

Standard procedures were used for plasmid manipulation (Sambrook et al., 1989).
Each chimeric PHA gene in Figure 2-1 contained the 35S promotor of CaMV, including
28 bp of the 5'UTR, followed by 12 bp of linker sequence (CAAGCTCAGATCTG). This
was fused to PHA sequences consisting of 3 bp of the 5'UTR, all of the coding region and

1.4 kb of 3' flanking sequence (ending at the Acc I site in the genomic clone Voelker et

77

al., 1990). PHA genes were introduced between the Sac I and Cla I sites of a pMON 505
derivative that also contained a 35S-GUS gene (as in Newman et al., 1993). Additional
details of the gene construction are available on request. For electroporation experiments,
the PHA genes were introduced between the Sac I and Sal I site of Bluescript II SK+
(Stratagene, La J olla, CA).

Nonsense codons were introduced using the Muta-gene kit (Biorad, Hercules, CA)
and the (antisense) oligonucleotides 5'GAT'I‘GGTEATCGTAACTG3' for the STOP
(20%) allele, 5'CAAAGGCAAGGCCClAGGCGGGTCC3' for the 40% allele,
5'AGTTCACGTCGAT1T_AAATATGACG3' for the 60% allele, and
5'GGAAGAACGCTCTAQAGTCCACTGT3' for the 80% allele. The introduced
nonsense codons are underlined. The mutagenic oligonucleotides for the 40%, 60% and
80% alleles also introduced a restriction endonuclease site that was used to identify
putative mutants. All mutant alleles were confirmed by dideoxy sequencing. The resulting
PHA alleles were cloned into pMON 505, as described above. After this work was
completed it was found that the binary vector containing the FS allele contains a small
(0.8kb) insertion outside the 3SS-PHA gene. This is unlikely to have an effect on its
expression because the behavior of the STOP gene which lacks this insertion is identical

to that of the FS gene.

BY-2 cell manipulations

Cell culture and transformation of Bright Yellow 2 (BY-2; N agata et al., 1992;

also known as NTl) tobacco cells was performed as described by Newman et al. (1993)

78

with the following modifications. All pMON505 derivatives were introduced into
Agrobacterium tumefaciens strain LBA 4404 by electroporation. Putative BY-2
transformants were screened for expression of the GUS gene by a histochemical assay
using the substrate X-Gluc (Clontech, Palo Alto, CA; Jefferson et al., 1987) and for PHA
expression by RNA gel blot analysis.

Protoplasts for electroporation were prepared by incubating BY-2 cells, three days
after subculture, for 3-5 hr, at 28° C, in 2% cellulysin (Calbiochem, La J olla, CA), 1%
cytolase (Genencor International, Rolling Meadows, IL) and 0.2% pectolyase (Karlan,
Santa Rosa, CA) in NT wash solution [3.4 g/l MS salts (Gibco BRL, Gaithersburg, MD),
30 g/l sucrose, 3 uM thiamine, 0.56 mM myoinositol, 1.3 mM KHZPO4, 54.1 g/l glycine
betaine and 10*5 M 2,4-D pH 5.7]. The protoplasts were washed 3 to 4 times in
approximately 20 volumes NT wash solution, and resuspended in electroporation buffer
(10 mM HEPES pH 7.2, 150 mM NaCl, 4 mM CaCl2 , 400 mM mannitol) to a
concentration of 4x 106 prot0plasts per ml. One half ml of electroporation buffer
containing 100 ug/ml of PHA encoding plasmid, 100 ug/ml of GUS encoding plasmid
(under control of the 35S promoter) and 100 ug/ml sheared single-stranded salmon sperm
DNA was added to 0.5 ml of protoplasts. Protoplasts were electroporated in a 24-well
plate on ice at 350 V, 500 uF, and a 6 mm distance between electrodes. The protoplasts
were harvested 14 hr after electroporation, and total RNA was isolated as described
below. A 20 u g aliquot of total RNA was treated with 1 unit DNase I (RQl; Promega,
Madison, WD for 15 min at 37° C to digest residual plasmid DNA.

Messenger RNA decay measurements following treatment with 100 ug/ml

actinomycin D (Act D) were as described by Newman et al. (1993), except that for

79

analysis of FS and STOP constructs time points were 15 min apart instead of 30 min
apart. Newman et al. (1993) showed that this concentration of actinomycin D inhibits

transcription by 94%.

Transgenic tobacco

Transformation and regeneration of transgenic tobacco plants were performed as
described by Newman et al. (1993). Plants were analyzed for GUS activity using 20 u g of
protein extract of leaves in enzymatic assays using 4-methylumbelliferyl-B-D—glucuronide
(United States Biochemical, Cleveland, OH; Jefferson et al., 1987) as substrate.
Transgenic plants with detectable GUS activity were grown to the 10-15 leaf stage and
RNA was isolated from a leaf near full expansion as described below. Statistical analyses

were carried out using the test of Wilcoxon-Mann-Whitney (Nap et al. 1993).

Transgenic Arabidopsis

Binary vectors were introduced into Agrobacterium strain C58C1(pMP90) by
electroporation. Transformation of Arabidopsis was done by vacuum infiltration, based
on the method of Bechtold et a1. (1993) with the following modifications. Plants of the
ecotype Columbia were grown under a regime of 16h light, 8h dark at 20° C until the
primary bolt was 5-15 cm long. Agrobacterium from a 500 ml overnight culture (of YEP
medium [10 g/l Yeast extract, 10 g/l bacto peptone, 5 g/l NaCl, pH 7.0] supplemented

with 50 mg/ml rifampicin, 25 mg/ml gentamycin and 100 mg/ml spectinomycin) was

80

pelleted and resuspended in l l of infiltration medium (0.5XMS salts [Gibco BRL,
Gaithersburg, MD], 100 pg/ml myo-inositol, 10 ug/ml thiamine-HCl, 1 ug/ml nicotinic
acid, 1 rig/ml pyridoxine-HC], 5% sucrose, 2.5 mM MES pH 5.7, 44 nM
Benzylaminopurine, and 200 ppm Silwet L-77 [OSI Specialties, Danbury, CT]).
Arabidopsis plants were infiltrated with this suspension under 450 mm Hg vacuum for
five min and returned to the growth chamber. Seeds of each plant were harvested
separately. Approximately 2000 seeds from each plant were plated on media containing
30 ug/ml kanamycin and 500 ug/ml vancomycin. Using this method >95% of the
infiltrated plants were transformed and 1-5% of the seed from an individual plant was
transgenic. To ensure that each analyzed transformant was from an independent event,
only one kanamycin resistant seedling from each infiltrated plant was used. This
transformant was transferred to soil and the seed was harvested. In the next generation
approximately 500 seeds from each independent transformant were plated on AGM plates
(Taylor et al., 1993) supplemented with kanamycin (30 ug/ml) and vancomycin (500
ug/ml). Seedlings were harvested 7 days after germination and immediately frozen in

liquid nitrogen.

RNA isolation and analysis

RNA was isolated by the method of Puissant and Houdebine (1990) with
modifications as described by Newman et al. (1993), with the following exceptions.
When RNA was isolated from protoplasts, the step involving grinding in liquid nitrogen

was replaced by vortexing in GTC buffer (Newman et al., 1993) for 1 min. When RNA

81

was isolated from Arabidopsis seedlings, the step involving grinding in liquid nitrogen
was replaced by lyophilizing the tissue and grinding it by vortexing in the presence of
about 20 three mm glass beads for 30 seconds before addition of GTC buffer. RNA was
separated on 1% agarose/formaldehyde gels, blotted, and hybridized as described by
Newman et a]. (1993). For the electroporation experiments, where a 35S-GUS construct
was co-electroporated with the PHA constructs, blots were probed with PHA, analyzed,
stripped, and then reprobed with GUS and analyzed again. The PHA probe consisted of
8bp of the 5' UTR and all of the coding region. The GUS probe was as described by
Newman et al. (1993). Both probes were labeled with 32P by the random priming method
(Feinberg and Vogelstein, 1983). RNA levels were quantified using a PhosphorImager
(Molecular Dynamics, Sunnyvale, CA) and half-lives determined by linear regression

analysis using Sigma Plot software.

ACKNOWLEDGMENTS

I am grateful to Drs. Toni Voelker and Maarten Chrispeels for providing the WT and FS
alleles, Drs. Donna Koslowsky, Michael Sullivan and Jay De Rocher and Jenifer Murphy
for comments on the manuscript, Scott Diehn and Pauline Bariola for optimizing the
electroporation conditions and Arabidopsis transformation, respectively, the members of
the Green lab for helpful discussions, and to Ms. Karen Bird for editorial assistance. The
technical assistance of André Dandridge is also appreciated. This work was supported by

grants from the USDA (930115) and DOE (FG02-91ER20021) to Pamela Green.

82

REFERENCES

Aharon, T. and Schneider, RI. (1993) Selective destabilization of short-lived mRNAs
with the granulocyte-macrophage colony-stimulating factor AU-rich 3' noncoding region
is mediated by a cotranslational mechanism. Mol. Cell. Biol. 13, 1971-1980.

Altmann, M. and Trachsel, H. (1993) Regulation of translational initiation and
modulation of cellular physiology. TIBS 18, 429-432.

Barker, GP. and Beeman, K. (1991) Nonsense codons within the Rous Sarcoma Virus
gag gene decrease the stability of unspliced viral RNA. Mol. Cell. Biol. 11, 2760-2768.

Becthold, N., Ellis, J. And Pelletier, G. (1993) In planta Agrobacterium mediated gene
transfer by infiltration of adult Arabidopsis plants. C. R. Acad. Sci. Paris, Sciences de la
vie/ Life sciences 316, 1194-1199.

Cheng, J ., Fogel-Pertrovic, M. and Maquat, LE. (1990) Translation to the distal end of

the penultimate exon is required for normal levels of spliced triosephosphate isomerase
mRNA. Mol. Cell. Biol. 10, 5215-5225.

Cleveland, D.W. (1988) Autoregulated instability of tubulin mRNAs: a novel eukaryotic
regulatory mechanism. TIBS 13, 339-343.

Dehesh, K., Franci, C., Parks, B.M., Seeley, K.A., Short, T.W., Tepperman, J .M. and
Quail, PH. (1993) Arabidopsis hy8 locus encodes phytochrome A. Plant Cell 5, 1081-
1088.

Dickey, L.F., Gallo-Meagher, M. And Thompson, W.F. (1992) Light regulatory
sequences are located within the portion of the fed] message sequence. EMBO J. 11,
23 1 1-23 17.

Dickey, L.F., Nguyen, T.-T., Allen, GO and Thompson, W.F. (1994) Light modulation
of ferredoxin mRNA abundance requires an open reading frame. Plant Cell 6, 1171-1176.

Edwards, DR. and Mahadevan, LC. (1992) Protein synthesis inhibitors differentially
superinduce c-fos and c-jun by three distinct mechanisms: Lack of evidence for labile
repressors. EMBO J. 11, 2415-2424.

Feinberg, AP. and Vogelstein, B. (1983) A technique for radiolabeling DNA restriction
endonuclease fragments to high specific activity (addendum). Anal. Biochem. 137, 266.

Hershey, J .W.B. (1991) Translational control in mammalian cells. Annu. Rev. Biochem.
60, 717-755.

83

Hoffman, L.M. and Donaldson, DD. (1985) Characterization of two Phaseolus vulgaris
phytohemagglutinin genes closely linked on the chromosome. EMBO J. 4, 883-889

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

J ofuku, K.D., Schipper, RD. and Goldberg, RB. (1989) A frameshift mutation prevents
Kunitz trypsin inhibitor mRNA accumulation in soybean embryos. Plant Cell 1, 427-435.

Leeds, P., Peltz, S.W., Jacobson, A. and Culbertson, MR. (1991) The product of the
yeast UPF] gene is required for the rapid turnover of mRNAs containing a premature
translational termination codon. Genes Dev. 5, 2303-2314.

Losson, R. and Lacroute, F. (1979) Interference of nonsense mutations with eukaryotic
messenger RNA stability. Proc. Natl. Acad. Sci. USA 76, 5134-5137.

Luehrsen, K.R., Taha, S. and Walbot, V. (1994) Nuclear pre-mRNA processing in higher
plants. Prog. Nucl. acids Res. Mol. Biol. 47, 149-193

Maquat, L.E., Kinniburgh, A.J., Rachmilewitz, EA. and Ross, J. (1981) Unstable B-
globin mRNA in mRNA-deficient [3° thalassemia. Cell 27, 543-553.

Maquat, LE. (1995) When cells st0p making sense: Effects of nonsense codons on RNA
metabolism in vertebrate cells. RNA 1, 453-465.

Naeger, L.K., Schoborg, R.V., Zhao, Q., Tullis, GE. and Pintel, DJ. (1992) Nonsense
mutations inhibit splicing of MVM RNA in cis when they interrupt the reading frame of
either exon of the final spliced product. Genes Dev. 6, 1107-1119.

Nagata, T., Nemoto, Y. and Hasezawa, S. (1992) Tobacco BY-2 cell line as the "HeLa"
cell in the cell biology of higher plants. International review of cytology 132, 1—30.

