. u . 461...”
”4‘33un 47'.» WW:
V . t. . $79....

3.

ban a

.u

'0. it”
I ‘ .D
o I
4 .9. .
. 23:35,;
.2. n x

 

5..
up;

3.3%".

 

‘5.bm‘. Al

mafia? r:

s

..!v04 .

.1... ix). .

. 1...? .. l n.

3.3.9:}.usxta u...

a q 533.21%]...1‘
uluv .
a...)

L .

7?. ' if

 

 

.1.
3 .‘...li.$

 

 

 

52:? .2. ..

 

TTiE'\ib

n . A
,,_CO+..

 

LIBRAHY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

Regulation and function of Arabidopsis
thaliana secreted ribonucleases

presented by

Nicole D. LeBrasseur

has been accepted towards fulfillment
of the requirements for

Ph . D . degree in Genetic s

W% 2/

Dated 2/14 101

MS U i: an Affirmative Action/ Equal Opportunity Institution 0 12771

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/Dateouopes-nts

 

REGULATION AND FUNCTION OF ARABIDOPSIS T HALIANA
SECRETED RIBONUCLEASES

By

Nicole D. LeBrasseur

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Genetics Program

2001

ABSTRACT
REGULATION AND FUNCTION OF ARABIDOPSIS THALIANA
SECRETED RIBONUCLEASES
By

Nicole D. LeBrasseur

While secreted ribonucleases (RNases) have been well studied at the enzymatic and
structural levels, little is known regarding their biological functions. One family of
secreted RNases, the RNase T2 family, is particularly widespread, with members
throughout various kingdoms. In recent years, many plant members of this superfamily
have been identified. Gametophytic self-incompatibility in several plant groups involves
the activity of S-RNases, one subfamily of plant T2 RNases. Another subfamily, the S-
like RNases, are found in self-compatible as well as self-incompatible plant species,
indicating that they have different functions than have S-RNases. Expression patterns of
S-like RNases in many species led to the suggestion that these enzymes are involved in
remobilization and recycling of phosphate from nucleic acid sources. The project
described in this thesis involves the study of secreted RNases in the model plant
Arabidopsis thaliana, which has five members of the S-like subfamily. I found that
several Arabidopsis RNases are induced in wounded tissues through an as-yet
unidentified signaling mechanism. One S-like member, RN81, is induced both locally
and systemically by wounding through a novel signal transduction pathway. The

regulation of RNSI expression was studied using several reporter constructs. I found that

transcriptional regulation accounts for the majority of the wound response of this gene.
Several other factors were shown to affect RNSI transcript levels, including abscisic acid,
salt stress, and heat shock.

In addition to regulatory mechanisms controlling RN51 transcription, the function of
RNSl was also studied. Tools used for these studies include plants with constitutively
high RNSl activity and T-DNA mutant lines with insertions in the RNSI gene. One of
these mutants, rnsI-Z, has little or no RNSl activity. Conditions that affect RNSI
transcript accumulation provided a starting point for the analysis of the mutant and
overexpressor. The observation that the overexpressors grew better than wild type when
supplied with RNA as the sole source of Pi supports the hypothesis that RNSl is involved
in P, remobilization. Analysis of the mutant and overexpressor also revealed that RNSl
activity levels affect root length in Arabidopsis. This effect was not dependent on P,
concentration and indicated that secreted RNases are involved in unexpected processes.
Certain phenotypes of the rnsI-Z plants resembled those of the my] A mutant. Rnyl, the
only RNase T2 enzyme found in Saccharomyces cerevz'seae, has been proposed to
function in regulating aspects of membrane stability and permeability. Therefore, I
propose that RNSl may also be involved in similar processes.

Finally, the analysis of an Arabidopsis mutant with altered activity levels of several
RNA-degrading enzymes is presented. Although one of the increased activities resembles
RNSl, I demonstrate that this activity is not RNSl. The altered activities are specific to
stems of the mutant, as are its morphological phenotypes. This mutant, arpl, provides an
insight into the involvement of RNases in stem-associated processes and stem

development.

ACKNOWLEDGMENTS

Numerous thanks are owed to many people who contributed in one way or another to this
work. First and foremost, I must extend my appreciation to Dr. Pam Green for the
innumerable ways in which she has helped me in the past 6 years, for taking the chance
on me as an inexperienced graduate student, and for supporting me even when I chose an
"alternative" scientific career. Not all mentors would have been so supportive. Thank you
to members of my guidance committee, Drs Hans Kende, Mike Abler, Mike Thomashow,
and Sheng Yang He, for helpful suggestions and moral support. Mike Abler was also
integral in the arpl story and in my earliest exposure to benchwork; I have never
forgotten his advice. Thanks to those who helped with experiments throughout this thesis,
including Julie Zwiesler—Vollick for the help with the bacterial growth curves, Miguel
Pérez-Amador for the microarray that got me started on wounding, Ted Farmer for great
discussions and several contributions to the wounding story, Christie Howard for RNSI
constructs, and Daniel Cook for blots of various stresses included in Chapter 3. For
photographic and figure-making services, Marlene Cameron and Kurt Stepnitz were both
patient and helpful. Thanks, Linda Danhof, for many activity gels, technical assistance,
and direction of an army of undergraduate slaves (whom I would also like to thank,
especially, Lindsey, Amanda, Robert, lance, and Courtney). Much of Chapter 4 would
not have been included without the talents of a special undergraduate, Tracey Millard, to
whom I am ever grateful for her abilities and keen interest in plant physiology.

To all Green lab members past and present, I can never thank you enough. You have

iv

been friends, colleagues, mentors, and surrogate family all in one. To Preet Lidder and
Rodrigo Gutierrez, thanks for being great roommates (the back room rules!). To Jay
De Rocher and Pauline Bariola, thanks for the advice, direction, and friendship in the
early years. Special thanks to Gustavo Mac lntosh, for help with experiments, for the
many papers lefi on my desk, for keeping me on the right track the past three years, and
for being my best friend. I wouldn't have made it without you. I also owe debt to many
PRL members, who have helped scientifically and otherwise. I have never found a more
friendly and productive department than the PRL. I would like to say a personal thank
you to my big sister, Antje Heese—Peck, who I miss very much. Thank you to the darts
crew, especially John Froelich, John Scott-Craig, Scott Peck, and Tom Newman;
Mondays will never be the same. Finally, thank you to my brother, Brian, and my
parents, Robert and Lauraly, for years of support, encouragement, and love across the

miles. I am lucky to have you.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. ix

LIST OF FIGURES ............................................................................................................. x

CHAPTER 1

INTRODUCTION: PLANT RNA-DEGRADING ACTIVITIES ...................................... l
Secreted ribonucleases and the RNase T2 family ........................................................ 2
Function: the S-RNases ............................................................................................... 3
The S-like RNase subfamily: general features ............................................................ 4
Plant nuclease l-type enzymes .................................................................................... 6
Regulation and putative functions of S-like and nuclease I enzymes in plants .......... 6
Dissertation topic and thesis overview ...................................................................... 14

CHAPTER 2

LOCAL AND SYSTEMIC WOUND INDUCTION OF RNASE AND NUCLEASE
ACTIVITIES IN ARABIDOPSIS: RNSl AS A MARKER FOR A JASMONIC ACID-

INDEPENDENT SYSTEMIC SIGNALING PATHWAY ............................................... 15
Abstract ......................................................................................................................... 16
Introduction ................................................................................................................... l 7
Results ........................................................................................................................... 19

Wounding induced several RNase activities in Arabidopsis .................................... 19
Multiple bifunctional nuclease activities are increased by wounding ....................... 22
RNSl activity is induced by wounding locally and systemically ............................. 24
The induction of RNSI and the nucleases is independent of j asmonic acid ............. 26
RNSI is not induced by the oligosaccharide pathway ............................................... 30
Discussion ..................................................................................................................... 32
Materials and Methods .................................................................................................. 38
Plant materials and treatments ................................................................................... 38
RNA extraction and RNA gel blot hybridization ...................................................... 40
Protein extraction and detection of RNase and DNase activities .............................. 41

CHAPTER 3

REGULATION OF RNSI EXPRESSION ........................................................................ 42
Abstract ......................................................................................................................... 43
Introduction ................................................................................................................... 44
Results ........................................................................................................................... 45

Run-on analysis of RNSI expression ........................................................................ 45
Regulation of RNSI expression in response to wounding ........................................ 47
Use of promoter constructs to analyze signals mediating RNSI induction ............... 52
Discussion ..................................................................................................................... 59
Sequence requirements for the wound induction of RNSI ........................................ 59
Response of RNSI to various biotic and abiotic stimuli ........................................... 61
Materials and Methods .................................................................................................. 64
Nuclei preparations and run-on analyses .................................................................. 64

vi

Cloning ...................................................................................................................... 66

Plant transformation .................................................................................................. 67
Plant treatments ......................................................................................................... 67
RNA and protein analyses ......................................................................................... 68
Luciferase analysis .................................................................................................... 68
CHAPTER 4
RNSl FUNCTION: ANALYSIS OF AN RNSI MUTANT AND OVEREXPRESSOR.69
Abstract ............................................................................................ 70
Introduction ................................................................................................................... 71
Results ........................................................................................................................... 72
Isolation of ms] mutants and RNSI overexpressing plants ...................................... 72
The rnsl-Z mutant does not have increased levels of anthocyanins ......................... 77
RNSl activity appears to affect plant growth on RNA ............................................. 79
RNSl activity affects plant size ................................................................................ 81
Absence of RNSl does not increase susceptibility of the plant to a bacterial
pathogen .................................................................................................................... 89
Discussion ..................................................................................................................... 92
Involvement of RNSl in phosphate remobilization .................................................. 92
Is RNSl involved in defense mechanisms? .............................................................. 94
RNSI activity negatively correlates with root length ............................................... 95
Materials and Methods .................................................................................................. 97
Plant materials ........................................................................................................... 97
Plasmid construction ................................................................................................. 98
Identification of ms] T-DNA insertional knock-out mutants ................................... 99
Protein and RNA analyses ....................................................................................... 100
DNA extraction and DNA gel blot hybridization ................................................... 100
Anthocyanin assays ................................................................................................. 101
Bacterial growth curves ........................................................................................... 101
CHAPTER 5
ANALYSIS OF AN ARABIDOPSIS T HALIANA MUTANT WITH AN ALTERED
RNASE PROFILE .......................................................................................................... 102
Abstract .......................................................................................... 103
Introduction ................................................................................................................. 104
Results ......................................................................................................................... 105
An RNase-activity gel screen for mutants affecting the Arabidopsis RNase profile
................................................................................................................................. 105
The 33-kDa RNase activities are bifunctional nucleases ........................................ 109
Genetic characterization of the arp mutants ........................................................... 109
Morphology of the arpI mutant .............................................................................. 110
The arp] mutation induces an RNase activity in addition to the 33-kDa nuclease
activities .................................................................................................................. l 12
The arp] RNase phenotype is stem associated ....................................................... 119
Discussion ................................................................................................................... 120
Mutants affecting single and multiple RNase activities .......................................... 120

vii

The 33-kDa activities are bifunctional nucleases .................................................... 122

Possible functions for the 33-kDa nucleases ........................................................... 123
arp] is a regulatory link between the 33-kDa nucleases and an RNase activity 125
The app] phenotypes are associated with stems ..................................................... 125
Materials and Methods ................................................................................................ 127
Plant materials ......................................................................................................... 127
RNase and DNase activity gels ............................................................................... 127
Mutant growth and development ............................................................................. 127
DNA and RNA extraction and gel blot analyses ..................................................... 128
Immunoblot analysis ............................................................................................... 128
CHAPTER 6
CONCLUSIONS AND FUTURE PROSPECTS ............................................................ 129
REFERENCES ................................................................................................................ 135

viii

LIST OF TABLES

Table 4—1. Statistical analysis of fresh weight of WT, msI-Z, and 35S-RNSI seedlings 85
Table 4-2. Statistical analysis of root length in ms] -2, WT Ws, 35S-RNS1 and WT Col

seedlings ............................................................................................................................ 86
Table 5-1. Statistical analysis of morphological differences between WT RLD and arpI
plants ............................................................................................................................... 113

ix

LIST OF FIGURES

Figure 2-1. Multiple RNase activities are induced by wounding ................................ 20
Figure 2-2. Bifunctional nuclease activities are induced by wounding. ...................... 23
Figure 2-3. RNSl activity and transcript increase transiently in wounded and in

systemic, non-wounded leaves. ............................................................................. 25
Figure 2-4. RNSl and nuclease induction by wounding is independent of jasmonic

iacid...........
Figure 2-5.
Figure 3-1.
Figure 3-2.
Figure 3-3.
Figure 3-4.

Figure 3-5.

Figure 3-6.

Figure 3-7.

......................................................................................................... 28
OGA-rich fractions do not induce RNase activities. ................................. 31
Run-on analysis of RNSI in response to P; starvation ............................... 46
Constructs used to examine the regulation of RNSI . ................................ 48
Differential response of RNSI transcribed sequences to wounding .......... 50

The RNS] promoter confers wound inducibility to reporter transcripts. .. 51

5' UTR sequences are not required for the wound responsiveness of the

RNSI promoter. ..................................................................................................... 53
Dehydrated plants do not induce the RNSI promoter. .............................. 54
RNSI transcript accumulates in response to salt stress. ............................ 54
ABA induces the RNSI promoter. ............................................................ 57

Figure 3-8.
Figure 3-9.
Figure 3-10.
Figure 4-1.
Figure 4-2.
Figure 4-3.
Figure 4-4.
Figure 4-5.

Figure 4-6.
medium..........

Mutants in ABA signaling and biosynthesis induce RNSl by wounding. 58

Heat shock inhibits the wound-induction of luciferase activity. ............... 60
ms] mutants and overexpressor constructs ............................................... 73
RNSl activity in the rnsl-I and ms] -2 mutants. ...................................... 75
DNA and RNA gel blot analyses of rnsI-2 mutant plants. ....................... 76
RNSl activity in overexpressing plants. ................................................... 78
Anthocyanin content in mutant ms lines ................................................... 80

Growth of rns‘I-Z and 3SS-RNSI seedlings on RNA-containing
. ................................................................................................... 82

Figure 4-7. Root phenotype of RNSI mutant and overexpressor ................................. 84
Figure 4-8. Absorption of dyes in wild-type Columbia (WT) and rnsI-Z mutant root

cells........... ......................................................................................................... 88
Figure 4-9. Growth of Pst DC3000 in leaves of 35S-RNS1 and rnsl-Z plants. .......... 91
Figure 5-1. arp mutants are altered in RNase activity profiles .................................. 107
Figure 5-2. The 33-kDa RNase activities also degrade DNA. ................................... 108
Figure 5-3. Morphological differences in the phenotypes of wild-type and arp] mutant

plants............ ................................................................................................... 111
Figure 5-4. The arpI mutant has increased levels of an activity that comigrates with

RNSl....... ................................................................................................... 115
Figure 5-5. RN51 transcript and protein is not detected in arp] stems ...................... 116
Figure 5-6. A previously unidentified 23-kDa band is increased in arpI .................. 118
Figure 5-7. The cup] RNase phenotype is stem specific. .......................................... 121
Figure 6-1. Model for the control of membrane permeability by RNases. ................ 132

xi

CHAPTER 1

INTRODUCTION: PLANT RNA-DEGRADING ACTIVITIES

The class of proteins known as ribonucleases (RNases) encompasses a diverse group of
enzymes whose catalytic activity involves the degradation and catabolism of RNA. There
are numerous ways in which one might categorize RNases, including primary structure
homology, cellular localization, and enzymatic cleavage mechanism or other biochemical
properties, to name a few. RNases are involved in many processes in cellular function.
Perhaps their most obvious function is the controlled processing and degradation of
mRNA as a mechanism of gene regulation, although to imagine that RNases are restricted
to such processes is likely a major underestimation of their firnctions. In fact, in most
systems, the genes encoding proteins involved in mRNA degradation are still unknown.
In contrast, many RNA-degrading activities that are not likely to be involved in the
processing or degradation of mRNA have been characterized and the genes encoding
them cloned, as is the case in the field of plant RNases. Likely the most well-
characterized of these activities are those that fall into the RNase T2 superfamily.

More than one review has been published in recent years in the field of plant RNases
and/or mRNA degradation (examples include Bariola and Green, 1997; Gutierrez et al.,
1999; and Johnson et al., 1998). Therefore, this chapter will not attempt to reiterate such
reviews. Instead, it will focus on recent findings in regard to the expression patterns and
putative functions of plant secreted RNases and the signaling mechanisms involved in
their regulation, with particular attention paid to the S-like RNases, one class of plant T2

RNases.

Secreted ribonucleases and the RNase T2 family

Biochemists have long used secreted RNases for structural and enzymatic studies, due to

the relative ease in their purification and the availability of large quantities of proteins.

IQ

RNase A, for example, is a low molecular weight, stable enzyme that has been purified in
large quantities from many vertebrate species. Perusing any review on pancreatic
RNases, which mostly focus on RNase A family members, will enlighten the reader on
several aspects of protein structure and enzymatic mechanisms (see, for example,
Cuchillo et al., 1997) and is therefore quite telling regarding the focus of past studies on
these particular subjects. While much is known on base-specificity, protein folding
structures, and role of specific amino acids in catalysis, little is known about the
physiological firnctions of these enzymes. Further, even fortuitous findings of "special"
biological actions of RNases (D'Alessio et al., 1991 ), including antitumor and angiogenic
activities, have not been explained biologically.

The T2 superfamily of RNases has also been well-studied at the enzymatic and
structural levels (Irie, 1997). First identified in the fungus Aspergillus oryzae (Sato and
Egami, 1957), RNase T2 and its relatives are secreted acid endonucleases with no base-
specificity, sharing two conserved stretches of amino acid residues important in the
enzymes' catalytic sites (Irie, 1997). Unlike the RNase A superfamily, members of which
have so far been identified only in vertebrates (Beintema et al., 1997), T2 enzymes are
incredibly widespread. To date, members of this family have been found in nearly every
system examined for their presence, encompassing viruses, bacteria, plants, and animals
(Irie et al., 1997). The ubiquitous distribution of T2 RNases may be interpreted as the

enzymes' having both an ancient origin and critical function(s) (Taylor and Green, 1991).

Function: the S-RNases

Despite the apparent necessity for the activity of these enzymes, very little has been

demonstrated regarding their biological functions. The exception is the S-RNascs, a class

of plant T2 RNases, whose activity is required for the process of self-incompatibility in
several plant families, including the Solanaceae (reviewed in McCubbin and Kao, 2000).
Briefly, when the S-allele of a pollen grain growing through the style of the mother plant
matches either of the two S—alleles in the pistil, growth of the pollen tube is inhibited.
RNase activity has been shown to be essential for this incompatibility system (Lee et al.,
1994). However, the mechanism through which the S-RNase imparts incompatibility is
still unknown. Available evidence indicates that the S-RNases somehow act as cytotoxins
in the incompatible tubes (McClure et al., 1989). The pollen component of the interaction
has not been identified, nor is it known at what site the two components interact. The two
leading models in the field predict that either 1) the S-RNases are specifically imported
into incompatible pollen tubes or 2) nonspecific S-RNase uptake occurs, with subsequent
specific inactivation of all compatible RNases or some other similar modification (Parry
et al., 1997). Recent evidence supports the latter hypothesis, in that S-RNases appear to
be taken up by pollen tubes in a genotype-independent manner (Luu et al., 2000).

In summary, the S-RNases are the sole members of the T2 RNase superfamily who
have been assigned a function. Yet, there is much to be understood regarding the

mechanism of the incompatibility reaction.

The S-like RN ase subfamily: general features

Soon after the identification of the S-RNases as members of the T2 family, several groups
reported similar proteins in self-compatible plant species, including Arabidopsis, tomato,
and Zinnia. Although these S-like RNases share certain structural features with the
S-RNases, conserved regions specific to each group can also be found (for a

comprehensive review of S-like RNases, see Bariola and Green, 1997). These enzymes

are obviously not involved in self-incompatibility, but appear to play important roles
throughout the plant kingdom, as S-like members have now been identified in both self-
compatible and self-incompatible species. The Arabidopsis thaliana genome contains
five genes, RNS] to RNS5, with high similarity to the S-like RNase gene family (Taylor
and Green, 1991; GO Maclntosh, N.D. LeBrasseur, and P.J. Green, unpublished), and
RNase activity has been demonstrated for the products of three of the RNS genes (Taylor
et al., 1993; Bariola et al., 1994).

Studies on S-like RNases have for the most part focused on gene expression or
activity patterns, as will be discussed below. Limited biochemical and structural
information is also available. As mentioned, the S-like RNases are secreted enzymes.
While many are extracellular, including RNSl (Bariola et al., 1999), some have been
shown to be retained intracellularly. These include RN82 (Bariola et al., 1999) and
several activities in tomato that are localized to the vacuole or endoplasmic reticulum
(Loffler et al., 1992; Lehmann et al., 2001). Alternate cellular localization indicates that
multiple members of the S-like subfamily in the same plant may carry out separate
functions or the same functions in different compartments.

The crystal structure of the tomato extracellular S-like RNase LE has been
determined at 1.65 A resolution (Tanaka et al., 2000). Other T2 enzymes that have been
crystallized include an S-like RNase from bitter gourd (Nakagawa et al., 1999) and the
fungal enzyme RNase Rh (Kurihara et al., 1996). Comparison of these and other T2
enzyme structures has revealed two structural groups: the animal/plant and the firngal
subfamilies (Tanaka et al., 2000). However, the groups appear to share catalytic

mechanisms despite these structural differences (Tanaka et al, 2000).

Plant nuclease I-type enzymes

Early classification of plant RNA-degrading activities based on biochemical properties of
the enzymes identified several groups: RNase l, RNase II, nuclease I, and exonuclease I
(Farkas, 1982; Wilson, 1982). While most T2 RNases appear to fall within the RNase I
class, many plant activities have been identified with characteristics placing them in the
nuclease I family, including activities in tobacco, barley, mung bean, zinnia, and rye
(reviewed in Bariola et al, 1997). Nuclease l enzymes are also extracellular heat-stable
proteins, but unlike T2 RNases, nucleases have the ability to degrade both RNA and
single-stranded (ss) DNA. In general, they also have higher molecular weights, in the
range of 31 to 42 kDa, are glycosylated, and have acidic pH optima. They are sensitive to
EDTA and also require Zn2+ for activity and stability (Fraser and Low, 1993).

Until recently, sequence information was not available for many plant nucleases.
Determination of the amino acid sequences of two fungal nucleases, P1 from Penicillium
citrum (Maekawa et a1, 1991) and S] from Aspergillus oryzae (lwamatsu et a1, 1991), has
allowed the cloning of genes encoding similar proteins in plants. Nuclease 1 cDNA
sequences are now available for daylily (Panavas et al., 1999), celery (Yang et al., 2000),

barley, zinnia, and Arabidopsis (Aoyagi et al., 1998; Pérez-Amador et al., 2000).

Regulation and putative functions of S-like and nuclease I enzymes in plants

To date, fluctuations in RNase activity levels or gene expression have provided the most
useful data for predicting RNase function. Through elucidating the biotic and abiotic cues
that correlate with increases in RNase activity, we have been able to postulate functions
that secreted RNases play in the plant. The following sections will briefly highlight

significant examples and recent findings concerning RNase and nuclease induction under

various environmental and developmental conditions and summarize conclusions drawn

from these expression patterns.

RNases and phosphate starvation

Induction of S-like RNases and the genes encoding them is strongly associated with
growth on low concentrations of inorganic phosphate (Pi). Pertinent examples include the
Arabidopsis RNS] and RN32 genes and proteins (Bariola et al., 1994, 1999), and intra-
and extracellular tomato activities RNases LX (Bosse and Keck, 1998) and LE
(Niimberger et al., 1990), respectively. In addition to the induction of enzymatic activity,
it was shown that the LE and LX mRNAs also accumulate upon P, starvation (Keck et al.,
1995). The genes encoding the Nicotiana alata RNase NE (Dodds et al., 1996), and the
Prunus dulcis RNase PD2 (Ma and Oliveira, 2000), both of which are highly similar to
RNS], are also responsive to low Pi.

ln-depth analyses of signal transduction pathways controlling Pi-sensing and
response to low P, have only recently begun. Studies in the tomato cell culture system
support the existence of both intracellular and extracellular mechanisms for sensing P,
levels and initiating RNase induction (Glund et al., 1990; Kock et al., 1998). Maintenance
of Pi homeostasis in plants is probably controlled by at least two signaling mechanisms,
one at the cellular level and another regulating multiple organs and likely originating
from the shoots (Raghothama, 2000). There is little direct evidence that plant hormones
are involved as primary signals in the Pi-response, but ethylene and auxin may be
involved in altering root architecture and root hair growth (Lynch and Brown, 1997).
However, cytokinins repress the expression of Pi-starvation inducible genes in

Arabidopsis, but do not affect root hair alterations that are controlled by local Pi

concentrations, indicating that cytokinins are possibly involved in the repression of
systemically controlled Pi-starvation responses (Martin et al., 2000). Cytosolic calcium
levels and the activity of Ca2+-ATPases are thought to be involved in regulating the
response and adaptation of plants to low P, (Lynch and Brown, 1997). Homeodomain
leucine zipper proteins have been shown to bind to the phosphate response domain of the
soybean VspB gene, placing these types of transcription factors as likely candidates for
effectors of gene induction (Tang et al., 2001). Recent screens have identified mutants in
the Pi-response pathway that should allow us to dissect the signal transduction pathways
in detail in the future. At least one of these mutants is unable to induce RNSl under low
Pi conditions (Chen et al., 2000). Future screens for constitutive Pi-starvation response
mutants, such as those proposed by Ticconi et al. (2001 ), will add to the tools available to
aid us in the understanding of signal pathways involved in plant molecular and
developmental responses to P, limitation.

The ever-growing number of RNases demonstrated to be responsive to low Pi has led
to the hypothesis that S-like RNases are part of a rescue system employed by plants to
recycle Pi when environmental pools are limiting (Goldstein et al., 1989). In conjunction
with acid phosphatases and phosphodiesterases, RNases could be involved in the
generation of Pi through the hydrolysis of RNA (Glund and Goldstein, 1993).
Extracellular secretion of a fungal phytase from Arabidopsis roots has been shown to
increase phosphorous nutrition in the plant (Richardson et al., 2001). Similarly, RNases
secreted from the root system to the rhizosphere could breakdown organic matter in the
surrounding soil into Pi. A similar function has been attributed to two secreted fungal

nucleases (Fraser and Low, 1993). Alternatively, RNases could be involved in recycling

cellular material from dead or dying cells, either internally (in the case of vacuolar and
ER-localized RNases) or externally (in the case of secreted RNases).

