PLACE II RETURN BOX to roman this choekout from your "cord.
TO AVOID FINES mum on or baton dd. duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Namath. AcfiorVEqml Opportunity Initiation
Wanna-m

 

THE IDENTIFICATION AND CHARACTERIZATION
OF THE NUCLEAR LOCALIZATION SEQUENCES OF
THE MAIZE R PROTEIN
By

Mark William Shieh

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Botany and Plant Pathology

1994

ABSTRACT

THE IDENTIFICATION AND CHARACTERIZATION
OF THE NUCLEAR LOCALIZATION SEQUENCES OF
THE MAIZE R PROTEIN

By

Mark William Shieh

Previous genetic and structural evidence indicates that the maize
R gene encodes a nuclear transcriptional activating factor. In-frame
carboxy- and amino-terminal fusions of the R gene to the reporter gene
B-glucuronidase (GUS) were sufficient to direct GUS to the nucleus of
transiently transformed onion epidermal cells. Further analysis of
chimeric constructs containing regions of the _R_ gene fused to the GUS
cDNA revealed three specific nuclear localization sequences (NLSs) that
were capable of redirecting the GUS protein to the nucleus. Amino-
terminal NLS A (aa 100-109, GDRRAAPARP) contained several
arginine residues; a similar localization signal is found in only a few
viral proteins. The medial NLS M (aa 419-428, MSERKRREKL) is an

SV40-type NLS and the carboxyl-terminal NLS C (aa 598-610,

MISEALRKAIGKR) is a Mata2 type. A deletion analysis of the three
localization signals indicated that the amino-terminal and carboxyl
terminal fusions of R and GUS were redirected to the nucleus only
when NLSs A and M, or C and M, were both present. These results
indicate that multiple localization signals are necessary for nuclear
targeting of this protein.

NLS-C is similar to the MataZ-type NLS because it contains
several hydrophobic residues and the basic amino acids are spaced
equally apart. In addition, when the conserved region of the MataZ
NLS (KIPIK) is substituted into the KAIGK region in NLS-C, the
hybrid NLS can still redirect GUS activity to the nucleus in onion
epidermal cells. To identify the essential amino acids for NLS-C
function, mutant NLS:GUS fusions were constructed. A mutant NLS
with the hydrophobic amino acids in NLS-C substituted with non-polar
hydrophilic residues partially redirected GUS activity to the nucleus.
Mutated NLSs, with either the two lysines of NLS-C substituted with
non-basic amino acids or the order of the amino acids in NLS-C
reversed, resulted in GUS activity in the cytoplasm. Therefore, in
NLS-C, the hydrophobic amino acids are important and the two lysines

are necessary for its targeting function. In addition, reversing the

order of the amino acids of NLS-C negated it ability to function as a
NLS, indicating a NLS requires more than the basic amino acid charge

density to function.

DEDICATION

For my mother Rosemary
and in memory of my father William

ACKNOWLEDGEMENTS

I would like to thank Phil McCrea, my high school advisor and
Biology club sponsor who encouraged my interest in science. I would
also like to thank Natasha Raikhel for her guidance and support. Her
lab has been an excellent environment to do research. I would like to
thank the members of my guidance committee; Dr. Natasha Raikhel,
Dr. Pamela Green, Dr. Ronald Patterson and Dr. John Wang as well
as Tracey Reynolds for their critical review of this manuscript and the
helpful discussions during my dissertation.

vi

TABLE OF CONTENTS

LIST OF TABLES .............................. ix
LIST OF FIGURES ............................. x
ABBREVIATION .............................. xii
CHAPTER 1 .................................. 1
INTRODUCTION ........................... 1
REFERENCES ........................... 28

CHAPTER 2 NUCLEAR TARGETING OF THE MAIZE R
PROTEIN REQUIRES TWO NUCLEAR

LOCALIZATION SEQUENCES ............... 36
ABSTRACT ............................. 37
INTRODUCTION .......................... 39
MATERIALS AND NIETHODS ................. 42
RESULTS ............................... 48
DISCUSSION ............................ 63
ACKNOWLEDGEMENTS .................... 74
REFERENCE ............................ 75

vii

CHAPTER 3 CHARACTERIZATION OF THE
CARBOXY-TERMINAL NUCLEAR LOCALIZATION

SEQUENCE OF THE MAIZE R PROTEIN ........ 80
INTRODUCTION .......................... 81
MATERIALS AND METHODS ................. 84
RESULTS ............................... 87
DISCUSSION ............................ 96
REFERENCES ........................... 103

CHAPTER 4 FUTURE RESEARCH PROSPECTIVES 105
REFERENCES ........................... 110

viii

LIST OF TABLES

Table 1.1 Summary of plant nuclear localization sequences identified
and the known function of their proteins. .............. 25

Table 3.1 Summary of the histochemical analysis for the mutated
NLS-C:GUS fusion protein. ....................... 92

ix

LIST OF FIGURES

Figure 1.1 Nuclear import of proteins in Xenopus oocytes is a two
step process, involving binding and translocation across the nuclear
envelope. .................................... 6

Figure 2.1. Histochemical localization of R:GUS (A) and GUS:R (B)
fusion proteins in onion epidermal cells. ............... 50

Figure 2.2. Cloning strategy for preparing R:GUS (A) and GUS:R
(B) fusions and results of localization experiments. ........ 52

Figure 2.3. Histochemical localization of three NLS regions of the R
protein fused to GUS (above) and schematic representation of
R:GUS fusion protein showing localization of three NLSs. . . . . 54

Figure 2.4. Effect of deletion of different NLSs on the histochemical
localization of R:GUS fusion proteins. Deletion of different NLSs
from the intact R protein fused to GUS showed that NLS A and M
or M and C are required for nuclear targeting. .......... 58

Figure 2.5. Summary of histochemical analysis of R:GUS fusion
proteins which identified NLSs that were necessary for nuclear
localization. ................................. 61

Figure 2.6. Amino acid comparison of R-Lc to other homologous
regulatory proteins. ............................ 67

Figure 3.1. Amino acid sequences of the mutated NLS-C polypeptides.
.......................................... 89

Figure 3.2. Histochemical localization of GUS activity of each of the
mutated NLS-C GUS fusion constructs. ................ 90

xi

ABBREVIATIONS

a.a. amino acid

ATP adenosine triphosphate

B-gal fl-galactosidase

C Celsius

CaMV Cauliflower mosaic virus

cDNA complementary deoxyribonucleic acids
DAPI 4’,6’ diamidino-2-phenylindoledihydroxychloride
DNA deoxyribonucleic acids

DEL delila

FGF fibroblast growth factor

G1 growth phase 1

GTP guanosine triphosphate

GUS fl-glucuronidase

HBV hepatitis virus

HSP 90 heat shock protein 90
I kappa ,82 Inhibitor kappaB

kDa kilodalton

Mata2 mating factor 012

MS Murashige and Skoog
NEM N-ethylmaleimide

NF kappafi nuclear factor kappa 62
NLS nuclear localization sequence
"In nanometer

NPC nuclear pore complex
NOS nopaline synthetase

nt nucleotide

NUP nuclear pore protein
P10 10 kDa factor

PRnls progesterone receptor nuclear localization sequence

xii

RNA
RN ase

SMHC-29
ssDNA
SV40
SW15

#8

um

WGA

ribonucleic acids

ribonuclease

synthesis (growth phase)

myosin heavy chain from rabbit
single stranded deoxyribonucleic acids
Simian virus 40

switch 15

microgram

micrometer

wheat germ agglutinin

xiii

CHAPTER 1

INTRODUCTION

2
INTRODUCTION

The increase in genome size from prokaryotes to eukaryotes
has led to the development of the nucleus. The nuclear
compartment organizes the genomic DNA and effectively separates
the processes of transcription and translation. RNA must be
exported to the cytoplasm to be translated and proteins such as
transcription factors require import to the nucleus to function.
Therefore, regulation of transport across the nuclear envelope serves
as an additional level of control over vital cellular processes.

The nuclear envelope consists of two lipid bilayer membranes.
The outer lipid bilayer is contiguous with the membrane of the
endoplasmic reticulum. On the nucleoplasmic side of the inner
membrane, there is a lattice of lamin intermediate filaments which
provide ridgity and shape to the nucleus. Spanning the inner and
outer membrane are large protein complexes termed nuclear pore
complexes (NPC). NPCs are approximately 120 nm in diameter and
are a round structure formed by eight protein complexes located on
both the nucleoplasmic and cytoplasmic sides of the envelope, with a

central protein complex in the center (Rout and Wente, 1994).

3

Extending from the pore complex are fibrial networks on the inside
of the nucleus. This interlaced pattern of the fibrials has led them
to being coined "fish baskets" (Jarnik and Aebi, 1991; Goldberg and
Allen, 1992). The nuclear pore complexes form aqueous channels, 9
nm in diameter, which connect the nucleoplasm to the cytoplasm.
Through the aqueous channels, small molecules are able to freely
diffuse (dextrans up to 50 kDa; for review, Forbes, 1992; Paine et.
al., 1975) and proteins and RNAs are transported.

Nuclear pores complexes were first identified as the site of
protein transport by using gold particles coated with nuclear
targeting signals which accumulated at the pore complex (for review,
Rout and Wente, 1994). Later studies demonstrated that both
protein import and RNA export occur through the same pores,
because gold particles coated with either RNA or NLS localized to a
single pore (Dworetzky and F eldherr, 1988a). Addition of wheat
germ agglutinin (W GA), a lectin with specificity for N-acetyl-
glucosamine, to in vitro protein transport systems in animal cells
blocked transport (Finlay et. al., 1987; for review, Forbes, 1992).
Immunocytochemical analysis revealed that WGA was bound to the

pore complexes, indicating that proteins of the pore complex are

4

modified in the cytoplasm with N-acetylglucosamine. This
carbohydrate modification was used to purify and isolate a number
of proteins which reside at the pore complex. Recently, N-
acetylglucosamine modified proteins have been discovered on nuclear
envelope proteins in plants, most are localized at the pore complex.
In addition, carbohydrate analysis of these plant proteins reveals
that, unlike the single O-linked carbohydrate N-acetylglucosamine
modification in animal cells, plants can have an oligosaccharide at a
single O-linked site (A. Heese-Peck, Cole, Hart and N. Raikhel,
unpublished).