Nap, J.-P., Keizer, P. and Jansen, R. (1993) First-generation transgenic plants and
statistics. Plant Molecular Biology Reporter 11, 156-164.

Newman, T.C., Ohme-Takagi, M., Taylor, CB. and Green, R]. (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, TC. and Green, R]. (1993) The effect of
sequences with high AU content on mRN A stability in tobacco. Proc. Natl. Acad. Sci.
USA. 90, 11811-11815.

Pelsy, F. and Lacroute, F. (1984) Effect of ochre nonsense mutations on yeast URAI
mRNA stability. Curr Genet. 8, 277-282

84

Peltz, S.W., Brewer, G., Bernstein, P., Hart, RA. and Ross, J. (1991) Regulation of
mRNA turnover in eukaryotic cells. In critical reviews in eukaryotic gene expression,
G.S. Stein, J.L. Stein and J .B. Lain, ed (Boa Radon: CC Press), pp 99-126.

Peltz, S.W., Donahue, J .L. and Jacobson, A. (1992) A mutation in the tRNA
nucleotidyltransferase gene promotes stabilization of mRNAs in Saccharomyces
cerevisiae. Mol. Cell. Biol. 12, 5778-5784.

Peltz, S.W., Brown, AH. and Jacobson, A. (1993) mRNA destabilization triggered by
premature translational termination depends on at least three cis-acting sequence elements
and one trans-acting factor. Genes Dev. 7, 1737-1754.

Puissant, C. and Houdebine, L.-M. (1990) An improvement of the single-step method of
RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction.
BioTechniques 8, 148-149.

Pulak, R. and Anderson, P. (1993) mRNA surveillance by the Caenorhabditis elegans
smg genes. Genes Dev. 7, 1885-1897.

Reed, J .W., Nagpal, P., Poole, D.S., Furuya, M. and Chory, J. (1993) Mutations in the
gene for the red/far-red light receptor phytochrome B alter cell elongation and
physiological responses throughout arabidopsis development. Plant Cell 5, 147-157.

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

Sambrook, J ., Fritsch, BF. and Maniatis, T. (1989) Molecular Cloning; a laboratory
manual. 2nd ed. Cold Spring Harbor Press, Cold Spring Harbor, NY.

Shyu, A.-B., Belasco, J .G. and Greenberg, ME. (1991) Two distinct destabilizing
elements in the c-fos message trigger deadenylation as a first step in rapid mRN A decay.
Genes Dev. 5, 221-231.

Simpson, SB. and Stoltzfus, CM. (1994) Frameshift mutations in the v-src gene of
Avian Sarcoma Virus act in cis to specifically reduce v-src mRNA levels. Mol. Cell.
Biol. 14, 1835-1844.

Smith, C.W.J., Patton, J .G. and Nadal-Ginard, B. (1989) Alternative splicing in the
control of gene expression, Annu. Rev. Genet. 23, 527-577.

Sullivan, ML. and Green, RI. (1993) Post-transcriptional regulation of nuclear-encoded
genes in higher plants: the roles of mRNA stability and translation. Plant Mol. Biol. 23,
1091-1104.

Surdej, P., Riedl, A. and J acobs-Lorena, M. (1994) Regulation of mRNA stability in
development. Annu. Rev. Genet. 28, 263-282.

85

Taylor, C.B., Bariola, P.A., delCardayre, S.B., Raines, RT. and Green, P.J. (1993) RN82:
A senescence-associated RNase of Arabidopsis that diverged from the S-RNases before
speciation. Proc. Natl. Acad. Sci. USA 90, 51 18-5122.

Urlaub, G., Mitchell, P.J., Ciudad, OJ. and Chasin, LA. (1989) Nonsense mutations in
the dihydrofolate reductase gene affect RNA processing. Mol. Cell. Biol. 9, 2868-2880.

Vancanneyt, G., Rosahl, S. and Willmitzer, L. (1990) Translatability of a plant-mRNA
strongly influences its accumulation in transgenic plants. Nucl. Acids Res. 18, 2917-
292 1 .

Voelker, T.A., Moreno, J. and Chrispeels, M.J. (1990) Expression of a pseudogene in
transgenic tobacco: A frameshift mutation prevents mRN A accumulation. Plant Cell 2,
255-261.

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

Zhang, S., Ruiz-echevaria, M.J., Quan, Y. and Peltz, SW. (1995) Identification and
Characterization of a sequence motif involved in nonsense-mediated mRNA decay. Mol.
Cell. Biol. 15, 2231-2244.

CHAPTER 3

POLYRIBOSOMAL PHA mRN A CAN BE DEGRADED BY

THE N ONSENSE-MEDIATED mRN A DECAY PATHWAY

86

87

ABSTRACT.

In every eukaryote examined, it has been found that mRN A containing premature
nonsense codons is preferentially degraded. Although general features of the nonsense-
mediated decay pathway are conserved among eukaryotes some characteristics vary.
Premature nonsense codon—containing transcripts from yeast and some mammalian genes
appear to be degraded in the cytoplasm, while other mammalian nonsense-containing
transcripts appear to be degraded while they are associated with the nucleus. Here I
describe experiments to localize the nonsense-mediated decay of PHA mRNA in tobacco
cells. A majority of nonsense-containing PHA mRN A was found to be associated with
polyribosomes, indicating that it is located in the cytoplasm. This indicates that PHA
mRNA containing premature nonsense codons is exported to the cytoplasm and

assembled into polyribosomes before it is degraded.

88

INTRODUCTION.

Although nonsense-mediated decay appears to be universal in eukaryotes,
important mechanistic differences among eukaryotes may exist. One of the most
important differences in nonsense-mediated decay relates to the subcellular location of
the pathway. In yeast and Xenopus nonsense-mediated decay appears to be cytoplasmic,
while conflicting evidence exists for the subcellular location of this decay pathway in
mammals. Some data support a nuclear nonsense-mediated mRNA decay process, while
other data suggests a cytoplasmic mechanism.

The best evidence for a cytoplasmic nonsense-mediated mRNA decay pathway
comes from microinjection experiments of Xenopus oocytes and embryos. Whitfield et
al. (1995) injected mRNAs that contained premature nonsense codons and several control
transcripts into the cytoplasm of Xenopus oocytes and embryos, and followed the decay of
these transcripts over time. In these experiments the nonsense codon-containing mRNAs
were degraded more rapidly than several control transcripts. This indicates that
cytoplasmic mRNA can be degraded in a nonsense codon-dependent manner.

A large body of evidence strongly suggests that nonsense-mediated mRNA decay
in yeast is also a cytoplasmic event, but alternative explanations for at least some of the
results are possible. First, UPFlp, a trans-acting factor involved in nonsense-mediated
decay has been localized in two independent studies (Peltz et al., 1993; 1994; Atkin et al.,
1995). In both cases the majority of UPFlp was localized to the cytoplasm and was
associated with polyribosomes. However, subsequently it was shown that UPFlp has two

genetically separable functions (W eng et al. 1996 a, b) One function of UPFlp is related

89

to translation, and thus it is not surprising that UPFlp localizes to polyribosomes. A
small amount of UPFlp has been localized to the nucleus (Peltz et al., 1994) and it seems
possible that the role of UPFlp in nonsense-mediated mRNA decay is carried out by this
subset. Second, translation, a cytoplasmic process, appears to be important for
recognition of nonsense codon-containing transcripts by the nonsense-mediated decay
pathway in yeast. Nonsense codon containing-mRNAs are stabilized both by expression
of suppressor tRNAs capable of translating the appropriate nonsense codon, but not by
other tRNAs (Losson and Lacroute, 1979; Gozalbo and Hohmann, 1990;) and by
treatment with cycloheximide (chx; Zhang et al. 1997). Similar observations have been
made for nonsense-mediated mRN A decay in mammalian cells (Takeshita et al., 1984;
Belgrader et al., 1993), yet other evidence (that is discussed below) suggests that in these
cases degradation occurs during export from the nucleus. Third, polysome-associated
nonsense-containing mRNA that accumulates during chx treatment is sensitive to
nonsense-mediated mRNA decay upon removal of the chx (Zhang et al., 1997). Other
data have also been used to argue that nonsense—mediated mRNA decay in yeast is a
cytoplasmic process (see Zhang et al., 1997), but the three lines of evidence mentioned
above are the strongest.

Nonsense codons in a wide variety of mammalian genes reduce the steady state
level of the mRN A by a post—transcriptional mechanism. While in some cases this can be
explained by increased turnover of cytoplasmic mRNA (e.g. Shyu et al., 1991; Kessler
and Chasin, 1996), in other cases there appears to be a nuclear mechanism to decrease the
abundance of nonsense-containing mRNAs. There are three main lines of evidence to

support this hypothesis. First, measurements of (cytoplasmic) mRNA decay rates of

90

some mRNAs with premature nonsense codons do not reflect the decrease in mRN A
abundance. Second, the level of mRNA that copurifies with nuclei is reduced for some
nonsense-containing mRN As. Finally, in some cases an intron is a required cis-acting
sequence for nonsense-mediated reduction of RNA levels.

Cytoplasmic mRN A decay rates in mammalian cells can be measured by
inhibiting transcription and purifying cytoplasmic RNA at regular time intervals.
Measurements of half-lives of total mRNA using actinomycin D (Act D; Maquat et al
1981) or cytoplasmic mRNA using a regulated promotor (Shyu et a1 1991), show an
increased degradation rate of nonsense-containing fl-globin mRNAs. For other genes the
half-lives are the same for nonsense-containing and wild-type transcripts (Urlaub et al.,
1989; Cheng et al., 1990; Baserga and Benz 1992; Carter et al., 1995). One possible
explanation is that nonsense codons affect the nuclear fate of the mRNA, while a small
fraction of the mRNA escapes to the cytoplasm and is stable. Alternatively, it is well
established that inhibition of global transcription can inhibit the decay of some mRNAs,
thus masking a difference in decay rates (Peltz et al., 1991; Ross, 1995; Abler and Green,
1996). This is very well demonstrated by equal stability of nonsense-containing and
wild-type aprt transcripts in the presence of actD, but a 4-fold difference in half—lives of
the same mRNAs when measured using a regulated promoter (Kessler and Chasin, 1996).
In contrast, Cheng and Maquat (1993) reported that nonsense codon-containing and wild-
type TPI mRNA were equally stable when measured using actD or a regulated promoter,
suggesting that the actD results are not an artifact.

A further indication that the degradation of nonsense-containing transcripts may

be a nuclear event comes from cell-fractionation studies. In a large number of studies

91

RNA was isolated from purified nuclei and purified cytoplasm and the amount of
nonsense-containing mRNA that copurified with nuclei was reduced relative to the wild
type control mRNA (Takeshita et al., 1984; Baserga and Benz, 1992; Cheng and Maquat,
1993; Belgrader and Maquat, 1994; Carter et al., 1996; Kessler and Chasin 1996).
Several controls indicated that this reduced copurification reflects an in vivo reduction in
nuclear RNA (Belgrader et al., 1994; Kessler and Chasin, 1996)

In mammalian cells the position of the nonsense codon relative to introns seems to
be important for nonsense-mediated decay. This has been used as supporting evidence
for a nuclear site of degradation because introns are removed from the RNA in the
nucleus, before export of the mRN A. In TPI, aprt, MUP and presumably other
transcripts, nonsense codons upstream of the most 3' intron of the coding region are
effective in triggering decreased mRN A accumulation, while nonsense codons in the final
exon of the coding region are ineffective (Urlaub et al., 1989; Cheng et al., 1990; 1994;
Kessler and Chasin, 1996; Belgrader and Maquat, 1994). Even though the accumulation
level of the mRNA depends on the presence of introns, it is the mature mRNA (or
partially spliced RNA), and not the pre-mRNA that is recognized and degraded (Cheng
and Maquat, 1993; Cheng et al., 1994; Carter et al., 1996; Zhang and Maquat, 1996). The
requirement of a downstream intron has been used as an argument for a nuclear nonsense
codon-dependent degradation mechanism, but the available data can also be explained by
hypothesizing that some factor associates with the mRNA in an intron-dependent manner
and remains associated with the mRNA after splicing. It may be this hypothetical protein
(and not the intron itself) that is required for nonsense-mediated mRNA degradation in

mammalian cells, which then may occur in any subcellular compartment.

92

The current model (Maquat, 1995; Weng et al., 1997) to explain how nonsense
codons can determine the nuclear fate of an mRNA is that (in mammalian cells) the 5' end
of the mRNA is exported first, and translation starts when the mRNA is still associated
with the nucleus. When a premature nonsense codon is encountered, most of the mRN A
is degraded before it completely dissociates from the nucleus. It is not clear whether this
reflects a fundamental difference in nonsense-mediated mRNA decay between yeast and
mammalian cells, whether the data can all be explained by minor variations on one
mechanism, or whether a nuclear and a cytoplasmic pathway operate in parallel in all
eukaryotes.