Experiments done in Arabidopsis have supported this theory. RNSl and RN82
activities were diminished using antisense techniques, and each line was found to have
increased anthocyanin contents. The phenotype was particularly dramatic when the
seedlings were starved for P. (Bariola et al., 1999). The authors suggested that this effect
might be due to the plants' reduced abilities to acquire Pi, especially under low Pi-levels,
although it could not be ruled out that the plants were somehow otherwise stressed,

leading to high levels of anthocyanins.

C ell-death pathways

Programmed cell death (PCD) is an integral part of many plant developmental programs
and environmental responses, including, but not limited to, senescence, the hypersensitive
response (HR) during defense against pathogens, pollination, and germination (for recent
reviews, see Beers and McDowell, 2001; Jones, 2001). There are numerous examples of
S-like RNase and nuclease l-type activities induced during such processes.

The final stages of xylogenesis have been a paradigm for studying PCD mechanisms
(for a review of xylogenesis, see Roberts and McCann, 2000). This process has been
well-studied in the zinnia cell culture system, in which mesophyll cells can be induced to
transdifferentiate into tracheary elements (TE; reviewed in Fukuda, 2000). This process is
accompanied by induction of nuclease activities, including a 43-kDa protein that is
apparently involved in autolysis, the final stage of vascular differentiation in plants
(Thelen and Northcote, 1989). Two cDNA clones, ZRNaseI and ZRNaseII, have been

isolated from differentiating zinnia cells and encode RNases with similarity to the S-like

RNase family. ZRNaseI is induced during late stages of TE differentiation (Ye and
Droste, 1996). The in vivo function, as well as the substrates, of nuclease and RNase
activities during TE formation is still a matter of speculation.

Nuclease and RNase activities are also associated with other cell death processes,
such as senescence and the hypersensitive response. Pollination-induced petal senescence
in petunia causes increases in RNase and ssDNase activities (Xu and Hanson, 2000).
Activities of the nuclease I family are induced in senescing tissues in Arabidopsis and
zimria (Pérez-Amador et al., 2000). In Arabidopsis, RN52 and RN53 transcripts
accumulate during leaf senescence (Taylor et al., 1993; Bariola et al., 1994). The gene
encoding RNase LX is also induced in senescing tomato leaves (Lers et al., 1998), and
the protein is expressed during xylem differentiation, germination, and senescence
(Lehmann et al., 2001). In tobacco, the hypersensitive response resulting from both TMV
and bacterial infection is associated with induction of extracellular nuclease activities
(Mittler and Lam, 1997). Barley aleurone cells contain several nuclease activities and an
S-like RNase that are induced by gibberellic acid and repressed by abscisic acid (Brown
and Ho, 1987; Path et al., 1999; Rogers and Rogers, 1999). The timing of induction of
these activities correlates with hydrolysis of nuclear DNA and the progression of cell
death. It has been hypothesized that the barley nucleases are involved in the degradation
of the endosperm in response to seed germination (F ath et al., 1999).

It is possible that RNase and nuclease activities are critical components of PCD,
required for the progression of the developmental pathway. It has been suggested that
nucleases could be involved in the degradation of nuclear DNA during PCD (F ath et al.,

1999; Mittler and Lam, 1997). However, most activities studied appear to be either

10

extracellular or associated with the vacuole or ER. Thus, they may be involved in late
stages of the process, after the commitment step of vacuolar lysis, for example (Obara et
al., 2001), or alter lysis of the cell itself. In such a role, they could be components of
recycling pathways engaged in order to mobilize useful molecules from dying cells to
growing parts of the plant.

A more interesting hypothesis has been proposed by Lehmann et al. (2001). After
demonstrating that RNase LX is retained in the ER, they suggested that hydrolytic
enzymes, such as proteases and RNases, found in distinct ER domains or ER-derived
structures such as ricinosomes play an important role in controlled degradation processes.
Upon swelling and rupture of ER- or ER-derived membranes following disruption of the
tonoplast (Fukuda, 1997), RNase LE and other hydrolytic enzymes would be brought in
contact with its RNA substrate in the cytoplasm (Lehmann et al., 2001). Similar
mechanisms may occur in other plant species during many degradative processes, such as

senescence, endosperm mobilization, and xylem differentiation.

RNases and defense

Increases in RNase and nucleases activities during pathogen invasion have been found in
several systems (for references, see Green, 1994). Two parsley pathogenesis-related (PR)
proteins induced by a fungal pathogen show homology to a ginseng RNase (Moiseyev et
al., 1994). These proteins are part of the large PR-lO family of intracellular plant PR
proteins, and are not related to the S-like RNase or nuclease I family, but do support a
role for RNases in defense responses. In Nicotiana tabacum, transcripts for the S-like
RNase NE are induced by inoculation of leaves with the oomycete Phytophthora

(Galiana et al., 1997). Interestingly. exogenous application of RNase A inhibited

ll

development of this pathogen on the leaves by as much as 90%, without eliciting general
defense responses such as HR. Additionally, RNase A inhibited the formation of lesions
caused by tobacco mosaic virus (TMV; Galiana et al., 1997). A 30-kDa protein sharing
similarity with a tomato RNase has been isolated from leaves of Engelmanm'a pinnatifida
and shown to have broad-spectrum antifungal activity (Huynh et al., 1996). An
extracellular RNase activity in rust-infected wheat leaves has also been reported (Barna et
al., 1989), and as many as 20 isoforrns of RNases are induced in pearl millet during SAR
after challenge with downy mildew disease (Shivakumar et al., 2000). RNSl antibody
(Bariola et al., 1999) appears to cross-react with a protein induced by SAR in this system
(P.D. Shivakumar and V. Smedegaard-Petersen, personal communication).

RNases are also induced by wounding in several systems. In tomato, for example,
the transcript for RNase LE accumulates in wounded leaves (Lers et al., 1998). RNase
NW, a tobacco protein with high sequence similarity to RNase NE, is also induced in
wounded leaves (Kariu et al., 1998). Since both pathogen challenge and mechanical
wounding have effects on RNase activity, it is possible that RNA-degrading enzymes
have roles in defense responses. During wounding, RNases may function in a manner
similar to their role in Pi-recycling, including recycling of nucleotides from damaged
cells. They may also be involved in rebuilding of damaged vascular tissue, consistent
with their induction during xylogenesis. RNases and nucleases may play a more specific
role as defensive proteins to protect damaged or particularly susceptible tissues from
pathogen invasion. This hypothesis is supported by the high expression levels of several
RNase transcripts in pistils. Despite the fact that these structures are rich in nutrients,

making them likely targets for pathogen invasion, they are rarely infected, possibly due to

the presence of various defense-related proteins (Atkinson et al., 1993). These defensive
proteins may include S-like RNases, as S-like genes are expressed in pistils and flower
structures of Arabidopsis (RN51, RN52, and RN53; Taylor et al., 1993; Bariola et al.,

1994), tobacco (NE; Dodds et al., 1996), and petunia (X2; Lee et al., 1992).

New ideas regarding secreted RNase functions

Recent results from our laboratory have prompted new ideas regarding functions for
secreted RNases. These ideas have come from the study of the only member of the RNase
T2 family found in Saccharomyees cerevisiae. Inactivation of this RNase, Rnyl, resulted
in unexpected phenotypes: the mutant cells are larger than wild-type, grow as cell-
aggregates, and are heat- and salt-sensitive. Complementation of the heat and salt-
sensitivity phenotypes was accomplished using Rnyl , three of the Arabidopsis RNSs, and
the structurally unrelated enzyme RNase A. Mutation of RNase A that caused a loss of
RNase activity but retained the protein's stability and structure resulted in the inability of
the protein to complement the growth phenotype (Maclntosh et al., 2001). These
phenotypes, in combination with recent experiments that demonstrated that specific
RNAs can bind to membranes and thereby alter membrane permeability (Khvorova et al.,
1999), led to the hypothesis that secreted RNases can control membrane stability and
permeability by regulating the amount of RNA in the membrane (Maclntosh et al., 2001).
More recent experiments have also shown that exogenous application of RNase A to the
culture medium complements the flocculation phenotype of the rnylA cells (G.C.
Maclntosh and P.J. Green, unpublished). Thus, it appears that an extracellular RNase

activity can affect membrane characteristics in yeast.

Dissertation topic and thesis overview

In summary, T2 RNases and nucleases I enzymes are widely distributed in nature and
tightly regulated in response to a variety of signals. Yet in most cases, their biological
roles and substrates remain elusive. The main objective of my researach was to determine
the biological significance of extracellular RNases in plants. To address this question,
several methods have been employed: studying the mode of regulation of the Arabidopsis
S-like RNase, RNS 1; identifying signals that direct the expression of RN51 ; analyzing the
effects of mutation and overexpression of RN51; and examining a mutant with altered
RNase activities. Identifying signals in addition to P. starvation that affect RN51
expression should help us deduce the processes in which RNS] activity is important.
Analysis of mutant and overexpressing plants during these processes may then reveal
functions of RNSl. Chapter 2 will detail my findings that RNSI and other RNA-
degrading activities are regulated by wounding in Arabidopsis through a novel signaling
pathway. The regulation of RNSl will be firrther analyzed in Chapter 3, including other
signals that control RN51 expression and their effects on the RN51 promoter. Functional
questions will be addressed in Chapter 4, in which I will summarize recent data derived
from the analysis of plants lacking RNSl and of plants with constitutively high RNSI
activity. Finally, in Chapter 5, I will present my contribution to the analysis of an

Arabidopsis mutant that has high levels of several RNase activities.

14

CHAPTER 2

LOCAL AND SYSTEMIC WOUND INDUCTION OF RNASE AND
NUCLEASE ACTIVITIES IN ARABIDOPSIS: RNS] AS A MARKER
FOR A JASMONIC ACID-INDEPENDENT SYSTEMIC SIGNALING

PATHWAY

A version of this chapter is in press:

LeBrasseur, N.D., Maclntosh, G.C., Pérez-Amador, M.A., Saitoh, M., and Green, P.J.
(2002) Local and systemic wound-induction of RNase and nuclease activities in

Arabidopsis: RN51 as a marker for a JA-independent systemic signalling pathway. Plant
J. In press.

15

ABSTRACT

Induction of defense-related genes is one way plants respond to mechanical injury. We
investigated whether RNases are involved in the wound response in Arabidopsis thaliana.
As in other plant systems, several activities are induced with various timings in damaged
leaves, stems, and seedlings in Arabidopsis, including at least three bifunctional
nucleases capable of degrading both RNA and DNA, as well as RNSI, a member of the
ubiquitous RNase T2 family of RNases. The strong induction of RN51 expression is
particularly interesting because it occurs both locally and systemically following
wounding. The systemic induction of this RNase indicates that members of this family
may be involved in defense mechanisms in addition to their previously hypothesized
functions in nutrient recycling and remobilization. Additionally, the systemic induction
appears to be controlled independently of j asmonic acid, and the local induction of RNSl
and the nuclease activities is independent of both JA and oligosaccharide elicitors.
Consequently, a novel systemic pathway, likely involving a third signal, appears to exist

in Arabidopsis.

16

INTRODUCTION

One of the mechanisms by which plants respond to wounding, either mechanical or
feeding-related, is the activation of transcription of a variety of genes. The products of
these genes include defensive proteins such as proteinase inhibitors (Boulter, 1993; Ryan,
1990), components of signal transduction pathways, for example, systemin in tomato and
potato (Ryan and Pearce, 1998), proteins that may be involved in wound healing
(Bowles, 1990), and other proteins whose functions in the wound response are as yet
unknown.

While the expression of some of these genes is proximal to the injury site, others are
expressed at greater distances and constitute the systemic response (Green and Ryan,
1972). In tomato, multiple signals appear to be involved in activating gene expression.
Local responses are thought to be controlled by carbohydrate signals released from
injured plant cell walls (Farmer and Ryan, 1992). Long-distance effects are initiated by
the small peptide systemin (Ryan and Pearce, 1998). Both oligosaccharides and systemin
initiate the de novo production of abscisic acid (ABA) and jasmonic acid (JA), which
leads to accumulation of the transcripts of wound-responsive (WR) genes (Leon et al.,
2001). Ethylene also appears to be involved in the amplification of systemin-activated
signaling (O’Donnell et al., 1996). In addition, electrical signals have been implicated in
the systemic response (Wildon et al., 1992). In general, wound signals in tomato are
thought to constitute a unified reaction that mounts both local and systemic defense and
repair responses (Leén et al., 2001), although a .IA-independent WR gene has been
identified (O’Donnell et al., 1998).

In Arabidopsis, the emerging picture has striking differences in comparison to the

17

tomato system. In Arabidopsis, two distinct signaling pathways can be identified. Some
genes whose transcripts accumulate upon wounding are independent of IA, while the
expression of others requires jasmonate synthesis and perception (Titarenko et al., 1997).
Recently, the JA-independent response was shown to be activated by oligosaccharides
(Rojo et al., 1999). It was suggested that the two pathways work in an antagonistic
manner, so that the inhibition of the JA-dependent pathway occurs in local wounded
leaves through the induction of ethylene by the oligosaccharide elicitors. Conversely,
only JA-dependent genes would be activated systemically, due to the limited diffusion
range of the oligosaccharides (Rojo et al., 1999). However, a JA-dependent WR gene has
been found that is strongly expressed both locally and systemically in Arabidopsis
(Kubigsteltig et al., 1999).

In other systems, RNases have been shown to accumulate in response to mechanical
wounding or pathogen attack, as discussed in Chapter 1. Although several wound-
responsive RNases have been identified, the signaling pathways responsible for RNase
induction are rarely addressed. In order to investigate the effects of wounding on the
expression of the S-like family of Arabidopsis RNases, we examined the RNase profile of
wounded plants. We found not only T2 enzymes, but also multiple activities increase in
damaged tissues in response to mechanical wounding, including several with both RNase
and DNase activity. Within the RNase T2 family, RNSl activity increases, and the RN51
transcript accumulates in both wounded and systemic unwounded leaves of wounded
plants. Unexpectedly, we found that the regulation of these activities by wounding is
independent of both jasmonic acid and oligosaccharides, indicating that a novel,

unidentified wound signaling pathway may be operating in Arabidopsis. Additionally, the

18

systemic induction of RN51 suggests that the enzyme may have a defensive function.

RESULTS

Wounding induced several RNase activities in Arabidopsis

To examine the effect of wounding on RNase expression in Arabidopsis, 2-week-old
seedlings, or stems or leaves of 4-week-old plants were wounded. Samples were
harvested at subsequent time points, and protein extracts were analyzed on RNase
activity gels (Figure 2-1). Significant alterations in RNase activities were seen in the
activity gels, including the activity of one well-characterized member of the RNase T2
family, RNS], and several other activities not likely to be related to this family. In Figure
2-1A, equivalent timepoints were taken from leaves of wounded and unwounded plants
to check for any possible light regulation of RNase activities. Several activities can be
seen even in unwounded leaves, the most prominent of which is a strong band of activity
at 25 kDa; no RNases appear to be regulated in by light (Figure 2-1A, control lanes).
Similar controls were taken for stems and seedlings, and there, too, no light/dark
regulation is evident (data not shown). In contrast, in leaves, stems, and seedlings
(Figures 2-1A, 2-lB, and 2-1C, respectively), multiple activities were induced by
wounding, including an activity that migrated at approximately 35 kDa (Figures 2-1A —
2-1C, top arrow), two at 33 kDa (middle arrow; separation of the two 33-kDa bands can
be seen in Figures 2-2A and 2-2B), and a fourth of approximately 23 kDa (bottom arrow,
labeled RNSl).

Past studies have indicated that the activity of a well-characterized RNase T2
enzyme, RNSl, is induced by phosphate starvation in seedlings and migrates at 23 kDa

on RNase activity gels (Bariola et al., 1994, 1999). In the wounded samples, a 23-kDa

l9

A Control Wounded
kD

 

MW1 36122436481 36122436 48h

   

43—

Figure 2-1. Multiple RNase activities are induced by wounding

Activity gels showing the increase in RNase activity in Arabidopsis at various time points
after wounding. Bands seen on the gel are areas in which RNA cast in the gel has been
degraded. Each lane contains 100 pg of protein from wounded leaves (A), stems (B), and
seedlings (C). Time, in hours from wounding to harvesting, is indicated above each lane
of the gels. Arrows to the right of the gels indicate increased activities. Control lanes
contain proteins from unwounded leaves. MW, molecular weight marker (Gibco-BRL).
Sizes of the MW markers in kDa are shown to the left in (A). hpw, hours post wounding.

activity was also induced (Figures 2-1A —- 2-1C; bottom arrow, labeled RNSl). This
activity was seen within 3 h and remained elevated for at least 48 h after wounding. RNA
gel blot analysis indicated that the RN51 transcript was induced in wounded leaves (see
below). Additionally, in an insertional knock out mutant of RN51, the 23-kDa activity
was not induced after wounding (see Chapter 4). Taken together, RNA gel blot analysis
and mutant phenotypes confirmed that the 23-kDa activity induced by wounding was
indeed RNSl. Thus, as in other plant systems, RNase T2 enzymes are also induced by
wounding in Arabidopsis. However, our results also demonstrate that RNase activities
not related to T2 enzymes respond to wounding in Arabidopsis.

A pair of 33-kDa activities induced in all three wounding experiments has a time
course of induction similar to that of RNS 1. In leaves, the doublet of activity at 33 kDa is
induced within 6 h (Figure 2-1A, middle arrow). In stem tissue and in seedlings, the
upper band of the doublet is evident even in unwounded samples. Still, both bands are
enhanced afier wounding, reaching an apparent maximum by 24 h (Figures 2-lB and 2-
1C), and high levels of these activities can be found as late as 48 h post wounding (hpw).

Interestingly, unlike the sustained induction of the 23- and 33-kDa RNases, a fourth
activity, at 35 kDa, is induced transiently. In leaves, this activity appears within 6 h of
wounding (Figure 2-1A). In stems and in seedlings, the 35-kDa activity appears more
rapidly, with induction evident within 3 h (Figures 2-1B and 2-1C). In all three cases, the
activity is no longer visible by 24 hpw. Unlike RNSl, these wound-inducible activities
exhibit characteristics that distinguish them from the RNase T2 family, as discussed

below.

Multiple bifunctional nuclease activities are increased by wounding

Previous studies using the gel assay system shown in Figure 2-1 demonstrated that a
doublet of bifirnctional nucleases capable of degrading both RNA and DNA migrate at 33
kDa; these nucleases are normally expressed at low levels but are upregulated in a
recently isolated mutant of Arabidopsis (M.L. Abler, N.D. LeBrasseur, L.R. Danhof,
D.M. Thompson and P.J. Green, manuscript in preparation; see Chapter 5). To investigate
whether the 33-kDa RNase activities increased by wounding are bifunctional nucleases,
protein extracts of wounded stem tissue were run on 9% acrylamide RNase and DNase
gels. While other activity gels in this study contained 11% acrylamide, the lower
acrylamide percentage used in these gels results in better separation of the two activities
that migrate at approximately 33 kDa, as can be seen by comparing Figures 2-1B and 2-
2A. Additionally, the increase in the intensity of the bands seen in the RNase activity gel
at 33 kDa is mirrored in the DNase gel (Figure 2-2A), confirming that this doublet has
bifunctional nuclease activity.

In addition to the 33-kDa doublet, another activity that increases upon wounding
appears to have bifunctional activity. The transient 35-kDa RNase activity seen in Figure
2-1 also degrades DNA (Figure 2-2A). Therefore, wounding causes an induction not only
of T2 RNases in Arabidopsis, but also of several nuclease activities, including a
previously unidentified nuclease at 35 kDa.

Past studies of RNase expression profiles in Arabidopsis have shown that cultivar
variation is present. Specifically, the lower band of the 33-kDa doublet can normally be
seen at 10w levels in the cultivar RLD but is absent in Columbia (M.L. Abler, L.R.

Danhof, and P.J. Green, unpublished). However, the lower band can be induced in

A 01361224hpw

 

 

B Unwounded Wounded
RLD Col RLD Col

 

 

Figure 2-2. Bifunctional nuclease activities are induced by wounding.

(A) Protein samples (100 ug for RNase gel; 75 ug for DNase gel) from wounded stems
were run on activity gels containing either RNA (upper panel) or DNA (lower panel) as a
substrate. Both gels contain 9% acrylamide in order to separate the two activities at 33
kDa. Arrows to the left of the gels indicate wound-regulated bifunctional activities. hpw,
hours post wounding.

(B) Stem protein samples (100 ug) fi'om both Columbia (Col) and RLD (RLD) ecotypes
were run on RNase activity gels. Unwounded samples are shown on the left, and stem
samples taken 3 h after wounding are on the right. The lower band (lower arrow) is not
found in unwounded Columbia stems but is present after wounding. Arrows indicate the
upper and lower bands of the 33-kDa nuclease doublet.

23

Columbia upon wounding (Figure 2b). Thus, wounding indicates that although this
activity is differentially regulated under nonwounded conditions in Columbia and RLD, it

is not altogether absent from Columbia.

RN 81 activity is induced by wounding locally and systemically

As seen in Figure 2-1, an RNase activity that migrates at approximately 23 kDa is
induced during wounding. During phosphate starvation in Arabidopsis seedlings, a 23-
kDa activity is also induced (Bariola et al., 1994). The increase in this activity is due to
an increase in the protein levels of RNS] (Bariola et al., 1999) and corresponds to an
increase in transcript accumulation (Bariola et al., 1994). Additionally, as mentioned,
wounded leaves of an RN51 T-DNA insertion mutant lack the 23-kDa activity, (see
Chapter 4). Therefore, we were interested in examining whether the RN51 transcript also
accumulates in response to wounding. Rosette leaves of adult plants were wounded and
then harvested at subsequent timepoints. RNA gel blot analysis demonstrated that the
RN51 transcript accumulated within 3 h and was still elevated as late as 48 hpw (Figure
2-3A). Immuno-blot analysis confirmed that RNS] protein accumulated with a time
course similar to that of the activity increase (data not shown).

We also investigated whether the Arabidopsis response to wounding includes
systemic induction of RNase or nuclease activities in non-wounded tissue. Unwounded
rosette leaves adjacent to wounded leaves were harvested, and RNA was examined by
RNA gel blot. The RN51 transcript was induced in unwounded leaves. The increase was
transient and rapidly disappeared between 6 and 8 hpw (Figure 2-3B), unlike the more
sustained increase of RN51 in wounded leaves (Figure 2-3A). Wounded seedlings (9

hpw; Figure 2-3 B, lane W) were included in the blot for a comparison of transcript levels

A 013 6 12243648 hpw

       

RN81

 

 

 

B
hpw 0 1 3 6 8 12 W
C Systemic Local

 

013681213612hpw

 

Figure 2-3. RNSI activity and transcript increase transiently in wounded and in systemic,
non-wounded leaves.

(A) Northern blot of wounded leaves. Arabidopsis leaves were wounded and harvested at
subsequent timepoints. RNA was extracted, and 10 pg was separated by electrophoresis,
blotted, and probed with RNSI. Blots were then stripped and reprobed with EF-Ia to
control for loading. hpw, hours post wounding.

(B) Systemic induction of the RN51 transcript shown by RNA gel blot analysis.
Unwounded rosette leaves adjacent to the wounded leaves were harvested at the
timepoints indicated above the lanes. The gel contained 10 ug total RNA per lane. RNA
from seedlings harvested 9 h after wounding was included for comparison (W). The blot
was stripped and reprobed with EF -1 a as a loading control. hpw, hours post wounding.
(C) RNase activity gel showing systemic induction of RNSI activity. Each lane contains
100 ug protein. The arrow indicates the RNSI band. No other induced activities were
detected in systemic leaves. Protein samples from wounded leaves were included for
comparison of RNSI activity levels.

25

induced locally and systemically. Additionally, protein extracts were examined by RNase
activity gel assay to determine whether RNSI or nuclease activities are induced in
unwounded leaves of wounded plants. In the unwounded leaves, the 23-kDa RNSI band
was induced (Figure 2-3C), indicating that RNases may play a role in the systemic
wounding response in Arabidopsis. As in wounded leaves, the systemic RNSI activity
increase could be seen slightly later than the transcript accumulation: in multiple
experiments, we observed a slight increase in this activity between approximately 6 and
12 hpw. Wounded samples were included in the gel for comparison (W, Figure 2-3 C). No
other bands aside from RNSI were induced systemically (not shown). Note that the
activity gel in Figure 2-3C was incubated at pH 6.0, instead of pH 7.0, to enhance RNSI
activity, since no other activities were seen to be induced systemically at pH 7.0 (see
Materials and Methods). As a result, the relative activity of RNSI is higher compared to
the constitutive 25-kDa band. It is interesting that the systemic activity increase does not
appear to reflect the amount of transcript induced (compare 3 and 6 h in Figure 2-3B with
6 and 12 h in Figure 2-3C). The discrepancy may be due to a possible posttranscriptional

regulation of RNS].

The induction of RN51 and the nucleases is independent of jasmonic acid

Wounding responses in Arabidopsis are controlled by jasmonic acid (JA)-dependent and
-independent signaling pathways (Titarenko et al., 1997). The presence of distinct
wounding pathways in Arabidopsis has led to the proposal that the function of multiple
signals is to control local and systemic gene expression differentially. Specifically, it has
been suggested that the JA-dependent pathway controls induction of systemic responses.

while an oligosaccharide-dependent pathway directs gene activation locally at the

Figure 2-4. RNSI and nuclease induction by wounding is independent of jasmonic acid.
(A) Top panel; activity gel showing induction of all four wound-responsive activities in
both the wild type (WT) and the JA-insensitive coil mutant (coil). Arrows to the right
indicate the induced activities, including the nuclease at 35 kDa (top arrow), the doublet
at 33 kDa (middle arrow), and RNSI (bottom arrow). Bottom panel; systemic induction
of RNSI in the coil mutant. hpw, hours post wounding.

(B) Northern blot analysis of wounded wild type (WT) and coil (coil) plants. The blot
contains RNA (10 pg) from unwounded leaves (U), local 6-h wounded leaves (L), and
unwounded leaves of wounded plants 3 h after wounding (S). The blot was probed with
RN51, stripped, and reprobed with EF -1 a as a loading control.

(C) Activity gels containing 100 pg of protein from leaves of JA-treated (JA) or untreated
(control) plants. Plants were treated with methyl jasmonate or buffer alone, and leaves
were harvested at 0, 3, 6, and 24-h timepoints. W, protein from leaves 24 h after
wounding. Arrows to the right of the gel indicate the running positions of wound-
inducible activities.

(D) Northern analysis of RNA (10 pg) from leaves of JA-treated (JA) or untreated
(control) plants. Plants were treated as in (c). 2- and 24-h timepoints are shown. Similar
results were seen at 3 and 6 h (data not shown). W, RNA from leaves 6 h after wounding.
Blots were probed with RNSI, then sequentially stripped and reprobed with A05 to
control for JA-treatment and with EF -1 a as a loading control.