The function of proteins located at the pore complex is
difficult to ascertain, although functions for some have been
proposed because they contain regions with homology to nucleotide
binding domains. The nuclear pore proteins, NUP 149, NUP 100,
and NUP 116 contain domains homologous to RNA binding motifs
(F abre et. al., 1994) and these proteins may be involved in the
export of RNA to the cytoplasm. The NUP 149 protein binds poly-
guanosine (ssDNA) and a temperature-sensitive mutant of NUP 149
accumulates polyadenylated messages in the nucleus (Fabre et. al.,

1994). Though the specificity of nucleotide binding for NUP 149 in

5

vitro does not match the phenotype of the mutant, the NUP 149
mutant does interfere with RNA export. The pore protein NUP 153
contains a zinc finger motif and may function as an anchor for
chromatin (Sukegawa and Blobel, 1993). Our understanding of the
nuclear pore is increasing rapidly and utilization of techniques such
as synthetic lethality are expanding our knowledge of the
interactions which occur between nuclear pore proteins and the

factors with which they associate (F abre et. al., 1994).

Molecular Mechanism of Nuclear Transport

In general, the import of proteins into the nucleus is both
energy and signal dependent, however, small macromolecules do
diffuse through the pore complexes (for review see Forbes, 1992;
Paine et. al., 1975). The molecular mechanism of nuclear transport
can be separated into two steps in Xenopus oocytes (Fig. 1.1). First,
the protein binds to the nuclear envelope; this is mediated by its
nuclear localization signal (NLS) and cytosolic factors. Then the
protein is translocated across the envelope. The transported protein

binds to the nuclear envelope with its NLS and this requires

 

NLS(s)-d ndent ATP-dependent
NEM sens ive ATPYS sensitive
A yrase sensitive
GA sensitive

 

 

Temperature sensitive

Figure 1.1 Nuclear import of proteins in Xenopus oocytes is a two
step process, involving binding and translocation across the nuclear

envelope.

7

cytosolic factors (Fig. 1.1; Richardson et. al., 1988, Newmeyer and
Forbes, 1988b). A NLS-containing protein will not bind to the
nucleus if the cytosol is treated with the sulfhydryl modifying agent
N-ethylmaleimide (NEM) (Fig. 1.1; Newmeyer and Forbes, 1990).
The second step, translocation, is an energy dependent process
requiring hydrolysis of ATP and, possibly, GTP (Fig. 1.1). The
requirement for ATP hydrolysis has been shown by blocking
transport with the non-hydrolyzable analog ATPyS and because
transport is inhibited when endogenous ATP is depleted with
apyrase, in in vitro reconstituted and semi-in viva (microinjection)
transport systems (Newmeyer and Forbes, 1988; Richardson et. al.,
1988). The requirement for GTP is implicated from studies
analyzing the cytosolic proteins necessary for transport. Blobel and
coworkers separated the cytosol into two fractions by ion-exchange
chromatography; one fraction was necessary for binding and the
other for translocation (Moore and Blobel, 1992). The cytosolic
protein in the translocation fraction was isolated and identified as
Ran/T C4, a GTP-binding protein with roles in DNA replication, cell
cycle checkpoint control, RNA synthesis, processing and export

(Moore and Blobel, 1993; Melchior et. al., 1993). Prior to these

8

studies, the role of GTP in nuclear transport had not been
thoroughly studied in Xenopus oocytes and no additional GTP is
required to reconstitute transport in vitro. The non-hydrolyzable
analog GTPTS inhibits but does not block nuclear import in vitro,
therefore, the function of GTP in nuclear transport is not clear
(Melchior et. al., 1993). Recently, a 10 kDa factor (p10) in the
translocation fraction of the cytosol was identified. The p10 protein,
with the Ran/T C4 protein, functions as well as the completed
translocation fraction of cytosol when combined with the binding
fraction in in vitro transport systems (Moore and Blobel, 1994).
Translocation across the envelope is also inhibited at lower
temperatures (4°C). Microinjection of cells with nucleoplasm protein
at 4°C results in no nuclear import; if the temperature is raised to
37°C, transport resumes (Richardson et. al., 1988). In addition,
microinjected WGA, which binds to nuclear pore proteins modified
with N-acetylglucosamines, blocks translocation of proteins and
export of RNA across the envelope (Fig. 1.1; Newmeyer and Forbes,
1988b; for review Rout and Wente, 1994).

Nuclear transport in systems other than the well studied

Xenopus oocytes differ in their behavior. As an example, in somatic

9

animal cells, there is no requirement for cytosol and nuclear
transport can be reconstituted in vitro with purified nuclei alone

(Markland et. al., 1987; Dean and Kasamatsu, 1994).

Nuclear Localization Sequences

NLSs are short polypeptides that are the sole determinant for
protein import to the nucleus. They are identified by their ability to
direct a reporter protein, normally localized in the cytoplasm, to the
nucleoplasm. The properties of the reporter proteins are that they
must be a soluble cytoplasmic-localized protein and larger than the
predicted exclusion limit for diffusion through the nuclear pore (40-
50 kDa; Garcia-Bustos, 1991). The reporter proteins used for yeast
and animal cells are human (bovine or chicken) serum albumin
(Dworetsky et. al., 1988), pyruvate kinase (Dingwall et. al., 1988;
Kalderon et. al.,1984) and fl-galactosidase (Hall et. al., 1990; Picard
and Yamamoto, 1987). fi-glucuronidase (GUS) is a common
reporter protein in plants (Shieh et. al., 1993; Raikhel, 1992).

In general, the ability of a polypeptide to function as a NLS

has been assessed by two methods. First, a gene fusion of the

10

putative NLS to a reporter cDNA is stably or transiently expressed
in the cells of interest. Localization of the reporter protein is
determined by immunocytochemistry or histochemical (GUS and B-
gal) assays (Opaque 2; Varagona et. al., 1992, R protein Shieh et.
al., 1993). Second, putative NLSs are fused to a reporter protein as
either synthetic peptides chemically crosslinked to a reporter
protein, or gene fusions are in vitro translated proteins or in vitro
transcribed RNA (SV40-NLS +pyruvate kinase DNA; Kalderon et.
al., 1984b) and are microinjected into cells. Localization, in
general, is determined by immunofluorescence as the reporter
proteins are generally labeled with fluorescent compounds.

Utilizing these methods, over a hundred NLSs have been
identified from plant, animal and yeast systems (for review see
Boulikas, 1993; Raikhel, 1992). The NLSs vary in amino acid
composition and length with no consensus sequence or structure and
therefore they must be identified empirically. Unlike targeting
signals to other organelles, the NLSs are found throughout a protein
and multiple targeting signals are frequently found in a single
protein (for review, Garcia-Bustos et. al., 1991). All NLSs contain

the basic amino acid(s) lysine and/or arginine (Garcia-Bustos et. al.,

11

1991; Boulikas, 1993). Based on amino acid composition and size,
NLSs can be separated into three groups: SV40 large T antigen-like
(Kalderon et. al., 1984a,b; Lanford and Butel, 1984), bipartite
(nucleoplasmin; Dingwall and Laskey, 1991) and MataZ-like (Hall
et. al., 1984). SV40-like signals are 7-20 amino acids (a.a.) in length
and enriched in basic amino acids, whereas bipartite signals contain
two basic amino acid enriched regions separated by 10-30 amino
acids. The NLS of SV40 large T antigen has been shown to function
in animal, yeast and plant cells (Lanford and Butel, 1984; Nelson
and Silver, 1989; Varagona and Raikhel, 1994; van der Krol and
Chua, 1991). Mata2-like (mating factor a2) NLSs contain several
hydrophobic residues and a single basic amino acid enriched region.
The polypeptide motif KIPIK in the amino terminal MataZ NLS is
conserved in several other yeast nuclear proteins, and is considered
to be a Mata2-like type of NLS. In addition, unlike other NLSs
which can function as targeting signals in animal, plant and fungal
systems, the yeast Mata2 NLS does not function as a targeting signal
in animal cell lines (Chelsky et. al., 1989; Lanford et. al., 1990).
Many of the NLSs contain a region enriched in basic amino

acids such as the SV40 large T antigen where five out of seven

12

residues are basic (Kalderon et. al., 1984a and for review see
Boulikas, 1993). However, a highly basic amino acid stretch does
not function as a NLS since a number of cytoplasmic proteins such
as SMHC-29 (rabbit myosin heavy chain; Nagiai et. al., 1988), 2-5A
dependent RNase (Murine; Zhou et. al., 1993) and porcine p11
(calcium binding protein; Mastakowski and Shooter, 1988) contain
basic a.a. enriched regions. Boulikas (1993) proposed that a
polypeptide containing at least four basic amino acids in a
hexapeptide would function as a NLS. This hypothesis was tested by
analyzing 117 transcription factors and 109 non-nuclear proteins for
the basic rich regions. The hypothesis was false as a number of
non-nuclear proteins contained regions with four of six basic amino
acids (Boulikas, 1994). The author proposed that the basic rich
regions may not function as a NLS because many of the proteins
were sequestered from the nuclear transport machinery in an
organelle or by association with a membrane. Also, the soluble non-
nuclear cytoplasmic proteins were thought to limit the exposure of
the basic rich regions on the surface of the protein therefore, hiding
the region from the nuclear transport machinery.

Studies with the SV40 large T antigen NLS and androgen

13

steroid hormone receptor indicate that the location of the NLS
within a protein has a strong influence on its function. To
determine if the SV40 NLS can function in different locations within
a protein, transcribed message from gene fusion constructs encoding
the SV40 NLS in different locations in the pyruvate kinase cDNA
were expressed in animal cells and their subcellular localization
determined. The NLS functions in four of the five regions tested
within pyruvate kinase and if inserted into amino acid position 125-
136, the SV40 NLS:pyruvate kinase remains cytoplasmic (Roberts et.
al., 1987). Deletion analysis of the androgen receptor indicated that
its bipartite NLS will not function when one of the two DNA binding
domains is deleted (Zhou et. al., 1994). Therefore, the context in
which a NLS is presented is crucial to its function.

Recently, it has been proposed that NLSs may also function in
the export of proteins. This is based upon the observation that, cells
microinjectioned with PRnls-fi-gal protein (progesterone receptor
NLS fused to fl-galactosidase) will import the fusion protein into
nucleus. However, upon energy depletion, induced by low
concentrations of sodium azide, the PRnls-B-gal protein accumulated

in the cytoplasm. When the same experiment is performed with B-

14

gal microinjected into the nucleus, the B—gal is not exported to the
cytoplasm. The effect of the energy depletion is reversible and
results in the nuclear import and nuclear localization of the export
PRnls-fi-gal protein (Guichon-Mantel et. al., 1994). These
experiments indicate that retention of proteins in the nucleus may be
an active process or that nuclear proteins may shuttle between the

two compartments in a NLS-dependent manner.