It has been shown that plants also contain a nonsense-mediated mRNA decay
pathway (van Hoof and Green, 1996), but little is known about the mechanism and
subcellular location of this pathway in plants. The existence of a nonsense-mediated
mRN A decay pathway in plants was proven by measurements of mRNA half-life made
following actD treatment of stably transformed tobacco cells. Although actD timecourses
are often assumed to measure cytoplasmic decay rates, this is not necessarily the case.
Such timecourses measure the decay of pre-existing mRNA and thus, if most of the
mRNA of interest is located in the cytoplasm at the beginning of the timecourse, the
cytoplasmic decay rate is measured. However if most of the mRNA of interest at the start
of the timecourse is in the nucleus the measured half-life reflects either the decay rate of
the mRNA in the nucleus, or the rate of export from the nucleus and subsequent decay in
the cytoplasm. To distinguish between these alternatives it is necessary to determine the
localization of the bulk of the mRN A at steady state (i.e. before the addition of actD). In

the experiments described here I have compared the subcellular location of PHA mRNA

93

destabilized by insertion of a nonsense codon at 60% of the normal coding region with
that of a stable control PHA mRNA with a nonsense codon at 80% of the normal coding
region. I have determined that the majority of the 60% mRN A copurifies with
polyribosomes, similar to the 80% mRNA and thus the 60% mRNA is likely to be
cytoplasmic. This indicates that nonsense-mediated decay of PHA mRNA is likely to be

a cytoplasmic event.

RESULTS.

It has been shown that insertion of a nonsense codon in the first 60% of the
normal coding region of PHA causes a reduction in the accumulation of the mRNA (van
Hoof and Green; 1996). Furthermore, this reduction is caused by an increase in the rate
of degradation of this mRN A. This was determined by measuring the decay rate of the
mRNA by treating stably transformed BY-2 cells with actD. PHA mRNA with a
nonsense codon at 60% (or 20% or 40%) of the coding region was 3-4 fold less stable
than PHA mRNA with a nonsense codon at 80% of the normal coding region or wild type
PHA mRNA. In addition, the difference in mRNA accumulation (about 4 fold) is fully
explained by the difference in half-lives. This is true when comparing accumulation of
PHA mRNA with a nonsense codon at 20% of the normal coding region with wild type
(van Hoof and Green; 1996) and when comparing the 60% and 80% constructs (data not
shown). Thus analysis of the subcellular distribution of the 60% mRNA at steady state
should allow us to localize the nonsense-mediated decay pathway in plant cells.

One way to address whether nonsense-mediated decay is a nuclear event is to

94

isolate RNA from purified nuclei and compare the accumulation of wild-type and
premature nonsense codon-containing mRN A levels. Experiments comparing nuclear
RNA levels are one of the main lines of evidence supporting a nuclear location of the
nonsense-mediated mRN A decay pathway in mammalian cells. Despite exhaustive
attempts I have been unable to purify nuclei of sufficient purity and containing intact
mRNA from BY-2 cells for analysis of nuclear RNA. In various attempts Ieither
obtained intact RNA from nuclei that were contaminated with large amounts of
cytoplasmic mRN A, or degraded RNA. This is probably caused by release of RNases
from the large central vacuole (which contains the majority of RN ase activity; Boller and
Kende, 1979) when the cells are lysed. This also prevents isolation of cytoplasmic RNA
simply by gently lysing the cells, as is routinely done with mammalian cells.

An alternative approach to measure the subcellular distribution of PHA mRNA is
to isolate polyribosomes and determine to what extent PHA mRNA copurifies with
polyribosomes. For these experiments polyribosomes were isolated from stably
transformed BY-2 cell lines expressing PHA mRNA with a nonsense codon at 60% of the
normal coding region. This RNA was chosen because it is susceptible to nonsense-
mediated decay, but still contains a fairly large coding region, and thus should be
associated with reasonably large polyribosomes. As a control cells expressing PHA
mRNA with a nonsense codon at 80% of the normal coding region were also analyzed.
Most published polyribosome isolation procedures include addition of chx to the cells or
lysate to stabilize the association of mRNA with ribosomes. However, I omitted chx
because it also inhibits the degradation of premature nonsense codon-containing PHA

mRNA (data not shown) and thus might artifactually increase the association of PHA

95

mRNA with polyribosomes. Typically about 50 to 70% of the total RNA copurified with
polyribosomes during centrifugation trough a 1.8 M (62% w/v) sucrose cushion. The
RNA from the polyribosomal fraction was further purified and analyzed by RNA gel
blotting. Figure 3-1 shows a typical RNA- gel blot of total and polyribosomal RNA. PHA
mRN A was slightly enriched in the polyribosomal RNA fraction relative to total RNA
and thus since a majority of total RNA is polyribosomal, a majority a the 60% PHA
mRNA appears to be polyribosomal.

To confirm that PHA mRNA is indeed associated with polyribosomes the isolated
polyribosomes were further separated on 15-60% w/v sucrose gradients. After
centrifugation the gradients were analyzed for the distribution of UV absorbing material
(i.e. RNA) and fractionated in 24 fractions (Figure 3-2). RNA was purified from each
fraction and analyzed by RNA gel blotting. Analysis of the UV absorption profiles
shown in figure 3-2B and ethidium bromide stained gels showed little RNA in fraction 1,
40S ribosomal subunits typically in fractions 2 and/or 3, 608 subunits in fractions 3
and/or 4, 80S monosomes in fractions 4 and/or 5, and polyribosomes in denser fractions.
Figures 3-2 A and C show typical UV absorption profiles for preparative sucrose
gradients. These preparative gradients were loaded with a larger amount of material than
the gradient in figure 3-2B and as a result the peaks of 40S, 60S and 80S are not well
resolved, but monosomes were easily identified as a shoulder on the large peak
representing subunits. Disomes and larger polyribosomes were always well separated
from monosomes and free subunits. The large peak containing subunits and monosomes
typically contained about 40% of the UV absorbing material and the polyribosome

fractions contained the remaining 60% of UV absorbing material.

96

60% (380% .3

of .8
\Q (Féo
w. “w

Gus» ”u 0...

Figure 3-1. PHA transcripts with premature nonsense codons copurify with
polyribosomes. Protoplasts were prepared from BY-2 cells expressing PHA alleles with a
nonsense codon at 60% or 80% of the normal coding region and polyribosomes were
isolated. Shown is a northern blot containing 20 u g of total or polyribosomal RNA
probed with a PHA probe (upper panel), stripped and reprobed with a GUS probe (lower
panel.

1000 --

800 --

A254

400 --

200 a

“WW

1000 --
800 --

600 --

A254

400-

200 -

Figure 3-2: Absorption profiles of sucrose gradients. Polyribosomes corresponding to
500 u g (A and C) or 100 u g (B) RNA were separated on 15-60% w/v linear sucrose
gradient and fractionated. A and B show typical profiles for a cell line expressing PHA
with a nonsense codon at 60% of the normal coding region. C shows a typical profile
from a cell line expressing PHA mRNA with a nonsense codon at 80% of the normal
coding region. The peaks corresponding to 408 and 608 ribosomal subunits and 808

600 --

 

 

 

Fractions

 

--l-l-1-+-1+H-l+l-+-l-l+l-++-l-H+l—

1 5

monosomes are indicated in B

11:0 15
ractlons

500

400

300
A254

200 ‘-

100 --

 

408 608

l/

1 5 1 0
Fractions

98

RNA gel blot analysis (Figure 3-3 and 3-4) of the fractions from these same
sucrose gradients showed that PHA mRNA was generally absent in the first three or four
fraction (i.e. those containing free RNA and free subunits) but was present in fractions 4
or 5 to 24 (i.e. those containing monosomes and polyribosomes). This was true for PHA
with a nonsense codon at 60% of the normal coding region, as well as for PHA with a
nonsense codon at 80% of the normal coding region. In addition upon reprobing the
RNA gel blots it was evident that GUS mRNA was present in the same fractions (data not
shown). Based on the spacing of ribosomes along mRNA, one might not expect to find
polyribosomes or PHA mRNA in the bottom part of the sucrose gradient (i.e. in fractions
denser than fraction 15). However, GUS mRNA and rRNA were also present in these
same fractions (data not shown) indicating that the material in these fractions is indeed
polyribosomal. All these data confirm that PHA mRNA was indeed associated with
polyribosomes. One possible explanation of finding polyribosomes in denser fractions
than predicted is that the polyribosome pellet was very tightly packed at the bottom of the
sucrose pad after an overnight centrifugation step. This pellet of polyribosomes could be
resuspended only partially, and the sample loaded onto the sucrose gradient may have
contained some nonspecific aggregates.

As a final confirmation that PHA mRNA is indeed polyribosome associated (and
not associated with some other very dense material) the sucrose gradients were repeated
under conditions that disrupt polyribosomes. Polyribosomes can be specifically disrupted
into ribosomal subunits and free mRN A by the omission of MgCl2 from the resuspension
buffer and gradient, and inclusion of chelating agents (Cox and Goldberg 1988).

Polyribosomes were solubilized as before and loaded onto duplicate sucrose pads. After

 

99

A. 6096
5 10 15 20

Bottom

8
'—

 

B 8096
10 15 20

Top
01
Bottom

 

Figure 3-3: RNA gel blot analysis of fractionated sucrose gradients. Polyribosomes
purified from BY-2 cell lines expressing PHA mRNA with a nonsense codon inserted
either at 60% (A) or at 80% (B) of the normal coding region were separated by sucrose
gradient centrifugation. 24 fractions were collected from the top of the gradient, RNA
was purified from each and analyzed by RNA gel blotting. Shown are typical RNA gel
blots probed with a PHA probe.

100

>

I 60%
80%

j

Nw-PMOQOO
IIIIIII

 

CDRNA levels (arbitrary units)

 

 

 

 

 

(immi'imm'lli ii:

3 5 7 911131517192123
fraction number

 

 

 

 

 

B
25 _ l60% +EDTA
g 80% +EDTA
520 —
>~.
5
-*:=15 -
.D
5;
£10 -
d)
>
£2
‘2‘ 5' .Wllilil l
M 0 L: I mmJ-IJ-IflflJ—IflflI-n—I i :u-III'lIr‘:

1 3 5 7 9 11 17 19 21
fraction number

Figure 3-4: Distribution of PHA mRNA on sucrose gradients. Signals from RNA gel
blots of fractionated sucrose gradients containing polyribosomes (A) or disrupted
polyribosomes (B) were quantitated using a PhosphorImager.

 

101

centrifugation the pellet under one sucrose pad was resuspended as before, while the
other pellet was resuspended under conditions that disrupt polyribosomes, and both were
loaded onto sucrose gradients. Figure 3-5 shows the superimposed UV absorption
profiles of both sucrose gradients. The polysome gradient shows the same distribution as
described above, while in the other gradient only one large peak can be seen. Whereas
polyribosome gradients never resulted in a large amount of RNA in fraction one (and to a
lesser extent fraction 2), the gradients with disrupted polyribosomes did contain a large
amount of RNA in these fractions. Presumably this is caused by disruption of the
polyribosomes, which yield free mRNA and free 408 and 608 subunits. RNA gel blot
analysis of fractions of these gradients revealed that PHA (and GUS) mRNA was shifted
to the top of the gradient: Most of the PHA mRNA was now present in fractions one to
five, while it was barely detectable in fractions 10 and higher. These same fractions now
contained rRNA as indicated by the ethidium bromide stained gel (data not shown), and
GUS mRNA (data not shown), confirming that PHA mRNA was indeed associated with

polyribosomes.

CONCLUSIONS

The subcellular location of the nonsense-mediated mRNA decay pathway in
plants is of considerable interest because there appears to be variation in this aspect
among eukaryotes. Nonsense-containing mRNAs in yeast and Xenopus appear to be
degraded in the cytoplasm while mammalian cells may have both a cytoplasmic and a

nuclear nonsense—mediated decay pathway. The location within the plant cell was

 

 

 

1000 --
900 --
800 --
7oo -
600 «-
A254 500 --
400 «-
300 --
200 -
100 -

 

o-

 

Figure 3-5. Absorption profiles of sucrose gradients. Polyribosomes (black line) or
disrupted polyribosomes (grey line) corresponding to 500 pg RNA were separated on 15-
60% w/v linear sucrose gradients and fractionated. Typical profiles are shown for a cell
line expressing PHA with a nonsense codon at 60% of the normal coding region.

 

 

103
investigated by localizing PHA mRNA that was destabilized by insertion of a nonsense
codon at 60% of the normal coding region. Aproximately 50% to 70% of the destabilized
PHA mRN A copurified with polyribosomes through a sucrose pad. Additional
fractionation of these isolated polysomes by sucrose gradient fractionation and analysis of
a control stable PHA mRNA strongly indicate that this copurification reflects an in vivo
association of PHA mRNA with polyribosomes. This indicates that a mayor fraction of
the unstable PHA mRNA is cytoplasmic and thus that its nonsense-mediated decay is
cytoplasmic.

Measurements of nonsense-mediated degradation rates of cytoplasmic or total
mRNA from some mammalian genes show that the mRNA is stable. This has been
suggested to indicate that a subset of the mRNA is completely exported from the nucleus
and is stable (Maquat, 1995). Measurements of degradation rates of nonsense-containing
PHA mRN A in tobacco cells and similar experiments in yeast do not show a significant
population of stable transcripts, which is consistent with a cytoplasmic degradation
pathway.