Local
0011

06240624hpw

 

 

  
 
 

Systemic
coi1
0 3 6 8 12 hpw

 

. RN81

 

C Control JA
0 3 6 24 3 6 24 W h

an {—33 kD

 

 

<—RNS1

Control JA

 

036243624W h

. RN81

a G... Aos

EF-10t

  

Figure 2-4. RNSI and nuclease induction by wounding is independent of jasmonic acid.

28

site of damage (Rojo et al., 1999). Under this assumption, RN51 would be expected to be
regulated by the JA-dependent pathway, as its transcript is induced systemically. In
contrast, the bifunctional nuclease activities, which are not systemically wound-
responsive, would be regulated independently of IA. The JA-insensitive mutant coil
(Feys et al., 1994) has been widely used to demonstrate dependence of the wound-
induction of specific transcripts on JA (e.g., Reymond et al., 2000; Titarenko et al.,
1997). We investigated whether the wound induction of RNSI or the nuclease activities
are regulated by the JA-dependent pathway by wounding leaves of coil plants. As
expected, like the wild type, wounded coil plants displayed increased levels of the 35-
and 33-kDa activities (Figrre 2-4A). The transient 35-kDa nuclease was seen in both
wild-type and mutant plants at 6 hpw, and the nuclease bands at 33 kDa were induced at
6 and 24 hpw in both wild-type and coil plants. However, wounded coil leaves
unexpectedly also had increased RNSI activity. In fact, in contrast to systemic responses
previously characterized in Arabidopsis, both local and systemic wound induction of
RN51 is independent of IA. RNSI activity was induced in mi] local wounded tissue
(Figure 2-4A, upper panel) and in unwounded leaves of the same plants (Figure 2-4A,
lower panel). RNA gel blot analysis demonstrated that the RN51 transcript also
accumulated in the coil mutant upon wounding (Figure 2-4B). Local 3-h and systemic 6-
h timepoints are shown (Figure 2-4B, L and S, respectively). Systemic RNSI levels at 6-
and 24-h timepoints were also similar in coil and wild-type plants (data not shown).

To confirm that JA does not affect the induction of RNSI or the nucleases, young
plants were treated with methyl jasmonate (MeJA) and examined for RNase activity

(Figure 2-4C). Efficacy of the MeJA treatment was shown by the accumulation of the

allene oxide synthase transcript (A05), a component of the jasmonate pathway known to
be induced by JA (Laudert and Weiler, 1998; Figure 2-4D). However, MeJA treatment
did not cause RN51 accumulation after 3, 6, or 24 h of treatment (Figure 2-4D). Likewise,
the 35- and 33-kDa nucleases, as well as RNSI, were not induced by MeJA at the activity
level (Figure 2-4C). Thus, we have demonstrated that the wound-inducible RNase and
nuclease activities are not controlled by the JA-dependent signaling pathway. Recent
results from an independent microarray experiment also demonstrated that the local
expression of RN51 is induced in a JA-independent manner (Reymond et al., 2000),

supporting our observations.

RN51 is not induced by the oligosaccharide pathway

It has been suggested that the JA-independent wound response in ArabidOpsis is
controlled by oligosaccharide elicitors. Transcripts whose accumulation after wounding is
not dependent on JA can be induced by oligogalacturonic acids (OGAs), proteinase-
inhibitor inducing factor (PIIF) fractions from tomato leaves, and chitosan (Rojo et al.,
1999), all of which are rich in oligosaccharides. To test if one of these known inducers of
certain JA-independent wound-inducible genes would also induce RN51 transcript
accumulation, an OGA-rich fraction known as 'TFA-PIIF' (generously provided by Dr.
E.E. Farmer; purified according to Bishop et al., 1984) was used to treat detached rosette
leaves. Leaves were floated in an MS medium either with or without the addition of TFA-
PIIF and harvested at subsequent timepoints. RN51 transcript levels were analyzed by
RNA gel blot analysis (Figure 2-5A).

As seen in Figure 2-5A, the slight wounding involved in detaching leaves caused a

30

A +PIIF -PllF

 

01.5361.536Wh

 

B +PIIF -Pl|F
o 3 s 3 s h

 

Figure 2-5. OGA-rich fractions do not induce RNase activities.

Leaves were detached from 4-week-old wild type plants and floated in buffer with
(+PIIF) or without (-PIIF) the addition of TEA-PIIF. Samples were harvested at
timepoints indicated above the lanes (in hours). '

(A) Northern blot containing 10 pg RNA per lane. W, RNA from leaves 6 h after
wounding. The blot was probed with RN51, then sequentially stripped and reprobed with
Choline kinase (CK) to control for PIIF-treatment and with ch-4A as a loading control.
(B) Activity gel containing 100 pg protein per lane. Arrows to the right of the gel indicate
the normal running positions of the wound-inducible activities.

31

higher than normal basal level of transcript in the samples after 3 h. However, in contrast
to most known Arabidopsis JA-independent wound-induced transcripts, the presence of
TFA-PIIF in the medium did not induce RN51 compared to the control (Figure 2-5A).
Blots were also probed with choline kinase (CK), an OGA-induced transcript (Rojo et al.,
1999), to demonstrate efficacy of the OGA treatment. The wounding involved in the
detachment technique induced CK expression at the 90 min timepoint, regardless of the
presence or absence of OGAs and consistent with the timing of CK wound-induction seen
previously (Rojo et al., 1999). However, as expected, the TFA-PllF-treated samples had
increased levels of CK compared to the control samples at both 3- and 6-h timepoints
(Figure 2-5A). Thus, while CK levels increased due to the treatment, RN51 levels were
not affected by the PIIF fraction. Interestingly, the wound-inducible nuclease activities,
as well as RNSI activity, were also not induced by the treatment, as seen by RNase
activity gel assay (Figure 2-SB). Throughout the time course, the only activity detected
was a major band at 25 kDa, which was not regulated by wounding. This activity was
also not affected by the TFA-PIIFS, and no other activities were induced, including the
wound-responsive 35- and 33-kDa nucleases and RNSI.

Thus, although the treatment was effective, induction of RNase activity by wounding
is not mediated by oligosaccharides. These results were unexpected, considering that all
the activities were also independent of J A. Therefore, we hypothesize that in addition to
oligosaccharides and JA, an additional signal exists that directs expression of nuclease

and RNase activities in wounded tissue.

DISCUSSION

In this chapter, we have demonstrated that several Arabidopsis RNase and nuclease

32

activities are coordinately regulated by wounding. Unlike other wound-inducible proteins
previously studied in Arabidopsis, these activities are not controlled by jasmonic acid
signaling or by oligosaccharide elicitors. The systemic induction of RNSI makes this
ribonuclease especially intriguing. The systemic induction of RNSI not only supports the
existence of a novel pathway for the regulation of systemic wound responses, but is also
suggestive of a defensive role for this RNase.

The increases in all the activities occur in a coordinated manner and provide us with
a unique perspective into Arabidopsis wound signaling mechanisms. Our understanding
of the wound response in Arabidopsis is currently highlighted by the presence of two
distinct, antagonistic pathways: JA-dependent and -independent. The JA-independent
pathway controls local induction of transcript accumulation and has been shown to be
regulated by OGA elicitors probably released from injured plant cell walls (Rojo et al.,
1999). The three nuclease activities and RNSI are strongly induced locally by wounding
(Figures 2-1 and 2-2). However, they are not induced by treatments with OGA-rich
fractions (Figure 2-5). In fact, in several repetitions of the OGA treatments, the slight
wound-induction of RNSI caused by detaching the leaves appeared to be inhibited by the
presence of the OGAs (for example, compare +PIIF and -PIIF at 6 h in Figure 2-5A).
Given that the RNase and nuclease activities are not induced by the PIIF treatments, we
conclude that another signal besides OGAs is likely operating to direct the local induction
of these activities upon wounding.

The local response of RNSI and the nucleases to wounding was also not controlled
by the JA-dependent signaling pathway, as shown by the strong wound-induction of these

activities in the coil mutant (Figure 2-4A). Further, RNSI transcript accumulated to high

levels in unwounded leaves of wounded plants. IA is an important systemic signaling
molecule during wound-induced gene induction in Arabidopsis (Titarenko et al., 1997).
However, the systemic induction of RN51 did not depend on JA (Figure 2-4B). To our
knowledge, RN51 is the first gene in Arabidopsis shown to be induced systemically by
wounding in a JA-independent manner and therefore indicates the existence of an
alternate long-distance signaling pathway.

It has been proposed that dehydration itself might be the causal induction factor of
some JA-independent wound-inducible genes, including RN51, which was shown by
microarray analysis to be induced by dehydration (Reymond et al., 2000). However, the
dehydration experiments performed involved detaching the rosette leaves from the roots,
which itself induces RN51 (see Figure 2-5). My results indicate that dehydration itself
does not induce RN51 expression but may potentiate the wound induction (see Chapter
3). The induction of certain wound-inducible genes in potato and tomato are affected by
ABA levels (Pena-Cortes et al., 1989). Therefore, ABA could be considered a possible
signal for the wound induction of RN51. The involvement of ABA in the wound response
will be discussed further in Chapter 3.

Given its involvement in the wound response in solanaceous plant species and in the
regulation of various programmed cell death processes (Jones, 2001), ethylene is another
candidate for the signal involved in the induction of RN51 both locally and systemically.
However, the wound induction of RN51 also occurs in the ein2 mutant, suggesting that
ethylene perception is not required (Reymond et al., 2000). Salicylic acid (SA) is required
for many plant-pathogen defense responses (reviewed in Alvarez, 2000), and interaction

between SA, J A, and ethylene pathways is emerging as an important regulatory method

34

for activating multiple resistance mechanisms (reviewed in Pieterse and Van Loon,
1999). However, SA does not appear to be involved in the regulation of RN51 by
wounding, as the transcript is still induced in SA-deficient, NahG-expressing plants (E.E.
Farmer, personal communication). Additionally, analysis of AF GC microarray data in the
Stanford Microarray Database available on the web (http://genome-www4.stanford.edu/
MicroArray/SMD/; Wisman and Ohlrogge, 2000) indicates that RN51 is not induced by
treatment with the SA-analog BTH or by various pathogens that induce systemic acquired
resistance, which is known to cause the accumulation of SA within the plant (Alvarez,
2000)

Other possible signals for the induction of RN51 include reactive oxygen species
(ROS). ROSs are commonly produced in plants in response to both pathogen and
herbivore attacks (Grant and Loake, 2000; Orozco-Cardenas and Ryan, 1999). In tomato,
H202 acts as a second messenger for the induction of defense—related genes induced
systemically at later timepoints than the earlier, signaling-related genes (Orozco-
Cardenas et al., 2001). The timing of RNase and nuclease responses coincides with these
later responses that are dependent on H202 signaling. The involvement of ROSS, as well
as other possibilities, for example, electrical signals, in RNase induction during the
Arabidopsis wound response will prove an important avenue of study for defining
alternate pathways regulating defense responses.

The systemic increase in RNSI appears to be regulated at multiple levels. Studies
with reporter constructs controlled by the RN51 promoter region indicate that the local
increase in RN51 transcript caused by wounding is likely due to transcriptional regulation

(see Chapter 3). However, while the transcript increases to significant levels systemically,

35

the activity increase is not as great (compare Figure 2-3B, 3 and 6 h, with Figure 2-3C, 6
and 8 h). It is possible that a post-transcriptional mechanism may also exist that regulates
the amount of RNSI protein translated or the activity of the translated protein in the
unwounded tissue.

In addition to its interesting regulatory properties, the systemic induction of RNSI
implies that the RNase may have an important function during the wound response. It is
striking that the RN51 transcript was the most highly induced in local wounded tissue of
all transcripts examined in two independent microarray experiments: one examined 150
genes enriched for those implicated in defense responses (Reymond et al., 2000), and the
second examined 600 genes, approximately half of which were hypothesized to be
involved in RNA metabolism and RNA turnover (M.A. Pérez-Amador and P.J. Green,
unpublished). The high level of transcript accumulation of RN51 may be indicative of an
important role for the protein product of this gene. Additionally, the transcript and the
activity are induced in non-damaged tissue, where recycling of nutrients and degradation
of bulk cellular nucleic acid, two of the proposed functions of secreted RNases, should
not be necessary. Instead, it is possible that RNSI has a defensive function. Normally,
RN51 is expressed solely in flowers (Bariola et al., 1994). The presence of RNases in the
pistil may contribute to protection of the structure from pathogens (Bariola et al., 1994).
In fact, it has been demonstrated that application of an extracellular RNase to tobacco
leaves inhibits growth of both TMV and a fungal pathogen (Galiana et al., 1997).
Likewise, local and systemic induction of RNSI could contribute to protection of the
plant against invasion of pathogens following wounding. Although it is not known how

RNases could achieve such defensive roles, wound signal molecules are known to

36

produce membrane-associated changes within the plant cells, including membrane
depolarization (Thain et al., 1995). Recently, it has been proposed that T2 RNases can
affect membrane stability or permeability (Maclntosh et al., 2001), suggesting a possible
link. The role of RNSI in defense mechanisms will be readily testable in a recently-
isolated RN51 T-DNA insertional mutant and in plants that overexpress RNSI activity
(see Chapter 4).

In contrast to the systemically-induced RNSI, the bifunctional activities were
induced only in local, damaged tissue. These activities may also have defensive roles in
the local tissue, or they may be functioning in other aspects necessary during the wound
response, such as recycling of nucleotides from damaged cells or rebuilding of damaged
vascular tissue. Cells damaged by wounding would contain phosphate and other
molecules that could be recycled for use in active, growing parts of the plant. Similar
functions have been hypothesized for T2 RNases that are induced by phosphate-starvation
and senescence in several species, including tomato and tobacco (reviewed in Howard et
al., 1998). Activities at 33 kDa, as well as RNSI, are induced in P,-starved seedlings (see
Figure 10 in Bariola et al., 1994), and bifunctional nucleases are induced by senescence
in Arabidopsis (Pérez-Amador et al., 2000). Possibly, RNSI and the nucleases are
induced during wounding as part of this wide-spread recycling mechanism.

It is possible that the systemic induction of RN51 could be explained as a sort of
"priming" of the plant for nucleotide recycling in as-of-yet undamaged tissue. However,
the nucleases, as well as RN52 and RN53, which encode two additional RNases shown to
be induced during conditions requiring nutrient recycling, such as P, starvation and

senescence (Bariola et al., 1994; Taylor et al., 1993), are not induced locally or

37

systemically by wounding, at least indicating that other hydrolytic enzymes thought to be
involved in P, recycling and remobilization are not generally activated by wounding. In
fact, RN51 is the only one of the five Arabidopsis genes related to the RNase T2 family
that is induced in either local or systemic tissue by wounding (not shown). Such
specificity indicates that the enzyme may have an important role in the wound response.
Although local induction of hydrolytic activities by wounding has been found in
several other plant species (e.g., Galiana et al., 1997; Lers et al., 1998), little is known
regarding the signal pathways involved in their regulation. We have now identified a
wide array of activities that are regulated by wounding in Arabidopsis, including RNase
and nuclease activities that show both sustained and transient patterns of induction. Given
its unique responses to wounding, independent of all known wound regulators, RN51 will
be a valuable marker for identifying novel signals that operate in the Arabidopsis wound
response. More importantly, to our knowledge, RNSI is the first RNase shown to be
induced systemically by wounding. This result points us toward new ideas regarding the
function of the widespread T2 family of RNases. Specifically, in addition to having a role
in P, recycling as has been previously hypothesized, RNSI may have a more direct
defense-related function, possibly involved in protection of the plant from firrther attack

after a wound-stimulus has been detected.

MATERIALS AND METHODS
Plant materials and treatments
Unless otherwise stated, the Columbia ecotype of Arabidopsis thaliana was used
throughout this study. Soil-grown plants were grown in chambers under 16 h of light in

50% relative humidity at 20°C. For seedling experiments, seeds were surface-sterilized

and germinated on Arabidopsis growth medium as described (Taylor et al., 1993). The
coil seeds were kindly provided by Dr. J.G. Turner (University of East Anglia, Norwich,
UK). Mutant coil plants were selected by germinating on MS medium supplemented
with 50 pM methyl jasmonate as described (Feys et al., 1994). The plants were then
transferred to soil and grown for an additional 4 weeks before wounding treatments were
performed.

Stems or leaves of 4- to 6-week-old plants or leaves of 14-day-old seedlings were
wounded using ridged flat-tipped tweezers and harvested at subsequent timepoints. For
non-wounded material, leaves on either side of the wounded leaf were harvested. All
samples were frozen in liquid N2 immediately after harvesting and stored at —80°C until
used for RNA or protein extractions. Wounding experiments were performed a minimum
of three times. Representative blots or gels are shown.

Jasmonic acid treatments were conducted on 4-week-old plants that were placed in
enclosed boxes. Methyl Jasmonate (MeJ A; Bedoukian Research, Inc., Danbury, CT) was
diluted 1:25 in ethanol to a final concentration of 190 mM. A cotton-tipped applicator
was soaked with 50 pL of the MeJA solution and placed in the container with the plants.
An applicator soaked with ethanol alone was placed in a separate container with control
plants. Four boxes were used in each of two experiments, for 2- and 24-h timepoints of
both MeJA-treated and control plants. Two additional experiments were conducted for 3-,
6-, and 24-h timepoints, for a total of four replications. Representative blots and gels are
shown in Figures 1-4c and l-4d.

For OGA-treatments, rosette leaves of 4-week-old plants were removed by slicing

the petiole with a razor blade. Approximately 20.1eaves were floated on 40 mL of MS

medium (Life Technologies, Rockville, MD) supplemented with 0.5% (w/v) sucrose in
each of two 250-mL flasks. One flask also contained 250 pg/mL of the OGA-rich TFA-
PIIF fraction (provided by Dr. E. Farmer, Universite' de Lausanne, Switzerland). Flasks
were shaken on a rotary platform at ~50 R.P.M. Six or seven leaves were removed from
the flasks at 1.5-, 3-, and 6-h timepoints and immediately frozen in liquid nitrogen. OGA
treatments were performed twice using this method, and the northern blot and activity gel
in Figure 2-5 are representative of this method. Additionally, two replicates of treatments
of seedlings grown in liquid culture, according to Rojo et al. (1999), were performed
using a PIIF fraction kindly provided by Dr. G. Howe (Michigan State University).
Similar results were obtained in all four repetitions; however, the greatest induction of the
positive control, choline kinase, was seen using the TFA-PIIF treatments of detached

leaves.

RNA extraction and RNA gel blot hybridization

Total RNA from Arabidopsis samples was extracted as previously described (Newman et
al., 1993). RNA (10 pg per lane) was separated by electrophoresis in 3% (w/v)
formaldehyde/1.2% (w/v) agarose gels and blotted to Nytran Plus nylon membrane
(Schleicher and Schuell, Keene, NH). The RNA blots were hybridized as described in
Taylor and Green (1991) using a 32P-labeled RN51 probe. To control for loading, the
same RNA blots were stripped in distilled water at 90-95°C for at least 20 minutes and
then hybridized with a 32P-labeled probe for the Arabidopsis translation elongation factor
EF -1 a (EST accession number R29806) or translation initiation factor eIF-4A (Taylor et
al., 1993). A choline kinase probe, kindly provided by Dr. J. Sétnchez-Seranno

(Universidad Autonoma de Madrid, Spain), was used as a positive control for OGA

40

treatments (Rojo et al., 1999). A probe for allene oxide synthase (Laudert et al., 1996)
was used as a positive control for MeJA treatments. Quantitation of hybridization was
performed using Phosphorimager (Molecular Dynamics, Sunnyvale, CA) analysis to
confirm increased CK and A05 levels in response to PIIF- and JA-treatments,
respectively, as well as increased RN51 levels in local wounded and systemic leaves (data

not shown).

Protein extraction and detection of RNase and DNase activities

Total protein was extracted as described (Maclntosh et al., 1996). Homogenates were
clarified by centrifugation, and soluble protein was quantified by the method of Bradford
(197 7). 100 pg total protein was loaded in each lane.

RNase and DNase activities were assayed using activity gels as described (Yen and
Green, 1991), with minor modifications. After the isopropanol wash, before incubation,
gels were washed in 100 mM Tris-HCl containing 2 pM ZnCl2 for 20 min in order to
restore Zn2+ required for certain RNase and DNase activities. Gels were washed and
incubated at pH 7.0, except for the RNase gels of systemic activity in Figures 2-3C and 2-
4A (lower panel), which were incubated at pH 6.0 to enhance resolution of RNSI
activity, which is most active between pH 5.0 and 6.0 (data not shown). Similar results
were seen at pH 7.0 (data not shown). Separating gels contained 11.3% (w/v) acrylamide,
except the gels in Figure 1-2, which contained 9% acrylamide for increased separation of

the doublet at 33 kDa.

41

CHAPTER 3

REGULATION OF RNSI EXPRESSION

ABSTRACT

RNSI activity and transcript have been shown to accumulate in response to wounding
(Chapter 2) and P, starvation (Bariola et al., 1994, 1999). I examined whether the
induction was a transcriptional or posttranscriptional process using nuclear run-on
analysis (in the case of P, starvation) and transgenic plants expressing promoter fusion
constructs (wounding). Whereas the run-on analyses were inconclusive, 1 demonstrated
that the RN51 promoter was responsive to wound stimuli, as shown by accumulation of
the reporter transcript. I then used the promoter constructs to examine other signals that
might direct transcription of RN51. Through analyses of these transgenic plants and RNA
blot analyses of endogenous RN51 transcript, I found that both ABA and NaCl induced
RN51 expression. Interestingly, however, ABA was not required for the wound induction
of RNSI activity. Additionally, heat shock appeared to have a negative effect on RN51
induction by wounding. In contrast to previous reports, 1 determined that dehydration did
not cause increases in RN51 expression, indicating that another signal directs the

induction of RN51 by wounding.

INTRODUCTION

As seen by results presented in Chapter 2 of this thesis, the levels of RN51 transcript are
responsive to external stimuli. In addition to wounding, RNSI also responds to the P,
status of its environment as well as to internal cues, as shown by the increase in RN51
transcript in senescing tissues (Bariola et al., 1994). In addition to RN51, genes encoding
RNases in many plant species are regulated in response to P, starvation, senescence, and
wounding.

Although numerous conditions that affect the levels of various RNase activities have
been identified in many plant species, little is known regarding the regulatory pathways
that control these responses. Previous studies addressed the regulation of RN51 in
response to P, starvation; in particular, run-on analyses and transcriptional fusions were
used to determine whether RN51 transcript accumulates as a result of transcriptional or
posttranscriptional regulatory mechanisms (Howard, 1996). No difference could be
detected by nuclear run-ons between seedlings grown on full and P,-deficient media.
However, these results were inconclusive, because RN51 transcript levels were not above
background. Transcriptional increases were detected by promoter fusion constructs, but
the increases were not sufficient to explain the large increase in RN51 levels. Therefore, I
repeated nuclear run-on analyses of P,-starved and P,-supplied seedlings to examine the
level of regulation of RN51 under these conditions.

Additionally, since RN51 transcript is so abundant afier wounding, I was interested
in how these levels are regulated during the wound response. I was also particularly
interested in other signals involved in regulating RN51 gene expression, given that known

Arabidopsis wound pathways are not involved in the response of this gene during

44

wounding (see Chapter 2).

RESULTS

Run-on analysis of RN51 expression

Run-on analyses were attempted to determine whether P, starvation leads to
transcriptional induction of RN51. Comparison of run-on transcription assays and RN51
transcript accumulation in seedlings grown in the presence or absence of P, should
indicate the contribution of transcriptional regulation to the increase in RN51 resulting
from P, starvation. Seedlings were plated on Arabidopsis germination medium (AGM) on
mesh circle inserts. On the day of germination (approximately 2.5 days after plating), the
mesh inserts were used to transfer seedlings to new plates that either contained (P+) or
lacked (P-) P,. After 7 d of growth, seedlings were harvested and used to isolate active
nuclei and total RNA. RNA was used for RNA gel blot analysis to confirm that RN51
transcript levels were increased (Figure 3-1 B).

Nuclei prepared from the two seedling samples were used for nuclear run-on
analyses. Labeled nascent transcripts were purified and used to probe slot blOts
containing ssDNA of plasmids of RN51, eIF4A, and empty Bluescript (BSK). The BSK
plasmid was used as a control for background hybridization, and eIF4A was used as a
loading control, as transcription of this gene is unaffected by the P, status of the
seedlings. As seen in Figure 3-lA, hybridization of RNSI, whose expression is regulated
in response to P, starvation, is below detection in samples from P,-starved and unstarved
seedlings. In multiple repetitions, the hybridization signals obtained were not above
background levels detected by the BSK control, although eIF 4A signals were significantly

above background. The results indicate that although the nuclei were active, RN51

45

p+ p-
M . eIF4A
RN81
BSK
B
p+ p.
$2.24 RN81

W W elF4A

Figure 3-1. Run-on analysis of RN51 in response to P, starvation.

(A) Nuclear run-on analysis was performed to compare transcriptional activity of the
RN51 gene in Arabidopsis seedlings grown on medium containing (P+) or lacking (P') P,.
(B) Accumulation of RN51 transcript is shown by Northern blot analysis of RNA from
the same samples used to isolate nuclei in (A).

46

transcription was not at sufficient levels to measure by run-on analysis. As a result, I
could not conclude anything concerning the transcriptional regulation of RN51 in

response to P, starvation.

Regulation of RN51 expression in response to wounding

Since RN51 is so strongly induced by wounding (see Chapter 2), I were interested in
whether this regulation is the result of transcriptional control of the RN51 promoter or of
a posttranscriptional mechanism. Several constructs were used to examine the regulation
of RN51 (Figure 3-2). The first, p1975 (referred to as nos-RNSchNA), included the
RN51 cDNA fused between the constitutive, low expressing nos promoter and the pea E9
terminator. This cDNA contains the open reading frame (ORF) plus approximately 60
bases of the 5' UTR (Bariola et al., 1994). The full 5' UTR, as determined by primer
extension analysis, contains an additional 40 bases (Howard, 1996). Therefore, I made a
second construct, p2019 (referred to as nos-preRNSl), in which intron sequences and the
missing bases of the 5' UTR not found in the cDNA were included. Transgenic plants
were made using Agrobacterium tumefaciens-mediated transformation via the method of
vacuum infiltration (Bechtold et al., 1993). Additionally, transgenic control plants were
made by transformation with a control plasmid, p1995, which consists of the nos
promoter driving the globin reporter coding region (nos-Globin). To examine the
regulation of the transgenes by wounding, leaves of the transformed plants were wounded
and harvested 3 h later. Unwounded leaves were also harvested to provide control
samples. RNA was prepared from the samples and used for RNA gel blot analysis. Blots
were probed with an oligo complementary to the transcribed nos sequences to distinguish

between RNSI transcribed from the transgene and the endogenous copy. The RNA gel

47

 

 

—_-- ‘-

p84811402 f ‘ 353 " ,’ Murals“
paw/1432 2mm; 59

p2081l2082

 

 

 

 
 

 

 

 

 

p1995 1 nos

 

 

p1975 1 nos

 

 

 

p2019 I n05 ‘3‘ preRNS1

 

Figure 3-2. Constructs used to examine the regulation of RNS l .