Dual Localization in the Nucleus and Secretory Pathway

Several proteins contain two different targeting signals, one is
a signal peptide and the other a NLS. These proteins are found in
the secretory pathway (Golgi and extracellular space) and also in the
nucleus. The FGF3 (mouse fibroblast growth factor) protein is an
example of a protein with two targeting signals. When translation is
initiated from the ATG codon, the FGF3 protein is secreted from the
cell (Keifer et. al., 1994). However, it was determined that
translation of the FGF3 protein is normally initiated from an
upstream CUG codon and this additional 29 a.a. results in FGF3

protein localized to the nucleus and extracellular space (via the

15
secretory pathway). The NLS in FGF3 is not located in the

additional amino acids but is contained within the amino-terminus of
the mature protein (Keifer et. al., 1994). Therefore, this may
represent a mechanism which regulates the subcellular localization of
FGF3 by the selection of different initiation sites. Another protein
which is dually-localized but not regulated in its localization, is the
hepatitis virus (HBV) precore P22 protein. P22 contains a 19 amino
acid secretory signal sequence and a NLS. The p22 is processed to
its mature form by a signal peptidase and localizes to both the Golgi
(30%) and nucleus (70%, Du et. al., 1989). The authors propose
that after cleavage of the signal sequence from P22 that some of the
protein is released back to the cytoplasm were it is imported to the
nucleus. However, the mature protein may have been secreted and
then reintroduced to the cells as part of a viral particle. Then
without a secretory signal the NLS on the mature P22 protein
targets it for import to the nucleus. NLSs can also reside in
normally secreted proteins without being utilized. The secreted
protein of simian sarcoma virus V-sis contains both targeting signals
but unlike FGF3 and P22, the V-sis protein is secreted. If the

secretory signal sequence is destroyed, the V-sis protein is then

16

imported to the nucleus (Lee et. al., 1987). These findings raise a
number of questions concerning the evolution of different targeting
signals and how selection of NLSs verses other signals occurs. Is
there an advantage to a protein containing different targeting signals
as opposed to having a gene family with a single type of signal?
Alternatively, are these examples of genes which, by selection, will
develop into gene families, each member with a single type of

targeting signal?

Modification of NLSs and Regulation of Transport

Regulation of the cell cycle, development and transcriptional
activation can occur by controlling the transport of proteins into the
nucleus. Two mechanisms have been identified which control
nuclear transport: modification of the protein to change the
efficiency of the NLS and retention of proteins in the cytoplasm by a
cytoplasmic—anchoring protein.

The efficiency of NLSs can be altered by the phosphorylation

state of the protein. The cell cycle is regulated by a number of

17

protein phosphorylation/dephosphorylation events that can change
the rate or time of nuclear transport. The SW15 protein of
Saccharomyces cervisiae is a transcription factor which regulates the
H0 gene, the protein involved in mating type switches. SW15
message is synthesized in the S phase of growth and transcription
stops before entering G1. The protein product accumulates in the
cytoplasm until the cell enters G1, then SW15 is transported into the
nucleus (Nasmyth et. al., 1990). Near the NLS are three CDC28
kinase sites that are phosphorylated when the cell enters the S phase
(Moll et. al., 1991). Mutation of the three phosphorylated serine
residues to non-phosphorylatable alanines results in nuclear
localization of the SW15 protein throughout the cell cycle. This
indicates that the unphosphorylated form of SW15 is nuclear
transport competent (Moll et. al., 1991). Therefore, nuclear import
of SW15 is prevented by phosphorylation and stimulated by the
unphosphorylated form of SW15.

In another protein, phosphorylation can increase the rate of
import to the nucleus. Rihs and coworkers (1991) identified two
putative casein kinase 11 sites within 15 amino acids of the SV40

large T antigen NLS. Substitution of either of the casein kinase II

18

site, with non-phosphorylatable amino acids, decreased the rate of
nuclear accumulation. This was done by microinjection of the
modified SV40 NLS B-gal fusion protein into cells (Rihs et. al.,
1991). Therefore, the phosphorylated form of SV40 large T antigen
NLS is imported at a much greater rate (15 fold) then the non-
phosphorylatable mutant; this increase in transport rate may be due
to conformational changes in the phosphorylated NLS (Rihs et. al.,
1991).

Phosphorylation of proteins also regulates protein-protein
interactions of the rel (proto)-oncogene family, NF kappaB (nuclear
factor) and dorsal. NF kappa B is a transcription factor for the
synthesis of immunoglobin Kappa-light chain. The dorsal protein is
a putative transcription factor that is essential for the Drosophila
embryo to develop a dorsal/ventral axis. Dorsal creates this
developmental polarity by establishing a gradient of nuclear
accumulation where the ventral pole has the highest level of dorsal
in the nucleus (Govind and Steward, 1991). Both NF kappaB and
dorsal are retained in the cytoplasm until, through a signal
transduction pathway from the plasma membrane, they are

phosphorylated and enter the nucleus. The proposed mechanism is

19

that the rel proteins are anchored in the cytoplasm by interaction
with another protein. It is known by genetic and in vitro binding
assays that the dorsal protein binds to a protein called cactus. Cells
with a mutant cactus protein have weakly ventrilized embryos with
dorsal protein in more nuclei of the dorsal pole (Isoda et. al., 1991).
Cactus protein resides in the cytoplasm and will not bind
phosphorylated dorsal, (Whalen and Steward, 1993) suggesting that
the cactus/dorsal complex is nuclear import incompetent and that
cactus, not phosphorylation of the dorsal NLS, is retaining dorsal in
the cytoplasm. A similar interaction is found between NF kappaB
and I kappaB; the binding affinity between these two proteins is also
disrupted by phosphorylation of NF kappaB (Beg et. al., 1993).
Sequence analysis of cactus and I kappaB showed that both proteins
contain ankryin-like repeats which are known to be involved in
protein-protein interaction. Therefore, the ankyrin-repeat is
thought to bind NF kappa B and dorsal or anchor the heterodimer
complex to the cytoskeletal proteins spectrin or ankyrin (Blank et.
al., 1992). The mechanism by which cactus and I kappaB block the
import of the rel proteins is not understood, but it is hypothesized

that they bind and mask the region containing the NLS in NF

20
kappaB and dorsal (Beg et. al., 1993). This hypothesis is supported

by studies which show that a protein which does not contain a NLS
can enter the nucleus when carried by another nuclear protein and
no cactus or I kappaB is found in the nucleus (Kang et. al., 1994).
An alternative hypothesis is that cactus and IkappaB are anchored
to the cytoskeleton and that the NLSs of dorsal or NF kappaB
though recognized by the transport machinary cannot facilitate the
nuclear import of the protein complex.

Another example of cytoplasmic retention is with members of
the steroid hormone receptors such as the glucocorticoid,
progesterone and estrogen receptors. In the absence of a steroid
hormone, they reside in the cytoplasm and addition of hormone
results in their import to the nucleus and transcriptional activation.
The receptors are known to bind to HSP90 (heat shock protein) in
the absence of hormone and it is believed that HSP90 retains the
hormone receptors in the cytoplasm. Once the receptor binds the
hormone, it is then imported into the nucleus. Recently, it was
shown that the glucocorticoid and progesterone receptors bind to
HSP90 in viva; when the two NLSs are deleted from the hormone

receptor and a NLS added to HSP90, the two proteins are localized

21
in the nucleus (Kang et. al., 1994). This indicates that HSP90 and

the hormone receptor are bound together in the cytoplasm. Since
HSP90 with a NLS is able to redirect the cytoplasmic form of the
hormone receptor to the nucleus, the HSP90 protein bound to the
hormone receptor must block the two NLSs on the hormone receptor
preventing the nuclear import of the heterodimeric protein.

The regulation of nuclear import adds an additional level of
control for transcriptional activity and it is interesting to speculate
why such a mechanism has developed. The advantage of retaining
transcriptional proteins in the cytoplasm until needed suggests that a
response to a stimuli should occur quickly and be faster than
initiating transcription and translation of the transcription factor.
Alternatively, the nuclear import process can be modulated, varying
the amount of transcription induced as in the case of dorsal which is
differentially imported along the axis of embryo development.
Perhaps irreversible developmental changes in cells, such as the
formation of the dorsal/ventral axis or initiation of immunoglobulin
production, require a unique form of regulation where the
transcription factors are produced at a single stage of cell

development and reside in the cytoplasm until activated.

22
Nuclear Transport in Plants

The study of nuclear transport in plants has only recently
begun and already differences are being discovered. The
identification of oligosaccharide modifications on nuclear envelope
proteins is unique to plants and N-acetylglucosamine residues have
not yet been discovered on yeast nuclear pore proteins. To begin,
the identification and characterization of the molecular mechanism
in plants, NLSs of plant proteins were identified. There have been a
number of reviews covering the NLSs of animal and yeast proteins
(Boulikas, 1993, Garcia-Bustos, 1991) but they do not include plant
NLSs, therefore a list of plants NLSs has been included in Table 1.1.
Similar to the targeting signals identified in other eukaryotes, plant
NLSs contain basic amino acids and they can be classified under the
three types of NLSs (SV40-like, bipartite, and Mata2-like).
However, plant NLSs have a stronger binding affinity for plant
nuclear envelope proteins (putative receptors) than mammalian
NLSs. When the binding affinity of NLSs for NLS-binding proteins
on isolated nuclei from tobacco or maize suspension cells the plant

Opaque 2 NLS had a stronger binding affinity and was a better

23
competitor than the SV40 large T antigen NLS (Hick and Raikhel,

1993). The nuclear NLS-binding affinity for peptides of the Opaque
2 bipartite NLS is 200 MM and the binding activity is proteinaceous
(Hicks and Raikhel, 1993).

When I began my thesis research virtually nothing was known
about nuclear transport in plants. Therefore, to study the molecular
mechanism of nuclear transport in plants, we chose to identify the
nuclear localization signals in the maize R protein; a proposed
regulator of anthocyanin biosynthesis. Anthocyanins are pigments
(red or purple) expressed in a tissue-specific manner in maize. Since
anthocyanins are an easily detectable phenotype, the genetics of its
biosynthetic pathway has been thoroughly studied and R expression
was found to be modulated by a number of genes and it may be
regulated at the level of import to the nucleus.

To identify the NLS in the R protein, gene fusion constructs
with the R cDNA or short segments of the R cDNA were fused to the
reporter gene GUS. These gene fusions were transiently expressed
in onion epidermal cells and the subcellular location of the GUS
activity was determined with a histochemical substrate using

Nomarski optics. Three NLSs were identified in the R protein.

24

Because there are multiple NLSs, those necessary for exclusive
nuclear localization of the full length R cDNA fused to GUS were
determined. In addition, the carboxy-terminal NLS (NLS-C) of R
has an unusual amino acid composition similar to the previously
uncharacterized yeast MataZ-like NLS. Therefore, NLS-C was
further analyzed in detail to determine which of its 13 a.a. are

essential for its function.