Reduced copurification of nonsense-containing mRNA with nuclei is one of the
main arguments supporting a nuclear location of the nonsense-mediated mRNA decay
pathway in mammals, and ideally one would want to compare similar experimental
designs to localize the pathway in different eukaryotes. Attempts to purify nuclear RNA
from BY—2 cells were unsuccessful because of technical limitations. Despite extensive
efforts nuclear RNA that was intact, and not contaminated with cytoplasmic RNA, could
not be obtained, probably because upon lysing the cells large amounts of RNase activity

were released from the vacuole. Experiments assaying copurification of nonsense-

 

104
containing mRNA with yeast nuclei have not been reported either, possibly because of
similar difficulties.

Recently Zhang et al. (1997) showed that polysome-associated mRNA is a
substrate for nonsense-mediated mRNA decay in yeast. One complication is that to show
association of nonsense—containing mRNA with polyribosomes the mRNA was stabilized
by treatment with chx. Polyribosomal RNA that accumulates during chx treatment is
subject to nonsense—mediated mRNA decay upon removal of the drug, indicating that
nonsense mediated mRNA decay is a cytoplasmic event. Copurification of mammalian
nonsense-containing messages with polyribosomeshas to my knowledge not been
reported. A direct comparison of results of similar experiments using yeast, plant and

mammalian cells is therefore not possible.
MATERIALS AND METHODS.

Construction of PHA alleles with nonsense codons at 60% or 80% of the normal
coding region, stable transformation of BY-2 cells, protoplast isolation, RNA isolation
and RNA gel blotting methods are described in Chapter 2.

Polyribosome isolation.
Polyribosomes were isolated from protoplasts essentially as described by Cox and

Goldberg (1988): A 50 ml culture of stably transformed BY—2 cells was used to make

protoplasts. One tenth of the protoplasts were frozen in liquid nitrogen and used for

 

105

isolation of total RNA. The remaining protoplasts were lysed in 20 ml of polyribosome
extraction buffer (0.2 M TRIS-HCl pH 9.0, 0.1 M KCl, 25 mM EGTA, 35 mM MgC12'
1% Triton X-100, 0.5% sodium deoxycholate, 0.5 M Sucrose, 1 mM sperrnidine, 5 mM
2-mercaptoethanol) and nuclei were pelleted (10 min 1300g). Membranes (and
membrane bound polyribosomes) were solubilized by addition of 1% each of Brij 35,
Tween 40 and Nonidet P40 and stirring for 30 min on ice. Insoluble material was
removed by pelleting (20 min l2000g). The supernatant was split in two and layered on 5
ml sucrose pads (1.8 M sucrose, 40 mM TRIS-HCl pH 9.0, 0.2 M KCl, 5 mM EGTA, 30
mM MgC12, 5 mM 2—mercaptoethanol) and polyribosomes were pelleted (16h 120000g in
a Beckman 70TI rotor). The pellet was either resuspended in GTC buffer (van Hoof and

Green 1996) for isolation of polyribosomal RNA, or gently resuspended on ice in 1 ml 40

 

mM TRIS-HCl pH 9.0, 0.2 M KCl, 5 mM EGTA, 30 mM MgC12, 5 mM, 2-mercapto-

ethanol for 2 hours for further fractionation of polyribosomes.

Polyribosome fractionation.

Insoluble material was removed from resuspended polyribosomes (10 min 120g)
and the CD. of the supernatant was measured at 260 nm. 500 pg RNA (assuming all
material absorbing at 260 nm was RNA) was loaded onto a 12 m1 15-60% w/v linear
sucrose gradient (in 40 mM TRIS—HCl pH 8.5, 20 mM KCl, 10 mM MgC12) and
centrifuged for 45 min at 300000g (in a Beckman SW40 rotor). For the gradient shown
in figure 3-2B only 100 pg RNA was loaded. The gradient was fractionated on an ISCO

640 gradient fractionator, Absorbance at 254 nm was measured during fractionation in a

 

106

flow cell (path length 5 mm) and 24 0.5 ml fractions were collected. Polyribosomes in
each fraction were precipitated by the addition of 10 pg glycogen (as a carrier), sodium
acetate pH 5.0 to 0.2 M, and 2 volumes of ethanol and incubated at -20°C overnight. The
pellet was resuspended in 0.5 ml urea lysis buffer (Brewer and Ross, 1990), extracted
twice with an equal volume of phenol/chloroform and once with an equal volume of
chloroform/isoamyl alcohol, and precipitated by the addition of sodium acetate pH 5.0 to
0.2 M and 1 volume isopropanol. Half of the resulting RNA pellet was analyzed by RNA
gel blotting as described.

For EDTA release gradients isolated polyribosomes were resuspended in buffer
lacking MgCl2 and fractionated on gradients containing 1 mM EDTA instead of MgClZ.

and analyzed exactly as described above for polysome fractionation gradients.

107

REFERENCES

Abler ML, Green PJ (1996): Control of mRNA stability in higher plants. Plant Mol. Biol.
32: 63-78.

Aharon T, Schneider RJ (1993): Selective destabilization of short-lived mRNAs with the
granulocyte-macrophage colony-stimulating factor AU-rich 3' noncoding region is
mediated by a cotranslational mechanism. Mol. Cell. Biol. 13: 1971-1980.

Atkin AL, Altamura N, Leeds P, Culbertson MR (1995): The majority of yeast UPFl
co-localizes with polyribosomes in the cytoplasm. Mol. Biol. Cell 6: 611-625.

Baserga SJ, Benz E] Jr (1992): B-Globin nonsense mutation: Deficient accumulation of
mRNA occurs despite normal cytoplasmic stability. Proc. Natl. Acad. Sci. USA 89:
2935-2939.

Belgrader P, Cheng J, Maquat LE (1993): Evidence to implicate translation by ribosomes
in the mechanism by which nonsense codons reduce the nuclear level of human
triosephosphate isomerase mRNA. Proc. Natl. Acad. Sci. USA 90: 482-486.

Belgrader P, Cheng J, Zhou X, Stephenson LS, Maquat LE (1994): Mammalian nonsense
codons can be cis effectors of nuclear mRNA half-life. Mol. Cell. Biol. 14: 8219-8228.

Belgrader P, Maquat LE (1994): Nonsense but not nrissense mutations can decrease the
abundance of nuclear mRN A for the mouse major urinary protein, while both types of
mutations can facilitate exon skipping. Mol. Cell. Biol. 14: 6326-6336.

Boller T, Kende H (1979): Hydrolytic enzymes in the central vacuole of plant cells. Plant
Physiol. 63: 1 123-1132.

Brewer G, Ross J (1990): Messenger RNA turnover in cell-free extracts. Methods
Enzymol. 181: 202-209.

Carter MS, Doskow J, Morris P, Li S, Nhim RP, Sandstedt S, and Wilkinson MF (1995):
A regulatory mechanism that detects premature nonsense codons in T-cell receptor
transcripts in vivo is reversed by protein synthesis inhibitors in vitro.. J. Biol. Chem. 270:

28995-29003.

Carter MS, Li SL, and Wilkinson MF (1996): A splicing-dependent regulatory
mechanism that detects translation signals. EMBO J. 15: 5965-5975.

Cheng J, Fogel-Petrovic M, Maquat LE (1990): Translation to near the distal end of the

penultimate exon is required for normal levels of spliced triosephosphate isomerase
mRNA. Mol. Cell. Biol. 10: 5215-5225.

108

Cheng J, Belgrader P, Zhou X, Maquat LE (1994): Introns are cis effectors of the
nonsense-codon-mediated reduction in nuclear mRNA abundance. Mol. Cell. Biol. 14:
63 17-6325.

Cheng J, Maquat LE (1993): Nonsense codons can reduce the abundance of nuclear
mRNA without affecting the abundance of pre-mRNA or the half—life of cytoplasmic
mRNA. Mol. Cell. Biol. 13: 1892-1902.

Cox KH, Goldberg RB (1988): Analysis of plant gene expression. In Shaw CH (ed):
"Plant Molecular Biology: A practical approach."(lst ed.) Oxford: IRL Press, pp. 1-35.

Gozalbo D, Hohmann S (1990): Nonsense suppressors partially revert the decrease of the
mRNA level of a nonsense mutant allele in yeast. Curr. Genet. 17: 77-79.

Kessler O, Chasin LA (1996): Effect of nonsense mutations on nuclear and cytoplasmic
adenine phosphoribosyltransferase RNA. Mol. Cell. Biol. 16: 4426-4435.

Losson R, Lacroute F (1979): Interference of nonsense mutations with eukaryotic
messenger RNA stability. Proc. Natl. Acad. Sci. USA 76: 5134-5137.

Maquat LE, Kinniburgh AJ, Rachmilewitz EA, Ross J (1981): Unstable B-globin mRN A
in mRNA-deficient [3° thalassemia. Cell 27: 543-553.

Maquat LE (1995): When cells stop making sense: Effects of nonsense codons on RNA
metabolism in vertebrate cells. RNA 1: 453-465.

Peltz SW, Brewer G, Bernstein P, Hart PA, Ross J (1991): Regulation of mRNA turnover
in eukaryotic cells. Crit. Rev. Euk. Gene Exp. 1: 99-126.

Peltz SW, Brown AH, Jacobson A (1993): mRNA destabilization triggered by premature
translational termination depends on at least three cis-acting sequence elements and one
trans-acting factor. Genes Dev. 7: 1737-1754.

Peltz SW, He F, Welch E, Jacobson A (1994): Nonsense-mediated mRNA decay in yeast.
Prog. Nucleic Acid Res. Mol. Biol. 47: 271-298.

Ross J (1995): mRNA stability in mammalian cells. Microbiol. Rev. 59: 423-450.
Shyu A-B, Belasco JG, Greenberg ME (1991): Two distinct destabilizing elements in the
c-fos message trigger deadenylation as a first step in rapid mRN A decay. Genes Dev. 5:

221-231.

Takeshita K, Forget BG, Scarpa A, Benz E] Jr ( 1984): Intranuclear defect in beta-globin
mRN A accumulation due to a premature termination codon. Blood 64: 13-22.

109

Urlaub G, Mitchell PJ, Ciudad CJ, Chasin LA (1989): Nonsense mutations in the
dihydrofolate reductase gene affect RNA processing. Mol. Cell. Biol. 9: 2868—2880.

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

Weng Y, Ruiz-Echevarria MJ, Zhang S, Cui Y, Czaplinski K, Dinman JD, Peltz SW
(1997): Characterization of the nonsense—mediated mRNA decay pathway and its effect
on modulating translation termination and programmed frameshifting. In
Post-transcriptional Gene Regulation. DR Morris and JB Harford eds. (New York: Wiley
and Sons)

Weng YM, Czaplinski K, Peltz SW (1996a): Genetic and biochemical characterization of
mutations in the ATPase and helicase regions of the Upf 1 protein. Mol. Cell. Biol. 16:
5477-5490.

Weng YM, Czaplinski K, Peltz SW (1996b): Identification and characterization of
mutations in the UPFI gene that affect nonsense suppression and the formation of the
Upf protein complex but not mRNA turnover. Mol. Cell. Biol. 16: 5491-5506.

Whitfield TT, Sharpe CR, Wylie CC (1994): Nonsense-mediated mRNA decay in
Xenopus oocytes and embryos. Dev. Biol. 165: 731—734.

Zhang J, Maquat LE (1996): Evidence that the decay of nucleus-associated nonsense
mRNA for human triosephosphate isomerase involves nonsense codon recognition after
splicing. RNA 2: 235-243.

Zhang S, Welch EM, Hogan K, Brown AH, Peltz SW, Jacobson A (1997):
Polysome-associated mRNAs are substrates for the nonsense-mediated mRNA decay
pathway in Saccharomyces cerevisiae. RNA 3: 234-244.

 

CHAPTER 4

RARE CODONS ARE NOT SUFFICIENT TO DESTABILIZE

A REPORTER GENE TRANSCRIPT IN TOBACCO.

Parts of this chapter will be published in Plant Molecular Biology.
Reference: A van Hoof and PJ Green (1997). Rare Codons Are Not Sufficient to
Destabilize a Reporter Gene Transcript in Tobacco. in Press.

110

111

ABSTRACT

In plants as in other eukaryotes, most synonymous codons of the genetic code are
not used with equal frequency. Instead some codons are preferred, whereas others are
rare. Circumstantial evidence led to the suggestion that rare codons have a negative
influence on mRNA stability. To address this question experimentally, rare codons
encoded by a B. t. toxin gene (cryIA(c)) or a synthetic sequence were introduced into a
phytohemagglutinin (PHA) reporter gene. In neither case was the mRNA stability
appreciably diminished in stably-transformed tobacco cell cultures nor was the
accumulation of mRNA in transgenic plants affected. Thus rare codons do not appear to
be sufficient to cause rapid degradation of the PHA mRN A and potentially other mRNAs

in plants.