Several constructs were used to transform wild-type Arabidopsis plants. Transgenic lines
were then used to analyze the expression of the reporters under various conditions. Clone
identification numbers are shown to the left. LUC, Luciferase coding region; GUS, ,8-
glucuronidase coding region; 355, CaMV 355 promoter; nos, nopaline synthase
promoter; RN51p, RN51 promoter; E9, 3' end of the pea E9 gene; preRNSI, transcribed
region ofRNSl ; The RN51 cDNA was isolated by Bariola et al. (1994).

48

blots indicated no difference between nos signal in wounded and unwounded leaves of
plants expressing the nos-RNSchNA construct, although endogenous levels of RN51
were clearly elevated by wounding, as shown by hybridization with the RN51 cDNA
(Figure 3-3; center panels). However, in the nos-preRNSl plants, a slight but
reproducible increase in nos signal was seen (Figure 3-3; left panels). It is therefore
possible that a certain amount of posttranscriptional regulation of RNSI mRNA exists,
which requires either the entire 5' UTR region or intron sequences, or both, whereas the
cDNA sequence alone is not sufficient. However, this slight increase of approximately
2.5-fold does not account for the high levels of RN51 transcript found in wounded tissues.

Since posttranscriptional regulation could not account for the increase of RNSI by
wounding, I also examined the ability of the RN51 promoter to confer wound-inducibility
to a reporter gene. Constructs were available (Figure 3-1) that contained 2.7 kb of RN51
promoter plus ~60 bases of RN51 5’ UTR sequence fused to either fi-glucuronidase
(GUS; p1401) or luciferase (LUC; p1432) reporter genes, as well as control constructs
with the same reporters driven by the constitutive CaMV 355 promoter (p848 and pl 402,
respectively) (Howard, 1996). These constructs were used to transform Arabidopsis to
analyze their regulation by wounding.

Regulation of the GUS constructs was analyzed by RNA gel blot. In unwounded
leaves of plants transformed with the RNSlp-GUS construct, little or no GUS transcript
could be detected. In contrast, 3 h after wounding, wounded leaves contained high levels
of GUS (Figure 3-4A). LUC lines revealed that luciferase activity also increased
following wounding. Transformed RNSlp-LUC lines were wounded and examined for

LUC activity 5 h alter wounding. Unwounded RNSlp-LUC and wounded and

49

nos—preRNSl—E9 nos-RNSchNA-E9 nos-Globin—E9

 

 

0h 3h 0h 31!

I"

    

HOS

 

RNSI

 

EF-I a

       

Figure 3—3. Differential response of RNSI transcribed sequences to wounding.

Pools of T2 Arabidopsis seedlings expressing either the entire RN51 transcribed region
(left panels), the RNSI cDNA (center), or the globin transcript (right) under the
constitutive nos promoter were wounded and harvested 3 h later. An oligonucleotide
corresponding to a transcribed portion of nos was used as a probe in order to distinguish
the transgene from endogenous RN51. Blots were then stripped and reprobed with RNSI
(to control for wounding) and EF-l a (to control for loading).

50

A RNSI p-GUS-E 9 3 SS-GUS-E9

 

     

 

0h 3h on 3h
1?an

 

 

Figure 3-4. The RNSI promoter confers wound inducibility to reporter transcripts.

(A) Leaves of transgenic Arabidopsis plants expressing the GUS reporter under the
control of either 2.7 kb of genomic sequence upstream of the RNSI transcription start site
plus 60 bases of 5' UTR sequence or the constitutive 35S promoter were harvested before
(0 h) or 3 h after wounding. Blots were probed with GUS, then stripped and reprobed
with RN51 (to control for wounding) and EF -1 a (to control for loading). 3SS-GUS plants
were used as controls to demonstrate that GUS is not stabilized by wounding.

(B) Images of wild type (WT) plants and plants expressing LUC under control of the
same upstream RN51 region as in (A) (RNSI-LUC). WND, wounded plants 5 h after
wounding. The lower left leaves of the rosette were wounded. Left panel, plants in visible
light; right panel, luciferase activity.

51

unwounded WT Col plants were used as controls, since no 3SS-LUC lines were ever
found to have LUC activity, either with or without wounding (data not shown). As seen
in Figure 3-4B, only wounded RNSlp-LUC plants have LUC activity. These results
indicate that RN51 is transcriptionally induced by wounding. However, since the
promoter constructs used contained approximately 60 bases of 5' UTR sequence, I
removed this sequence to confirm that the promoter sequence alone confers wound-
inducibility to the GUS and LUC reporters. These constructs, p2081 and p2082, are
shown in Figure 3-2. As expected, plants transformed with these constructs also
displayed wound-inducibility of the transgenes. RNA gel blot analysis revealed increased
GUS levels 3 h following wounding of leaves of transformed plants (Figure 3-5A), while
LUC activity was again seen only following wounding of p2082 lines (Figure 3-5B). My

results confirm that wounding causes a transcriptional induction of RN51.

Use of promoter constructs to analyze signals mediating RNSI induction

One unique aspect of the wound-induced expression of RN51 is the fact that JA and
OGAs are not required (Chapter 2 and LeBrasseur et al., in press). The LUC reporters
offered us the opportunity to screen for various signals that might induce RN51
expression. I used the lines to examine several aspects of RN51 expression, including
response to hormone treatments, interaction of the wound-induction with these hormones.
and effects of other external stimuli.

Since dehydration has been proposed to be the signal for the wound induction of J A-
independent genes (Reymond et al., 2000), I first assessed whether dehydration caused
induction of RN51. Figure 3-6 shows that dehydration did not cause induction of LUC

activity in adult plants left unwatered for over two weeks, until leaves of the plants

A 2081.2 2081.3

 

0h 3h 0h 3h

RN51-LUC

 

 

Figure 3-5. 5' UTR sequences are not required for the wound responsiveness of the RN51
promoter.

(A) Northern blot analysis of RNA from unwounded (0 h) and wounded (3 h) leaves of
T2 plants (2081.2 and 2081.3) expressing the GUS reporter controlled by 2.7 kb of the
RN51 promoter with no 5' UTR sequence.

(B) Luciferase activity (bottom panel) in control and wounded (WND) wild-type (WT)
and transgenic plants expressing the LUC reporter under control of the RN51 promoter
(RNSI-LUC) as in (A).

53

Control

Wounded

 

Figure 3-6. Dehydrated plants do not induce the RN51 promoter.

Plants were dehydrated as described in the text. Luciferase activity (right panel) is seen
only after wounding the dehydrated plants. No RNS] transcript was detected by Northern
blot analysis of dehydrated WT seedlings (not shown).

 

Figure 3-7. RN51 transcript accumulates in response to salt stress.

Ten-day-old wild-type Arabidopsis seedlings were transferred to plates containing
250 mM NaCl and subsequently harvested at the time points indicated. RNA (15 pg) was
analyzed by Northern blot using an RN51 probe.

This blot was kindly provided by Daniel Cook.

54

appeared purple from high levels of anthocyanins (not shown). Wounded dehydrated
plants were used as a control and had high LUC activity. The activity in the wounded
dehydrated plants was the strongest seen in any of the LUC experiments performed in
this work. Therefore, I speculate that although dehydration does not cause RN51
induction, it may potentiate the wound-induction. To confirm that RN51 does not respond
to dehydration, I probed a blot containing RNA from dehydrated seedlings. No RN51
signal was detected on the blot (not shown), which contained RNA from time points 0-8
h of dehydration, although typical dehydration-responsive transcripts were induced (D.
Cook, personal communication). Since both the LUC reporter and endogenous RN51
transcript did not respond to the dehydration treatment, I conclude that dehydration is not
the signal that causes induction of RN51 by wounding.

As discussed briefly in Chapter 1, 5. cereviseae has only one member of the RNase
T2 family, Rnyl. Since the RNYI transcript accumulates in response to salt stress
(Maclntosh et al., 2001), I were interested in examining the response of RNSI to a similar
treatment. RNA blots containing RNA samples from Arabidopsis seedlings transferred to
plates containing 250 mM NaCl were probed with RN51. As seen in Figure 3-7, like
RNYl, RN51 also accumulated in response to high NaCl concentrations. Transcript was
detected by 2 h of treatment and remained high at 8 h.

Recently, a mutant deficient in ABA signaling has been isolated that affects an
mRNA cap-binding protein and is suggested to affect RNSI levels through a
posttranscriptional mechanism (Hugouvieux et al., 2001). In an attempt to identify any
effects of ABA on RN51 expression, I treated the LUC lines with ABA. LUC activity was

seen 5 h after addition of ABA to seedlings growing in liquid‘medium (Figure 3-8A). I

55

also probed a blot containing RNA from ABA-treated seedlings and found that
endogenous RN51 transcript accumulated beginning 2 h after transfer of seedlings to
plates containing 100 pM ABA (Figure 3-8B). RN51 levels remained high after 8 h of
treatment. Thus, ABA caused induction of RN51. Most likely, this induction is
transcriptional, since I detected LUC activity upon ABA-treatment. These results point
toward a possible signal for the wound induction of this gene.

Since ABA induced RN51, I examined whether ABA signaling is required for the
wound induction of this gene. Several different mutants in the ABA pathway were
wounded and analyzed for RNase activity. Two of the mutants, abiI and abi2, are
blocked in ABA signal transduction, and a third, abaI-I, is deficient in ABA
biosynthesis, (for a review on ABA signal transduction, see Leung and Giraudat, 1998).
Twenty-four hours following wounding, all three mutants had elevated levels of RNSI
activity (Figure 3-9). Thus, although ABA is able to induce RN51 transcription, it is not
required for the wound induction. The 33-kDa activities were also induced in the
mutants, indicating that the bifunctional nucleases are also upregulated independently of
ABA.

Previous studies in this laboratory indicated that wounding and heat shock may
interact to regulate RN51 expression. I used the LUC and GUS reporters to examine the
effects of these two stimuli on RN51 transcription. In multiple repetitions, LUC activity
was not seen after wounding when combined with treatment at high temperatures (Figure
3-10A). High temperature alone had no effect on LUC activity. Since it has been reported
that high temperature can affect RNase activity posttranslationally (Chang and Gallie,

1997), I tested whether the RN51 transcript was affected by heat shock.

56

 

0 1 2 4 6.8,.-.“

 

Figure 3-8. ABA induces the RNSI promoter.

(A) Luciferase activity (bottom panel) in seedlings grown in liquid culture 5 h after
addition of 5 pM ABA (+ABA) or buffer alone (-ABA).

(B) Northern analysis of RNA isolated from seedlings treated with ABA for the indicated
times.

RNA gel blot in (B) provided by Daniel Cook.

57

WT abi1 abi2 aba1-1
0 24 0 24 0 24 0 24 h

 

m <— 33 kD

 

Figure 3-9. Mutants in ABA signaling and biosynthesis induce RNSI by wounding.

Wild type (WT) and mutants abi1, abi2, and abal-l were examined for RNase activities
in unwounded (0) and wounded (24) 2-week-old seedlings. Induction of RNSI and the
33-kDa activities is seen in all four lines.

58

For this experiment, stems of wild-type plants were wounded, then cut and floated in a
buffer at either 22 or 42°C. The wound induction of RN51 is diminished by the heat
treatment (Figure 3-IOB). These results indicate that heat shock inhibits the wound

response of RN51 .

DISCUSSION

The findings in Chapter 2 of this thesis revealed that RNSI responds to wounding through
a pathway not previously identified in Arabidopsis. This has led to the obvious question
as to what signal transduces the wound stimulus to cause increased levels of RN51? l
addressed this question by examining the level of regulation of this gene and found that
RN51 is transcriptionally activated by wounding. I next examined several other factors
that might affect RN51 transcription and found that, although ABA induces RNSI

expression, it is not required for the wound induction.

Sequence requirements for the wound induction of RN51

The accumulation of RN51 following wounding occurs within 2 h in Arabidopsis leaves
(refer to Figure 3-IOB). In fact, previous microarray analyses of wounded leaves found
that the fold-induction of RNSI at 2 h was higher than any other of the 600 spots
examined (M.A. Pérez-Amador and P.J. Green, unpublished). In the unwounded state,
this transcript normally can not be detected or is at very low levels in leaves of adult
Arabidopsis plants (Bariola et al., 1994). Thus, wounding apparently causes a strong
response of RN51. The accumulation of RN51 could be due to either a transcriptional
induction or a posttranscriptional mechanism, such as an increase in stability of the
transcript. Posttranscriptional regulation of this transcript has been suggested by previous

experiments that showed that although P, starvation caused increased transcription of

WND 22°C

WND 42°C

 

.. Control
22°C
- 42°C

'_ h 7, ,1;- RNS1

Figure 3-10. Heat shock inhibits the wound-induction of luciferase activity.
(A) Transgenic plants expressing the RNS lp-LUC construct were wounded and placed at

either 22 or 42°C for 5 h. Luciferase activity (right panel) was seen only in the wounded
plants left at 22°C. Similar results were seen in seedlings (not shown).

(B) RNSI accumulation in unwounded (control) and wounded stems incubated at 22 or
42°C for 2 h after wounding.

Experiment in (B) was performed by Gustavo Maclntosh

60

RN51 promoter fusion reporters, transcriptional regulation could not account fully for the
levels of RNSI under these conditions (Howard, 1996).

In contrast to these results, I found that increased transcription of RN51, as directed
by 2.7 kb of promoter sequence, causes the majority of the increased levels of this
transcript by wounding. However, I did detect small effects of the transcript alone upon
wounding. This effect was only seen when the introns and full 5' UTR sequence were
included in the reporter; the cDNA sequence alone did not confer wound-inducibility. On
the other hand, the full 3' UTR is not required for the wound induction (see Chapter 4). In
conclusion, the RN51 promoter contains the sequences most critical for the regulation of

the induction of the RNSI enzyme by wounding.

Response of RNSI to various biotic and abiotic stimuli

Previous work had shown that RN51 transcript accumulates in P ,-starved seedlings and in
senescent leaves (Bariola et al., 1994). Work in this thesis has shown that multiple other
signals direct expression of this gene. Wounding, as seen here and in Chapter 2, strongly
influences RN51 transcription. This effect is not directed by the known wound signals
jasmonic acid and OGA elicitors. Thus, I was interested in what other internal or external
cues alter RN51 levels. To examine this question, I used the LUC and GUS lines to screen
many different stimuli.

Using this screen, I found that ABA causes increased transcription of RN51 and the
LUC reporter. ABA has been suggested to affect RN51 posttranscriptionally (Hugouvieux
et al., 2001). A posttranscriptional mode of regulation was suggested because the authors
isolated a mutant in a gene that encodes a cap-binding protein. The mutant shows ABA-

hypersensitivity and down-regulates several transcripts implicated in ABA signaling.

61

RN51 is one of the transcripts with lower basal levels in this mutant. These results may be
slightly misleading, since RN51 transcript is usually just at or below the level of
detection, suggesting that the wild-type plants used for Northern and microarray analyses
in their study may have been stressed. Nonetheless, if RN51 is affected in this mutant, it
is logical to infer that this transcript is normally posttranscriptionally regulated by ABA. I
found that ABA causes transcriptional induction of a reporter controlled by the RN51
promoter (Figure 3-8A); however, I cannot rule out the possibility that certain ABA
effects are carried out through a posttranscriptional mechanism. Treatment of the nos-
RNSchNA or nos-preRNSl lines with ABA should address this possibility.

Although ABA does affect transcription of RNSI, it is not required for the wound
induction of RNSI activity. Similarly, the wound-induction of the 33-kDa activities does
not require ABA signaling or biosynthesis (Figure 3-9). It is possible that local RNSI
induction is independent of ABA, while systemic induction requires ABA. The systemic
regulation should be addressed in future experiments.

One interesting result derived from the dehydration experiments was a particularly
strong induction of the LUC activity in wounded dehydrated plants. Dehydration is
known to induce ABA production in plants (Thomashow, 1999). Since ABA caused
increased LUC activity in the transgenic lines, it is possible that the combined effect of
wounding and increased ABA levels may have an amplification effect on the induction of
the reporter. This effect may explain the increased RN51 transcript levels identified in the
dehydration experiment by Reymond et al. (2000), since dehydration was achieved by
detaching rosette leaves from the roots, possibly causing a weak wound response that was

amplified by the dehydration. It appears, however, that the dehydration-elicited ABA

increase alone is not sufficient for RN51 induction, since both luciferase activity in the
transgenic lines (Figure 3-6) and RN51 transcript in WT plants (not shown) did not
accumulate in response to dehydration.

The differential control of RNSI by ABA and dehydration is particularly interesting
since overexpression of CBF! or CBF2 results in accumulation of RN51 (M.F.
Thomashow, personal communication). CBF] and CBF2 are transcription factors rapidly
induced by cold temperature in Arabidopsis (Gilmour et al., 1998), and overexpression of
CBF] results in increased cold tolerance (J aglo-Ottosen et al., 1998). Interestingly, RNSI
is not induced by cold treatment (not shown), nor is it induced by overexpression of
CBF3 (M.F. Thomashow, personal communication), another in the cold-induced CBF
family (Gilmour et al., 1998). Therefore, cross-talk between pathways that control
dehydration, cold, and ABA responses appears to regulate RNSI accumulation in a highly
specific manner. Alternatively, overexpression of the transcription factors may cause
secondary effects that result in increased RNSI accumulation. ’

While dehydration may potentiate the wound-induction of RN51, we found that heat
shock inhibited this response. Posttranslational effects on RNase activity by heat shOck
have been suggested to be one mechanism by which plants reduce mRNA turnover until
recovery has occurred (Chang and Gallic, 1997). RNSI is an extracellular enzyme and
therefore is not expected to be involved in mRNA decay. Additionally, this effect is
apparently not posttranslational, according to the reduced RNSI transcript levels by heat
shock (Figure 3-IOB). The wound-induced 35-kDa nuclease activity was also inhibited
by the heat treatment (G.C. Maclntosh and P.J. Green, unpublished). I am currently using

the GUS reporter lines to determine whether RNSI is transcriptionally affected by heat

63

shock. Further, effects on the 33-kDa nuclease activities will be examined.

During the screen, I also examined several other factors that did not cause increased
LUC activity (not shown). These include auxin and cytokinin. Additionally, phosphatase
and kinase inhibitors did not affect the induction of the LUC reporter by wounding. It is
possible that the treatments were not effective and that further studies will uncover
regulation by these factors; however, it currently appears that these two hormones and
inhibitors of other Arabidopsis wounding signal transduction pathways (Rojo et al., 1998)
are not involved in the wound induction of RN51.

In conclusion, transcription of the RN51 gene is regulated by several factors.
Wounding, senescence, and P,-starvation were already known to control RNSI levels
(LeBrasseur et al., in press; Bariola et al., 1994). I have now shown by promoter fusion
analysis that the wound induction is largely a transcriptional response. ABA and NaCl are
other factors that regulate RNSI transcription, while stresses such as cold and dehydration
do not. Although dehydration does not cause RN51 induction, it may amplify the
induction by wounding. In contrast, heat shock has an inhibitory effect on RN51 wound-
induction. There appear to be multiple networks controlling the expression of RN51.
These may be important for tightly regulating RNSI activity levels under various stress
conditions. Understanding of the conditions regulating RNSI activity is important in
order to elucidate the function of this and related enzymes in plants. This question will be

addressed in the following chapter.

MATERIALS AND METHODS

Nuclei preparations and run-on analyses

Seedlings for nuclei purification were grown on Pi and PT media as described (Bariola et

64

al., 1994). Nuclei were purified as follows. Frozen 7-day-after-transfer seedlings were
ground in Nuclear Isolation Buffer (NIB; 1.0 M sucrose, 5 mM MgCl2, 10 mM Tris pH
7.2) to a powder. The ground tissue was filtered through two layers of miracloth, Triton
X-100 was added to 0.5% (v/v) to lyse the chloroplasts, and the filtrate was spun at
1,000 X g at 4°C for 10 min. The pellets were resuspended in 0.5 ml NIB plus 0.5%
Triton and spun as above, for 15 min. Pellets were then resuspended in Nuclei Storage
Buffer (50% glycerol, 5 mM MgCl2, 2 mM DTT, 20 mM HEPES pH 7.2) and frozen at
~80°C until used for run-on analysis. An aliquot of nuclei from each preparation was
DAPI-stained, and nuclei were counted to estimate the number of nuclei purified.

Nuclear run-on reactions were begun by adding each of the following to a 50-pl
aliquot of nuclei for a final volume of 100 pl: 1 x transcription salts (10 mM MgCl2,
100 mM ammonium sulfate, 20 mM HEPES pH 7.9), 10 pl glycerol, 1 pl 0.1 M DTT,
0.5 mM ATP, GTP, and CTP, 70 units of RNase inhibitor (Promega), and 100 pCi
32P-UTP (3000 Ci/mmol). After 20 min, 5 pl DNase RQI (Promega) was added to each
tube, and reactions were placed at 30°C for 5 min. Reactions were stopped with 300 pl
phenol/chloroform (50:50 v/v) and 200 pl stop solution (1% SDS, 20 mM EDTA,
100 mM LiCl, 10 mM aurintricarboxylic acid). The tubes were spun in a microcentrifuge
for 10 min at 4°C. The supernatant was transferred, and 75 pl of 10 M ammonium acetate
and 940 pl EtOH was added. After precipitation, the transcripts were resuspended and
used to probe slot blots containing 2.5 pg ssDNA of the appropriate plasmids.

ssDNA was purified using a procedure from Stratagene. Briefly, a single colony of
bacteria (E. coli strain DH5a F ') carrying the appropriate clone was inoculated into 5 ml

of 2 x YT medium containing 50 pg/ml ampicillin and l x 108pfu/ml VCM13 helper

65

phage and grown overnight at 37°C. Cells were pelleted, and 1 ml of supernatant plus 150
pg 20% PEG and 2.5 M NaCl was precipitated on ice for 15 min. After centrifugation,

the pellet was resuspended in 400 pl 0.3 M sodium acetate (pH 6.0) and 1 mM EDTA.

The ssDNA was extracted with phenol/chloroform and precipitated.

Cloning

The RN51 cDNA was digested with Smal and blunt-end ligated into a nos-globin gene
construct, p2031, with the globin gene insert removed. This construct, p1966, was then
digested with Pstl and ligated into the Pstl site of the plant transformation vector
pCambia 1301, which has the hygromycin resistance plant selection marker. This clone
was named p1975. The nos-globin control in the plant vector pCambia 2301 was named
pl 995 and confers kanamycin resistance to transformed plants. To clone the entire RN51
transcribed region, including the full 5' UTR and introns, primers (PG749 and PG750)
were designed for PCR amplification using wild-type genomic DNA. Each primer
contained a BamHI recognition sequence. The PCR product was isolated from an agarose
gel (Qiagen) and ligated into pGEM-T Easy (Promega). The insert was sequenced to
confirm the sequence of the PCR product and ligated with BamHl/BglII-digested p203].
The orientation of the insert was confirmed, the nos-RN51 insert was retrieved by
digestion with Pstl and ligated into the Pstl site of pCambia 2301.

Constructs p848, pl401, p1402, and p1432 were made previously (Howard, 1996).
To remove 5' UTR sequences from these constructs, primers corresponding to the
transcription start site (3' primer PG866) and to a site in the promoter approximately 420
bp upstream of the transcription start site (5' primer PG853) were synthesized. The 3'

primer inclUded a BglII restriction site. The PCR product was isolated and ligated into

66

l
pGEM-T Easy (p2028), sequenced to confirm the PCR product sequence, and digested
with BgIII and SpeI to isolate the correct 3' end of the promoter fragment. This 383-bp
fragment, which lacks the 5' UTR sequences present in constructs pl401 and p1432, was
used to replace the 475-bp BgIII/Spel fragment in the original vectors. These pMON-

derivative plasmids carry a kanamycin resistance gene.

Plant transformation

All plasmids were transformed into Agrobacterium tumefaciens strain GV3101 C58C1
Rif (pMP90) (Koncz and Schell, 1986) by electroporation, using a Gene-Pulser
apparatus (Bio-Rad) as per the manufacturer's instructions. Arabidopsis thaliana ecotype
Columbia plants were transformed by vacuum infiltration according to the method of
Bechtold et a1. (1993). Seeds from transformed plants were plated on selection medium
containing either kanamycin at 50 pg/ml or hygromycin at 30 pg/ml. Resistant seedlings
were transferred to soil after 2 weeks. Expression of the transgene was tested in adult
plants by Northern analysis (not shown). Lines expressing the transgenes were used for

further studies.

Plant treatments

Plants used for various treatments were selected on medium with appropriate drugs and
transplanted to soil. Seedlings grown in liquid culture were sterilized for 7 min. in 50%
bleach/0.02% Triton X-100, washed three times in ddH2O, and resuspended in 0.1%
agarose. Approximately 25 — 30 seeds were added to each well in 12-well ELISA plates.
Each well contained 4 mL of MS medium (Life Technologies, Rockville, MD)
supplemented with 0.5% (w/v) sucrose. Plates were shaken on a rotary platform at ~50

R.P.M. in a growth chamber at 16 h light and 22°C. Treatments were performed on 10- or

67

14-day-old seedlings.

ABA was added to the liquid medium from a 5-mM stock solution dissolved in IN
NaOH. Final ABA concentration was 5 pM. Controls were treated with equal amounts of
1N NaOH. Wounding was done using ridged flat-tipped tweezers. Approximately 30% of

the leaf surface was wounded.

RNA and protein analyses

RNA and protein extraction, RNase activity assays, and Northern hybridization were

performed as described in Chapter 2.

Luciferase analysis

After each set of treatments, LUC-expressing and control plants were sprayed with the
substrate luciferin (Promega; 5 pg/ml in 0.011% triton). In some cases, luciferin was
added directly to the liquid medium instead of being sprayed onto the plants. Luciferase
activity was examined using a photonic camera 40 min after addition of the luciferin

substrate.

68

CHAPTER 4

RNSI FUNCTION: ANALYSIS OF AN RNS] MUTANT AND
OVEREXPRESSOR

69

ABSTRACT

I used a T-DNA insertional mutant, rnsl-2, and plants overexpressing the RN51 gene to
analyze the function of the RNSI ribonuclease. The rnsl-2 mutant lacked detectable
RNSI activity, even in wounded leaves. Plant morphology was examined. Unlike
antisense RNSI lines, rnsI-2 plants did not have elevated anthocyanins. However, in
support of a previous hypothesis regarding the function of S-like RNases in P,-
remobilization (Goldstein et al., 1989), the overexpressors were better able to utilize
RNA as a source of P,. I also found an unexpected effect of RNSI activity on root length
in the plants. While the mutant had longer roots, the overexpressing plants had shorter
roots than the wild type. These results indicate that RNSI is involved in processes other
than recycling P,. since the phenotypes were seen independently of P, concentrations in

the medium.