25

Table 1.1 Summary of plant nuclear localization sequences identified
and the known function of their proteins. The bold letters denote

basic amino acids.

tein and Function

   

26
Table 1.1

Nuclear

  

 

Reference

 

 

 

 

 

 

 

 

Localization NLS
Sequence
Nla amino-terminus - GKE‘JQKHKLK SV40- Carrington
Tobacco etch virus, M Like et. al.,
catalyzes proteolysis of 1991
the viral polyprotein in
five places and is a
genomic VPg
Nla central signal KRKGTTRGMG Bipartite Carrington
AKSRKFINMYG et. al.,
FDPTDFSYI 1991
Opaque 2 NLS-A - MEEAVTMAPA SV40- Varagona
Maize transcriptional AVSSAVVGDPM like et. al.,
regulator of a zein EYNAILRRKLE 1992
biosynthesis EDLE
Opaque 2 NLS-B MPTEERVRKR Bipartite Varagona
KESNRESARRS et.al, 1992
RYRKAAHLKE
L
R NLS-A -Maize GDRRAAPARP SV40- Shieh et.
transcriptional like al., 1993
regulator of ‘
anthocyanin
biosynthesis
R NLS-M -As above MSERKRREKL SV40- Shieh et.
like al., 1993
R NLS-C -As above MISEALRKAIG Mata2- Shieh et.
KR like al., 1993

 

 

 

 

 

27
Table 1 .l (cont’d)

 

 

 

 

 

 

 

Protein and Function Nuclear Type of Reference
Localization NLS
Sequence
TGA-lA -transcription MAKPVEVLRR Possible van der
factor with homology LAQNREAARKS Bipartite Krol and
to the CREB protein RLRKKAYVQQ Chua, 1991
LENSKLKLIQL
EQELEQILERA
RKQGMCVGGG
VDASQLSYSGT
ATRGSPGGQSL
TGA-lB -transcription MAEKKRARLV Bipartite van der
factor with homology RNRESAQLSRQ Krol and
to the CREB protein RKKHV ----- IS Chua, 1991
VirD2 amino terminus YISRKGKLEL SV40- Tinland et.
-Agrobacterium; VirD2 like al.,1992
binds at the 5’ end of
the T-DNA strand and
may facilitate the
nuclear import of the
T-DNA
VirD2 carboxyl- KRPREDDDGEP Bipartite Howard et.
terminus SERKRER al.,1992;
Tinland et.
al., 1992
VirE2 -Agrobacterium KLRPEDRYIQT Proposed Citovsky
ssDNA binding protein EKYGRR-49a.a. Bipartite et. al.,
coats the intermediate spacer-- maybe 1992
T-DNA strand of TKYGSDTEIKL two
Agrobacterium T-DNA KSK SV40-
and may facilitate the like
T-DNA transport to the NLSs

nucleus

 

 

 

 

 

28

REFERENCES

Beg AA, Ruben SM, Scheinman RI, Haskill S, Rosen CA,
Baldwin AS Jr. (1992) I-kappaB interacts with the
nuclear localization sequences of the subunits of NF-

kappa B: a mechanism for cytoplasmic retention. Genes
and Dev. 6: 1889-1913

Blank V, Kourilsky P, Israel A (1992) NF kappaB and related
proteins: Rel /dorsal homologies meet ankyrin-like
repeats. T.I.B.S. 17: 135-140

Boulikas T (1993) Nuclear localization signals. CRC Crit. Rev.
Euk. Gene Expr. 3(3): 193-227

Boulikas T (1994) Putative nuclear localization signals (NLS)
in protein transcription factors. J. Cell. Biochem. 55: 32-
58

Carrington JC, Freed DD, Leinicke AJ (1991) Bipartite signal
sequence mediates nuclear translocation of the plant potyviral
Nla protein. Plant Cell 3: 953-962

Chelsky D, Ralph R, Jonak G (1989) Sequence requirements for
synthetic peptide-mediated translocation to the nucleus. Mol.
Cell. Biol. 9: 2487-2492

Citovsky V, Zupan J, Warnick D, Zambryski P (1992) Nuclear
localization of Agrobacterium VirE2 protein in plant cells.
Science 256: 1802-1805

Dean DA, Kasamatsu H (1994) Signal- and energy-dependent
nuclear transport of SV40 Vp3 by isolated nuclei.

Establishment of a filtration assay for nuclear transport.
J. Biol. Chem. 269: 4910-4916

Dingwall C, Laskey RA (1991) Nuclear Targeting sequences- a
consensus? TIBS 16: 478-481

29

Dingwall C, Robbins J, Dilworth SM, Roberts B, Richardson
WD (1988) The nucleoplasm nuclear location sequence is

larger and more complex than that of SV40 large T
antigen. J. Cell. Biol. 107: 841-849

Dworetsky SI, Feldherr CM (1988a) Translocation of RNA-
coated gold particles through the nuclear pores of
oocytes. J. Cell. Biol. 106: 575-584

Dworetsky SI, Lanford RE, Feldherr CM (1988b) The effect
of variations in the number and sequence of targeting
signals on nuclear uptake. J. Cell. Biol. 107: 1279-1287

Fabre E, Boelens WC, Wimmer C, Mattaj IW, Hurt EC
(1994) Nup145p is required for nuclear export of mRNA

and binds homopolymeric RNA in vitro via a novel
conserved motif. Cell 78: 275-289

Finlay DR, Newmeyer DD, Price TM, Forbes DJ (1987)
Inhibition of in vitro nuclear transport by a lectin that
binds to nuclear pores. J. Cell. Biol. 104: 189-200

Forbes D (1992) Structure and function of the nuclear pore
complex. Ann. Rev. Cell Biol. 8: 495-527

Garcia-Bustos J, Heitman J, Hall MN (1991) Nuclear protein
localization. Biochim. Biophys. Acta. 1071: 83-101

Goldberg MW, Allen TD (1992) High resolution scanning
electron microscopy of the nuclear envelope:
demonstration of a new, regular, fibrous lattice attached
to the baskets of the nucleoplasmic face of the nuclear

pores. J. Cell. Biol. 119: 1429-1440

Govind S, Steward R (1991) Dorsoventral pattern formation in
Drosophila. Trends In Gen. 7: 119-125

30

Guiochon-Mantel A, Delabre K, Lescop P, Milgrom E (1994)
Nuclear localization signals also mediate the outward

movement of proteins from the nucleus. Proc. Natl.
Acad. Sci. USA 91: 7179-7183

Hall MN, Craik C, Hiraoka Y (1990) Homeodomain of yeast
repressor 022 contains a nuclear localization signal. Proc.
Natl. Acad. Sci. 87: 6954-6958

Hall MN, Hereford L, Herskowitz I (1984) Targeting of E.coli B-
galactosidase to the nucleus in yeast. Cell 36: 1057-1065

 

Hick GR, Raikhel NV (1993) Specific binding of nuclear
localization sequences to plant nuclei. Plant Cell 5: 983-
994

Howard EA, Zupan JR, Citovsky V, Zambryski PC (1992)
The VirD2 protein of A. tumerfaciens contains a C-
terminal bipartite nuclear localization signal:

implications for nuclear uptake of DNA in plant cells.
Cell 68: 109-118

Isoda K, Roth S, Nusslein-Volhard C (1992) The functional
domains of the Drosophila morphogen dorsal: evidence
from the analysis of mutants. Genes and Dev. 6: 619-630

Jarnik M, Aebi U (1991) Toward a more complete 3-D
structure of the nuclear pore complex. J. Structural
Biol. 107: 291-308

Kalderon D, Richardson WD, Markham AF, Smith AE (1984a)
Sequence requirements for nuclear location of simian virus 40
large T antigen. Nature 311: 33-38

Kalderon D, Roberts BL, Richardson WD, Smith AE (1984b) A
short amino acid sequence able to specify nuclear location.
Cell 39: 499-509

31

Kang KI, Devin J, Cadepond F, J ibard N, Guiochon-Mantei
A, Baulieu E, Catelli M (1994) In vitro functional
protein-protein interaction: nuclear targeted hsp90 shifts

cytoplasmic steroid receptor mutants into the nucleus.
Proc. Natl. Acad. Sci. 91: 340-344

Kiefer P, Acland P, Pappin D, Peters G, Dickson C (1994)
Competition between nuclear localization and secretory

signals determines the subcellular fate of a single CUG-
initiated form of FGF3. EMBO J. 13 (17): 4126-4136

Lanford RE, Butel JS (1984) Construction and characterization of
an SV40 mutant defective in nuclear transport of T antigen.
Cell 37: 801-813

Lanford RE, Feldherr CM, White RG, Dunham RG, Kanda P
(1990) Comparison of diverse transport signals in

synthetic peptide-induced nuclear transport. Exp. Cell
Res. 186: 32-38

Lee BA, Maher DW, Hannink M, Donoghue DJ (1987)
Identification of a signal for nuclear targeting in platelet-

derived-growth-factor-related molecules. Mol. Cell. Biol.
7: 3527-3537

Markland W, Smith AE, Roberts BL (1987) Signal-dependent
translocation of simian virus 40 large-T antigen into rat

liver nuclei in a cell-free system. Mol. Cell. Biol. 7:
4255-4265

Masiakowski T, Shooter EM (1988) Nerve growth factor
induces the genes for two proteins related to a family of

calcium-binding proteins in PC12 cells. Proc. Natl. Acad.
Sci. 85: 1277-1281

Melchior F, Paschal B, Evans J, Gerace L (1993) Inhibition of
nuclear protein import by nonhydrolyzable analogues of
GTP and identification of the small GTPase Ran/TC4 as
an essential transport factor. J. Cell Biol. 123: 1649-1659

32

Moll T, Tebb G, Surana U, Robitsch H, Nasmyth K (1991)
The role of phosphorylation and the CDC28 protein
kinase in cell cycle-regulated nuclear import of the S.
cerevisiae transcription factor SW15. Cell 66: 743-758

Moore MS, Blobel G (1992) The two steps of nuclear import,
targeting to the nuclear envelope and translocation
through the nuclear pore, require different cytosolic

factors. Cell 69: 939-950

Moore MS, Blobel G (1993) The GTP-binding protein
Ran/TC4 is required for protein import into the nucleus.
Nature 365 661-663

Moore MS, Blobel G (1994) Purification of a Ran-interacting
protein that is required for protein import into the
nucleus. Proc. Natl. Acad. Sci. USA 91: 10212-10216