112

INTRODUCTION

It was shown more than 30 years ago that the genetic code is degenerate (Crick et
al., 1961; Nirenberg et al., 1963), Le. several codons are available to dictate the
incorporation of a particular amino acid residue into a protein. The usage of synonymous
codons does not appear to be random, since certain codons are used less frequently than
others. The underlying reason for this biased codon usage is not clear, but it is thought to
be related to the abundance of isoaccepting tRN As. tRN As corresponding to rare codons
are less abundant in Escherichia coli and yeast than tRNAs corresponding to preferred
codons (lkemura , 1982; lkemura and Ozeki, 1983). The choice of codons has been
suggested to influence the rate of translation and mRNA degradation, although little
evidence exists that the latter is indeed the case. One possible mechanism by which rare
codons can affect the rate of translation is that a ribosome may pause when encountering
a rare codon, because it may take longer for a rare isoaccepting tRNA to enter the A-site
of the ribosome. A shortage of a specific charged tRNA in bacteria can cause the
ribosome to pause at codons translated by that tRNA. As few as one (Hsu et al., 1985) or
two (Yanofsky, 1981) codons for which there is little charged tRNA available can cause
ribosomal pausing. More recently it was suggested that a stretch of rare codons causes
the ribosome to stall on the yeast matal mRNA (Caponigro et al., 1993; Hennigan and
Jacobson, 1996). The possible effect of codon usage on mRNA stability is not only
intriguing from a purely scientific view, but also relevant to plant biotechnology. In
several cases it has been problematic to express foreign genes in plants (Diehn et al.,

1996). Often, these problems can be overcome by redesigning the genes to alter the

113

codon usage.

Evidence for a general role of rare codons in regulating mRNA stability in
eukaryotes comes mainly from two types of experiments. First, in yeast there is a
statistical correlation between the presence of rare codons and transcript instability
(Herrick et al., 1990). This might be an indication that rare codons are sufficient to
trigger the rapid decay of an mRNA, but this observation could equally well be explained
by two separate evolutionary pressures on the genes for abundant proteins. That is, there
may have been evolutionary selection for mRNAs that are both stable and efficiently
translated (Herrick et al., 1990). Second, several studies have addressed the effect of rare
codons on transcript stability in yeast (e. g. Hoekema et al., 1987; Kotula and Curtis,
1991) and plants (e. g. Perlak et al., 1991), by either altering codons in highly expressed
mRNAs to change them to rare codons, or, more commonly, by altering codons to more
typical codon usage in low abundance mRNA, and measuring the resulting change in
mRNA abundance. In general, the results are consistent with the hypothesis that rare
codons have an effect on mRNA stability, but alternative explanations can not be
excluded. For example the rare codons introduced into the yeast PGKI gene, resulting in
decreased mRNA accumulation (Hoekema et al., 1987), were in a region later shown to
contain a transcriptional enhancer element (Caponigro et al., 1993; Mellor et al., 1987).

Additional examples of changes in RNA abundance as a result of large scale
alterations in codon usage have been found for engineered Bacillus thuringiensis (B. t.)
toxin genes expressed in plants (reviewed in Diehn et al., 1996). However, elimination of
rare codons in B. t. toxin genes significantly increased the GC content, thereby eliminating

AU-rich sequences that may be responsible for improper recognition of introns

114

(Filipowicz et al., 1995) and polyadenylation sites (Hunt, 1994), as well as removing
instability determinants (Ohme-Takagi et al., 1993). Thus, the effect of changing the B. t.
toxin gene may be to eliminate aberrant RNA processing. In only one case were the
stabilities of the modified and nonmodified versions of B. t. toxin mRNA compared. In
this case, the altered B. t. toxin transcript was more stable than the wild type control, but
the difference in stability was not large enough to fully explain the effect on steady state
mRNA levels (De Rocher and Green, personal communication). Moreover, these
experiments do not address whether codon usage changes or RNA sequence changes
cause the observed difference in B. t. toxin mRNA stability.

In only one case has it been clearly shown that rare codons can influence mRNA
metabolism in eukaryotes. In yeast, rare codons enhanced by about 2-fold the effect of a
downstream instability determinant from the matal gene (Caponigro etal., 1993).
Several other sequences rich in rare codons could substitute for the matal rare codons to
enhance mRN A decay, whereas control sequences did not. However, rare codons by
themselves were not sufficient to cause mRNA instability. Additional matal sequences
were also required to trigger rapid mRN A degradation. The role of the rare codons
appears to be to stall the ribosome, to allow interaction with these additional sequences

(Caponigro et al., 1993; Hennigan and Jacobson, 1996).

115

RESULTS

Insertion of rare codons does not affect PHA mRNA stability in BY—2 cells

To test whether rare codons are sufficient to destabilize a reporter transcript in
plants, two sequences rich in rare codons were inserted into the phytohemagglutinin
(PHA) gene from Phaseolus vulgaris. PHA is well-suited to serve as a reporter gene for
this study. Unlike more frequently used reporter genes (such as GUS or luc), PHA is a
plant gene that has a typical plant codon usage. In addition, stopping translation early in
the PHA coding region leads to increased degradation of the mRNA (van Hoof and
Green, 1996). As discussed above, rare codons have been suggested to act by stalling the
ribosome. If this is indeed the case, an effect of rare codons may be more readily
detectable in a reporter gene that can be destabilized by stopping its translation. The
sequences that were inserted into the PHA gene were chosen not only because they were
rich in rare codons, but also because they contained an alternative reading frame that was
both open and showed a typical plant gene codon usage. Insertion of the sequence in one
reading frame would lead to the incorporation of a stretch of rare codons in the reporter
transcripts (Figure 4-1). If a given construct caused rapid degradation of the PHA
mRNA, then insertion of the same sequence in the alternative frame could be used to
distinguish between an effect caused by the presence of rare codons and an effect caused
by insertion of the RNA sequence per se.

The first construct was designed to test whether a stretch of rare codons from a

B. t. toxin gene could trigger rapid mRNA decay. It is firmly established that codon usage

116

 

Insert in frame 1: RARE CODONS

\/

PHA CODING REGION

/\

Insert in frame 2: CONTROL

 

 

 

 

 

 

 

 

Figure 4-1. Experimental design used to test the effect of rare codons on mRNA stability.
The same fragment of DNA can be inserted into the PHA coding region in two different
reading frames. In the first reading frame it contains multiple rare codons, while in the
second reading frame it contains more typical plant codons. DNA fragments containing
rare codons were inserted into a Sty I restriction site 123 nt from the beginning of the 825
nt-long coding region.

117

in B. t. toxin genes poorly matches the preferred codons in plants, and it has been
suggested that this difference may (in part) cause the low abundance of B. t. toxin
transcripts in transgenic plants (reviewed in Diehn et al., 1996). Specifically, two copies
of codons 317 to 339 from a cryIA(c) gene (Adang et al., 1985) were inserted into the Sty
I site of the PHA gene under control of the CaMV 35$ promoter. This chimeric gene
(BT-PHA) was stably transformed into tobacco BY2 cells, and the stability of the mRNA
was measured as previously described (van Hoof and Green, 1996). Insertion of the B. t.
toxin segment, even in two copies, did not significantly alter the stability of the mRNA
(Figure 4-2). Because insertion of the fragment from B. t.-toxin had no effect, analysis of
the alternative reading frame control was not necessary. One possible explanation for the
lack of effect of the B. t. toxin rare codons is that only very rare codons cause an effect.
To test this possibility, six very rare codons were inserted into the Sty I site of PHA. If
rare codons stall translation, six very rare codons should be sufficient because a limited
number of very rare codons have been shown to stall translation in bacteria (Hsu et al.,
1985; Yanofsky, 1981), and only four rare codons were needed to pause ribosomes in
yeast (Caponigro et al., 1993; Hennigan and Jacobson, 1996). CGA is one of the rarest
codons in plants: Only 7% of the 2915 arginine codons in 207 plant genes are CGA
(Murray et al., 1985). However, inserting six copies of a CGA codon would also
introduce a fairly stable stem-loop structure. Instead, I introduced the sequence
(CGA)4(CUA)2. CUA is used only for 8% of 5285 leucine codons (Murray et al., 1985).
The stability of the resulting mRNA was slightly reduced (1.5 fold, Figure 4-2), but this
difference was much smaller than the effect of previously—characterized bonafide

instability determinants. Instead, it was comparable to the effect of inserting certain

118

BT-
PHA

Ve
Barr‘s

 

 

100

PHA mRNA
Half-llfe
(min)

50

 

  

0

 

WI' BT-PHA Very
Flare

Figure 4-2. Insertion of rare codons does not affect PHA mRNA stability in BY-2 cells.
Tobacco BY—2 cells were stably transformed with either the WT-PHA gene or a PHA gene
containing rare codons (BT-PHA or Very Rare). At least two independent transformed
cell lines were used for each construct. Messenger RNA half—lives were determined as
described in chapter 2. (A) Shown are RNA gel blots from representative half-life
measurements for wild type PHA and two chimeric genes containing rare codons. (B)
The histogram represents the average half-life calculated from at least four experiments
and the standard error.

119

random control sequences into a reporter gene (Newman et al., 1993).

Insertion of rare codons does not affect PHA mRNA levels in transgenic tobacco plants

To confirm that the results obtained in cultured tobacco cells were representative
of what occurs in intact plants, 12 to 21 transgenic tobacco plants expressing either the
wild type PHA gene or the two versions with rare codons inserted were generated. RNA
was isolated from a leaf of each plant, and the abundance of PHA mRN A was determined
by RNA gel blotting and quantitation with a Phosphorlrnager. As shown in figure 4-3A,
the average level of the BT-PHA and WT-PHA RNA were virtually identical, while the
average RNA level for the Very Rare gene was slightly reduced. Statistical analysis of
these data by the Wilcoxon-Mann-Whitney test, showed that this difference was not
statistically significant. An RNA gel blot showing the difference in average PHA mRNA
levels is shown in figure 4-3B. To generate this blot, equal amounts of RNA from each
transgenic plant were mixed and analyzed by RNA gel blotting. Thus neither the BT-
PHA nor the VERY RARE gene showed a significant reduction in mRNA stability in BY2
cells or in mRNA abundance in tobacco plants. This is unlikely to be due to technical
limitations because effects of nonsense codons on PHA mRNA stability and mRNA
abundance were easily detected in both systems in previous studies (van Hoof and Green,
1996). In addition, immuno gel blot analysis indicated that both wild-type and rare
codon-containin g PHA transcripts are translated in BY2 cells (data not shown). Thus, the
failure to find an effect of rare codons on mRNA stability cannot be explained by a failure

of these mRN As to be translated.

120

 

11111111

0
° 0
106 g a g g
PHA mRNA 3 ° 9 5
Abundance — o 9
(arbitrary units) — a,
105 9

o 0000
00 go

I llllllll

 

 

4 —
1° WI' BT- Very
PHA Rare

BT- Very
— WI' PHA Rare

1‘ 777????

PHA >

 

Figure 4-3. Insertion of rare codons does not affect PHA mRNA levels in transgenic
tobacco plants. Tobacco leaf explants were stably transformed with either the WT-PHA
gene or the PHA genes containing rare codons. For each construct at least twelve
independent plants expressing PHA were regenerated and transferred to soil. Plants were
grown to the 10-15 leaf stage and total RNA was isolated from one leaf near full
expansion. PHA mRNA levels were quantitated by RNA gel blot using a
PhosphorImager. (A) Shown is the mRNA level for each individual plant (circles) as
well as the average for each construct (bar). Statistical analysis (van Hoof and Green,
1996) showed that there was no significant difference in RNA levels. (B) Shown is an
RNA gel blot of pooled RNA from these same plants. Equal amounts of RNA from each
individual plant were mixed and 10 pg of each pool was analyzed by RNA gel blot.

121

Rare codons may affect translation rates

In addition to affecting mRNA stability, rare codons may also reduce the rate of
translation, resulting in reduced levels of protein. I addressed this hypothesis by
measuring PHA synthesis rates in cell lines stably transformed with either WT-PHA or
one of the two PHA genes with rare codons inserted. Because WT-PHA, BT-PHA and
Very Rare-PHA likely decay at different rates, a difference in protein accumulation can
not be assumed to be caused by a difference in the rate of synthesis. Instead, translation
rates were measured directly by labeling newly-synthesized protein with tritiated leucine,
extracting proteins, and immunoprecipitating PHA with either a polyclonal serum
previously shown to be effective in immunoprecipitation (Kjemtrup et al., 1994; Bollini
and Chrispeels, 1979) or commercially available immunopurified polyclonal antibodies
(Pierce, Rockford IL). Neither of these antibodies immunoprecipitated detectable
amounts of tritiated PHA even though the cells expressed WT-PHA as determined by
RNA gel blot and immunoblot analysis (data not shown). It is not clear whether this was

caused by insufficient labeling of PHA or by a failure to immunoprecipitate PHA.

CONCLUSIONS

The results described in this chapter indicate that rare codons are not sufficient to
destabilize a transcript. It seems likely that if there is a general effect of rare codons on
transcript stability, then the experiments described above should have detected it.