70

INTRODUCTION

For decades, RNases have been used as models for structural and enzymatic studies.
RNase A in particular has been extensively analyzed, and crystallized structures of wild
type as well as enzymatic mutants have been determined (Gilliland, 1997). Despite the
knowledge gained on enzymatic activity, little is known concerning physiological
functions of secreted ribonucleases. In the T2 family, the exception is the S-RNases
involved in self-incompatibility in the Solanaceae (Parry et al., 1997). Previously,
antisense methods were used to reduce Arabidopsis S-like RNase activity to address the
function of members of this family. These experiments revealed that seedlings with low
RNSI or RNS2 activity had high anthocyanin content, an effect that was amplified by P,
starvation (Bariola et al., 1999). These results are consistent with the hypothesis that S-
like RNases are involved in the mobilization and recycling of P, during P, starvation and
senescence. However, in the antisense experiments, RN51 transcript levels were never
reduced below 10% of that of the wild type (Bariola et al., 1999). Also, the segregation of
kanamycin resistance in the antisense lines indicated that they may have had silencing or
T-DNA rearrangement complications (P.A. Bariola, G.C. Maclntosh, and P.J. Green,
unpublished). Recent results, including the systemic induction of RNSI during wounding
(LeBrasseur et al., in press) and phenotypes of the insertional mutant of the yeast T2
RNase (Maclntosh et al., 2001), have indicated new directions regarding the functions of
secreted RNases. This chapter will focus on the use of insertional mutagenesis (Krysan et
al., 1999) to examine the function of the RNSI enzyme. Along with this project, I also
overexpressed RN51 in plants to examine the effect of increased RNSI activity. Previous

attempts to overexpress RN51 in roots did not lead to increased levels of the transcript

71

beyond that induced by P, starvation (Bariola, 1996). In this study, I expressed the RN51

cDNA throughout the plant under the control of the strong cauliflower mosaic virus

(CaMV) 355 promoter.

RESULTS

Isolation of ms] mutants and RNSI overexpressing plants

To get a more accurate assessment of the function of RNSI, I screened the University of
Wisconsin T-DNA-tagged lines for insertions within the RN51 gene. PCR primers
complementary to sequences in the promoter region of RNSI (650 bp upstream of the
transcription start site) and 450 bp 3' of the end. of the cDNA were synthesized Two
putative insertion lines identified in the primary screen were pursued further. The
locations of the insertions are shown in Figure 4-1. The alleles were named rnsl-I and
rnsl-2. The rnsl -1 allele lies within the 3' UTR, while the rnsI-2 allele is 85 bp 5' of the
translation start site in the 5' UTR. PCR analysis revealed that the left border (LB) of the
T-DNA flanked both ends of the insertion in both alleles (not shown). Plants homozygous
for each allele were identified by PCR analysis. To analyze RNSI function firrther, I
transformed plants with the RNSI gene under the control of the strong constitutive CaMV
355 promoter (Figure 4-1).

Once plants homozygous for the T-DNA insertions were identified, I examined RNSI
activity in the two mutants. Protein extracts from unwounded and 6-h wounded leaves
were run on RNase activity gels. The rnsl-I plants still had significant levels of wound-
induced RNSI activity (Figure 4-2A), as might be expected since the T-DNA insertion
lies in the 3' UTR. Not only was there still activity, but also regulation of the gene

appeared to be unaffected by the insertion: there was no activity in the absence of

72

 

 

RN51 1" l

 

 

 

 

Figure 4-1. rnsl mutants and overexpressor constructs.

Two RN51 T-DNA insertion alleles were identified. The first insertion, rnsI-l, is
downstream of the RNSI polyadenylation site (asterisk). The second, rns1-2, lies in the
5' UTR, 85 bases 3' of the translation start site (arrow). The bent arrow indicates the
transcription start site. Boxes represent exons, and lines represent introns. To overexpress
RN51, the cDNA was cloned between the CaMV 355 promoter and the pea E9 3' end.
This construct was named p2020.

73

wounding, and strong activity after wounding. The regulation in the mutant is consistent
with sequences required for the wound-induction, as demonstrated in Chapter 3.

In contrast, rnsI-2 mutants lacked RNSI activity, even after wounding (Figure 4-
28). I therefore studied this allele further. To confirm that the rnsI-Z line was
homozygous for the T-DNA insertion, DNA gel blot analysis was performed. Genomic
DNA digested with EcoRI was subjected to DNA gel blot analysis using an RNSI probe.
While the probe hybridized to a 5-kb band in wild-type DNA, only a 2-kb band was seen
in the rnsl-2 lane (Figure 4-3 B). The absence of the S-kb band in this lane indicates that
the rnsl-Z allele is in the homozygous state in this line. The blot was also hybridized with
a T-DNA probe to determine the number of T-DNA insertions in the genome. Sizes of
the expected bands, based on restriction maps of the T-DNA vector and genomic region
surrounding RN51, are 1.0, 1.7, 4.6, and 5.1 kb. Bands of these sizes were observed
(Figure 4-3A), and no other bands were found on the blot, indicating that the T-DNA had
inserted at only one site in the rnsI-Z line. However, at least two copies of the T-DNA
were present at this allele, since PCR fragments were obtained using primers
corresponding to the LB of the T-DNA and to either the 5' or 3' ends of the RN51
genomic region (not shown). The exact number of T-DNA copies inserted was not
determined.

Since this insertion lies in the 5' UTR of the gene and the gene might therefore
produce transcripts including the full open reading frame, I performed RNA gel blot
analysis of wounded and unwounded leaves from the mutant. Several aberrantly sized
bands could be detected using an RN51 probe (Figure 4-3C). These were found only upon

wounding, confirming previous results that the 5' UTR is not required for wound

74

WT rns1-1

 

 

WT rns1-2
03 62403 624hpw

 

w‘m“ “YT“ . . If <~“35 kD

«RN51

 

Figure 4-2. RNSI activity in the rnsI-l and rnsl-2 mutants.

(A) Wounded (7 hours post wounding, hpw) and unwounded (0) leaves of 4-week-old
rnsl-l and wild-type (WT) Ws Arabidopsis plants were analyzed by RNase activity gel
assay.

(B) Leaves of WT Ws and rns1-2 mutants were wounded, harvested at timepoints
indicated, and analyzed by RNase activity gel assay. RNSI activity (bottom arrow) can
be found only in WT wounded samples. Induction of the 35-kDa RNase activity (top
arrow) provides a positive control for the wound response of the rnsl-Z leaves. hpw,
hours post wounding.

75

A

[WT rnst-Z
I“ — 5.1

 

—— 1.7

. ——1.0

C WT rns1-2

 

0 3 624 03 624hpw

 

Figure 4-3. DNA and RNA gel blot analyses of rns1-2 mutant plants.

Genomic Southern analysis of wild-type (WT) and rnsl-2 plants using the T-DNA (A) or
RNSI (B) probe. Estimated sizes of hybridizing bands are shown to the right.

(C) Northern analysis of wounded WT and rnsl-2 leaves harvested at the indicated
timepoints using an RNSI probe. hpw, hours post wounding.

76

induction of RN51 (see Chapter 3). The various sizes could be due to irregular splicing
within the T-DNA and RN51. One of the smaller bands was approximately the same size
as the wild-type transcript; however, RT—PCR analysis of RNA from wounded rns1-2
leaves did not detect any polyadenylated transcript that included the start of the open
reading frame (not shown). Additionally, RN51 transcript induced in the wild type was
significantly more abundant than the aberrant bands produced. Thus, although it is
possible that a small amount of wild-type transcript is made and properly translated in
this mutant, I predict the amount to be minimal.

Plants with high RNSI activity in the absence of wounding were identified from a T,
population of the 3SS-RNS1 lines. Leaves were harvested and analyzed by activity gel
assay. Multiple T. plants had high RNSI activity; two examples are shown in Figure 4-4.

Seeds from a homozygous T2 progeny plant of line 2020.2 were used in further studies.

The rnsI-2 mutant does not have increased levels of anthocyanins

Since RNSI antisense lines have high levels of anthocyanins, particularly when grown on
P,‘ medium, I looked for a similar effect in the rns1-2 mutant, which had lower RNSI
activity induced by wounding than found in antisense plants subjected to P,-starvation
(compare Figure 6 in Bariola et al., 1999 and Figure 4-2b of this thesis). Thus, I expected
a similar or more severe anthocyanin phenotype to be associated with the rns1-2 mutant.
The anthocyanin content of rns1-2 seedlings was measured after 7 d of growth on
medium either containing (P+) or lacking ( P‘) P,. Antisense RN52 lines, which have high
anthocyanin levels similar to antisense RN51 plants (Bariola et al., 1999), and wild-type

Ws seedlings were used as positive and negative controls, respectively. As expected, the

77

M 3SS-RNS1
1 2

<— RNS1

 

Figure 4-4. RNSI activity in overexpressing plants.

Protein extracts of unwounded leaves of wild-type (WT) and two T, 35S-RNS1 plants
were analyzed by RNase activity gel assay. High RNSI activity (arrow) is seen only in
the overexpressing plants.

78

antisense line had a modest increase in anthocyanins when grown on Pi medium and a
stronger increase on' P" medium. However, the rnsl-2 mutant had anthocyanin contents
similar to wild-type seedlings (Figure 4-5). This unexpected result was seen regardless of
whether the rnsl-2 seedlings were grown on kanamycin-containing selection medium
(kanamycin selection was used for the antisense lines in Bariola et al., 1999). Thus, it
appears that the antisense lines have high anthocyanins for a reason other than the

absence of RNSI or RNS2 activity.

RNSI activity appears to affect plant growth on RNA

Expression patterns of RN51 during P,-starvation indicated that the RNSI enzyme may be
involved in nutrient recycling during P, stress (Bariola et al, 1999). Therefore, Iexamined
the effect of low P, concentrations in the growth medium on rnsI-Z and 3SS-RNSI
seedlings. No obvious morphological difference between wild-type and transgenic
populations grown on concentrations of P, varying from 0 to 1250 pM was found (not
shown).

Further studies were conducted to examine the ability of the mutant lines to grow
when provided with RNA as the sole source of phosphate. In the initial experiment, seeds
were plated densely on AGM medium with either 1250 pM P, (P+) or 0.6 mg/ml RNA.
On RNA-containing medium, the 35S-RNS1 seedlings were larger and bolted and
flowered sooner than the WT Col seedlings, while the rnsI-2 mutants were smaller and
lagged developmentally compared to WT Ws (Figure 4-6A). I then repeated the analysis
of the 3SS-RNSI and WT Col seedlings with only 12 seedlings per plate to increase
spacing between the'plants. Once again, the overexpressing lines grew larger on the RNA

plates (Figure 4-6B). While 35S-RNS1 seedlings were larger on RNA medium, on PL

79

Anthocyanin levels of ms mutants

ElP+'
"Pi

Relative anthocyanin levels

 

rns1-2 AIS rns'2 WT
Line '

Figure 4-5. Anthocyanin content in mutant ms lines.

Anthocyanin contents of rns1-2, wild-type Ws (WT), and antisense RNS2 (A/S rnsZ)
lines grown in the presence (white bars) or absence (gray bars) of P,. Like antisense
RNSI seedlings, the antisense RN32 line had high anthocyanin levels, particularly in the
absence of P,. The rnsl-Z line had anthocyanin levels similar to that of the wild type.

80

medium, as determined by fresh-weight analysis, the 35S-RNS1 seedlings were smaller
than WT Col (Table 4-1; see below for more on this effect), indicating that high levels of

RNSI activity aid the overexpressor line in growing on RNA.

RN S1 activity affects plant size

When I first examined the rns1-2 mutant, I expected the mutant to be less able to grow on
medium containing low concentrations of P,. I did not observe such effect, at least based
on morphological examination (not shown). However, during a test in which I plated the
wild type and mutant on various concentrations of P,, I noticed that the roots of the
mutant appeared different. I then plated the lines again and grew them vertically, to allow
the roots to grow downward. I plated on two different concentrations of P,: 1250 pM
(concentration of P, in P+ plates) and 62.5 pM. The latter concentration allowed the plants
to grow but caused severe starvation phenotypes, including high anthocyanin levels and
severely reduced growth. rnsI-2 mutants grown on both of these P, concentrations
appeared to have longer roots than the wild type Ws seedlings (Figure 4-7A). To
quantitate these differences, roots were measured and analyzed by t-test (Table 4-2). In
each of five repetitions, the differences in root length were statistically significant
(P<0.001). They were also different independent of the P, concentration in the medium,
indicating that this phenotype is not related to the putative function of RNSI in P,
remobilization.

Since RNSI activity affected root length in the mutant, I also examined the
35S-RNS1 roots. In two repetitions, the 3SS-RNSI seedlings had significantly shOrter
roots than the Columbia wild type (Figure 4-7B; P<0.001). Therefore, it appears that

levels of RNSI activity negatively correlate with root length in Arabidopsis.

81

 

' 3SS-RNS1

 

 

Figure 4-6. Growth of msI-Z and 35S-RNS1 seedlings on RNA-containing medium.

(A) 35S-RNS1, wild-type Columbia (WT Col), ms] -2, and wild-type Ws (WT Ws) seeds
were grown on P,‘ plates containing 0.6 mg/ml RNA. After 34 days, the transgenic lines
exhibited differences from their wild-type counterparts. While the WT Ws plants had
bolted, rns1-2 plants had not. Similarly, the WT Col seedlings, which normally bolt later
than Ws, had still not yet bolted, while many of the 35S-RNS1 plants had bolted and
flowered. Arrowheads point to flowers in the Ws and 35S-RNS1 plants.

(B) 3SS-RNSI and WT Col seedlings grown on P,’ plates containing 0.6 mg/ml RNA.

82

Figure 4-7. Root phenotype of RN51 mutant and overexpressor.

(A) Root length of 12-day-old rnsl-2 and wild-type Ws (WT Ws) seedlings grown on
vertical plates containing either 1250 or 62.5 pM P,.

(B) Root length of l2-day-old 3SS-RNSI and wild-type Columbia (WT Col) seedlings
grown on 1250 pM P,.

83

WTWs

ms 1-2

 

M
u.
5
2
6

 

 

WT Col

35S-RNS1

 

 

Figure 4-7. Root phenotype of RNSI mutant and overexpressor.

84

 

Table 4-1. Statistical analysis of fresh weight of WT, rnsI-2, and
35S-RNSI seedlings'

 

 

Line Ave. weight per seedling t P
(mg)

wr Col P+ 12.9 2.3 <0.05

35S-RNS1 P+ 8.8

WT WS P+ 9.6 4.3 <0.01

rnsI-2 P+ 12.9

WT Col RNA 10.8 4.8 <0.01

35S-RNS1 RNA 15.7

 

IThe t-test was used to determine statistical significance of the
weight differences between 9-day-old WT C01 and 35S-RNS1 and
WT Ws and rns1-2 seedlings grown on P+ medium and between 37-
day-old WT C01 and 3SS-RNSI seedlings grown on P' medium
containing 0.6 mg/ml RNA. Results shown are from one
experiment. A P-value of 0.05 or less is considered statistically
significant.

 

85

 

Table 4-2. Statistical analysis of root length 1n ms] -,2 WT Ws, 35S-RNS1
and WT Col seedlings

 

Age Ave Root

 

 

 

 

 

 

Test Line P, (pM) (days) length (mm) t P

rns1-2 10.0

WT Ws 62.5 6.6 12.5 <0.001
1 rnsl-2 13 48 3

WT Ws 1250 31.0 12.6 <0.001

rnsI-2 7.1

WT Ws 62.5 3.1 8.1 <0.001
2 ms! 2 12 37 4

WT Ws 1250 29.5 8.0 <0.001
rnsl-2 4.5

WT Ws 62.5 3.6 6.0 <0.001
3 rnsI-2 7 l6 9

WT Ws 1250 13.5 5.3 <0.001
rns1-2 36.8

4 WT Ws 1250 8 29.1 5.3 <0.001
35S-RNSI 41.2

5 WT C01 1250 12 50.3 4.4 <0.001
rnsI-Z 24.6

6 WT Ws 1250 9 18.3 5.2 <0.001
35S-RNS1 ’ 19.5

WT C01 1250 24.3 5.2 <0.001

 

IThe t-test was used to determine statistical significance of the differences in
root lengths in seedlings grown on two different P, concentrations. A P-
value of 0.05 or less is considered statistically significant.

86

In addition to the roots, the size of the green tissue of the rnsl-2 and overexpressor
seedlings appeared to differ from the size of the wild types. I quantitated this difference
by growing the four lines on P+ medium and weighing the seedlings. I removed the roots
to eliminate any effect of the differences I observed in root tissue. T-test analysis of the
weights indicated that rnsl-2 seedlings were significantly larger than WT Ws seedlings
(P<0.01), and 35S-RNS1 seedlings were smaller than WT Columbia (P<0.05) (Table
4-1). These results are preliminary but are consistent with the phenotype seen in root
lengfli

The difference in root length could be due to cell size, cell number, or a combination
of the two. To address this issue, I examined the roots by laser confocal microscopy.
Roots of the four lines were fixed and stained using two dyes, DAPI to stain nucleic acid
and calcofluor to stain the cell wall. I did not observe a large increase in size of the
rnsl-2 cells or smaller 35S-RNS1 cells (not shown), but a small difference may be
difficult to visualize using this technique. A more involved experiment, including
measurements of hundreds of cells, will be a more thorough method to assess any
differences.

During the analysis of these roots, I noticed a difference in the rate at which the
rnsl-2 and wild-type roots absorbed the dyes. Five minutes after adding dyes to the WT
roots, the intensity of the stain was low (Figure 4-8). After 45 min, the staining was
stronger, and a larger portion of the root could be viewed. However, even at this point,
the intensity of the WT root was not as high as the rnsI-2 root only 5 min after the dyes
were added. These results are preliminary and require repetitions to ensure that the effect

was not an artifact of the staining preparation procedure. However, as will be discussed

87

 

Figure 4-8. Absorption of dyes in wild-type Columbia (WT) and rns1-2 mutant root
cells.

The dyes calcofluor and DAPI were added to fixed intact roots and photographed at the
indicated times. Staining of the mutant root cells was greater in 5 min than that of the WT
cells even after 45 min.

These photos were taken by Gustavo Maclntosh and Kirk Czymmek.

88

below, the differential penetration of the dyes was consistent with the phenotype seen in a
yeast mutant in the RNYI gene, the only homolog of RN51 in 5. ccreviseae, and supports
the hypothesis that RNase T2 enzymes are involved in regulation of membrane

permeability (Maclntosh et al., 2001).

Absence of RN S1 does not increase susceptibility of the plant to a bacterial pathogen

As discussed in Chapter 2, the induction of RNSI activity by wounding in unwounded
leaves is not easily explained as part of a nucleotide-recycling mechanism, since this part
of the plant has not been injured and should not require such a system. I hypothesized
that RNSI might also play a role in defense of the plant from opportunistic pathogens
that invade after wounding. Thus, as a first examination of this putative function, I tested
the ability of a bacterial pathogen, Pseudomonas syringae pv tomato (Pst) DC3000 to
infect the rnsI-Z and 3SS-RNSI plants in comparison to their respective WT.

Leaves of the rns1-2, WT Ws, WT C01, and 3SS-RNS1 plants were inoculated with
Pst, and growth of the pathogen was monitored for three days. In two repetitions, growth
of the pathogen on the rnsI-2 mutant was similar to that on WT Ws (Figure 4-9A).
Unlike rns1-2, the overexpressor appeared to have an effect on the growth of the bacteria;
in four independent experiments, I consistently detected a lower Pst grth rate in 35S-
RNSl compared to that in WT Col (Figure 4-9B). However, error bars indicated that
variation between growth on the two lines was too great to make a statistically valid
conclusion. Additionally, analysis of variance indicated no significant difference between
growth of the bacteria on the 2nd day after infection. Further repetitions Of this
experiment, particularly focusing on day two, may indicate valid differences. However, I

currently cannot conclude that RNSI activity affected grth of this pathogen.

89

Figure 4-9. Growth of Pst DC 3000 on leaves of 3SS-RNS1 and rnsl-2 plants.

(A) Leaves of 4-week-old wild-type Ws (Ws) and rnsl-2 mutants were infiltrated with P.
syringae. Starting on the day of infiltration (day 0), bacterial growth was measured until
day 3. Representative results of one oftwo experiments are shown.

(B) Growth of P. syringae on wild-type Columbia (Col) and 355-RNSI plants was
monitored as in (A). Averaged results from four independent experiments are shown.

90

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

P. syringae Growth Curve
7
A 6
‘3
t ” "
-- 4 ‘
C 4+ -,
S 3 1/‘. Q +WS
a, / ~--» rns1-2
2 2 '.
.5 1 ,X
3 ll/
,_ .
o o .
'1 T T r I
0 1 2 3 4 5
Days
Combined Growth Curve
6l — — 4 — — 1
5
a
e
a 4 .
§ ' / ..- .— T— .... ~..
g 3 v ‘V f N a J
a f, “'" IL
2 /”
5 2
i / l
o ,T J .
i
n.—4——m_ ,._ - ___ ,
o i 2 3 4 ;
Days j

 

Figure 4-9. Growth of Pst DC3000 in leaves of 35S-RNS1 and rns1-2 plants.

91

DISCUSSION

The vast majority of studies done on extracellular RNases, including the RNase T2
family, have focused on structural or enzymatic studies, or have addressed regulatory
questions. Gene expression or protein induction patterns have been particularly usefirl in
providing insights into the function of S-Iike RNases, but many questions still remain. I
have attempted to gain further information on the function of plant secreted RNases
through basic plant biology techniques, namely, overexpression and gene knock out of
RN51. The phenotypes of these plants have not only reinforced previous ideas concerning
the role of this family of enzymes, but also indicate that they may perform unexpected

functions in the plant.

Involvement of RNSI in phosphate remobilization

Previous data, including gene expression patterns and the phenotype of antisense RN51
plants, have been consistent with the idea that RNSI is involved in the remobilization of
P, from nucleic acid sources, either in the extracellular space or possibly from an
intracellular source (Bariola et al., 1999). I therefore expected similar insights to be
gained by studying the rnsI-2 null mutant and RN51 overexpressor. One phenotype of the
plants studied here is consistent with this previously proposed function: the
overexpressing plants appear better able to grow when provided with RNA as the only
source of phosphate. This phenotype is interesting because it is the first evidence that
RNSI may be secreted from the roots in order to degrade organic material in the
surrounding rhizosphere to provide the roots with P, for uptake. Introduction and
secretion of an Aspergillus phytase activity into Arabidopsis has been shown to improve

the plants' ability to utilize phosphorus from phytate, a major component of soil organic

92

phosphorus (Richardson et al., 2001 ). By extension, induction of secreted RNases may be
an endogenous method plants use to acquire P, from other organic sources. It will be
important to study this effect further, to quantitate the differences in plant weight and to
ensure that this effect is specific to grth on RNA by using DNA-containing medium as
a control. Although I did not have the opportunity to quantitate differences, the observed
growth effects on RNA are in contrast to the sizes of the plants grown on high P,, in
which case the 3SS-RNS1 plants appear to be smaller than the WT, indicating that the
effect of the RNA overcomes the differences in grth on medium containing high P,.

I did not detect any differences in the sizes of the plants grown on low concentrations
of P, that are consistent with a role for RNSI in P,-recycling. One possible interpretation
of this result is that the presence of RNS2, which is also induced by P, starvation,
provides the rns1-2 seedlings with sufficient RNA-degrading activity. However, the ms]-
2 mutant was less able to grow on RNA plates than the wild type. The different
localizations of the two enzymes may cause these effects; RNSI is extracellular, and
RNS2 appears to be intracellular, most likely either vacuolar or ER-associated (Bariola et
al., 1999). Therefore, RNS2 may be unable to degrade RNA in the medium because the
substrate is not accessible.

I was also surprised that the rns1-2 mutant did not have high levels of anthocyanins,
either in the presence or absence of P,. The antisense lines had significantly higher
anthocyanin contents, and this was shown to be specific for reduced RNSI or RNS2
activity; the vector controls did not have high anthocyanin content (Bariola et al., 1999).
Therefore, there appears to be another reason for the accumulation of anthocyanins in the

antisense lines. For example, as mentioned in the introduction to this chapter, kanamycin

93

segregation indicates that there may be an unusual silencing of the transgenes in the
antisense lines. The reduced kanamycin resistance may somehow result in increased
anthocyanins, even in the kanamycin resistant seedlings. Alternatively, the sequences
used to silence RN51 or RN52 may have an effect on the accumulation of RN54 and/or
RN55, which may then cause increased anthocyanins. Expression of RN52 was not
affected in the antisense RN51 lines, and vice versa (Bariola et al., 1999), but RN54 and
RN55 had not been identified at the time of the analysis of the antisense lines. The

expression of these genes in the antisense lines was therefore not explored.

Is RNSI involved in defense mechanisms?

One interesting hypothesis borne from the observation that RNSI is induced systemically
upon wounding is the idea that this enzyme may be involved in defense of the plant. My
initial results do not support this hypothesis. The rns1-2 mutant did not have increased
susceptibility to Pst. However, RN51 is not normally induced by infection with this
pathogen (not shown). Therefore, the mutant would not be expected to have increased
sensitivity. Additionally, Pst is virulent in Ws, and therefore, I may not be able to observe
further susceptibility. Interestingly, though, the RN51 overexpressing plants appeared to
have an increased resistance to Pst. This effect is only preliminary; further replications
must be done for this to be considered a reliable result.

Other pathogens may be more likely to reveal that the presence or absence of RNSI
can have an impact on defense mechanisms. For instance, it was shown that application
of RNase A in the extracellular space of tobacco leaves prevented the growth of the
oomycete Phytophthora parasitica (Galiana et al., 1997). Inoculation of tobacco leaves

with this pathogen also induced expression of the gene encoding RNase NE, another S-

94

like RNase. In the future, a pathogen more closely related to P. parasitica, such as

Peronospora parasitica, may be a better tool to analyze the role of RNSI in defense.

RNSI activity negatively correlates with root length

Possibly the most surprising phenotype that resulted from altering the expression levels
of RNSI was the difference in root length. This effect may extend to the green tissue of
the plant as well, as preliminary evidence suggests that the weight of this tissue is also
increased in the mutant and reduced in the 35S-RNS1 plants (not shown). These effects
are particularly unexpected because RN51 transcript was not detected in roots or leaves
by Northern blot (Bariola et al., 1994). Under non-inducing conditions, RNSI was
detected only in flowers. Most likely, RNSI is expressed in pistils, not petals, based on
microarray analysis of the apetalaZ and agamous mutants (SMD experiment ID 12302;
http://genome-wwwS.stanford.edu/MicroArray/SMD/). However, RN51 can be detected
in leaves by RT-PCR (not shown), indicating that some activity may also be present in
tissues other than flowers. Perturbation of this small amount of activity in roots and
possibly throughout the plant is evidently sufficient for the phenotypes I observed.
Similarly, although the yeast enzyme Rnyl cannot be detected by activity gel assay in
wild-type cells, the knock-out still shows the enlarged cell-size phenotype (Maclntosh et
al., 2001).