Nagai R, Larson DM, Persiasamy M (1988) Characterization
of a mammalian smooth muscle myosin heavy chain

cDNA clone and its expression in various smooth muscle
types. Proc. Natl. Acad. Sci. 85: 1047-1051

Nasmyth K, Adolf G. Lydall D, Seddon A (1990) The
identification of a second cell cycle control on the HO

promoter in yeast: cell cycle regulation of SW15 nuclear
entry. Cell 62: 631-647

Nelson M, Silver P (1989) Context affects nuclear protein
localization in Saccharomyces cerevisiae. Molec. and Cell.
Biol. 9: 384.389

Newmeyer DD, Forbes DJ (1988) Nuclear import can be
separated into distinct steps in vitro: nuclear pore
binding and translocation . Cell 52: 641-653

33

Newmeyer DD, Forbes DJ (1990) An N-ethylmaleimide-
sensitive cytosolic factor necessary for nuclear protein

import: requirement in signal-mediated binding to the
nuclear pore. J. Cell. Biol. 110: 547-557

Ou JH, Yeh CT, Yen TS (1989) Transport of hepatitis B virus
precore protein into the nucleus after cleavage of its
signal peptide. J. Virol. 63: 5238-5243

Paine PL, Moore LC, Horowitz SB (1975) Nuclear envelope
permeability. Nature 254: 109-114

Picard D, Yamamoto KR (1987) Two signals mediate hormone-
dependent nuclear localization of the glucocorticoid
receptor. E.M.B.O. J. 6: 3333-3340

Raikhel NV (1992) Transport of proteins to the nucleus. Plant Phys.
100: 1627-1632

Restrepo-Hartwig MA, Carrington JC (1992) Regulation of
nuclear transport of a plant potyvirus protein by
autoproteoiysis. J. Virol. 66: 5662-5666

Richardson WD, Mills AD, Dilworth SM, Laskey RA,
Dingwall C (1988) Nuclear protein migration involves
two steps: rapid binding at the nuclear envelope followed

by slower translocation through nuclear pores. Cell 52:
655-664

Rihs H, Jans DA, Fan H, Peters R (1991) The rate of nuclear
cytoplasmic protein transport is determined by the casein

kinase 11 site flanking the nuclear localization sequence
of the SV40 T-antigen. EMBO J. 10: 663-661

Robbins J, Dilworth SM, Laskey RA, Dingwall C (1991) Two
interdependent basic domains in nucleoplasmin nuclear
targeting sequence: Identification of a class of bipartite nuclear
targeting sequence. Cell 64: 615-623

34

Roberts BL, Richardson WD, Smith AE (1987) The effect of
protein context on nuclear location signal function. Cell
50: 465-475

Rout MP, Wente SR (1994) Pores for thought: nuclear pore
complex proteins. Trends in Cell Biol. 4: 357-365

Shieh MW, Wessler SR, Raikhel NV (1993) Nuclear targeting
of the maize R protein requires two nuclear localization
sequences. Plant Physiol. 101: 353-361

Sukegawa J, Blobel G (1993) A nuclear pore complex protein
that contains zinc finger motifs, binds DNA, and faces
the nucleoplasm. Cell 72: 29-38

Tinland B, Koukolikova-Nicola Z, Hall MN, Hohn B (1992)
The T-DNA linked VirD2 protein contains two distinct

functional nuclear localization signals. Proc. Natl. Acad.
Sci. USA 89: 7442-7446

van der Krol AR, Chua N-H (1991) The basic domain of plant
B-ZIP proteins facilitates import of a reporter protein
into plant nuclei. Plant Cell 3:667-675

Varagona MJ, Schmidt RJ, Raikhel NV (1992) Nuclear localization
signal(s) required for nuclear targeting of the maize regulatory
protein, Opaque-2. Plant Cell 4: 1213-1227

Whalen AM, Steward R (1993) Dissociation of the dorsal-
cactus complex and phosphorylation of the dorsal protein

correlate with the nuclear localization of dorsal. J. Cell.
Biol. 123: 523-534

Zhou A, Hassel BA, Silverman RH (1993) Expression cloning
of 2-5-A-dependent RNAse: a uniquely regulated
mediator of interferon action. Cell 72: 753-765

35

Zhou Z, Sar M, Simental JA, Lane MV, Wilson EM (1994) A
ligand-dependent bipartite nuclear targeting signal in the

human androgen receptor. J. Biol. Chem. 269 (18):
13115-13123

CHAPTER 2

NUCLEAR TARGETING OF THE MAIZE R

PROTEIN REQUIRES TWO NUCLEAR

LOCALIZATION SEQUENCES

Reference: Shieh M.W., Wessler S.R., Raikhel N.V. (1993) Plant

Physiol. 101: 353-361

36

37
ABSTRACT

Previous genetic and structural evidence indicates that the maize
R gene encodes a nuclear transcriptional activating factor. In-frame
carboxy- and amino-terminal fusions of the R gene to the reporter gene
B-glucuronidase (GUS) were sufficient to direct GUS to the nucleus of
transiently transformed onion epidermal cells. Further analysis of
chimeric constructs containing regions of the R gene fused to the GL8
cDN A revealed three specific nuclear localization sequences (NLSs) that
were capable of redirecting the GUS protein to the nucleus. Amino-
terminal NLS A (a.a. 100-109, GDRRAAPARP) contained several
arginine residues, a similar localization signal is found in only a few
viral proteins. The medial NLS M (a.a. 419-428, MSERKRREKL)
is an SV40-type NLS, and the carboxyl-terminal NLS C (a.a. 598-610,
MISEALRKAIGKR) is a MAT a2 type. NLSs M and C are
independently sufficient to direct the GUS protein to the nucleus when
fused at the amino-terminus of GUS, while NLS A fused to GUS
partitioned between the nucleus and cytoplasm. Similar partitioning
was observed when localization signals NLS-A and NLS-C were

independently fused to the carboxy-terminal portion of GUS. A

38

deletion analysis of the three localization signals indicated that the
amino-terminal and carboxyl terminal fusions of R and GUS were
redirected to the nucleus only when NLSs A and M, or C and M, were
both present. These results indicate that multiple localization signals
are necessary for nuclear targeting of this protein. The conservation
of the localization signals within the alleles of R and similar proteins

from other organisms are also discussed.

39
INTRODUCTION

In eukaryotic cells, proteins can be targeted to a variety of
subcellular compartments such as the end0plasmic reticulum,
mitochondrion, chloroplast, peroxisome, glyoxisome, or nucleus. The
import of proteins into the nucleus, which has been examined
extensively in mammalian, amphibian and yeast systems, can be
distinguished from transport into other organelles because proteins and
small molecules traverse the nuclear envelope through a
macromolecular complex known as the nuclear pore (for review, see
Nigg et. al., 1991; Wagner et. al., 1990). The nuclear pore complex
forms a large aqueous channel across the nuclear membrane that
allows diffusion of small molecules, yet tightly regulates the movement
of larger molecules (for review, see Dingwall and Laskey, 1986;
Newmeyer et. al., 1986). Unlike the amino-terminal signal sequences
that direct proteins from the cytoplasm to the endoplasmic reticulum,
mitochondrion, and chloroplast, the import of nuclear proteins is
mediated by nuclear localization sequences (NLSs) that may be located
at any position within a protein (Garcia-Bustos et. al., 1991). In
addition, NLSs are not proteolytically cleaved from the protein, which

allows nuclear proteins to re-enter the nucleus after cell division.

40

There is no consensus sequence for NLSs, however they are
characterized as short amino acid regions that are rich in basic
residues (Garcia-Bustos et. al., 1991). The known NLSs can be
categorized into three classes based upon their composition and
structure: the SV40 large T type antigen (Kalderon et. al., 1984a,b;
Lanford and Butel, 1984), MAT (12 (Hall et. al., 1984) and bipartite
signal structure (nucleoplasmin; Dingwall and Laskey, 1991).
Recently, several NLSs have been identified in plantsand these are
similar to the mammalian and yeast NLSs (see Raikhel, 1992 for
review).

For our localization studies in higher plants, we have chosen to
utilize the maize R protein. Prior genetic analysis indicates that R
protein controls where and when the anthocyanin biosynthetic pathway
is expressed in plant tissues (Ludwig et. al., 1990). Consistent with a
proposed regulatory role was the finding that the R gene encodes a
protein with the structural features of a transcriptional activator
including large acidic and basic regions and a basic helix-loop-helix
domain (Ludwig et. al., 1989). As a transcriptional activator, the R
protein should localize to the nucleus. However, the predicted

molecular mass of the R protein is 66 kDa which exceeds the size limit

41

for the diffusion of gold particles through the nuclear pore complex
(Paine et. al., 1975). Thus, the R protein is a reasonable choice for the
study of nuclear protein import in higher plants, since it should possess
at least one NLS.

The goal of this study was to identify NLSs in the R protein and
to determine whether or not they were sufficient and necessary for
nuclear transport. To facilitate the localization of the protein within
plant cells, the reporter gene GL8, was fused to the cDNA of an allele
of the R gene called Q (leaf color). The gene fusions were transiently
expressed in onion epidermal cells following introduction of the DNA
by particle bombardment. Using this system, three NLSs were
identified in the maize R protein. We have also determined that at
least two of the NLSs are necessary and sufficient to target the R:GUS
fusion protein to the nucleus in onion cells. These results may be of
broad significance, since they constitute the first reported instance
where multiple NLSs are required for competent transport of a plant

regulatory protein.

42
MATERIALS AND METHODS

Materials

The white onions were purchased locally, stored at 4°C in the
dark and used within two weeks. Oligonucleotides were synthesized by
the MSU Macromolecular Facility (#1-3, #9-13) or by CIBA-GEIGY
Biotechnology (#4-8, Research Triangle Park, NC). The enzymes used
in the restriction digests were purchased from Boehringer Mannheim
Biochemicals (Indianapolis, IN) and enzymes used for other molecular
manipulations were purchased from New England Biolabs (Beverly,
MA). The supplies for the helium biolistic gun transformation system

(Dupont, Wilmington, DE) were from Bio-Rad (Richmond, CA).

Constructs

All standard recombinant DNA protocols were obtained from
Sambrook et. al. (1989). The protocol for site-directed mutagenesis
was performed as described by Kunkel et. al. (1987). After
mutagenesis constructs were sequenced to verify their integrity and
completed constructs were subcloned into the expression vector

pGA643 (An et. al., 1988), except for the R:GUS 598-610 construct

43

which was ligated into the pMF 6 expression vector (Goff et. al., 1990).
Expression vectors pGA643 and pMF6 expressed the gene fusions at
the same relative level as determined by histochemistry (data not
shown). The allele of the R gene used in this study was Lg (leaf color;

Ludwig et. al., 1989).