However, rare codons can have an indirect effect in certain specific cases. For example

122

the yeast matal instability determinant is clearly stimulated by upstream rare codons
(Caponigro et al., 1993; Hennigan and Jacobson, 1996), and similar situations may occur
in some plant genes. It is possible that for rare codons to have a direct effect on mRN A
stability, they need to be present at a specific location within the coding region (for
example just downstream of the start codon), or at a high frequency throughout the
transcript (as they are in the B. t. toxin gene). Testing these hypotheses would require the
insertion of rare codons at many different positions within the coding region or
construction of two full-length genes that are nearly identical in nucleotide sequence, but
differ by the reading frame that is read.

Whether rare codons reduce the translation rate of a transcript remains an
unresolved question. In general initiation of translation is the slowest step (about 6.5
seconds) in translation, while elongation is more than 10-fold faster (about 1/8 to 1/3
second; Mathews et al., 1996). Thus for rare codons to affect translation rates they would
have to have a very large effect on the elongation step. In addition rare codons that are
translated by rare isoacccepting tRNAs may not be translated much slower because they
compete with fewer other codons for these tRNAs. In any case, an effect of rare codons
on protein synthesis rates would be most easily addressed by constructing a set of genes
coding for identical proteins and producing similar steady state levels of mRNA, but
differing in their codon usage. In the process of analyzing the cause(s) of the low
accumulation of B.t.-toxin mRNA, such constructs have fortuitously been made. De
Rocher et al. (1997) constructed a synthetic B. t. -toxin gene with preferred codon usage,
and several derivatives with segments of the nonmodified B. t-toxin gene replacing the

corresponding synthetic segment. Genes containing certain segments of the nonmodified

123

B. t.-toxin gene accumulate as much mRNA as fully synthetic B. t.-toxin genes (De Rocher

et al., 1997), and thus are ideally-suited to test the effect of rare codons on translation.

ACKNOWLEDGMENTS

This work was in part funded by grants from the USDA (9301155 and 9600798), DOE
(FG02-91-ER200210) and the Consortium for Plant Biotechnology Research and
matching funds to Pam Green. I would like to thank Dr. E. Jay De Rocher for helpful

comments on the manuscript, Lisa Poduje for her help with some of the experiments and

 

members of Pam Green's laboratory for helpful discussions.

124

REFERENCES

Adang MJ, Staver MJ, Rocheleau TA, Leighton J, Barker RF, Thompson DV (1985):
Characterized full-length and truncated plasmid clones of the crystal protein of Bacillus

thuringiensis subsp. kurstaki HD—73 and their toxicity to Manduca sexta. Gene 36:
289-300.

Bollini R, Chripeels MJ (1979): The rough endoplasmic reticulum is the site of reserve-
protein synthesis in developing Phaseolus vulgaris cotyledons. Planta 146: 487-501.

Caponigro G, Muhlrad D, Parker R (1993): A small segment of the MATad transcript
promotes mRN A decay in Saccharomyces cerevisiae: A stimulatory role for rare codons.
Mol Cell Biol 13: 5141-5148.

Crick FHC, Leslie-Barnett FRS, Brenner S, Watts-Tobin RJ (1961): General nature of the
genetic code for proteins. Nature 192: 1227-1232.

 

De Rocher EJ, Diehn SH, Vargo TC, Green PJ (1997): Manuscript in preparation.

Diehn SH, De Rocher EJ, Green PJ (1996): Problems that can limit the expression of
foreign genes in plants: Lessons to be learned from B. t. toxin genes. Genetic Engineering
18: 83-99.

Filipowicz W, Gniadkowski M, Klahre U, Liu H (1995): Pre-mRNA splicing in plants.
in Lamond AI (ed) Pre-mRN A processing. pp. 65-77 R.G. Landes Company,
Georgetown, TX.

Hennigan AN, Jacobson A (1996): Functional mapping of the translation-dependent
instability element of yeast MATocl mRNA. Mol Cell Biol 16: 3833-3843.

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

Hoekema A, Kastelein RA, Vasser M, de Boer HA (1987): Codon replacement in the
PGKI gene of Saccharomyces cerevisiae: Experimental approach to study the role of
biased codon usage in gene expression. Mol Cell Biol 7: 2914-2924.

Hsu J -H, Harms E, Umbarger HE (1985): Leucine regulation of the ilvGEDA operon of
serratia marcesens by attenuation is modulated by a single leucine codon. J Bact 164:
217-222.

Hunt AG (1994): Messenger RNA 3' end formation in plants. Annu Rev Plant Physiol
Plant Mol Biol 45: 47-60.

125

Ikemura T (1982): Correlation between the abundance of yeast transfer RNAS and the
occurrence of the respective codons in protein genes; Differences in synonymous codon
choice patterns of yeast and Escherichia coli with reference to the abundance of the
isoaccepting transfer RNAs. J Mol Biol 158: 573-597.

Ikemura T, Ozeki H (1983): Codon usage and transfer RNA contents: Organism-specific
codon-choice patterns in reference to the isoacceptor contents. Col Spring Harb Symp
Quant Biol 47: 1087-1097.

Kjemtrup S, Herman EM, Chrispeels MJ (1994): Correct post-translational modification
and stable vacuolar accumulation of phytohemagglutinin engineered to contain multiple
methionine residues. Eur J Biochem 226: 385-391

Kotula L, Curtis PJ (1991): Evaluation of foreign codon optimization in yeast: Expression
of a mouse Ig kappa chain. Bio/Technology 9: 1386-1389.

Mathews MB, Sonenberg N, Hershey JWB (1996): Origins and targets of translational
control. p 1-29. In Translational control. Hershey JWB, Mathews MB, Sonenberg N
(eds) Cold Spring Harbor Laboratory Press, Plainview NY.

Mellor J, Dobson MJ, Kingsman AJ, Kingsman SM (1987): A transcriptional activator is
located in the coding region of the yeast PGK gene. Nucleic Acids Res 15: 6243-6259.

Murray EE, Rochelau T, Eberle M, Stock C, Sekar V, Adang M (1985): Analysis of -
unstable RNA transcripts of insecticidal crystal protein genes of bacillus thuringiensis in
transgenic plants and electroporated protoplasts. Plant Mol. Biol. 16: 1035-1050.

Newman TC, Ohme-Takagi M, Taylor CB, Green P] (1993): DST sequences, highly
conserved among plant SAUR genes, target reporter transcripts for rapid decay in
tobacco. Plant Cell 5: 701-714.

Nirenberg MW, Jones OW, Leder P, Clark BFC, Sly WS, Pestka S (1963): On the coding
of genetic information. Cold Spring Harbor Symp. Quant. Biol. 28: 549-557.

Ohme-Takagi M, Taylor CB, Newman TC, Green PJ (1993): The effect of sequences
with high AU content on mRNA stability in tobacco. Proc. Natl. Acad. Sci. USA 90:
11811-11815.

Perlak FJ, Fuchs RL, Dean DA, McPherson SL, Fischhoff DA (1991): Modification of
the coding sequence enhances plant expression of insect control protein genes. Proc. Natl.
Acad. Sci. USA 88: 3324-3328.

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

126

Yanofsky C (1981): Attenuation in the control of expression of bacterial operons. Nature
289: 751-758.

CHAPTER 5

CONCLUSIONS AND FUTURE PROSPECTS

127

128

N ON SENSE-MEDIATED mRN A DECAY

At the outset of the research described in this thesis, little was known about the
regulation of mRN A stability in plants, and only slightly more was known for other
eukaryotes. Since then, significant progress has been made in identifying cis-acting
sequences that can confer rapid degradation on reporter transcripts. The characterization
of mRNA degradation triggered by one of these elements (premature nonsense codons) is
described in this thesis (chapters two and three). In addition, rare codons had been
suggested to confer rapid degradation, and this was investigated directly in the
experiments described in chapter four.

The data described in chapter two provide the first conclusive evidence that plants
contain a nonsense-mediated mRNA decay pathway. Direct measurement of decay rates
shows that PHA mRN A is destabilized by nonsense codons in the first 60% of the normal
coding region. Some of the characteristics of nonsense-mediated mRNA decay are
conserved in all eukaryotes, while others vary. One of the interesting differences is the
subcellular location of this turnover pathway (W eng et al., 1997). My results on the
subcellular location of the pathway in tobacco cells are described in chapter three. Cell
fractionation demonstrates that a substantial amount of PHA mRNA destabilized by a
nonsense codon at 60% of the normal coding region is associated with polyribosomes.
This suggests that this mRNA is cytoplasmic, and thus, that its nonsense-mediated decay
is likely also cytoplasmic.

A second difference in nonsense-mediated mRNA decay can be found in the

signals that mediate discrimination between normal and premature nonsense codons.

 

129

Yeast recognizes a premature nonsense codon by its position relative to additional cis-
acting sequences (W eng et al., 1997), while mammalian cells use the position of nonsense
codons relative to introns (Maquat, 1995). N onsense-mediated mRNA decay in plants
does not require introns in contrast to the process in mammalian cells. However, it is not
clear whether plants and yeast use similar cis-acting sequences; the region between 60%
and 80% of the normal PHA coding region that appears to be involved in discriminating
between those nonsense codons that cause mRNA instability and those that don't has no
obvious resemblance to a functionally similar well-characterized element from yeast.
Further characterization of the putative PHA element is required to prove that it indeed
functions by destabilizing transcripts when it is inserted downstream of a nonsense
codon. Additional explanations of the differential degradation of mRNAs from different
PHA alleles, such as the difference in length of the coding region and/or 3' UTR, can not
be ruled out based on the results presented here. Additional alleles of PHA are needed to
resolve this.

Recognition (or degradation) of premature nonsense codons in yeast and C.
elegans requires specific trans-acting factors not required for degradation of typical
mRNAs. None of the cloned genes shown to code for nonsense-mediated mRNA decay
factors (i.e. UPF], UPF2, UPF3 and smg2) have strong sequence similarity to known
Arabidopsis sequences. Recently, He and Jacobson (1995) used the two-hybrid screen to
identify additional yeast factors that interact with UPFlp. One of these factors, named
NMD3, has significant sequence similarity with ESTs from Arabidopsis (119F15T7) and
maize (W 21741 and P38861). The Arabidopsis EST represents a partial cDNA clone

which contains 708 bp, including 513 bp of coding region (data not shown). The deduced

 

130

amino acid sequence is 38% identical and 49% similar to NMD3 (data not shown), but it
has not been shown that NMD3 is involved in nonsense-mediated decay, and thus the role
of the Arabidopsis gene is not clear. Sequencing of the complete Arabidopsis genome
will probably identify trans-acting factors involved in nonsense-mediated decay.

Identifying novel trans-acting factors involved in nonsense-mediated decay in
plants may be difficult. In principle one could use both biochemical and genetic methods.
Biochemical methods to isolate trans-acting factors involved in mRNA degradation have
not been very successful in most systems. Some in vitro systems for mRNA decay have
been developed using mammalian cell extracts (Ross, 1995), but it is not clear to what
extent mRN A degradation in these systems mimics in vivo mRNA tum-over.
Additionally, these systems are generally not translationally competent, and thus one
would not expect to be able to reconstitute nonsense-mediated mRNA decay. On the
other hand wheat germ extract is able to translate added RNAS, but PHA mRNA is stable
in wheat germ extract during an overnight incubation regardless of wether it contains a
premature nonsense codon or not (data not shown).

Currently mutants defective in AUUUA- or DST-mediated mRNA decay are
being isolated in Pam Greens laboratory. This can be achieved by destabilizing the
transcript coding for a selectable marker, and selecting mutants that are able to grow on
higher concentrations of the corresponding antibiotic. The results described in this thesis
indicate that it may be possible to adapt this strategy to find mutants in nonsense-
mediated mRNA decay. The most likely explanation of the results in chapter two is that
the region between 60% and 80% of the normal PHA coding region destabilizes

transcripts when it is present downstream of a nonsense codon. This would predict that a

 

131

transcript coding for a selectable marker could be destabilized by inserting this region of
PHA into the 3' UTR and that the same transcript would be stable in any mutants in
nonsense-mediated mRNA decay. The hypothesis that a selectable marker mRNA can be
destabilized by this putative cis-acting sequence is certainly testable.

Perhaps one of the most interesting questions concerning nonsense-mediated
mRNA decay in plants concerns the events following recognition for rapid decay. In both
yeast and mammals, shortening of the poly(A) tail appears to be the first hydrolytic event
in the degradation of a typical mRN A (Tharun and Parker, 1997). One prominent
exception to this is the degradation of nonsense-containing transcripts (Tharun and
Parker, 1997). In both systems, deadenylation rates of nonsense-containing mRNAs are
slow, while their decay is fast. It will be interesting to see whether this difference holds
true for mRNA decay in plants.