Examination of the cells in the roots did not reveal gross changes in the cell size. It is
possible that there is a small difference, e.g., 10%, in the cell size between the different
lines. Differences on this order would be difficult to detect by eye. As such, more in
depth analysis of the size of the cells is required.

I hypothesized that cells in rns1-2 would be larger than WT cells based on results

95

seen in a mutant in the RNYI gene of 5. cereviseae. In this case, the mutant cells were
approximately lO-fold larger in diameter than the WT cells and had enlarged vacuoles
(Maclntosh et al., 2001). Thus, I postulated that a similar effect may be causing the rnsl-
2 roots to be larger. However, the presence of the cell wall or factors regulating cell-cycle
may limit the expansion of the rns1-2 cells. As such, the cells may not be enlarged, but
may instead divide more rapidly. We plan to compare cell division rates in the mutant
and overexpressor lines by microscopic video analysis of the root cells during growth.
During the analysis of the root cells, we also noticed that rns1-2 cells seemed to
absorb the dyes more rapidly than the WT cells (Figure 4-8B). This effect has also been
seen in the rnylA mutant using syrol3, a dye that does not normally enter cells, but
which is taken up by the rnyIA cells (G.C. Maclntosh and P.J. Green, unpublished
observations). These results are also preliminary; however, two of the phenotypes in the
yeast mutant are consistent with my observations of the plant knock out: the yeast rnyl A
cells and the rns1-2 mutant are bigger under normal grth conditions, and both appear
to be more permeable to stains. Maclntosh et a1. (2001) hypothesized that natural RNAs
may exist in the membrane, possibly as an ancient mechanism to control membrane
stability and permeability in the "prebiotic world". They further postulated that secreted
ribonucleases may be used to control the amount of RNA in the membrane. In the my] A
mutant, this control is lacking, resulting in excess RNA in the membrane and a
corresponding increase in membrane permeability. The cells may compensate for this
change by increasing in size and enlarging the vacuole. Similarly, the rns1-2 mutant is
lacking an extracellular RNase activity in the same family as Rnyl. The cells may be

more permeable, and consequently, the plant is larger.

96

My results indicate that several phenotypes result from altering the expression of
RN51. It will be important in the future to confirm these results by complementing the
phenotypes using a genomic fragment of RN51. The rnsl-Z mutant has already been
transformed with a construct that consists of the RN51 coding region plus 1.4 kb of
upstream promoter sequence and 1.8 kb of sequence downstream of the stop codon, and
primary transforrnants have been selected. In the near future, seeds from these plants and
from plants transformed with the vector control should be plated for analysis of root
length. If lack of RNSI causes longer roots, the complemented plants should have roots
of WT length. This result is expected, since the 35S-RNS1 roots are shorter than roots of
Columbia wild type.

Since the rnsI-2 line has a minor ap2 phenotype as a result of the T-DNA used in the
mutagenesis, we are in the process of isolating additional alleles. The T-DNA insertions
in these rnsl-3 and rnsI-4 alleles lie in the RN51 coding region (+602) and promoter
(-80), respectively. Since the T-DNA lies between the two conserved histidine residues
highly conserved in S- and S-like RNases, the rns1-3 allele may prove to be a valuable
null mutant. Any phenotypes seen in rnsl-2 should also be present in new alleles

identified.

MATERIALS AND METHODS

Plant materials

Throughout this chapter, Arabidopsis thaliana ecotypes Columbia or Ws were used. The
Ws ecotype was used as a WT control for the rnsl mutants, since Ws was used to
generate the Wisconsin T-DNA collection. The 35S-RNS1 construct was transformed

into the Columbia ecotype; therefore, Columbia was used as a control for experiments

97

involving these lines.

For plating, seeds were sterilized as described (Pérez-Amador et al., 2000). Selection
medium contained 50 pg/ml kanamycin. RNA-containing medium was made according
to Chen et al. (2001).

Root length was determined by growing seedlings vertically in a chamber (Percival)
at 22°C. Lengths were measured at different ages, including 9, 12, and 14 days. Four
repetitions of rnsl-2 versus WT W5 and two repetitions of 35S-RNS1 versus WT
Columbia were performed. The first three sets of rnsl-Z/Ws seedlings were grown on
1250 and 62.5 pM P,. In total, more than 320 seedlings each of rnsI-2 and WT Ws and
more than 40 seedlings each of 35S-RNS1 and WT Columbia lines were measured and
analyzed for a statistically significant variation by t-test. Values of t for each of the nine
sets exceeded the cut-off value for a probability of 0.1% (P<0.001).

Fresh weight of green tissue was determined by excising this tissue from the root
base. Seedlings were measured in sets based on the plate they were grown on. The total
weight per line per plate was measured and a per seedling weight was calculated for each
plate. A t-test was used to determine whether the differences between the weights of the
four lines were significant. Using at least 37 seedlings per line, P<0.05 for the 35S-RNS1

vs. WT Columbia set and P<0.01 for the rns1-2 vs. WT Ws set.

Plasmid construction

A clone containing the RN51 cDNA fused between a doubly enhanced c0py of the
cauliflower mosaic virus 355 promoter and the nos terminator was constructed previously
(Bariola, 1996). This plasmid was digested with Xbal to remove the entire

promoter/cDNA/terminator cassette, and the fragment was ligated into the Xbal site of

98

the plant transformation vector pCambia 2301. WT ArabidOpsis thaliana ecotype
Columbia plants were transformed with this construct, p2020, by vacuum infiltration

(Bechtold et al., 1993).

Identification of rns1 T-DNA insertional knock-out mutants

The University ofWisconsin T-DNA insertion line collection was screened using primers
PG904 (positioned at —572 from the RN51 transcription start site) and PG905 (+1645).
Primers were tested on WT Ws genomic DNA and resulted in the amplification of a 2.2-
kb fragment. These primers were sent to the University of Wisconsin for PCR-screening
of their T-DNA insertion line collection using each of the two primers and a primer
corresponding to the left border (LB) of the T-DNA. An aliquot of each reaction was
electrophoresed on 0.8% (w/v) agarose gels and transferred to Nytran Plus nylon
membrane (Schleicher and Schuell, Keene, NH). The RNA blots were hybridized as
described in Taylor and Green (1991) using a 32P-labeled RN51 probe. The RN51 probe
hybridized strongly with the PCR products from two reactions: LB/PG904 reaction 15
and LB/PG905 reaction 19. Bands corresponding to these signals were gel-isolated and
sequenced to confirm the position of the insertion. The LB/PG904 insertion (rns1-1) lay
20 bases 3' of the stop codon, and the LB/PG905 insertion (rns1-2) caused a deletion of
21 bases in the 5' UTR, 65 to 86 bases upstream of the RN51 translation start site.

Seeds were obtained from Wisconsin for the appropriate 25 pools of 9 insertion lines
each. Seeds were grown and harvested as pools for PCR analysis to identify the pools that
contained the desired insertions. Seedlings from those two pools were then transferred to
soil, and individual plants were screened by PCR analysis to identify single plants

harboring at least one copy of the rnsl alleles. At least one heterozygous plant was

99

identified for both alleles. These plants were allowed to self-fertilize. and their progeny
were analyzed similarly to identify a homozygous line. Seeds from the homozygous lines

were used for further analyses.

Protein and RNA analyses

Protein and RNA purification and analyses were performed as described in Chapter 2.

DNA extraction and DNA gel blot hybridization

For genomic Southei'n analysis. DNA was extracted from rosette leaves of mature plants
using Plant DNAzol (GibcoBRL) according to the manufacturer's instructions. DNA was
quantified by ethidium staining and comparison against a known amount of Lambda
DNA. After digestion with EcoRI, 3 pg DNA was separated by electrophoresis in a 1.0%
(w/v) agarose gel and blotted and hybridized as above using a 32P-labeled RN51 or
T-DNA probe. The T-DNA probe was a PCR product of the entire transferred region of
the T-DNA in the rnsl-2 allele.

For genomic DNA extraction for PCR, a shorter extraction protocol was used to
isolate DNA from one to two leaves of young plants. Tissue was pulverized and
incubated for 30 min at 65°C in 500 pl CTAB buffer [2% (w/v) CTAB (Sigma); 1.4 M
NaCl; 0.2% (v/v) B-mercaptoethanol; 20 mM EDTA; 100 mM Tris, pH 8.0]. The solution
was extracted twice with chloroform:IAA [24:1 (v/v)], precipitated, and resuspended in
100 pl ddH2O containing RNase A (l pg/pl). PCR was performed according to the
University of Wisconsin web site (www.biotech.wisc.edu/Arabidopsis/Guidelines.html)

using 1 pl of DNA for each reaction.

100

 

Anthocyanin assays

Anthocyanin contents of WT Ws, rns1-2, and RN52 antisense lines were measured
according to Bariola et al. (1999). Levels of rnsl-2 and WT anthocyanins were
determined twice, by growing the lines once on medium lacking selection and once on
kanamycin selection medium (except WT). The RN52 antisense line was included in the
latter as a positive control for high anthocyanin content. Representative anthocyanin
content of one experiment is shown; results from the other were very similar. Error bars

are not included because the RNS2 antisense line was examined only once.

Bacterial growth curves

Inoculations with Pseudomonas syringae pv. tomato DC3000 were performed as
described by Li et a1. (2000). Samples of inoculated tissue from four different leaves of
each line were taken daily by excision with a cork borer (area/leaf = 0.250 cmz). Bacteria
inside the leaf discs were released by grinding the tissue in a microfuge tube in sterile
water and plated on LB medium. The bacterial population was determined based on the
numbers of colonies formed on LB plates, as described by Bertoni and Mills (1987).
Mean results from four repetitions of the 35S-RNS1 and WT Col lines are shown.

Representative results of two repetitions of the rns1-2 and WT Ws lines are shown.

101

CHAPTER 5

ANALYSIS OF AN ARABIDOPSIS THALIANA MUTANT WITH AN
ALTERED RNASE PROFILE

ABSTRACT

Plants produce a complex series of ribonucleases (RNases), but little is known about their
biological function and regulation at the molecular level. In Arabidopsis, stem tissues are
particularly rich in RNases, suggesting that RNase regulation and function may be linked
at the molecular level to stem growth and development. To address this possibility and to
generate tools to study RNase function, we screened stem extracts of Arabidopsis to
identify mutants with altered RNase profiles (mp mutants). Several mutants affecting
different size classes of RNA-degrading enzymes were isolated. Of particular interest was
arpl, which overproduces a doublet of 33-kDa RNases. Based on our analysis, we were
able to demonstrate genetically that these activities are bifunctional nucleases that
degrade both RNA and DNA. Additionally, arpl overproduces a previously unidentified
small RNase of approximately 23 kDa that comigrates with the well-characterized
RNase, RNSI. The elevated RNase activity caused by the arpl mutation is observed only
in stem tissue and is not found in leaves or in seedlings. Other phenotypes of arpl are
also stem-associated, including overproduction of the 33-kDa nucleases, short stature,
and increased branching. These data indicate that the ARPI locus constitutes a novel
regulator of the production of several RNase and nucleases in stems and that ARPl may

be an important factor in stem growth and development.

INTRODUCTION

Although RNases have been used for decades as models for protein structure studies, the
biological roles of these enzymes and the mechanisms by which they are regulated in
plants and in other higher eukaryotes are largely unknown. In particular, functional
analyses of secreted RNases that do not appear to be involved in mRNA or rRNA
processing are limited. To date, fluctuations in RNase activity levels or gene expression
have provided the most useful data for predicting RNase function, although the
machinery involved in these regulatory pathways has not been elucidated. Additionally,
organ-specific expression patterns of RNases can be found that may lend clues to their
functions in vivo. For instance, the most complex pattern in mature Arabidopsis plants is
found in stem extracts (Yen and Green, 1991). Our knowledge of the molecular pathways
controlling stem-specific processes in plants, e.g., branching, or other prOcesses
prominent in stem tissue, such as tracheary element formation (for reviews, see Schmitz
and Theres, 1999, and Roberts and McCann, 2000, respectively), is limited. However, the
high level of activity found in stems suggests that RNases could play important roles in
such processes. The induction of multiple nuclease activities that occurs in zinnia cells
undergoing differentiation into tracheary elements (Fukuda, 2000) is consistent with this
argument.

In contrast to the many studies on the expression patterns of RNases, very little is
known concerning the factors that regulate these pathways at the molecular level. For
example, while P, starvation in Arabidopsis induces RN51 and RN52, but not RN53
(Bariola et al., 1994) or BFNl (Pérez-Amador et al., 2000) expression, senescence

induces all four of these activities (Taylor et al., 1993; Bariola et al., 1994; Pérez-Amador

104

et al., 2000). Evidently, there is a high degree of specificity of RNase regulation by
different stimuli, but the molecular components controlling these patterns are not known,
neither in plants, nor in most higher eukaryotic systems. The identification of regulators
is the first step in the elucidation of the mechanisms plants use to mobilize certain sets of
enzymes in response to developmental or environmental cues.

To further our understanding of the regulation of RNases and thereby gain insight
into their potential functions, we designed a screen to isolate mutants of Arabidopsis with
altered RNase profiles. Through a gel-based screen of EMS-mutagenized M2 plants, we
identified several mutants that show altered expression of at least one RNase activity. In
particular, one mutant with altered levels of at least three activities provides a useful tool
for the elucidation of regulatory pathways involved in RNase expression in Arabidopsis.
This alteration of RNase patterns appears to be specifically associated with the stem, and
physiological phenotypes of the mutant are also related to modifications in stem
morphology, possibly offering new insights into processes involved in stem development

and differentiation.

RESULTS

An RN ase-activity gel screen for mutants affecting the Arabidopsis RN ase profile

To screen for mutants with altered RNase profiles, a gel-based assay for RNase activity
(Yen and Green, 1991) was used to analyze protein extracts from the stems of EMS-
mutagenized M2 Arabidopsis plants. Stem extracts were assayed because the stem RNase
profile is more complex. containing more bands of RNase activity, than the leaf RNase
profile (Yen and Green, 1991). It was previously reported that up to 13 bands of RNase

activity could be distinguished on a gel containing Arabidopsis stem extracts (Yen and

105

Green, 1991). Eight bands of RNase activity were consistent and intense enough to
screen for alterations in their activities. Protein extracts from 2,500 M2 plants were scored
on the RN ase activity gels to identify plants whose RNase profiles differed from those of
wild-type stems. Figure 5-1 depicts an example of the identification of a this type of
mutant (compare the intensity of bands at approximately 33 kDa in lane 2657 with their
intensities in other lanes). In the primary screen, 112 putative mutants with alterations in
the Arabidopsis RNase profile were identified. The phenotypes of 60 of the 112 putative
mutants could be reproduced on subsequent gels. To determine whether the phenotypes
observed in such potential mutants were heritable, several progenies (M3) of each were
grown, stem extracts were prepared, and the extracts were examined using the activity gel
assay. Those mutants with heritable phenotypes were studied further.

In total, nine gltered RNase profile (arp) mutants that showed the original phenotypes in
all M3 progeny have been isolated. This chapter focuses on those mutants that affect the
activity of a doublet of 33-kDa RNase activities for several reasons. Six of the nine
mutants isolated display alterations in the levels of these activities. Previously, 33-kDa
activities were seen on both RNase and DNase activity gels (Yen and Green, 1991). The
doublet is also interesting because the RNase activities are induced in seedlings starved
for inorganic phosphate (P,) (see Figure 10 in Bariola et al., 1994) and by wounding
(Chapter 2). One of the six mutants affecting the 33-kDa activities was plant 2657, first
identified in Figure 5-1 as showing increased RNase activity of the doublet. The
phenotype of 2657 was seen more clearly in a lower percentage acrylamide gel that
resolved the two bands of the doublet (Figure 5-2A). In contrast to the increased activity

of the 33-kDa RNases in line 2657, relative to the wild-type activity (first and last lanes),

1 ()6

2654
2655

OFNM
mmmm
£0000
NNNN

 

Figure 5-1. arp mutants are altered in RNase activity profiles

Primary identification of an M2 RNase activity mutant. Stems of plant 2657 showed
increased levels of two bands of RNase activity at 33 kDa. Plant numbers are shown
above the gels. Approximate molecular weights (kDa) are indicated to the left. Arrows on
the right point to the RNase activities affected in the mutant extract. 100 pg of protein
was loaded in each lane. An extract from aerial portions of wild-type plants (leaves,
stems, flowers), indicated by "A", was included as a standard.

This experiment was performed by Michael Abler and Linda Danhof.

107

N
1.0
‘0
N

3270
2048
2438

VVT

3433

 

VVT

2657
2972
3270
3433
2048
2438
VVT

 

Figure 5-2. The 33-kDa RNase activities also degrade DNA.

(A) RNase activity gel of stem samples from the six lines in which the 33-kDa RNase
activities are affected. Plant numbers are shown above the gels and approximate
molecular weights (kDa) are indicated to the left. WT, wild-type RLD stem extracts.

(B) DNase activity gel of the six mutants in which the 33-kDa RNase activities are
affected. Denatured single-stranded DNA was substituted for RNA as the substrate in the
gel.

This experiment was performed by Michael Abler.

108

plants 2972, 3270, 3433, 2048, and 2438 all greatly diminish or abolish the activity of the

lower band of the doublet (Figure 5-2A).

The 33-kDa RN ase activities are bifunctional nucleases

Previous data from our laboratory indicated that there is also a doublet of DNase activity
at 33 kDa, but it was not known whether both the RNase and DNase activities were
derived from the same enzyme(s) (Yen and Green, 1991). The availability of several
plants affecting the 33-kDa RNase activities allowed us to test this hypothesis by
analyzing the DNase activities in these mutants.

Figure 5-2A shows the RNase gel phenotypes for the six plants affected in the
activity of the 33-kDa doublet. Shown in Figure S-ZB, protein extracts from the same
plants were run in a gel containing denatured single-stranded DNA as the substrate for
the enzyme activity. The profile of DNase activities in Arabidopsis stem extracts is far
simpler than the RNase profile, and the 33-kDa doublet is the most abundant of the
DNase activities in wild-type stems. In each of the plants, the alteration in the DNase gel
phenotype corresponds with that in the RNase gel for the 33-kDa activities. Because all
of the mutations affect the DNase and RNase activities in the same manner, these data
demonstrate genetically that the 33-kDa doublet corresponds to a pair of bifunctiOnal

nuclease activities, capable of degrading either RNA or DNA.

Genetic characterization of the arp mutants

Genetic analyses of the plants affected in the 33-kDa activities were performed. Each line
was backcrossed to WT RLD to determine the inheritance patterns of the mutations. The
RNase profiles of the F , plants from each of the crosses were indistinguishable from that

of the wild type, thereby indicating that all nine mutations were recessive (Abler et al.,

109

manuscript in preparation). Segregation of the mutant phenotype in the F2 generation
showed that the mutations were inherited in a Mendelian fashion, fitting a 3:1 ratio of
wild type to mutant (Abler et al., manuscript in preparation).

Complementation tests of the six mutants affecting the 33-kDa nuclease activities
indicated that the mutations correspond to two loci and that the 2657 mutation affects a
locus distinct from that affected in the other four mutants (Abler et al., manuscript in
preparation). Accordingly, the mutant locus generating the gel phenotype exhibited by
plant 2657 (increased 33-kDa activities) was designated arpl. During subsequent
mapping experiments, it was determined that the gel phenotype of mutant lines 2048,
2438, 2972, 3270, and 3433 resembled the pattern of stem RNase activities in the
Columbia ecotype of Arabidopsis (not shown and LeBrasseur et al., in press). From this
observation and subsequent RFLP analysis, we concluded that these lines were in fact
Columbia contaminants in the screen, which had been performed using the ecotype RLD.
Nevertheless, our results demonstrate that a single locus confers ecotype-specific control

of the lower band of the 33-kDa doublet in wild-type stems.

Morphology of the arpl mutant

Examinations of growth and development indicated that a visible phenotype is associated
with altered expression of the 33-kDa nuclease activities in arpl. As shown in the
example in Figure 5-3A, arpl plants appear shorter and more branched than wild type.
To quantify these visible phenotypic differences, two blind studies were performed in
which approximately 40 plants each (wild type and arpl) were scored. Seeds from the

arpl line and from wild-type plants were planted in pots, and the order of the pots was

110

LLLI
III

000

40 --

883%“

Plant Height (cm)

 

20 -»

 

 

Figure 5-3. Morphological differences in the phenotypes of wild—type and arpl mutant

plants.
(A) Six-week-old arpl and wild-type (WT) plants are shown. The arpl plants appear

shorter and more branched than wild type plants.

(B) Scatter plot of final plant height of wild-type and arpl plants. Individuals in the
respective populations are represented by circles. The average plant heights for the
populations are represented as horizontal bars.

This experiment was performed by Michael A bier.

lll

randomized. The time for each plant to reach certain developmental stages (e.g.,
germination, flowering) was measured, as were physical (e.g., plant height)
characteristics of the plants. Differences between populations were analyzed by the
Mann-Whitney rank-sum test.

Significant growth and developmental differences between wild-type Arabidopsis
and arpl plants were confirmed in the blind study. Figure 5-3B depicts the differences in
height between arpl and wild type. The data in Table 5-1 demonstrate statistically that
arpl plants were more heavily branched and had shorter siliques than did wild type. The
differences between the wild-type and arpl plants were evident in plant height, number
of branches, and silique length. Developmental processes (days to bolting, appearance of
first flower and branch flower) did not differ between the populations. The altered
morphological phenotypes of the arpl mutants are likely due to the arpl mutation itself,
since these phenotypes have always cosegregated with the gel phenotypes through four

generations of backcrosses (not shown).

The arpl mutation induces an RNase activity in addition to the 33-kDa nuclease
activities

Another difference observed in the RNase profiles of arpl and wild-type inflorescence
stems was the appearance of an approximately 23-kDa band of RNase activity in arpl
plants. Like the morphological phenotypes, the 23-kDa band of RNase activity
consistently cosegregated in plants showing the increased 33-kDa nuclease arpl
phenotype through four generations of backcrosses. Initially, this 23-kDa band was not
scored because it was not seen consistently in wild-type Arabidopsis plants. However, as

shown in Figure 5-4, RNSI produced in yeast comigrates with the 23-kDa RNase activity

 

Table 5-1. Statistical analysis of morphological
differences between WT RLD and arpl plants“

 

 

arpl vs. WT
Parameterb T P
Plant height (final) 6.89 <0.000001
6.93 <0.000001
Number of branches 3.56 0.00037
4.26 0.000020
Silique length 9.03 <0.000001

1 1.51 <0.000001

 

aThe Mann-Whitney rank-sum test was used to
determine the statistical significance of the
morphological differences between populations.
Parameters with reproducibly significant differences are
listed in the first column.

Ir’Other parameters (measured in days from planting) that
were not significantly different include germination, 2,
4, and 6 rosette leaves, bolting, first flower, lO-cm bolt,
15-cm bolt, branch flower, and axillary stem flower.

 

This analysis was performed by Michael A bier

that is elevated along with the 33-kDa doublet in arpl plants. Both the original mutant
(arpl M6) and a mutant recovered after four generations of backcrosses (arpl BC4®)
have RNase activities that comigrate with yeast-produced RNSI. Neither the WT plants
(WT), nor plants from the BC4® generation with a wild-type ARPI allele (ARPI BC4®),
have activities that comigrate with RNSI (Figure 5-4).

Since the 23-kDa band comigrates with RNSI, RNA gel blot and immunoblot
analyses were performed to determine whether RN51 mRNA or protein levels were
affected in arpl. For RNA blot analysis, RNA was isolated from stems of 6-week-old
arpl and WT RLD plants. Poly(A)+ RNA was purified from total RNA preparations so
that small effects might be detected. RN51 transcript was not detected in either total or
poly(A)+ RNA in WT RLD or arpI stems by RNA gel blot analysis using an RN51 probe
(Figure 5-5A). RNA from wounded seedlings was included as a positive control for RN51
hybridization [Wnd and Wnd Poly(A)+].

Since we did not detect increased RNSI mRNA levels in arpl stems, we were interested
in whether the RNSI protein was elevated because of a possible posttranscriptional effect
on RNSI. To examine this possibility, proteins from arpl and WT RLD stems and
seedlings were separated by SDS-PAGE and subjected to immunoblot analysis (Figure 5-
SB). RNSI purified from yeast cells and protein from P,-starved and P,-supplied
seedlings were included as controls (RNSI, P-, and P+, respectively). Although the
RNSI antibody (Bariola et al., 1999) detects the yeast-produced RNSI and RNSI in P,-
starved seedlings, no RNSI was detected in arpl stems. Therefore, significant increases

in RN51 transcript as well as protein could not be found in arpl stems compared to WT.

114

 

Figure 5-4. The arpl mutant has increased levels of an activity that comigrates with
RNSI.

Approximate molecular weights are shown to the left of the gel. 100 pg of protein was
loaded in the arpl and wild-type RLD (WT) lanes. Lane RNS] contains an aliquot of
culture medium from yeast cells expressing the Arabidopsis RN51 gene from a yeast
secretion vector as described (Bariola et al., 1994). ARPl BC4®, F l progeny plant of the
fourth backcross of arpl with wild-type RLD displaying wild-type levels of the 33-kDa
nucleases; arpl BC4®, F, progeny plant of the fourth backcross of arpl with wild-type
RLD displaying increased levels of the 33-kDa nucleases; arpl M6, sixth generation of
mutant line 2657.

This experiment was performed by Michael Abler.

115

 

 

 

 

A $.21,
<<A
355-
002‘
O.ILO
222::
58338;
RN81
B
10
Gift)
:5)
gee
8848
mmfig
Email)“;
2+.
85351:an
RN81 “i

 

 

Figure 5-5. RNSI transcript and protein is not detected in arpl stems.

(A) RNA gel blot analysis of RNA isolated from wild-type (WT) and mutant (arpl)
stems. Total and poly(A)+ RNA samples from wild-type wounded (9 h) seedlings (Wnd)
are included as positive controls. Lanes contain 20 pg total RNA or 1 pg poly(A)+ RNA.
(B) Immunoblot analysis of 100 pg of protein from stems and seedlings of wild-type and
mutant lines. An aliquot of RNSI purified from yeast (RNSI) and 30 pg protein from P,-
supplied (P+) and P,-starved (P') seedlings were used as controls.