R:GUR-A SacI restriction site was inserted before the stop codon (nt
1830) and a Smal restriction site was inserted after the stop codon of
the Le cDNA by site-directed mutagenesis. The GRS cDNA (pB1101.3,
Jefferson et. al., 1987) was then subcloned in front of the stop codon

of the Q cDNA.

GILS:_R -_L_c cDNA was modified to include a XhoI and Smal site in
frame before the first initiating AUG codon by site-directed
mutagenesis. Also, a XhoI restriction site was inserted in frame in
front of the stop codon in GUS (nt 1807) by site-directed mutagenesis.

The modified GUS gene was then subcloned in front of the Q cDNA.

R:GUS 82-610, 598-610 - Restriction enzymes BglII and SacI were

 

used to construct the R:GUS gene fusions encoding a.a. 82-610 and

44
598-610 from the R:GUS construct.

GUS:R 1-109 -Restriction enzymes NaeI and Seal were used to

construct GUS:R 1-109 from the GUS:R construct.

 

 

To facilitate subcloning of the Re cDNA deletion constructs a set
of restriction enzyme digest sites encoding a KpnI site followed by an
ATG and Xhol site and were introduced into both R:_G_U_S and _G_I_J_S:_R.
By adding this set of restriction sites at positions before nts encoding
a.a. 411 (nt 1231), 457 (nt 1368) and 512 (nt 1533) it was possible to
subclone the fragments as a KpnI and EcoRI (R:GUS) fragments,
thereby making the constructs R:_(_}U_S 411-610, R_:G_ILS_ 457-610 and

R:GUS 512-610. To construct GUS:R 1-411, GUS:R 1-457 and GUS:R

 

 

1-512 the same set of restriction sites were added. However these
_G_U§:R constructs were subcloned as Xbal (5’ of GUS) and KpnI
fragments into a Xbal and KpnI site which had a stop codon in frame
after the KpnI site.

When two sets of the restrictions sites (Xhol-ATG-Kpnl) were
added at nts encoding a.a. 411 and 457 (1231 and 1368) or 411 and 512

(nts 1231 and 1533), they allowed the isolation and cloning of a Xhol

45
(5’) to Kpnl (3’) fragment. To subclone the R Xhol (5’) to Kpnl (3’)

fragment to GUS, an additional Kpnl sites was added to either R:GUS

at nt 1831 of R or GUS:R at nt 1831 of R_. In subcloning, the R gene

 

is removed before the fragments encoding a.a. 411-457 or 411-512 are

inserted.

R:GUS 128-411 -Utilizing the construct R:GUS 1-411, a NaeI and Kpnl
fragment (encoding a.a. 128-411) was inserted into a Smal and Kpnl
(Kpnl site added by site-directed mutagenesis at nt 1831, the Smal and

Kpnl cut drops out the R gene) cut R:GUS construct.

 

R:GUS 1-109 and 82-109 -R:GUS, with the additional Kpnl site at nt

 

1831, was cut with NaeI and Kpnl restriction enzymes. T4 DNA
polymerase was then used to make blunt ends which were ligated
(R:GUS 1-109). The R:_G_US 1-109 construct was then restriction
digested with BglII (leaving the first ATG at nt 246, a.a. 82) and
EcoRI to and ligated into pUCll8 into a BamHI and EcoRI site to

construct R:GUS 82-109

 

To construct the R:GUS and GUS:R fusions which encoded a.a.

 

46
100-109, a Kpnl site was introduced at nt 300 of R in the R:GUS 1-109

construct encoding nts 1-327. Then, the fragment encoding a.a. 100-
109 was subcloned into pUC118 as a Kpnl and EcoRI fragment. The
constructs encoding a.a. 419-428 were constructed by adding a Kpnl
site after the codon for a.a. 428 (nt 1284) to the m 411-457 and
M 411-457 constructs. The nucleotides encoding a.a. 429-457
were then excised from the clone as a Kpnl fragment.

The deletion constructs outlined in Figure 2.4 were also
constructed by using site-directed mutagenesis on the IgG—US and
M constructs. When each NLS encoding region was deleted a
specific restriction site was inserted or created for confirmation that
the sequence was deleted. The site-directed mutagenesis removed nts
300-327 for NLS-A (NaeI), nts 1257-1284 for NLS-M (Ach), or nts
1794-1830 for NLS-C (Kpnl). By utilizing three, two or one of the
deletion mutations in a single construct, the different combinations of

NLSs could be deleted.

Transformation of Onion Cells
Onion epidermal layers were placed inside up on a petri plate

containing MS basal media [per liter; 4.2 gm MS salts (Gibco-BRL,

47

Gaitherburg, MD), 1 mg thiamine, 10 mg myo-inositol, 180 mg
KHZPO4 (Miller I), 30 gm sucrose, pH 5.7] (Murashige and Skoog,
1962) with the antifungal agent amphotericin B (2.5 mg/L; Sigma, St.
Louis, MO) and 6% agar. Plasmid DNAs were prepared using either

CsCl2 gradient purification (Sambrook et. al., 1989) or column

purification (Qiagen, Chatsworth, CA). The plasmids (2.5 pg) were

precipitated onto 1.6 pm gold particles (1.25 pg) as described by the

manufacturer (Dupont, Wilmington, DE). DNA-coated particles were

washed with 180 [I] of 100% ethanol and then resuspended in 30 ul of

100% ethanol. Vortexing and then sonication (cup horn probe, 60%

power, 5 s) were used to resuspend the particles before loading 10

[Ll/disc (3 times) of the suspension onto particle delivery discs. Petri

plates of onion epidermal cells were transformed with the three particle
delivery discs (two discs on one plate and one disc on another plate) via
the helium biolistic gene transformation system. Rupture discs of 1300
PSI were optimal for onion cell transformation. Transformed cells

were incubated at 28°C in the dark for 24 or 48h.

48

Histochemical Analysis

The colorimetric substrate X-gluc was used to determine the
location of the enzymatic activity of the R:GUS and GUS:R fusion
proteins. The protocol for the addition of substrate to the onion cells
was described in Varagona et. al. (1992). The DNA specific nuclear
stain DAPI was included in the mounting solution for each sample
(V aragona et. al., 1991). Intracellular localization of the blue
precipitate was determined using a Zeiss Axiophot microscope with
Nomarski optics. Location of the blue precipitate was compared with
the location of DAPI stained nuclei using fluorescence optics. The
subcellular localization of each fusion protein was determined from two
to four separate transformations. The minimum number of cells

analyzed for each construct was three and the maximum was thirty.

RESULTS

The R Protein Redirects GUS to the Nucleus

To determine whether the R protein is imported into the nucleus,

the R (Re allele) and GUS cDNAs were ligated to form a gene fusion.

49

Since an active GUS enzyme might sterically alter the R protein, the
coding region of R was ligated both 5’ and 3’ of the GU_S gene to
increase the probability that putative targeting signals would be
properly exposed for recognition by the nuclear targeting apparatus.
The fusion constructs were then ligated into the expression vector
pGA643 between the CaMV 358 promoter and the NOS terminator
sequences. The constructs w (R 5’ of _G_U_S_), GU_S$ Q 3’ of
GQS) and GUS were then transformed into a monolayer of onion
epidermal cells by particle gun bombardment. Subcellular localization
of the fusion proteins was determined with the histochemical substrate
X-gluc which, when processed by GUS, forms a blue precipitate.
When the GUS protein was expressed in onion cells, the blue dye
remained in the cytoplasm (results not shown, Varagona et. al., 1992).

However, when R is fused to GUS (R:GUS or GUS:R fusion

 

constructs) and expressed in onion cells, GUS activity was redirected

 

to the nucleus (Fig. 2.1A and B, respectively). The conclusion from
these experiments was that the R protein was sufficient to redirect the
reporter protein GUS to the nucleus indicating that the R protein

contained at least one NLS.

50

 

figure 2.1. Histochemical localization of R:GUS (A) and GUS:R (B)
fusion proteins in onion epidermal cells. Tissues were simultaneously
analyzed using both X-gluc histochemical staining (A and B) and
nuclei-specific DAPI staining (A1 B‘). Nomarski optics were used in A

and B and fluorescence optics in Al and B‘. Bars= 10 um.

51
R Protein Contains Three NLSs

The strategy used to identify the NLSs in the R protein was to
construct gene fusions in which coding regions from either the 5’ or
3’end of the R gene were deleted (Fig. 2.2A,B). Thus, putative NLSs
could be identified by the process of elimination. Initially, the R:GUR
construct was modified with deletions at the 5’ terminus (Fig. 2.2A),

and the GUS:R construct was modified with 3’ deletions (Fig. 2.2B).

 

In addition, constructs were specifically designed around a.a. 411-457
because this region is enriched in basic amino acids, characteristic of
NLSs, and contains the helix-loop-helix region (a.a. 420-462; Ludwig
et. al., 1989). Upon completion, the constructs were ligated into
expression vectors as described in Material and Methods.

The R_:_GU_S deletion constructs were expressed in onion
epidermal cells and the subcellular locations of the resulting proteins
were determined by assaying for GUS activity (Fig. 2.2A). The series
of deletions from the amino-terminus contained a.a. 82-610, 411—610,
457-610, 512-610, and 598-610 and revealed NLS-C (NLS in the
carboxyl-terminus). The thirteen amino acids encoded at position 598-
610 (NLS-C) of R were sufficient to redirect GUS to the nucleus (Fig.

2.3). The localization of GUS by NLS-C was exclusively to the nucleus

52

Figure 2.2. Cloning strategy for preparing R:GUS (A) and GUS:R (B)
fusions and results of localization experiments. The upper construct
in Figure 2.2A represents the amino-terminal fusion of coding
sequences of the R cDNA clone (open box) and the GUS cDNA clone
(wavy lined box), and in Figure 2.2B, the carboxyl-terminal fusion of
QUS cDNA clone to the R cDNA clone. The position of first and last
deduced amino acids in the R cDNA clone are indicated above the
constructs in 2.2A and 2.2B. The amino acids of R protein used to
prepare amino- (2.2A) and carboxyl- (2.2B) terminal fusions to GUS
are indicated on the left. The results of subcellular localizations

determined by histochemical assays for GUS activity are indicated on

the right.