Nonsense codons trigger rapid mRN A decay in both prokaryotes and eukaryotes
(reviewed by Weng et al., 1997), indicating that there is an evolutionary advantage to
maintaining this pathway. In C. elegans, certain nonsense mutations that are recessive in
a wild type background become dominant when combined with a mutation in nonsense-
mediated mRN A decay (Pulak and Anderson, 1993). This indicates that one function of
nonsense-mediated mRNA decay may be to prevent these mutations from having a
deleterious effect on the survival of the individual (Pulak and Anderson, 1993). From an
evolutionary standpoint, survival of an individual carrying a deleterious mutation may not
be an advantage. Nonsense-mediated mRNA decay may be advantageous for the survival
of the species because it allows for genotypic variation to be maintained in the

heterozygote, while it suppresses a negative phenotypic effect. This genotypic variation

 

132

may be used for the generation of new advantageous genes or alleles by further mutation.
The machinery that degrades nonsense codon-containin g transcripts in yeast does
not appear to be required for degradation of typical mRNAs (Reviewed by Weng et al.,
1997). However, higher eukaryotes may use the nonsense-mediated mRNA decay
machinery to degrade certain specific wild—type transcripts as well as mutant transcripts.
One possible example of a wild-type transcript that is degraded by the nonsense-mediated
mRN A decay pathway is that of the mammalian UHG gene (Tycowski et al. 1996). The
functional products of the UHG gene are snoRNAs that are processed from the excised

introns. The spliced mRNA does not appear to code for a protein or other functional

 

product and is rapidly degraded. The degradation of UHG mRNA was suggested to be
mediated by the nonsense-mediated mRN A decay pathway because the mRN A contains
many AUG codons followed closely by nonsense codons, is associated with
polyribosomes and is increased in abundance in response to protein synthesis inhibitors
(Tycowski et al. 1996). Similar natural targets for the nonsense-mediated mRNA decay
pathway may be the GU T1 5 transcripts from tobacco and Arabidopsis, described in the
appendix of this thesis, and related transcripts. Degradation of GUT15 transcripts by the
nonsense-mediated mRN A decay pathway is consistent with the facts that they contain
many nonsense codons in all three reading frames, are rapidly degraded, are
polyadenylated, and are induced by cycloheximide.

Another category of transcripts that may be degraded by the nonsense-mediated
mRNA decay pathway is antisense RNA introduced for research or biotechnological
goals. As described in chapter one introduction of an antisense copy of a gene often leads

to a reduction in expression of the corresponding gene. It seems reasonable to postulate

133

that an artificial antisense transcript is unstable because it contains nonsense codons. It
may be possible to increase the abundance of the antisense RNA by removing these
nonsense codons. This, in turn, may lead to an increase in the effectiveness of antisense
technology, although given that the mechanism of antisense inhibition is not clear, an
increase in antisense RNA abundance may not result in an increased inhibition of gene
expression.

Insufficient data have been gathered to completely describe how the cell targets
some mRNA for nonsense-mediated mRNA decay at the molecular level, but the

following model does account for most of the essential features. mRNA does not occur

 

as an isolated molecule in the cell, but is associated with various proteins. During nuclear
processing certain proteins are associated with the mRNA, while other proteins are
associated with the mRN A in the cytoplasm. The essential part of my model for
nonsense-mediated mRNA decay is that the ribosome plays an essential cytoplasmic role
in removing certain proteins from the coding region of the mRNA. These proteins would
associate with the mRNA in the nucleus, and the mRNA would be exported as an mRNP
particle. During translation by the first ribosome these proteins would be removed. Thus,
if a transcript contains a premature nonsense codon, these nuclear proteins would not be
removed from part of the normal coding region. The cell would recognize this as an
abnormal mRNP and degrade the RNA. This model is appealing because it does not
require very many new trans-acting factors. Nuclear and cytoplasmic mRNP proteins
have been described. In addition it explains why very many different mutant transcripts
are degraded by the nonsense mediated decay pathway. Under this model the required

additional cis-acting downstream sequences are degenerate binding sites for specific

134

nuclear proteins. TGYYGATGYYYYY could be one such binding site for a protein in
yeast, while in mammals specific proteins could remain associated with the junction of
exons after splicing has been completed. In both systems (and in other systems such as
plants) other cis-acting sequences that bind other nuclear proteins may still remain to be

discovered.

mRN A DEGRADATION IN PLANTS

Obviously, a thorough understanding of mRNA degradation is required for an

 

understanding of links between translation and mRNA degradation. The next challenge
in the field of plant mRNA degradation will be the identification and characterization of
the trans-acting factors that are important players in mRNA degradation and the
mechanisms by which specific transcripts are recognized and degraded in higher
eukaryotes. The only systems where significant numbers of trans-acting factors have
been identified are E. coli, yeast, and chloroplasts (see Sugita and Sudiura 1996,
Nicholson, 1997, and Tharun and Parker, 1997 for reviews). The main methods of
identifying trans-acting factors have been classical biochemical methods of purifying
RNases and RNA-binding proteins and/or identifying mutants defective in mRN A
turnover. Both of these approaches should also be valuable tools to identify trans-acting
factors involved in the degradation of cytosolic transcripts in plants. Currently both
genetic and biochemical means are being used in Pam Green's lab to identify trans-acting
factors involved in AUUUA- and DST-mediated decay. Specific RNA-binding proteins

are being sought through gel-shift experiments and the yeast three hybrid system

135

(Sengupta et al., 1996), Arabidopsis mutants with increased levels of a normally unstable
selectable marker transcript are being selected, and RNA degrading activities will be
purified and characterized.

Both genetic and biochemical approaches for the characterization of nonsense-
mediated mRNA decay are complicated by the dual effect of nonsense codons on
translation and on mRN A stability. However, any factors and mechanisms identified to
be important for general mRNA degradation can be used to check whether similar
mechanisms and the same trans-acting factors are involved in nonsense-mediated mRN A

decay. It seems likely that some trans-acting factors (for example those conferring

 

specificity) are only required for recognizing (and degrading) mRNAs with one specific
instability determinant. In contrast, some of the actual mRNA-degrading enzymes might
be involved in the decay of a larger number of transcripts. Once mutants defective in
AUUUA- and DST-mediated decay have been obtained, these ideas will be easily testable
by introducing genes containing premature nonsense codons into a mutant background by
crossing or by Agrobacterium-mediated transformation.

In addition to the methods mentioned above, large-scale Arabidopsis genome and
cDNA sequencing projects will help to identify putative trans-acting factors. Several
RNA-degrading activities (and the corresponding genes) from E. coli are well-
characterized, and at least three homologous genes are present in the dbEST database.
EST ATTSO840 is similar to RNase III (N .D. LeBrasseur, M.L. Abler, and P.J. Green,
personal communication), which is an endoribonuclease specific for certain dsRNA
secondary structures. RNase III participates in rRNA processing and some rate-limiting

cleavages of mRNA in E. coli (Nicholson, 1997). Currently, research is in progress to see

 

136

whether the Arabidopsis EST has similar functions. The other Arabidopsis ESTs with
interesting similarities to E. coli RN ases are 143N9T7 and 116G11T7, which are similar
to polyribonucleotide phosphorylase (PNPase) and RNase PH respectively. Both of these
enzymes are 3' to 5' exoribonucleases from E. coli. EST 143N9T7 may be involved in
chloroplastic mRNA degradation, since it is more closely related to PNPase from spinach
chloroplasts (GenBank accession number U52048) than to any other PNPase (69%
identity over 68 amino acid residues and 88% identity over an additional 9 amino acid

residues; data not shown). Unfortunately, the EST clone does not contain the 5' end of

 

the coding region, so it is unknown whether it contains a chloroplast targeting signal.
RNase PH is one of the least well-characterized RNases of E. coli, and its exact role in
vivo is not clear (Nicholson, 1997). EST 116G11T7 is more closely related to
hypothetical proteins from C. elegans, S. pombe and yeast than to prokaryotic RNase PHs
(data not shown), suggesting that it probably does not function in the chloroplast.
However, this cDNA is also a partial clone, so it is unknown whether it contains an N-
terminal targeting signal. This gene may be involved in any of a number of different RNA
metabolism pathways. If it is involved in mRNA degradation, its possible roles include
shortening of the poly(A) tail and/or degrading the body of the mRNA.

In addition to homologs of E. coli enzymes, dbEST also contains sequences
similar to yeast mRNA decay factors. The most interesting of these are ESTs H4B8T7
and H4B9T7 which likely represent one Arabidopsis gene with strong sequence similarity
to a group of 5' to 3' exoribonucleases (J .P. Kastenmayer, A. van Hoof and P.J. Green,
unpublished data) of which Xmlp from yeast is the best-characterized example. Xrnlp is

the main enzyme responsible for the degradation of at least some mRNAs in yeast

137

(Muhlrad and Parker, 1994), and its known substrates include premature nonsense codon-
containing mRNAs. In addition, the biochemical characteristics of XRN 1p, and the less
well-characterized 5' exonuclease 2 (Stevens and Poole, 1995; Poole and Stevens, 1995)
are similar to the classically defined "exonuclease I" class of enzymes from plants
(Bariola and Green 1997; J .P. Kastenmayer, A. van Hoof and P.J. Green, unpublished
data). Clearly, experimental data are required to show whether any of these Arabidopsis
genes play a role in (nonsense-mediated) mRN A degradation.

In addition to the ESTs mentioned above, dbEST also contains Arabidopsis

 

sequences with low, but possibly significant, similarity to other genes thought to be
involved in mRNA decay, such as yeast DCPl (Beelman et al., 1996) and zebrafish narI
(Gaiano et al., 1996), numerous EST s with high sequence similarity to RNA-binding
proteins and helicases, and ESTs with high sequence similarity to secretory RNases (data
not shown). This latter class includes EST 62B4T7, which is similar to fungal”
bifunctional nucleases (i.e. capable of degrading both RNA and DNA) (ML. Abler, D.M.
Thompson, A. van Hoof, N.D. LeBrasseur and P.J. Green, manuscript in preparation)
and likely represents the first cloned representative of the biochemically-defined
"nuclease I" enzymes (reviewed in Bariola and Green, 1997).

It seems clear that it will be relatively easy to identify RNases and RNA-binding
proteins (using a combination of biochemistry and sequence analysis). The real challenge
will be to differentiate activities involved in mRNA degradation from those involved in
other processes, and to identify the specific role of each trans-acting factor. Some of the
cloned genes mentioned above are being mapped (L.R. Danhof and P.J. Green, personal

communication) and thus may be found to be identical to genes identified by the genetic

138

selection described above. In addition, it is now theoretically possible to identify
insertion mutants in any cloned gene and this approach is also being pursued (N .D.

LeBrasseur and P.J. Green personal communication).

 

1 39
REFERENCES

Bariola PA, Green PJ (1997): Plant Ribonucleases. In Riordan JF, D'Alessio G (eds):
Ribonucleases: Structure and Function. Orlando, FL: Academic Press, pp. 163-190.

Beelman CA, Stevens A, Caponigro G, LaGrandeur TE, Hatfield L, Fortner DM, Parker

R (1996): An essential component of the decapping enzyme required for normal rates of
mRNA turnover. Nature 382: 642-646.

Gaiano N, Amsterdam A, Kawakami K, Allende M, Becker T, Hopkins N (1996):

Insertional mutagenesis and rapid cloning of essential genes in zebrafish. Nature 383:
829-832.

He F, Jacobson A (1995): Identification of a novel component of the nonsense—mediated
mRNA decay pathway by use of an interacting protein screen. Genes Dev. 9: 437-454.

Maquat LE (1995): When cells stop making sense: Effects of nonsense codons on RNA
metabolism in vertebrate cells. RNA 1:453-465.

Muhlrad D, Parker R (1994): Premature translational termination triggers mRNA
decapping. Nature 370: 578-581.

Nicholson AW (1997): Escherichia coli ribonucleases: Paradigms for understanding
cellular RNA metabolism and regulation. In Riordan JF, D'Alessio G (eds):
Ribonucleases: Structure and Function. Orlando, FL: Academic Press, pp. 1-49.

Peltz SW, Brown AH, and Jacobson A (1993): mRNA destabilization triggered by
premature translational termination depends on at least three cis-acting sequence elements
and one trans-acting factor. Genes Dev. 7: 1737-1754.

Poole TL, Stevens A (1995): Comparison of features of the RNase activity of 5'-

exonuclease-l and 5'-exonuclease-2 of Saccharomyces cerevisiae. Nucleic Acids Symp.
Ser. 33: 79-81.

PUIak R, Anderson P (1993): mRNA surveillance by the caenorhabditis elegans smg
genes. Genes Dev. 7: 1885-1897.

ROSS J (1995): mRNA stability in mammalian cells. Microbiol. rev. 59: 423-450.

Sengupta DJ, Zhang BL, Kraemer B, Pochart P, Fields S, Wickens M (1996): A

three-hybrid system to detect RNA-protein interactions in vivo. Proc. Natl. Acad. Sci.
USA 93: 8496-8501.

Stevens A, Poole TL (1995): 5'—exonuclease-2 of Saccharomyces cerevisiae. Purification
and features of ribonuclease activity with comparison to 5'-exonuclease-1. J. Biol. Chem.

140

270: 16063-16069.

Sugita M, Sugiura M (1996): Regulation of gene expression in chloroplasts of higher
plants. Plant Mol. Biol. 32: 315-326.

Tharun S, Parker R (1997): Mechanisms of mRNA turnover in eukaryotic cells. In
Morris, DR Harford, JB (eds): Modern Cell Biol 17: 1-81-199. Post-transcriptional Gene
Regulation. New York, NY: Wiley and Sons.