116

These results raised the possibilities that the increased activity at 23-kDa was due to
either a posttranslational effect on RNSI or an increase in another, unknown RNase
activity that comigrates with RNSI. Biochemical characteristics of the 23-kDa band, such
as pH optimum, EDTA-sensitivity, Zn-requirements, and the ability to degrade poly(A)+
RNA, resembled those of RNSI (data not shown). To determine whether the small RNase
was indeed RNSI, we crossed the arpl mutant with a T-DNA insertional mutant of the
RN51 gene. The T-DNA in this mutant allele, rnsl-2, lies in the 5' UTR of the RN51 gene
(see Figure 4-1 of this thesis). Plants homozygous for this insertion do not produce
RNSI, as shown by activity gel assay of wounded leaves, a condition known to induce
RNSI transcript and activity (Figure 4-28 of this thesis and LeBrasseur et al., in press).

The arpl mutant, in the RLD background, was crossed with rns1-2, which is in
ecotype Ws, and stem samples of segregating F3 progeny were analyzed by activity gels.
Three F3 progeny lines in particular, lines 8, 13, and 24, were analyzed further to
determine RN51 zygosity. DNA gel blot analysis was used to confirm that these three
progenies were homozygous for the rns1-2 mutation. Genomic DNA was extracted from
wild type, rnsl-2, and progeny plants 8, l3, and 24, digested with EcoRI, and analyzed
by DNA gel blot analysis. A 5-kb band is seen in wild-type DNA, while only the
expected 2-kb band can be seen in the rns1 -2 mutant as well as in progeny plants 8, 13,
and 24 (Figure 5-6A). Since the 5-kb band is not seen in these plants, we conclude that
the progeny plants are homozygous for the rns1-2 mutant locus. Next, we analyzed the
gel phenotype of these lines (Figure 5-6B). The position of RNSI (arrow) is shown in
wounded wild-type Ws stems (Ws wnd). This band is absent in unwounded Ws and RLD

stem samples (Ws and RLD). Progeny plant 8 has at least one copy of the wild-type

ll7

F3 #3
F3 #13
F3 #24

 

—5kb
i2};
2...... “”—2kb
B
1: '2
‘37- 3 mg

,_

.. mEBESEtfi
ZBKENKLLLLU.

 

Figure 5-6. A previously unidentified 23-kDa band is increased in arpl.

(A) Southern analysis demonstrating that the F3 progeny examined have no wild-type
allele of RN51. EcoRI-digested DNA was probed with an RN51 fragment. Only WT
DNA hybridizes to the expected 5-kb fragment.

(B) RNase activity gel of stem protein extracts from wild type (Ws and RLD), rnsI-Z
(rnsl), arpl, and F3 progeny of a cross between arpl and rns1-2 (F3 #8, F3 #13, and F3
#24). Wounded stem samples were harvested 7 h after wounding (W s wnd and rnsl wnd)
and are shown as controls for RNSI levels in the rns1-2 mutant. F3 progeny 13 and 24 are
homozygous for the arpl mutation, as shown by the increase in the 33-kDa nucleases
(upper arrow). Although 13 and 24 are homozygous for the rns1-2 allele, they still have
high levels of the 23-kDa band (lower arrow).

118

ARP] allele, as seen by the normal levels of the 33-kDa nuclease doublet (F 3 #8). In
contrast, progeny plants 13 and 24 are homozygous for the arpl mutation, resulting in an
increase in the 33-kDa doublet (F3 #13 and F3 #24). Like the arpl mutant (arpl), stem
samples of plants 13 and 24 still display increased levels of a 23-kDa activity, although
they lack a normal copy of the RNSI gene. Therefore, we conclude that the 23-kDa band
that is increased in arpl stems is not RNSI. Wounded stems of the rns1-2 mutant have
slight levels of an activity at 23 kDa (Figure 5-6B, compare ms] and ms] wnd). Two
explanations could account for this result. It is possible that the 23-kDa band in arpl
stems is also wound~inducible, but is not normally seen due to the high levels of RNS] ,
which runs at the same position. Alternatively, slight amounts of wild-type RNSI
transcripts may be made and translated in wounded rnsI-2 stems. ,Small amounts of
aberrantly sized wound-induced transcripts are detected by Northern blot in the mutant;
however, we were unable to detect full-length polyadenylated RNSI transcript in rns1-2
plants by RT-PCR (data not shown). Thus, although we cannot exclude the latter
possibility, we believe the former is more likely. Nevertheless, the levels of the 23-kDa

arp] band are significantly higher than the wound-induced 23-kDa band in rns1-2,

indicating that the two bands are different RNase activities.

The arpl RN ase phenotype is stem associated

Because the visible phenotypes of the arpI plants are affected predominantly in aspects
of stem morphology, such as plant height and branching, the RNase profiles of several
tissues in the mutant were examined to determine whether this phenotype was stem-
specific. The 33-kDa doublet and 23-kDa RNase activities, which are increased in the

stems of arp], are at wild-type levels both in leaves and in IO-day-old seedlings of the

119

mutant (Figure 5-7). Phosphate-starved seedlings, in which RNSI and the 33-kDa
doublet activities are induced, have been included for comparison of the running
positions of these activities. Thus, it appears that the arpl RNase phenotype, along with

the morphological arpl phenotypes, are associated with the stem.

DISCUSSION

Similar to other eukaryotes, plants produce a number of major RNase activities that have
been implicated in various processes on the basis of their regulation. However, the genes
involved in the regulation of eukaryotic RNases are largely unknown. In this study, we
used a novel brute-force genetic screen to identify several Arabidopsis mutants with
alterations in the RNase profile (arp mutants). Some of these mutants affect multiple
RNase activities, indicating that they have defects in RNase regulatory components rather
than RNase structural genes. The mutant characterized in the most detail, arpl, has
unusually high levels of two 33-kDa activities, which degrade both RNA and DNA, and a
smaller RNase. These data together with the recessive character of arpl prompt us to
suggest that the mutation may inactivate a regulatory component common to the control
of these three RNA-degrading activities. An interesting and unique feature of arpl is that
the elevated RNase levels and other morphological phenotypes of this mutant are
prominent in the stem. Therefore, the arpl mutation may alter a repressor of these RNase

activities and other functions required for normal stem growth and development.

Mutants affecting single and multiple RNase activities

The screen of 2,500 M2 plants ultimately yielded four plants with different heritable

alterations in the stem RNase profile and identified an ecotype polymorphism in

arp1 seedlings

ID
U!
.5
'0
d)
d}
0)

WT leaves
arp1 leaves
P-seedlings
WT stem
arp1 stem

 

Figure 5-7. The arpl RNase phenotype is stem specific.

The arpl increase in RNSI and the 33~kDa activities is seen in stems but not in seedlings
or leaves. Pi-starved seedlings are shown as a control for the running position of RNSI
(arrow). 100 pg protein was loaded in all lanes except P‘, in which 30 pg was loaded.
RNSI purified from RNSI-expressing yeast (Bariola et al., 1994) is shown to the far
right. Approximate molecular weights are indicated on the lefi. WT, wild-type RLD
samples.

121

Columbia compared to Ws. RNase activities of different molecular weights were affected
in the mutants. 1n particular, plant 2657 exhibited an increase in both activities of the 33-
kDa doublet, whereas in the Columbia polymorphism represented by plants 2972, 3270,
3433, 2048, and 2438, the lower 33-kDa activity was decreased or eliminated. The
varieties of RNase phenotypes identified in this study differ markedly from previous
plant mutants shown to affect RNases. Both the variegated immutans mutant of
Arabidopsis and the Opaque2 mutant of maize, which were isolated on the basis of other
phenotypes (Redei, 1967; Wilson and Alexander, 1967; Dalby and Davies, 1967) appear
to elevate all major RNase activities in the gel profile (Wetzel et al., 1994; Wilson and
Alexander, 1967; Dalby and Davies, 1967). Because all of the mutants isolated in this
work affect only a subset of major Arabidopsis RNases, they have the potential to

provide insight about specific regulatory networks and functional associations.

The 33-kDa activities are bifunctional nucleases

Previous evidence indicated that a doublet of RNase activity was the same activity as
seen in DNase activity gels. In addition to the fact that the doublets run at approximately
33 kDa on both gels, both degradative activities had similarities in EDTA-sensitivity, pH
optima, and cation requirements (Yen and Green, 1991). Still, it was possible that the
activities were in fact products of separate genes. The isolation of arp] and the
identification of the Columbia polymorphism in this study provide firm evidence that the
33-kDa activities are bifunctional nucleases that can degrade both RNA and ssDNA.
Each plant examined affected in one or both of the bands of RNase activity also shows
the same alteration in the DNase activity at the same position. It is theoretically possible

that arpl and the Columbia polymorphic allele both affect separate DNase and RNase

activities, which coincidentally are similarly sized, in the same manner. However, this
seems extremely unlikely, and we therefore conclude that the RNase and DNase activities
are derived from the same protein products. The potential relationship between the upper
and lower bands of the 33-kDa doublet is unknown. Other experiments have determined
that both bands are wound-inducible and that the lower band normally missing in
Columbia but found in RLD can be induced by wounding in Columbia (LeBrasseur et al.,

in press).

Possible functions for the 33-kDa nucleases

Very little is known about the biological functions of RNases in plants. To investigate the
possibility that the 33-kDa activities were involved in mRNA degradation, mutant plants
were transformed with constructs containing globin reporter genes. The mutant plants
showed little or no effect on the accumulation of globin compared to wild type, whether
or not globin was destabilized (Abler et al., manuscript in preparation). Thus, it appears
the mutations do not affect global mRNA turnover or the turnover of specific unstable‘or
stable reporter transcripts. These results are not unexpected, as most of the RNase activity
in plants is localized to the vacuole (Boller and Kende, 1979; Abel and Glund, 1986), or
secreted from the cell, and therefore is not likely to function in the cytoplasmic
degradation of mRNA (Bariola and Green, 1997).

Certain characteristics of the activities in the 33-kDa doublet can aide in their
classification. For instance, their size, acidic pH optima, sensitivity to EDTA, and
requirement for divalent cations (Yen and Green, 1991), as well as ability to degrade both
RNA and DNA, place them within the nuclease I-type class of enzymes (Wilson, 1982).

In general, nuclease l enzymes are secreted glycoproteins (Fraser and Low, 1993).

Although it is not known whether the 33-kDa activities are secreted, extracellular washes
of stem and leaf extracts show an increase of activity in this range (M.L. Abler and P.J.
Green, unpublished). Additionally, while we do not have evidence that the 33-kDa
activities are glycosylated, one explanation of the doublet of activity is differential
glycosylation of the same protein. In this case, the ecotype-specific polymorphism and
the arpl effect on the 33-kDa activities may be explained by altered regulation of the
glycosylation of these proteins. Analyses of other plant nuclease I enzymes provide
insight into possible functions for the 33-kDa activities. Nuclease I enzymes have been
characterized biochemically in many species, and cDNAs encoding these enzymes have
been isolated recently from several species (see Chapter 1).

Putative functions for nuclease I enzymes can be hypothesized based on their
expression patterns. Plant nuclease I enzymes and the genes encoding them are induced
during various plant growth and developmental conditions, including germination,
tracheary element formation, senescence, and various stress responses (for a review, see
Bariola and Green, 1997). The 33-kDa activities altered in the arpl mutant are strongly
induced by wounding (LeBrasseur et al., in press) and phosphate starvation (see Figure
10 in Bariola et al., 1994). These activities appear to have many characteristics of
nuclease I-type enzymes and may be functioning similarly to other members of this
family, which have been proposed to be involved in remobilizing the constituents of
nucleic acids from dying tissue for use in other parts of the plant (Bariola and Green,
1997). Additionally, they may play a role in vascular formation, either in killing the cell
directly, in degrading components left afier cell lysis to clear the xylem channel, or in

nuclear degradation during PCD of tracheary elements (Bariola et al., 1997; Fukuda,

124

2000). The fact that altering these activities in am] correlates with stem-associated
phenotypes is consistent with the induction of these enzymes during tracheary element

formation.

arpl is a regulatory link between the 33-kDa nucleases and an RNase activity

The arpl mutation is particularly interesting because several activities are affected. In
addition to the 33-kDa nucleases, a previously unidentified 23-kDa RNase activity is also
increased. Since multiple activities are upregulated, ARPI is not likely to encode a
structural RNase, but rather a regulatory protein. It is interesting to consider in what
pathways ARPI might regulate RNases. An obvious possibility is the Pi-starvation
pathway, as both 33-kDa and smaller RNase activities are increased in seedlings starved
for P, (Bariola et al., 1994). However, in the case of Pi-starvation, RNS] activity and
transcript are induced to high levels, a situation not seen in arpl . Additionally, the visible
phenotypes of arpl plants do not resemble those associated with Pi—deficiency and the Pi-
loading defective phoI mutant, which include increased anthocyanin levels and delayed
flowering (Marschner, 1996; Poirier et al., 1991). The arpl plants and seedlings do not
have high levels of anthocyanins (not shown), nor is flowering time affected (refer to
Table 5-1). Finally, the Pi-starvation response is clearly visible in seedlings, whereas the
arpl RNase phenotype is not seen in seedlings (Figure 5-7). Thus, it is more likely that
the ARP] locus is involved in the regulation of RNases in other pathways, as will be

discussed further below.

The arpl phenotypes are associated with stems

Insight into the function of ARPl is provided by the phenotypes seen in the mutant. The

mutation obviously causes pleiotropic effects. Interestingly, many phenotypes seem to be

125

associated with the stem; the increased activities can be seen in the stem tissue, while
leaves and seedlings do not exhibit this increase, stems of arpl are shorter than wild-type,
and branching is increased in arpl stems.

Considering the obvious stem-related mutant phenotypes seen in arpl, an intriguing
possibility is that ARPl is involved in the control of developmental pathways in stems.
RNase and nuclease activities are abundant in stems and may constitute an important part
of stem growth and development. The induction of nucleases during zinnia tracheary
development (Thelen and Northcote, 1989) implies that nucleases are involved in stem-
specific growth and maturation processes. It appears that normal function of ARPl
results in low activity of the three enzymes in stems under standard growth conditions.
The product of the ARPI gene may be involved in upregulating these enzymes only under
specific conditions, for instance, during stem growth and tracheary element formation.
The increase in levels of the 23- and 33-kDa activities may be indicative of deviations
from the normal developmental patterns in the stern. Cross sections of arpl and wild-type
stems did not reveal any gross alterations in cell structure or organization of the tracheary
elements (M.L. Abler and P.J. Green, unpublished), but further analysis may reveal
abnormalities. The ARPI locus has been mapped to the top of chromosome 5 (LR.
Danhof, M.A. Johnson, and P.J. Green, unpublished), and the cloning of this gene should
provide us with an interesting tool with which to study stem-related developmental
processes, an important area of study that remains largely unexplored. Additionally, this
mutant will be a useful tool for isolating activities to provide us with more plant RNase
sequences. Progeny of the arpl and rns1-2 cross will be particularly useful, since RNSI

will not interfere with purification of the new 23-kDa activity in these lines.

126

MATERIALS AND METHODS

Plant materials

Arabidopsis thaliana ecotype RLD was used in this study. Ethyl methanesulfonate-
mutagenized M2 seed was provided by Lehle Seeds (Round Rock, Texas). Plants were
grown under conditions of 16 h light/8 h dark and 50% RH at 20°C and 100 pE light
intensity. Seedlings were grown on AGM plates as described (Taylor et al., 1991). P,-
starved seedlings were germinated on complete medium and transferred to P" medium as
described (Bariola et al., 1994). A lateral branch was harvested from each M2 plant to

provide tissue samples and still allow the plant to grow and set seed.

RNase and DNase activity gels

Protein samples were prepared essentially as described (Yen and Green, 1991) with the
exception that frozen plant tissues were homogenized on ice. Soluble protein
concentrations were measured by the method of Bradford (1976) and an equal amount
(100 pg) of total protein was loaded into each lane of the polyacrylamide gels, except for
P.-starved seedlings, for which 30 pg of protein was loaded. Polyacrylamide gels
containing RNA or DNA as substrate were prepared as described (Yen and Green, 1991),
except the concentration of acrylamide in the separating gels of Figures 5-2 and 5—4 was
reduced from 11% to 9% in order to better separate the 33-kDa bands of activity. The
lower acrylamide concentration necessitated a higher initial RNA concentration in the

gel, so 5.0 mg/ml of RNA was used in the 9% gels instead of 2.4 mg/ml RNA.

Mutant growth and development

Pots containing arp] or wild—type seeds were randomly assigned to three flats. Individual

127

pots were identified only by number, in order to perform a blind study. Parameters
measured (in days from planting) for each plant included germination, 2, 4, and 6 rosette
leaves, bolting, first flower, lO-cm bolt, lS-cm bolt, branch flowering, and axillary stem
flowering. Physical characteristics measured for each plant included final plant height,
number of branches, and silique length. Silique length was determined by measuring
length of the fifth, sixth, and seventh siliques on the primary inflorescence stem for each
plant. The Mann-Whitney rank sum test was used to determine significant morphological

differences between the wild type and mutant populations.

DNA and RNA extraction and gel blot analyses

RNA methods were performed as described in Chapter 2. Genomic DNA procedures
were performed as described in Chapter 4, except DNA was extracted from 2-week—old

seedlings.

Immunoblot analysis

Protein for Western analysis was extracted as described in Maclntosh et a1. (1996) and
quantified using the Bradford assay (Bio-Rad). Protein extracts were mixed with loading
buffer and boiled for 5 min before separation on 11% SDS-PAGE gels (Laemmli, 1970).
Lanes contained 100 pg of stem or seedling extracts from arpl or wild-type plants or
75 pg of protein extracts from seedlings grown on P+ or P“ medium. An aliquot of RNSl
produced in yeast (Bariola et al., 1994) was included as a positive control. For RNS]
signal detection, proteins were transferred to PVDF membrane (Immobilon-P, Millipore),
and membranes were processed as described (Bariola et al., 1999), except detection was
performed using an ECL+Plus Western blotting detection system (Amersham) according

to the manufacturer's instructions.

 

CHAPTER 6

CONCLUSIONS AND FUTURE PROSPECTS

129

When I began this thesis project, the general consensus in the field was that S-like
RNases were involved in degrading nucleic acids in order to supply the plant with P,. The
nucleic acid, it was thought, might be from within the plant itself, or perhaps from
external sources. This more hypothetical idea that RNases could be secreted from roots
into the surrounding soil to degrade organic compounds for uptake by the plant was
dubbed by our lab as the "dead bird hypothesis." While many, including even members of
my lab, are skeptical of this idea, it seems that results in this thesis are consistent with
this concept. The fact that plants overexpressing RNSI grow larger on medium containing
RNA as the only phosphate source (while on Pg-supplied medium, these plants are
smaller) seems most readily explained by secretion from the roots of the RNSI enzyme,
which is then able to degrade the RNA and thereby provide the seedlings with P,. While I
cannot rule out alternative theories, I can only conceive of those that are less intuitive
than the "dead bird." For instance, the increase in RNSI might cause membrane
permeability issues in the plant that somehow affect the ability of the roots to utilize
complex forms of phosphate.

Given the wide range of processes with which induction of S-like RNases correlates,
it certainly seems likely that they are involved in recycling and remobilization processes.
In fact, findings in this thesis that nucleases and RNSl are induced during wounding
might at first seem mundane, another example of RNases used as maid-service for the
plant. However, this family of enzymes may be involved in more than that. The
similarities between the ms! and rnyIA mutants indicate that the lost activities must have
a common fimction in both systems, which does not appear to include recycling

phosphate.

Although the model proposed by Maclntosh et al. (2001) has been discussed in
Chapters 1 and 4, I would like to extend the discussion here. Their proposal is integral to
the extension of hypothesized functions for RNS 1. Figure 6-1 summarizes their model for
the function of secreted RNases in controlling membrane permeability. In this model, the
amount of RNA in the membrane is one factor that determines the amount of water and
ions that enter the cells. This idea is based on the finding that specific, synthetic RNAs
can bind both artificial and natural membranes and in the process, alter the permeability
of those membranes (Khvorova et al., 1999). The use of RNA to alter membrane
characteristics may have been a function of this nucleic acid in the hypothetical RNA
world. If RNA was once used in this capacity, one could envision that during evolution,
an extracellular RNase might have been recruited to control the amount of RNA in the
membrane; the widespread existence of T2 RNases supports an early evolutionary origin
and basic requirement for these enzymes. During stress conditions, including heat shock
and osmotic stress, the RNase activity is induced by the cell in order to degrade some of
the RNA in the membrane. As a result, the cell is less permeable, more "closed" to the
environment. In an RNase mutant, such as rnyIA, this control is missing. Therefore,
RNA accumulates in the membrane, causing it to be more permeable. To compensate, the
cell expands, and the vacuole is enlarged, two phenotypes seen in the yeast mutant.

Similar phenotypes are seen in the plant mutant, as discussed in Chapter 4. Namely,
the roots, and possibly the above-ground tissue, of rnsI-Z seedlings are larger than'the
WT. Interestingly, the overexpressors have the opposite phenotype: smaller roots. The

other phenotype the yeast and plant mutants may have in common is an increase in

131

 

WT cells, no stress WT cells, STRESS

I
,,. f:

 

 

r) .
b RN59 wamr ' s .. L" .i :-

-'~'.'r.-,-.'a-T' , -
‘m Ions .
RNA

 

 

 

Figure 6-1. Model for the control of membrane permeability by RNases.

In this model, modified from one provided by G.C. Maclntosh and P.J. Green, RNA can
normally be found in the membrane in a given amount. Under conditions of stress,
extracellular RNase activity increases, degrading some of the RNA in the membrane and
reducing cell permeability. In the mutant cells, such as rnyIA and rnsI-2, the secreted
RNase is missing, leading to high amounts of RNA in the membrane. In the case of the
yeast mutant, the cell compensates by increasing the size of the vacuole and overall cell-
size. n, nucleus; v, vacuole.

permeability. Dyes used to stain either cell type seem to be taken up by the mutant cells
more readily than in their WT counterparts. Additionally, the regulation patterns of the
two enzymes share some similarities; for instance, both are induced by salt stress. As yet,
the rnsI-Z mutant does not appear sensitive to high salt, as is the yeast mutant (Maclntosh
et al., 2001). [The effect of osmotic stress on the ms] -2 mutant should be tested in the
future by transferring the mutant to medium containing various concentrations of sorbitol,
for example, and monitoring the response of the root cells to the changes. If the mutant
cells are more permeable, they should also be more sensitive to external osmotic
changes] Additionally, RNS] is not induced by heat stress, another variation from the
yeast enzyme. Thus, differences exist. Obviously, we do not imagine that RNA is the
only, or even the most important, control of membrane permeability; protein ion transport
channels play the critical role in this process. Even so, remnants of an ancient use of
RNA may still exist. The yeast mutant phenotypes identified were specific to one strain
of S. cereviseae. It is possible that this strain is somehow compromised in another aspect
of membrane regulation, making it particularly susceptible to alterations in levels of RNA
in the plasma membrane (Maclntosh et al., 2001). Similarly, Arabidopsis has various
mechanisms to control membrane traits, and the loss of RNSI may not have a strong
effect in many circumstances. However, the loss of whatever small amount of activity is
normally present in roots is sufficient to cause a significant increase in the length of the
roots.

Interestingly, we can also reconcile this membrane-permeability hypothesis with the
systemic induction of RNSI upon wounding. We have not yet identified the signal

involved in regulating RNSI by wounding. Electrical signals have been proposed to cause

133

systemic gene induction in tomato (Stankovic and Davies, 1996) and could affect RNS]
as well. Not only might the induction of RNSI be mediated through electrical stimuli, it is
conceivable that by altering the permeability of cell membranes, RNSI could be involved
in either potentiating or shutting down the plants' response to the signal.

One key question to be addressed in order to support this theory will be the
identification of natural RNAs that are present in membranes. The yeast mutant will
provide a useful tool for these searches, which are currently being conducted by G.C.
Maclntosh and P.J. Green. If they exist, these RNAs are most likely highly specific
species produced by the cell and targeted to the membrane. Current searches in the lab to
identify Arabidopsis RNAs that do not encode proteins as their final product, but instead
function as RNA, may provide a fortuitous overlap with the function of secreted RNases,
in that some of the so-called non-coding RNAs identified may turn out to be membrane-
associated RNAs.

As mentioned briefly in Chapter 1, secreted RNases have several "special" actions
that have not been explained at a biological level. For example, bovine seminal RNase
exhibits a wide range of cytotoxic activities, including immunosuppressive, antiviral,
embryotoxic, and antitumor effects (reviewed in D'Alessio et al., 1997). Angiogenin, a
relative of RNase A, has the unique ability to induce blood vessel formation (reviewed in
Riordan, 1997). How extracellular RNases cause such responses has not been elucidated,
but the field might one day benefit from a reevaluation of special biological actions of

RNases with these new ideas in mind.

REFERENCES

Abel, S., and Kock, M. (2001) Secretory acid ribonucleases from tomato, Lyc0persicon
escutentum Mill. Methods Enzymol. 341, 351-368.

Abel, S., and Glund, K. (1987) Ribonuclease in plant vacuoles: purification and
molecular properties of the enzymes from cultured tomato cells. Planta, 172, 71-78.

Abler, M.L., LeBrasseur, N.D., Danhof, L.R., Thompson, D.M., and Green, P.J. Altered
RNase profile (arp) mutants of Arabidopsis thaliana: Multiple RNase and nuclease
activities are controlled by the ARPI locus. Manuscript in preparation.

Alvarez, ME. (2000) Salicylic acid in the machinery of hypersensitive cell death and
disease resistance. Plant Mol. Biol., 44, 429-42.

Aoyagi, S., Sugiyarna, M., Fukuda, H. (1998) BEN] and ZEN] cDNAs encoding Sl-type
DNases that are associated with programmed cell death in plants. F EBS Lett., 429,
134-138.

Atkinson, A.H., Heath, R.L., Simpson, R.J., Clarke, A.E., and Anderson, M.A. (1993)
Proteinase-inhibitors in Nicotiana alata stigmas are derived from a precursor protein
which is processed into 5 homologous inhibitors. Plant Cell, 5, 203—213.

Bariola, PA. (1996) The RNS family of S-like ribonucleases of Arabidopsis thaliana:
Structures, expression and functions. (Thesis dissertation).

Bariola, RA. and Green, P.J. (1997) Plant ribonucleases. In Ribonucleases: Structures
and Functions (D'Alessio, G. and Riordan, J .F., eds). New York: Academic Press, pp.
163-190.

Bariola, P.A., Howard, C.J., Taylor, C.B., Verburg, M.T., Jaglan, V.D. and Green, P.J.
(1994) The Arabidopsis ribonuclease gene RNSI is tightly controlled in response to
phosphate limitation. Plant J., 6, 673-685.

Bariola, P.A., Maclntosh, G.C., and Green, P.J. (1999) Regulation of S-like ribonuclease
levels in Arabidopsis. Antisense inhibition of RNSl or RN82 elevates anthocyanin
accumulation. Plant Physiol., 119, 331-342.