2A

1-‘10
.2-‘10
‘11-‘10
‘57-‘10
512-‘10
‘11-512
‘11-‘57
1-109
D2-109

120-‘11

008
100-109
‘19-‘20

590-610

28

1-‘10
1-512
1-‘57
1-‘11
1-10’
‘11-512
‘11-‘57
008
100-109

‘19-‘28

53

‘10

  

 

  

&

 

 

 

 

 

 

 

 

 

 

iiififiili. iiiiiéffiifizlii’fifiifiifiiiééifiififilfifiififfiffiiifif:

iffiffiifiifitfiifiifiifiifiifififiififififiéfimfii

 

 

 

 

  
 

H/C

l/C

H/C

H/C

U/C

54

figure 2.3. Histochemical localization of three NLS regions of the R
protein fused to GUS (above) and schematic representation of R:GUS
fusion protein showing localization of three NLSs (below). Positions of
amino acids are indicated above the construct; the acidic domain of R
protein (striped box), helix-loop-helix domain (stippled box) and three
NLSs (NLS-A -orange circle, NLS-M -yellow circle and NLS-C -green
circle) are indicated. Amino acid sequences of three NLSs of the R
protein are shown under corresponding photomicrographs. Tissues
were stained using X-gluc histochemical staining and analyzed with the
Nomarski optics. ‘The brown particles on pictures with NLS-A and

NLS-M result from gold precipitation. Bars: 10 um.

55

wall

own so

 

 

 

56

and exhibited subcellular localization similar to the intact R protein
fused to GUS (Fig. 2.1A).

The deletion constructs were also used to examine the amino-
terminus of the GUS:R fusions (Fig. 2.2B). The series of deletions
from the carboxyl-terminus contained a.a. 1-512, 1-457, 1-410 and 1-
109 and revealed NLS-A (NLS in the amino-terminus). NLS-A was
further defined by constructs containing a.a. 82-109 and 100-109 (Fig.
2.2A,B). GUS activity of the fusion protein NLS-A+ GUS (a.a. 100-
109) partitioned between the nucleus and cytoplasm (Fig. 2.3).

Since the R:GUS and GUS:R deletion constructs described could

 

not distinguish any NLSs in the region of a.a. 109-598, a second set of
constructs was designed (Fig. 2.2A,B). . The central region of the R
protein (a.a. 109-598) was subdivided into constructs containing the
basic helix-loop-helix motif (a.a. 411-457 and a.a. 419-428) and the
non-basic residue rich region (a.a. 128-411). Amino acids 128-411 were
unable to redirect GUS to the nucleus and remained in the cytoplasm
(Fig. 2.2A); this was not analyzed in the GUS-R orientation. However,
a.a. 411-512, a.a. 411-457 and 419-428 (NLS-M) were sufficient to
redirect GUS to the nucleus (Fig. 2.2A, and 2.3). NLS-M was located

in the amino-terminus of the helix-loop-helix motif and, unlike NLS-A,

57

was as efficient as NLS-C in localizing GUS activity exclusively to the
nucleus. The GUS-NLS-M fusion protein (GUS:R orientation) resulted
in GUS activity partitioned in the cytoplasm and nucleus (Fig. 2.2B).
Therefore, for this study, the amino-terminal GUS fusions displayed
stronger redirection of GUS activity to the nucleus. In conclusion, the
R protein contained three NLSs (A, M, C) each of which were
sufficient to redirect the reporter protein GUS to the. nucleus of onion

epidermal cells (Fig. 2.3).

Two NLSs are Necessary for Transport of R:GUS to the Nucleus

 

The identification of three NLSs in the R protein that were
sufficient to redirect the GUS reporter protein to the nucleus prompted
our investigation of the role of these NLSs in the full-length protein.
To determine which NLSs were functional and necessary for the import
of intact R protein, site-directed mutagenesis was used to delete the

NLSs from the fusion constructs of R:GUS and GUS:R. This strategy

 

 

resulted in either none, one or two NLSs in the R protein (Fig. 2.4).
The constructs were then subcloned into expression vectors and
transiently expressed in onion epidermal cells, as in the previous

experiments.

58

Figure 2.4. Effect of deletion of different NLSs on the histochemical
localization of R:GUS fusion proteins. Deletion of different NLSs from
the intact R protein fused to GUS showed that NLS A and M or M and
C are required for nuclear targeting. Several examples of the
histochemical localizations for R-GUS fusion proteins are shown. The
main features of the R protein are the same as in Figure 2.3, except
intact R protein was fused to GUS with deletions of specific NLSs. (1)
R. protein containing NLS-A (orange circle) and NLS-M (yellow circle)
is indicated. (2) R protein containing NLS-A and NLS-C (green circle)
is indicated. (3) R protein containing only NLS-M. (4) All three NLSs
deleted from R protein. Tissues were simultaneously analyzed using
both X-gluc histochemical staining (1-4) and nuclei-specific DAPI
staining (ll-4‘). Tissues were stained and analyzed as in Figure 2.1.

Bar = 10 pm.

59

 

60
When all three NLSs (A,M,C) were deleted from R:GUS and

GUS:R fusion proteins, GUS activity was retained within the cytoplasm
[Fig. 2.4 (4)]. This indicated that all NLSs in the R protein
wereidentified. However there is the formal possibility exists that the
deletion of the NLSs could sterically hinder an unidentified signal.
These results also showed that the strongest determinants of each
targeting signal were within the identified NLSs.

To determine whether or not any single NLS was capable of
targeting the fusion protein, two of the three NLSs were deleted from
R:GUS and GUS:R fusion proteins in each of three possible
combinations (Fig. 2.5). NLS-A, in the intact R protein, was able to
function as an NLS but it was inefficient as a signal and resulted in
GUS activity in both the nucleus and cytoplasm (Fig. 2.5). Therefore,
in both the intact R:GUS and NLS-A:GUS protein, NLS-A was an
inefficient NLS (Figs. 2.3 and 2.5). Both NLS-M and NLS-C also
retained their functions as NLSs and conferred partitioned localization.
However, their expression in the nucleus was visually greater than in
the cytoplasm [Figs. 2.4(2.3) and 2.5]. Despite the ability of the
polypeptide encoding sequences for NLS-M and NLS-C to redirect GUS

activity to the nucleus, when those signals were present in the R

61

Am... NLS Mum, NLS cm“, NLS Locuzzanon
_ R-GUS
- - c
t - — N/C
- + - N/C
- + file
+ + - N
+ - + N/C
- + + :1

Figure 2.5. Summary of histochemical analysis of R:GUS fusion

proteins which identified NLSs that were necessary for nuclear

localization .

62
protein minus the two other NLSs, NLS-M and NLS-C were incapable

of conferring exclusive nuclear localization to the R:GUS fusion
protein. The constructs which retained two of the three NLSs
displayed different subcellular locations depending upon the orientation
of R protein to GUS [Figs. 2.4(1),2.2 and 2.5]. Since the R:GUS
fusions exhibited stronger nuclear localization than the GUS:R fusions
(Fig. 2.5), conclusions were drawn from the R:GUS fusion proteins.
If NLS-A (Fig. 2.5) or NLS-C [Figs. 2.4 (1) and 2.5] were deleted, the
fusion protein localized to the nucleus (Fig. 2.5). Therefore, either
combinations of NLS-A and NLS-M, or NLS-C and NLS-M were
sufficient for nuclear localization. However, if NLS-M was deleted, the
fusion protein partitioned between the nucleus and cytoplasm [Fig.
2.4(2)]. Our conclusion from this data was that two NLSs, one of
which must be NLS-M, were sufficient and necessary for the strong

localization of R:GUS protein to the nucleus.

63
DISCUSSION

To identify the NLSs of the maize R protein, a transient
expression system was developed utilizing onion cells. Onion epidermal
cells were used because their large size facilitated subcellular
localization and provided a useful transformation system for particle
gun bombardment (Klein et. al., 1987). Furthermore, the results of
subcellular localization in onion cells were shown to correlate with the
localizations determined by stable transformation of Opaque2-GUS
fusion proteins in tobacco plants (V aragona et. al., 1992). In that
study, cellular fractionation and histochemical analysis of the
transgenic tobacco cells was used to determine the location of the
fusion proteins. It was shown that the subcellular locations of the GUS
enzymatic activities correlated with those determined by the transient
expression assays in onion epidermal cells. Therefore, transformation
of onion cells by particle bombardment is a rapid and efficient system
for studying nuclear localization.

The full length R protein fused to GUS yielded efficient nuclear
localization in both amino and carboxyl-terminal orientations.

However, only amino-terminal fusion proteins were efficiently

64

transported to the nucleus when smaller regions of the R protein were
fused to GUS, indicating that the position of the NLS in the
transported protein is important. A similar conclusion was drawn
when the bipartite NLS of Opaque2 protein was analyzed (V aragona
et. al., 1992).

Three nuclear localization signals were identified in the R protein
(NLS-A, M, C) utilizing the onion system. Two of the NLSs, NLS-M
(419-428) and NLS-C (598-610) are intact signals because they
redirected GUS activity exclusively to the nucleus (Fig. 2.3). The third
signal, NLS-A (100-109), partially redirected GUS to the nucleus,
partitioning the fusion protein between the nucleus and cytoplasm.
Since several larger constructs including NLS-A, encoding a.a. 82-109
and 1-109 (Fig. 2.2), also partially redirected GUS, this inefficient
targeting may be due to intrinsic weakness of the targeting signal or it
is possible that amino acids following a.a. 109 are part of the signal,
but this was not analyzed. The identification of the three NLSs was
confirmed when the gene fusion constructs containing the full-length R
protein with the three NLSs deleted (A,M,and C) were retained in the
cytoplasm. Deletion of two of the three NLSs revealed that not all

three NLSs were required for nuclear localization and that each signal

65

could function independently. However, localization of R-GUS or
GUS-R constructs containing individual NLSs was less efficient than
localization of constructs containing all three or two of the three
signals.

The NLSs of R were dissimilar in their amino acid composition
and may confer different specificities to the nuclear import machinery.
NLS-A had the most intriguing composition because it contained
arginines and no lysines. This is a characteristic of some viral NLSs.
Examples of viral proteins with NLSs containing no lysines are
influenza nucleoprotein and N81 (Davey et. al., 1985), adenovirus pTP
(Zhao and Padmanabhan, 1988) and human immunodeficiency virus
REV (Malim et. al., 1989). NLS-C was enriched in hydrophobic amino
acids that were interspersed within its basic residues. One of the few
NLSs which has a high content of hydrophobic amino acids is the yeast
Mata2 protein (KIPIK; Hall et. al., 1984) which is similar to NLS-C
(MISESLRKAIGKR).

NLS-M, located within the amino-terminus of the helix-loop-helix
homologous motif, contained more basic amino acids than NLSs A or
C, with five arginines and one lysine within the ten amino acid signal.