Tycowski KT, Shu M—D, Steitz J A (1996): A mammalian gene with introns instead of
exons generating stable RNA products. Nature 379: 464-466.

Weng Y, Ruiz-Echevarria MJ, Zhang S, Cui Y, Czaplinski K, Dinman JD, and Peltz SW
( 1 997): Characterization of the nonsense-mediated mRN A decay pathway and its effect

on modulating translation termination and programmed frameshifting. In Morris DR,
Harford JB (eds.): Modern Cell Biol 17: 241-263. Post-transcriptional Gene Regulation

New York, NY: Wiley and Sons.

Yen Y, Green P] (1991): Identification and properties of the major ribonucleases of
Arabidopsis thaliana. Plant Physiol. 97: 1487-1493.

 

APPENDIX

APPENDD(

GUT15 CDNAS FROM TOBACCO AND ARABIDOPSIS
CORRESPOND TO TRANSCRIPTS

WITH UNUSUAL METABOLISM AND A SHORT CONSERVED ORF.

Parts of this appendix have been previously published in Plant Physiology.

Reference: A van Hoof, JP Kastenmayer, CB Taylor and PJ Green (1997). GUT15
cDNAs from tobacco (Accession No. U84972) and Arabidopsis (Accession No.
U84973) correspond to transcripts with unusual metabolism and a short conserved ORF
(PGR97-048). Plant Physiol. 113: 1004

141

 

142

An important aspect of gene expression is regulation exerted at the level of
mRNA stability. The majority of transcripts in eukaryotic cells are stable, with half lives
on the order of hours. In contrast to these stable messages, a subset of transcripts is
rapidly degraded, with half lives on the order of minutes (Sullivan and Green, 1993).
Previously, tobacco Genes with Unstable Transcripts (GUTs) were isolated by differential
hybridization. The most unstable transcripts found in this study were those encoded by
the GUT15 gene (Taylor and Green, 1995).

The complete sequence of the original tobacco GUT15 cDNA was determined in
an effort to elucidate the function of the GUT15 gene (GenBank accession Number.
U84972). The GUT15 cDNA is unusual in that it does not contain a long open reading
frame (ORF); the longest ORF in any of the three reading frames is only 76 codons. To
determine whether the original GU T1 5 cDNA represented the functional GUT15
transcript, a transcript from a pseudogene, or was a cloning artifact, additional cDNA
clones were isolated from the same library. Eleven clones were partially sequenced and
could be divided into two classes. One class consisted of cDNAs that were identical in
sequence to the original clone, although their 5' and 3' ends were different. The other
class contained some small insertions and deletions and base substitutions (Taylor and
Green, personal communication), indicating that there are at least two GUT15 genes in
tobacco (confirming the results of Southern blotting; Taylor and Green, 1995). All of this
information combined suggested that the original cDNA was not a cloning artifact and
was likely to represent the functional transcript. 11 ESTs in dbEST that represent a
putative GUT15 homolog in Arabidopsis thaliana were also identified. Although the

overall degree of sequence similarity between the genes from Arabidopsis and tobacco is

 

143

not very high, they share many common characteristics. In addition, analysis of the EST
sequences combined with Southern blotting indicated that Arabidopsis contains only one
GUT15 gene. Therefore further analysis was concentrated on Arabidopsis, although most
of the results were confirmed in tobacco.

RNA gel blot analyses of Arabidopsis RNA using the GUT15 cDNA as a probe
resulted in two bands of approximately 1.2kb and 1.6kb of similar intensity (data not
shown). A comparable pattern was observed in tobacco, in which transcripts of
approximately 1.7kb and 1.9kb were detected (confirming the data of Taylor and Green,
1995). This pattern of transcripts was also seen when tobacco or Arabidopsis RNA
enriched for poly(A)+ RNA was examined (Figure 6-1 and data not shown), indicating
that transcripts of both sizes are polyadenylated. Analysis of all the Arabidopsis and
tobacco cDNA clones suggested that the larger of the two transcripts retained an intron
relative to the smaller transcripts. These putative introns have the characteristic features
of plant introns, such as a high AU content, and conserved splice sites. The intron is
conserved in sequence as well as position between Arabidopsis and tobacco. Preliminary
data from RNase H cleavage experiments (Kleene et a1, 1984) using oligonucleotides
complementary to the intron sequence also indicated that the putative intron was present
in the 1.6kb species seen in RNA blot analysis in Arabidopsis, but absent from the 1.2kb
species (Kastenmayer and Green, personal communication).

Taylor and Green (1995) reported that cycloheximide (chx) stabilized the smaller
GUT15 mRNA species in tobacco cells. chx also induces the smaller of the two
transcripts in Arabidopsis. The differential effect of chx on the two transcripts may be

indicative of a difference in their degradation pathways. Alternatively, the larger

 

 

Figure 6-1. GUT15 probes hybridize to two polyadenylated transcripts. Shown is an
RNA gel blot of 20 pg polyA' and 0.4 pg polyA+ RNA from tobacco BY2 cells probed
With a cDNA clone of the tobacco GUT 15 gene.

 

145
transcript may be a precursor of the smaller transcript that is slowly processed. Both
explanations are consistent with the disappearance of the larger transcript during
actinomycin D time courses as observed by Taylor and Green (1995).

Sequence analysis of the longest Arabidopsis cDNA clone (GenBank accession
number U84973) showed that it also lacks a long ORF, but the longest ORF in tobacco is
conserved in Arabidopsis (Figure 6-2A). The putative peptides do not show significant
sequence similarity with known proteins. In all three genes this ORF is located 3' of
multiple AUGs, which is unusual for an eukaryotic transcript. Therefore the encoded
peptide would not be expected to be produced in large amounts by the normal scanning
mechanism of translation (Kozak, 1989). However, it is conceivable that this ORF is
translated by an alternative translation mechanism (Jackson and Kaminski, 1995). If the
ORF is translated, the peptide may function as a signaling molecule that is needed in very
small amounts (as has been described for the product of the ENOD40 gene; van de Sande
et a1, 1996). To test this hypothesis, transgenic Arabidopsis plants were generated to
overexpress the peptide. Specifically, the GUT15 ORF was inserted between a double
enhanced 358 promoter and NOS polyadenylation signals, and the sequence just upstream
of the AUG was changed to the preferred context for plants. However, none of the 40
independent transgenic plants that were generated displayed an obvious phenotype. Thus
the available data provides no support for the hypothesis that the conservation of the
small ORF in the GUT15 gene is biologically important. It seems possible that the
GUT15 transcript itself, or an intron within it (Tycowski et a1, 1996), may be the
functional gene product, and conservation of the ORF may reflect underlying

conservation of important nucleotides.

146

A
Tobacco MMGALAWQV. . . TIPNKVSIFCFG. RGDGTGILPG. .APLFVSS
||l||||| -: -| ==-== l=-| ==|| -|| 111
Arabidopsis MSGALAWQVKKLILNKKKILWVLERRSHGLLFFPGISSPLCVSS
Tobacco RLLFSSL ..... FPRYYTQDQYHQERHIRLL
| || |......... .
Arabidopsis CLCSISLPALSHFISFLNAHIHSKTDHKQSL
B
Clone Species

GUT 1 5 tobacco CCGACCUUUGCCAUGAUGGGUGCGCUCGCAUGGCAGGUCA
GUT 1 5 Arabidopsis C CGACCUUUGC CAUGUCAGGUGC GCUUGCAUGGCAGGUCA
CR20 cucumber CCGACCUUUGCCAUGACAGGUGCGCUUGCAUGGCAGGUCA
CR20 A rabidopsis CCGACCUUUGCCAUGACAGGUGCGCUUGCAUGGCAGGUCA

St'c 1 a potato C C GGCC UUUGC CAUGACGGGUGC GCUCGCAUGGCAGGUCA
sre 1 b potato CCGGCCUUUGCCAUGCCGGGUGCGCUCGCAUGGCAGGUCA
sre 1 c potato C C GGC CUUUGUCAUGCCGGGUGCGCUCGCAUGGCAGGUCA
Consensus CCGRCCUUUGCCALENCRGGUGCGCUYGCAUGGCAGGUCA

Figure 6-2. Comparison of the conserved regions in GUT15 genes. A. Comparison of the
conserved open reading frame in tobacco and Arabidopsis GUT15 genes. The deduced
amino acid sequences were aligned using the bestfit program (genetics computer group,
Madison, WI), using the default parameters. B. Alignment of a 40 nucleotide segment of
GUT15 genes from tobacco and Arabidopsis and related sequences from CR20 genes of
Arabidopsis (D79218) and cucumber (D79216) and from three cDNA clones that
probably represent different alleles of one gene from potato (W egener and Scheel,
personal communication). The consensus shows nucleotides that are identical in at least
six of the seven sequences. R indicates A or G, Y indicates U or C, N indicates a
nonconserved nucleotide. The AUG underlined in the consensus corresponds to the
initiation codon for the open reading frame show in A. The two nucleotides deviating
from the consensus are indicated in bold.

147

As a further step toward functional analysis the Arabidopsis GU T 1 5 gene has
been mapped to chromosome 2, ch distal of marker m216 (LR Danhof, A van Hoof and
PJ Green, unpublished data), using the recombinant inbred lines developed by Lister and
Dean (1993). Several mutants for which the corresponding genes have not been cloned
map close to this location. Further research is needed to investigate the possibility that
the phenotype of one of these mutants is caused by a mutation in the GUT15 gene.

Recently Teramoto et al. (1996) reported the isolation of genomic and cDN A
clones for the CR20 gene from cucumber that was isolated by a differential screen for
cytokinin repressed transcripts. Similar clones were also isolated from potato during a
differential screen for transcripts induced during pathogen attack (W egener and Scheel,
personal communication), but preliminary results indicate that the Arabidopsis GUT15
gene is not induced during pathogen attack (Kastenmayer, Lawton, van Hoof, Ryals and
Green, unpublished data). The cucumber and potato clones share many of the unusual
features of GUT15 clones from tobacco and Arabidopsis. In particular, several transcripts
were detected when these clones were used as probes on RNA gel blots (Teramoto et al.,
1996; Wegener and Scheel, personal communication). Several cDNA clones from
cucumber were isolated and appeared to differ by the presence or absence of one or more
introns. Teramoto et al. also isolated a cDNA from Arabidopsis (AtCRZO-I) that is
distinct from, but similar to the Arabidopsis GUT15 gene. AtCRZO-I and GUT15 share
many of the same characteristics and very limited sequence similarity. The sequence
similarity is not sufficient for probes for GUT15 to cross-hybridize to AtCR20-1 since
only one gene was detected in Southern blotting. These data combined indicate that

GUT15 genes from Arabidopsis and tobacco are members of a large class of genes. The

148
ORF that is conserved in Arabidopsis and tobacco GUT15 genes is not conserved very
well in some of the cucumber and potato clones (although the AUG codon is conserved in
all clones). A small segment of these RNAs is highly conserved (Figure 6-2B) and may
be (an essential part of) the functional gene product. A similar sequence is not present in
any other sequences in GenBank, indicating that this noncoding RNA may be unique to
plants. Outside this 40nt area the clones do not share extensive sequence similarity. The
presence of a small region of sequence conservation in the RNA sequence, combined
with the absence of a highly conserved ORF suggests that these genes are transcribed into
noncoding RNA. The function of this RNA remains unknown and additional

experiments will be needed to resolve this.

149

REFERENCES

Jackson R.J., and Kaminski A. (1995). Internal initiation of translation in eukaryotes:
The picomavirus paradigm and beyond. RNA 1: 985-1000.

Kleene K.C., Distel R.J., and Hecht NE. (1984). Translational regulation and
deadenylation of a protanrine mRNA during spermiogenesis in the mouse. Devel. Biol.
105: 71-79.

Kozak M. (1989). The scanning model for translation: an update. J. Cell Biol. 108: 229-
241.

Lister C. and Dean C. (1993). Recombinant inbred lines for mapping RFLP and
phenotypic markers in Arabidopsis thaliana. Plant J. 4: 745-750.

Sullivan ML, and Green P.J. (1993). Post-transcriptional regulation of nuclear-encoded
genes in higher plants: the roles of mRNA stability and translation. Plant Mol. Biol. 23:
1091-1104.

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

Teramoto H., Toyama T., Takeba G., and Tsuji H. (1996). Noncoding RNA for CR20, a
cytokinin-repressed gene of cucumber. Plant Mol. Biol. 32: 797-808.

Tycowski K.T., Shu M-D., and Steitz J .A. (1996). A mammalian gene with introns
instead of exons generating stable RNA products. Nature 379: 464-466.

van de Sande K., Pawlowksi K., Czaja 1., Wienke U., Schell J ., Schmidt J., Walden R.,
Matvienko M., Wellink J ., van Kammen A., Franssen H., and Bisseling T. (1996).

Modification of phytohormone response by a peptide encoded by ENOD40 of legumes
and a nonlegume. Science 273: 1234-1236.

 

 

 

1111111111111111111111111111111111111111111111111111111111