Bama, B., Ibenthal, W.D., and Heitefuss, R. (1989) Extracellular RNase activity in
healthy and rust-infected wheat leaves. Physio]. Mol. Plant Path., 35, 151-160.

Bechtold, N., Ellis, J., and Pelletier, G. (1993) In planta Agrobacterium-mediated gene

transfer by infiltration of adult Arabidopsis thaliana plants. C. R. Acad. Sci. Paris
316,1194-1199.

135

 

Beers, E.P., McDowell, J .M. (2001) Regulation and execution of programmed cell death
in response to pathogens, stress and developmental cues. Curr. Op. Plant Biol., 4,
561-567.

Beintema, J.J., Breukelman, H.J., Carsana, A., and Furia, A. (1997) Evolution of
vertebrate ribonucleases: Ribonuclease A superfamily. In Ribonucleases: Structures
and Functions (D'Alessio, G. and Riordan, J.F., eds). New York: Academic Press, pp.
245-265.

Bertoni, G. and Mills, D. (1987) A simple method to monitor growth of bacterial
populations in leaf tissue. Phytopathology, 77, 832—835.

Bishop, P.D., Pearce, G., Bryant, IE. and Ryan, CA. (1984) Isolation and
characterization of the proteinase inhibitor-inducing factor from tomato leaves.

Identity and activity of poly- and oligogalacturonide fragments. J. Biol. Chem., 259,
13172-13177.

Boller, T., and Kende, H. (1979) Hydrolytic enzymes in the central vacuole of plant cells.
Plant Physiol., 63, 1123—1132.

Boulter, D. (1993) Insect pest control by copying nature using genetically engineered
crops. Phytochemistry, 34, 1453-66.

Bosse, D., and chk, M. (1998) Influence of phosphate starvation on phosphohydrolases
during development of tomato seedlings. Plant Cell Environ, 21, 325-332.

Bowles, DJ. (1990) Defense-related proteins in higher plants. Annu. Rev. Biochem., 59,
873-907.

Bradford, M.M. (1976) A rapid and sensitive method for the quantification of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem.,
72, 248-254.

Brown, RH, and Ho, T.-H.D. (1987) Biochemical properties and hormonal regulation of
barley nuclease. Eur. J. Biochem., 168, 357-364.

Campbell, RC. (1989) Statistics for Biologists 3rd edn. Cambridge: Cambridge
University Press.

Chang, SC, and Gallic, DR. (1997) RNase activity decreases following a heat shock in

wheat leaves and correlates with its posttranslational modification. Plant Physiol.,
113,1253-1263.

Chen, D.L., Delatorre, C.A., Bakker, A., and Abel, S. (2000) Conditional identification of
phosphate-starvation-response mutants in Arabidopsis thaliana. Planta, 211, 13-22.

Cuchillo, C.M., Vilanova, M., and Nogués, M.V. (1997) Pancreatic Ribonucleases In
Ribonucleases: Structures and Functions (D'Alessio, G. and Riordan, J .F ., eds). New

136

York: Academic Press, pp. 272-297.

Dalby, A., and Davies, 1. (1967) Ribonuclease activity in the developing seeds of normal
and opaque-2 mutant endosperm in maize. Science, 155, 1573-1575.

D'Alessio, G., Di Donato, A., Mazzarella, L., and Piccoli, R. (1997) Seminal
ribonuclease: The importance of diversity. In Ribonucleases: Structures and

Functions (D'Alessio, G. and Riordan, J.F., eds). New York: Academic Press, pp.
383-425.

D'Alessio, G., Di Donato, A., Parente, A., and Piccoli, R. (1991) Seminal RNase: A
unique member of the ribonuclease superfamily. Trends Biochem. Sci. 16, 104-106.

Dodds, P.N., Clarke, A.E., and Newbigin, E. (1996) Molecular characterisation of an
S-like RNase of Nicotiana alata that is induced by phosphate starvation. Plant Mol.
Biol., 31, 227-238.

Farkas, G.L. (1982) Ribonucleases and ribonucleic acid breakdown. Encycl. Plant
Physiol., 14B, 224-262.

Farmer, BE and Ryan, CA. (1992) Octadecanoid precursors of jasmonic acid activate
the synthesis of wound-inducible proteinase-inhibitors. Plant Cell, 4, 129-134.

Fath, A., Bethke, RC, and Jones, R.L. (1999) Barley aleurone cell death is not apoptOtic:
characterization of nuclease activities and DNA degradation. Plant J ., 20, 305-315.

F eys, B.J.F., Benedetti, C.E., Penfold, ON. and Turner, J .G. (1994) Arabidopsis mutants
selected for resistance to the phytotoxin coronatine are male sterile, insensitive to
methyl jasmonate, and resistant to a bacterial pathogen. Plant Cell, 6, 751-759.

Fraser, M.J., and Low, R.L. (1993) Fungal and mitochondrial nucleases. In Nucleases, 2"d
Edn. (Linn, S.M., Lloyd, RS, and Roberts, R.J., eds). Cold Spring Harbor, NY: Cold
Spring Harbor Laboratory Press, pp. 171-207.

Fukuda, H. (1997) Tracheary element differentiation. Plant Cell, 9, 1147-1156.

Fukuda, H. (2000) Programmed cell death of tracheary elements as a paradigm in plants.
Plant Mol. Biol., 44, 245-253.

Galiana, E., Bonnet, P., Conrod, 8., Keller, H., Panabieres, F ., Ponchet, M., Poupet, A.
and Ricci, P. (1997) RNase activity prevents the growth of a fungal pathogen in
tobacco leaves and increases upon induction of systemic acquired resistance with
elicitin. Plant Physiol., 115, 1557-1567.

Gilliland, G.L. (1997) Crystallographic studies of ribonuclease complexes. In

Ribonucleases: Structures and Functions (D'Alessio, G. and Riordan, J.F., eds). New
York: Academic Press, pp. 305-341.

137

Gilmour, S.J., Zarka, D.G., Stockinger, E.J.. Salazar, M.P., Houghton, J.M., and
Thomashow, M.F. (1998) Low temperature regulation of the Arabidopsis CBF family
of AP2 transcriptional activators as an early step in cold-induced COR gene
expression. Plant J. 16, 433-442.

Glund, K., and Goldstein, AH. (1993) Regulation, synthesis, and excretion of a
phosphate starvation inducible RNase by plant cells. In Control of Plant Gene
Expression (Verma, D.P.S., ed). Boca Raton, FL: CRC Press, pp. 31 1-323.

Glund, K., Niimberger, T., Abel, S., Jost, W., Preisser, J., and Komor, E. (1990)
Intracellular Pi compartmentation during phosphate starvation-triggered induction of
an extracellular ribonuclease in tomato cell culture. In Progress in plant cellular and
molecular biology (Nijkamp, H.J.J., Van Der Plas, L.W.H., and Van Aartrijk, J., eds).
Boston: Kluwer Academic Publishers, pp. 338-342.

Goldstein, A.H., Baertlein, DA. and Danon, A. (1989) Phosphate starvation stress as an
experimental system for molecular analysis. Plant Mol. Biol. Rep, 7, 7-16.

Golz, J .F., Clarke, AB. and Newbigin, E. (1995) Self-incompatibility in flowering plants.
Curr. Opin. Genet. Dev., 5, 640-645.

Grant, J .J . and Loake, G]. (2000) Role of reactive oxygen intermediates and cognate
redox signalling in disease resistance. Plant Physiol., 124, 21-9.

Green, TR. and Ryan, CA. (1972) Wound-induced proteinase inhibitor in plant leaves: a
possible defense mechanism against insects. Science, 175, 776-777.

Green, P.J. (1994) The ribonucleases of higher plants. Annu. Rev. Plant Physiol. Plant
Mol. Biol., 45, 421-445.

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

Howard, C]. (1996) Identification and characterization of ribonucleases in Arabidopsis
thaliana. (Thesis dissertation).

Howard, C.J., LeBrasseur, N.D., Bariola, RB. and Green, P.J. (1998) Control of
ribonucleases in response to phosphate limitation: Induction of RNSI in Arabidopsis.
In Ribonucleases: Structures and Functions (D'Alessio, G. and Riordan, J .F., eds).
New York: Academic Press, pp. 204-214.

Hugouvieux, V., Kwak, J .M., and Schroeder, J .I. (2001) An mRNA cap binding protein,
ABHl, modulates early abscisic acid signal transduction in Arabidopsis. Plant Cell,
106, 477-487.

Huynh, Q.K., Borgrneyer, J.R., Smith, C.E., Bell, L.D. and Shah, D.M. (1996) Isolation
and characterization of a 30 kDa protein with antifungal activity from leaves of

138

Engelmannia pinnatifida. Biochem. J., 316, 723-727.

Irie, M. (1997). RNase T1/T 2 family RNases. In Ribonucleases: Structures and Functions
(G. D'Alessio. and J .F . Riordan, eds. New York, Academic Press, pp. 101-130.

lwamatsu, A., Aoyama, H., Dibo, G., Tsunasawa, S., and Sakiyama, F. (1991) Amino
acid sequence of nuclease $1 from Aspergillus oryzae. J. Biochem., l 10, 151-158.

Jaglo-Ottosen, K.R., Gilmour, S.J., Zarka, D.G., Schabenberger, O., and Thomashow,
M.F. (1998) Arabidopsis CBF] overexpression induces COR genes and enhances
freezing tolerance. Science, 280, 104-106.

Johnson, M.A., Baker, E.J., Colbert, J.T., and Green, P.J. (1998) Determinants of mRNA
stability in plants. In A Look Beyond Transcription: Mechanisms Determining
mRNA Stability and Translation in Plants (Bailey-Serres, J. and Gallic, D.R., eds).
Rockville. MD: American Society of Plant Physiologists, pp. 40-53.

Jones, AM. (2001) Programmed cell death in development and defense. Plant Physiol.,
125, 94-97.

Karin, T., Sano, K., Shimokawa, H., ltoh, R., Yamasaki, N. and Kimura, M. (1998)
Isolation and characterization of a wound-inducible ribonuclease from Nicotiana
glutinosa leaves. Biosci. Biotechnol. Biochem., 62, 1144-1151.

Y
Khvorova, A., Kwak, Y.G., Tamkun, M., Majerfeld, l., Yarns, M. (1999) RNAs that bind
and change the permeability of phospholipid membranes. Proc. Natl. Acad. Sci., 96,
10649-10654.

Kock, M., Loffler, A., Abel, S., and Glund, K. (1995) cDNA structure and regulatory
preperties of a family of starvation-induced ribonucleases from tomato. Plant Mol.
Biol., 27, 477-485.

Kock, M., Theierl, K., Stenzel, 1., and Glund, K. (1998) Extracellular administration of
phosphate-sequestering metabolites induces ribonucleases in cultured tomato cells.
Planta, 204, 404-407.

Koncz, C., and Schell, J. (1986) The promoter of TL-DNA gene 5 controls the tissue-

specific expression of chimaeric genes canied by a novel type of Agrobacterium
binary vector. Mol. Gen. Genet, 204, 383-396.

Krysan, P.J., Young, J.C., and Sussman, MR. (1999) T-DNA as an insertional mutagen
in Arabidopsis. Plant Cell, 11, 2283-2290.

Kubigsteltig, 1., Laudert, D. and Weiler, E.W. (1999) Structure and regulation of the
Arabidopsis thaliana allene oxide synthase gene. Planta, 208, 463-471.

Kurihara, H., Nonaka, T., Mitsui, Y., Ohgi, K., Irie, M., and Nakamtrra, K.T. (1996) The
crystal structure of ribonuclease Rh from Rhizopus niveus at 2.0 A resolution. J. M01.

139

Biol., 255, 310-320.

Laudert, D., Pfannschmidt, U., Lottspeich, F., Hollander-Czytko H. and Weiler, E.W.
(1996) Cloning, molecular and functional characterization of Arabidopsis thaliana
allene oxide synthase (CYP74), the first enzyme of the octadecanoid pathway to
jasmonates. Plant Mol. Biol., 31, 323-335.

Laudert, D. and Weiler, E.W. (1998) Allene oxide synthase: a major control point in
Arabidopsis thaliana octadecanoid signalling. Plant J., 15, 675-684.

LeBrasseur, N.D., Maclntosh, G.C., Pérez-Amador, M.A., Saitoh, M., and Green, P.J.
(2002) Local and systemic wound-induction of RNase and nuclease activities in
Arabidopsis: RNSI as a marker for a JA-independent systemic signalling pathway.
Plant J. In press.

Lee, H.-S., Huang, S., and Kao, T.-H. (1994) S proteins control rejection of incompatible
pollen in Petunia inflata. Nature, 367, 560-563.

Lee, HS, Singh, A., and Kao, T.H. (1992) RNase X2, a pistil-specific ribonuclease from
Petunia inflata, shares sequence similarity with Solanaceous S proteins. Plant Mol.
Biol., 20,1131-1141.

Lehmann, K., Hause, B., Altrnann, D., Keck, M. (2001) Tomato ribonuclease LX with
the functional endoplasmic reticulum retention motif HDEF is expressed during

programmed cell death processes, including xylem differentiation, germination, and
senescence. Plant Physiol, 127, 436-449.

Leon, J., Rojo, E. and Sanchez-Serrano, J.J. (2001) Wound signalling in plants. J. Exp.
Bot., 52, 1-9.

Lers, A., Khalchitski, A., Lomaniec, E., Burd, S. and Green, P.J. (1998) Senescence-
induced RNases in tomato. Plant Mol. Biol., 36, 439-449.

Leung, J., and Giraudat, J (1998) Abscisic acid signal transduction. Annu. Rev. Plant
Phys, 49, 199-222.

Li, Q., Von Lanken, C., Yang, J., Lawrence, CB. and Hunt, AG. (2000) The yeast
polyadenylate binding protein (PABl) gene acts as a disease lesion mimic gene when
expressed in plants. Plant Mol. Biol., 42, 335—344.

Loffler, A., Abel, 8., lost, W., Beintema, J .J ., and Glund, K. (1992) Phosphate-regulated
induction of intracellular ribonucleases in cultured tomato (Lycopersicon esculentum)
cells. Plant Physiol., 98, 1472-1478.

Luu, D.-T., Qin, X., Morse, D., and Cappadocia, M. (2000) S-RNase uptake by
compatible pollen tubes in gametophytic self-incompatibility. Nature, 407, 649-451.

Lynch, J., and Brown, K.M. ( 1997) Ethylene and plant responses to nutritional stress.

140

Physiol. Plant, 100, 613-619.

Ma, R.C., Oliveira, M.M. (2000) The RNase PD2 gene of almond (Prunus dulcis)
represents an evolutionarily distinct class of S-like RNase genes. Mol. Gen. Genet,
263, 925-933.

Maclntosh, G.C., Bariola, P.A., Newbigin, E. and Green, P.J. (2001) Characterization of
Rnyl, the Saccharomyces cerevisiae member of the T2 RNase family of RNases:

Unexpected functions for ancient enzymes? Proc. Natl. Acad. Sci. USA, 98, 1018-
1023.

Maclntosh, G.C., Ulloa, R.M., Raices, M. and Te’llez-lfron, MT (1996) Changes in
calcium-dependent protein kinase activity during in vitro tuberization in potato. Plant
Physiol., 112, 1541-1550.

Maekawa, K., Tsunasawa, S., Dibo, G., and Sakiyama, F. (1991) Primary structure of
nuclease Pl fiom Penicillium citrinum. Eur. J. Biochem., 200, 651-661.

Marschner, H. (1986) Function of mineral nutrients: macronutrients: phosphorous. 1n
Mineral nutrition of higher plants (Marschner, H., ed). London: Academic Press, pp
265-277.

Martin, A.C., del Pozo, J .C., lglesias, J ., Rubio, V., Solano, R., de la Pefia, A., Leyva, A.,
and Paz-Ares, J. (2000) Influence of cytokinins on the expression of phosphate
starvation responsive genes in Arabidopsis. Plant J ., 24, 559-567. ‘

Matousek J ., Tmena, L., Svoboda, P., Oriniakova, P., and Lichtenstein, CF. (1995) The
gradual reduction of viroid levels in hop mericlones following heat therapy: A
possible role for a nuclease degrading dsRNA.

McClure, B.A., Haring, V., Ebert, P.R., Anderson, M.A., Simpson, R.J., Sakiyama, F.,
and Clarke, A.E. (1989) Style self-incompatibility gene products of Nicotiana alata
are ribonucleases. Nature 342, 955-957.

McCubbin, A.G., and Kao, T.H. (2000) Molecular recognition and response in pollen and
pistil interactions. Annu. Rev. Cell Dev. Biol., 16, 333-364.

Mittler, R., and Lam, E. (1997) Characterization of nuclease activities and DNA

fragmentation induced upon hypersensitive response cell death and mechanical stress.
Plant Mol. Biol., 34, 209-221.

Moiseyev, G.P., Beintema, J.J., Fedoreyeva, LL, and Yakovlev, G.I. (1994) High
sequence similarity between a ribonuclease from ginseng calluses and fungus-elicited

proteins from parsley indicates that intracellular pathogenesis-related proteins are
ribonucleases. Planta, 193, 470-472.

Nakagawa, A., Tanaka, I., Sakai, R., Nakashima, T., Funatsu, G., and Kimura, M. (1999)
Crystal structure of a ribonuclease from the seeds of bitter gourd (Momordica

141

charantia) at 1.75 angstrom resolution. Biochim. Biophys. Acta Prot. Struc. Mol.
Enzymol., 1433, 253-260.

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

Nilmberger, T., Abel, S., Jost, W., and Glund, K. (1990) Induction of an extracellular
ribonuclease in cultured tomato cells upon phosphate starvation. Plant Physiol., 92,
970-976.

Obara, K, Kuriyama. H., and Fukuda, H. (2001) Direct evidence of active and rapid

nuclear degradation triggered by vacuole rupture during programmed cell death in
zinnia. Plant Physiol., 125, 615- 626.

O'Donnell, P.J., Calvert, C., Atzom, R., Wastemack, C., Leyser, H.M.O. and Bowles,
DJ. (1996) Ethylene as a signal mediating the wound response of tomato plants.
Science, 274, 1914-1917.

O'Donnell, P.J., Truesdale, M.R., Calvert, C., Dorans, A., Roberts, R. and Bowles, DJ.
(1998) A novel tomato gene that rapidly responds to wound- and pathogen-related
signals. Plant J., 14, 137-142.

Orozco-Cardenas, M. and Ryan, CA. (1999) Hydrogen peroxide is generated
systemically in plant leaves by wounding and systemin via the octadecanoid pathway.
Proc. Natl. Acad. Sci. USA, 96, 6553-6557. '

Orozco-Cardenas, M., Narvaez-Vasquez, J. and Ryan, C. (2001) Hydrogen peroxide acts
as a second messenger for the induction of defense genes in tomato plants in response
to wounding, systemin, and methyl jasmonate. Plant Cell, 13, 179-191.

Panavas, T., Pikula, A., Reid, P.D., Rubinstein, 3., Walker, EL. (1999) Identification of
senescence-associated genes from daylily petals. Plant Mol. Biol., 40, 237-248.

Parry, S.K., Liu, Y., Clarke, A.E., and Newbigin, E. (1997) S-RNases and other plant
extracellular ribonucleases. In Ribonucleases: Structures and Functions (D'Alessio, G.
and Riordan, J.F., eds). New York: Academic Press, pp. 192-208.

Pefia—Cortéz, H., Sénchez-Serrano, J.J., Merens, R. and Willmitzer, L. (1989) Abscisic
acid is involved in the wound-induced expression of the proteinase inhibitor 11 gene
in potato and tomato. Proc. Natl. Acad. Sci. USA, 86, 9851-9855.

Pérez-Amador, M.A., Abler, M.L., De Rocher, E.J., Thompson, D.M., van Hoof, A.,
LeBrasseur, N.D., Lers, A. and Green, P.J. (2000) Identification of BFNl, a

bifunctional nuclease induced during leaf and stem senescence in Arabidopsis. Plant
Physiol., 122, 169-179.

Pieterse. CM. and van Loon, LC. (1999) Salicylic acid-independent plant defence

pathways. Trends Plant Sci., 4, 52-58.

Poitier, Y., Thoma, S., Somerville. C., and Schiefelbein, J. (1991) A mutant of
Arabidopsis deficient in xylem loading of phosphate. Plant Physiol., 97, 1087-1093.

Raghothama, KG. (2000) Phosphate transport and signaling. Curr. Op. Plant Biol., 3,
182-187.

Redei, GP. (1967) Biochemical aspects of a genetically detemiined variegation in
Arabidopsis. Genetics, 56, 431-443.

Reymond, P., Weber, H., Damond, M. and Farmer, BE. (2000) Differential gene
expression in response to mechanical wounding and insect feeding in Arabidopsis.
Plant Cell ,12, 707-719.

Richardson, A.E., Hadobas, P.A., and Hayes, J .E. (2001) Extracellular secretion of
Aspergillus phytase from Arabidopsis roots enables plants to obtain phosphorus from
phytate. Plant J ., 15, 641-649.

Riordan, J .F. ( 1997) Structure and function of angiogenin. In Ribonucleases: Structures
and Functions (D'Alessio, G. and Riordan, J .F., eds). New York: Academic Press, pp.
446-466.

Roberts, K., and McCann, MC. (2000) Xylogenesis: the birth of a corpse. Curr. Op.
Plant Biol., 3, 517-522.

Rogers, S.W., and Rogers, JG (1999) Cloning and characterization of a gibberellin-
induced RNase expressed in barley aleurone cells. Plant Physiol., 119, 1457-1464.

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

Rojo, E., Titarenko, E., Leon, J ., Berger, 8., Vancanneyt, G., and Sénchez-Serrano, J.J.
(1998) Reversible protein phosphorylation regulates jasmonic acid-dependent and
-independent wound signal transduction pathways in Arabidopsis thaliana. Plant J .,
13, 153-165.

Ryan, CA. and Pearce, G. (1998) Systemin: a polypeptide signal for plant defensive
genes. Annu Rev Cell Dev. Biol., 14, 1-17.

Ryan, CA. (1990) Protease inhibitors in plants: Genes for improving defenses against
insects and pathogens. Annu. Rev. Phytopathol., 28, 245-249.

Sato, K., and Egami, F. (1957) Studies on ribonuclease in Takadiastase I. J. Biochem.,
44, 753-767.

Schmitz, G., and Theres, K. (1999) Genetic control of branching in Arabidopsis and

143

tomato. Curr. Op. Plant Biol.. 2, 51-55.

Shivakumar, P.D., Vasanthi, N.S., Slretty, H.S., Smedegaard-Petersen, V (2000)
Ribonucleases in the seedlings of pearl millet and their involvement in resistance
against downy mildew disease. Eur. J. Plant Path. 106, 825-836.

Stankovic, B., and Davies, E. (1996) Both action potentials and variation potentials
induce proteinase inhibitor gene expression in tomato. FEBS letters, 390, 275-279.

Stennis, M.J., Chandra, S., Ryan, CA. and Low, RS. (1998) Systemin potentiates the
oxidative burst in cultured tomato cells. Plant Physiol., 117, 1031-1036.

Tanaka, N., Arai, J., Inokuchi, N., Koyama, T., Ohgi, K., Irie, M., Nakamura, K.T.
(2000) Crystal structure of a plant ribonuclease, RNase LE. J. Mol. Biol., 298, 859-
873.

Tang, Z., Sadka, A., Morishige, D.T., and Mullet, J.E. (2001) Homeodomain leucine
zipper proteins bind to the phosphate response domain of the soybean VspB tripartite
promoter. Plant Physiol., 125, 797-809.

Taylor, C.B., Bariola, P.A., DelCardayré, S.B., Raines, RT. and Green P.J. (1993)
RNS2: A senescence-associated RNase of Arabidopsis that diverged from the S-
RNases before speciation. Proc. Natl. Acad. Sci. USA, 90, 5118-5122.

Taylor, CB. and Green, P.J. (1991) Genes with homology to fungal and S-gene RNases
are expressed in Arabidopsis thaliana. Plant Physiol., 96, 980-984.

Thain, J .F., Gubb, LR. and Wildon, DC. (1995) Depolarization of tomato leaf-cells by
oligogalacturonide elicitors. Plant Cell and Environment, 18, 211-214.

Thelen, M.P., and Northcote, DH. (1989) Identification and purification of a nuclease
from Zinnia elegans L.: A potential molecular marker for xylogenesis. Planta, 179,
181-195.

Thomashow, M.F. (1999) Plant cold acclimation: freezing tolerance genes and regulatory
mechanisms. Annu. Rev. Plant Physiol. Plant Mol. Biol., 50, 571-599.

Ticconi, C.A., Delatorre, C.A., and Abel, S. (2001) Attenuation of phosphate starvation
responses by phosphite in Arabidopsis. 127, 963-972.

Titarenko, E., Rojo, 13., Leon, J. and Sénchez-Serrano, J .J. (1997) JA-dependent and -
independent signalling pathways control wound-induced gene activation in
Arabidopsis thaliana. Plant Physiol., 1 15, 817-826.

Wetzel, C.M., Jiang, C.Z., Meehan, L.J., Voytas, D.F., and Rodermel, SR. (1994)
Nuclear-organelle interactions: the immutans variegation mutant of Arabidopsis is
plastid autonomous and impaired in carotenoid biosynthesis. Plant J ., 6, 161-175.

144

Wildon, D.C., Thain, J.F., Minchin, P.E.H., Gubb, I.R., Reilly, A.J., Skipper, Y.D.,
Doherty, H.M., O'Donnell, P.J. and Bowles, DJ. (1992) Electrical signalling and
systemic proteinase-inhibitor induction in the wounded plant. Nature, 360, 62-65.

Wilson, CM. (1982) Plant nucleases: Biochemistry and development of multiple
molecular forms. Isozymes: Curr. Top. Biol. Med. Res, 6, 33-54.

Wilson, C.M., and Alexander, DE. (1967) Ribonuclease activity in normal and opaque-2
mutant endosperm ofmaize. Science, 155, 1575-1576.

Wisman, E. and Ohlrogge, J. (2000) Arabidopsis Microarray Service Facilities. Plant
Physiol., 124, 1468-1471.

Xu, Y., and Hanson, MR. (2000) Programmed cell death during pollination-induced
petal senescence in petunia. Plant Physiol., 122, 1323-1333.

Yang, 8., Wen, X., Kodali, N.S., Oleykowski, C.A., Miller, C.G., Kulinski, J ., Besack,
D., Yeung, J .A., Kowalski, D., Yeung, AT (2000) Purification, cloning, and
characterization of the CEL I nuclease. Biochem., 39, 3533-3541.

Ye, Z.H. and Droste, D.L. (1996) Isolation and characterization of cDNAs encoding
xylogenesis-associated and wounding-induced ribonucleases in Zinnia elegans. Plant
Mol. Biol., 30, 697-709.

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

145