The high concentration of basic amino acids in NLS-M is similar to the

66
SV40 large T antigen NLS (Kalderon et. al., 1984a) in which five of

the seven amino acid signal are basic. Another transcription factor,
myoDl, which shares homology to the helix-loop-helix domain, also
contains an NLS in this motif. However, the NLS was defined to 34
a.a. of the helix-loop-helix and it is unknown if the NLS of myoDl is
in the amino-terminus (the first 10 a.a. of the 34 a.a. signal identified)
of the helix-loop-helix domain (Tapscott et. al., 1988). A comparison
of NLS-M to the amino terminus of other DNA binding helix-loop-helix
domains revealed a conserved region (Fig. 2.6). It is logical, in
evolutionary terms, to retain an NLS within an essential domain of a
transcriptional activator and it would be interesting to determine
whether the import and DNA-binding functions are separable.

Two NLSs were required for efficient transport of the R:GUS
fusion proteins. Combinations of NLS-A and NLS-M or NLS-M and
NLS-C conferred exclusive nuclear localization to the fusion proteins.
This requirement of two NLSs for efficient transport to the nucleus is
known to occur in other nuclear proteins and was proposed to be a
consensus structure, termed bipartite, by Dingwall and Laskey
(Dingwall and Laskey, 1991). Bipartite signals contain two regions

enriched in basic amino acids separated by more than 4 amino acids

Figure 2.6. Amino acid comparison of R-Lc to other homologous
regulatory proteins. Alignments are made to maximize homology with
the NLSs of R. Identical amino acids are marked by vertical lines and
the conservative substitutions by two dots. The sequences shown are
for: maize R-Lc (Ludwig et. al., 1989), maize R-S (Perrot and Cone,
1989), maize B-Peru (Radicella et. al., 1991), Antirrhinum DEL
(Goodrich et. al., 1992), L-myc (DePhino et. al. 1987), N-myc (Kohl et.
al. 1986), myogenin (Edmondson and Olson, 1989), CBF-l (Cai and
Davis, 1990) , AP-4 (Hu et. al., 1990), human E3 (Beckman et. al.,

1990), and human E47 (Voronova and Baltimore, 1990).

68

.nu>_>__u¢ 3....
a...___a__~.H an
.n.:_.__:u on:
.z___.._>u Can
._ua_au_¢a nan-mo»:
2:22.: 3....
_a“_____.... auxin
M: _a__;_> .________.. __e...a:..._ a...
_______u.___ ____._____ o>e¢_.__u as.-.
_ _ _

88>
_
_

IKOH‘KIA‘IQHS dlfllflllflm! tits—((BIOO Quiz

69
(Robins et. al., 1991). The NLSs of the R protein are bipartite, but

they do not fit the model proposed by Dingwall and Laskey (1991).
First, unlike the model signal in which both basic regions are required
for efficient targeting of a reporter protein to the nucleus, two NLSsof
the R protein, NLSs M and C, independently and efficiently redirected
the reporter protein GUS to the nucleus. Second, although two NLSs
are necessary for targeting of R-GUS protein to the nucleus, the
spacing between NLSs A, M, and C (at least 170 a.a.) is greater than
the spacing found in the NLSs examined by Dingwall and Laskey
(1991). Also, the potyviral protein Nla (Carrington et. al., 1991) has
a long spacer (32 a.a.) separating the two basic regions which are
involved in nuclear localization. Although the significance of NLS
repetition is not understood, this phenomenon has been reported in
many proteins: glucocorticoid steroid hormone receptor (Picard and
Yamamoto, 1987), Agobacterium VirE2 (Citovsky et. al., 1992) and
Ea mys 02 (V aragona et. al., 1992) are examples. One study
examined the effect of multiple NLSs upon the import of peptide-coated
gold particles (Dworetzky et. al., 1988). Increasing amounts of SV40
large T antigen NLS were covalently linked to coat gold particles,

which were microinjected into Xenopus laevis oocytes. The results

70

showed that larger diameter gold particles require several NLSs to
enter the nucleus.

To determine which amino acids of the NLSs might be important
for function, we searched for conserved amino acids in members of the
R gene family (g, R-_s_, Rm) and an R homolog from Antirrhinum
majus (DEL, Fig. 6). Two of the alleles (Ric, R-_S) of the R gene are
cloned and they are 95% homologous in their amino acids. Therefore,
the regions corresponding to R NLSs are equally conserved (Fig. 2.6).
However, the maize B (Radicella et. al., 1991) and Antirrhinum Del
(Goodrich et. al., 1991) proteins share 78% and 25% amino acid
homology to R-Lc and encode sequences similar to the NLSs of R R
g was used in this study). The greatest homology was retained for
NLS-M, which was associated with the helix-loop-helix domain (Fig.
2.6). NLS-A was the least conserved and was the weakest of the
signals that we have identified. Though NLS-C was not highly
conserved (Fig. 2.6), the presence of two lysines and overall
hydrophobicity content of the carboxyl-terminus are retained. Since
the R:GUS fusion protein required two NLSs for exclusive nuclear
localization, the conservation of NLS-M and NLS-C indicates that they

may be the NLSs utilized in the R protein.

71

NLS-M represents a second function for the helix-loop-helix
domain, to serve in both DNA binding and nuclear targeting. Since
NLS-M is absolutely necessary for efficient targeting of R and is also
the most conserved region among transcriptional activators carrying
helix-loop-helix motifs, the dual function of the R protein’s basic helix-
loop-helix may be conserved in other transcriptional activators with
helix-loop-helix domains. A similar hypothesis was proposed for the b-
ZIP proteins (V aragona et. al., 1992; Raikhel, 1992) and for steroid
hormone receptors which contain zinc-finger motifs (Picard and
Yamamoto, 1987).

Possible functions for the multiple NLSs of R could be to act as
developmentally regulated or tissue-specific signals. Recently, a
developmentally regulated NLS was identified in the adenovirus type
5 Ela protein (Standiford and Richter, 1992). Standiford and Richer
(1992) identified the second of two NLSs in Ela, termed drNLS, which
is not constitutively utilized as a signal for nuclear transport. It has
been shown using developing Xenopus oocytes that the drNLS alone
resulted in transport to the nucleus until oocytes reach the late gastrula
stage when the drNLS Ela protein is retained in the cytoplasm. None

of the R protein’s NLSs share homology to the drNLS of Ela.

72

However, the possibility exists that these multiple NLSs function at
different developmental stages. Another possibility is that multiple
NLSs are required in the R protein to regulate tissue-specific
expression, since different alleles of the R gene are expressed in
different tissues (Styles et. al., 1973; Coe, 1985). The Q allele used in
this study is expressed in a number of tissues including pericarp, ligule,
midribs, coleoptiles, anthers, silks, and brace roots; whereas another
allele of R, R111 is expressed only in the scutellum, coleoptiles and
brace roots. One proposal is that tissue specificity is regulated by
different promoters. However, it is possible that the NLSs of R may
function differentially, with each NLS providing different efficiencies
for transport in a tissue-specific manner.

The most striking feature of the different NLSs of the R protein
was their varying compositions. NLS-A contained no lysine residues,
a characteristic that has been observed only in viral proteins. NLS-M
possessed the greatest density of charged residues, with seven of the ten
amino acids being basic. NLS-C was enriched with hydrophobic
residues which also affect the charge density of the NLS. Though it is
not surprising that the compositions of the signals are different, as

NLSs lack a consensus sequence, it is obvious that the import

73

machinery has to recognize some general features of the NLSs.
Therefore, signals which are as divergent in charge and hydrophobicity
as those in the R protein could be useful in the identification of

different NLS binding proteins.

74
ACKNOWLEDGEMENTS

We would like to thank Drs. Glenn Hicks, Marguerite Varagona, ,
and Susannah Gal for many helpful discussions and critical reading of

this manuscript.

75

REFERENCE

An G, Ebert PR, Mitra A, HA SB (1988) Binary vectors. Plant
Mol Biol Man A3: 1-19

Beckmann H, Su LK, Kadesch T (1990) TFE3: a helix-loop-helix
protein that activates transcription through the immunoglobulin
enhancer uE3 motif. Genes Dev 4: 167-179

Cai M, Davis RW (1990) Yeast centromere binding protein CBFl, of
the helix-loop-helix protein family, is required for chromosome
stability and methionine prototropy. Cell 61: 437-446

Carrington J, Freed DD, Leinicke A (1991) Bipartite signal sequences
mediates nuclear translocation of the plant potyviral Nla Protein.
Plant Cell 3: 953-962

Citovsky V, Zupan J, Warnick D, Zambryski P (1992) Nuclear
localization of Agrobacterium VirE2 protein in plant cells.
Science 256: 1802-1805

Coe EH Jr (1985) Phenotypes in corn: control of pathways by alleles,
time and place. In Plant Genetics, UCLA Symposia on Molecular
and Cellular Biology, M. Freeling, ed (New York: Alan R Liss)

'Vol 35: 509-521

Davey J, Dimmock NJ, Colman A (1985) Identification of the sequence
responsible for the nuclear accumulation of the influenza virus
nucleoprotein in Xenopus oocytes. Cell 40: 667-675

DePhino RA, Hatton KS, Tesfaye A, Kohl NE, Yancopoulos GD, Alt
FW (1987) The human my; gene family: structure and activity
of L-myc and an L-mjg psuedogene. Genes Dev. 1:1311-1326

Dingwall C, Laskey RA (1986) Protein import into the cell nucleus.
Ann Rev Cell Biol 2: 367-390

76

Dingwall C, Laskey RA (1991) Nuclear Targeting sequences- a
consensus? TIBS 16: 478-481

Dworetsky SI, Lanford RE, Feldherr CM (1988) The effect of
variations in the number and sequence of targeting signals on
nuclear uptake. J Cell Biol 107: 1279-1287

Edmundson DG, Olson N (1989) A gene with homology to _n_1yc
similarity of MyoDl is expressed during myogenesis and is

sufficient to activate the muscle differentiation program. Genes
Dev 3: 628-640

Garcia-Bustos J, Heitman J, Hall MN (1991) Nuclear protein
localization. Biochim Biophys Acta 1071: 83-101

Goff SA, Klein TM, Roth BA, Fromm ME, Cone KC, Radicella
JP, Chandler VL (1990) Transactivation of anthocyanin

biosynthetic genes following transfer of R regulatory genes
into maize tissue. EMBO J 9: 2517-2801

Goodrich J, Carpenter R, Coen ES (1992) A common gene regulates
pigmentation pattern in diverse plant species. Cell 68: 955-964

Hall MN, Hereford L, Herskowitz I (1984) Targeting of E.coli B-
galactosidase to the nucleus in yeast. Cell 36: 1057-1065

Howard EA, Zupan JR, Citovsky V, Zambryski PC (1992) The VirD2
protein of A. tumefaciens contains a C-terminal bipartite nuclear

localization signal: Implications for nuclear uptake of DNA in
plant cells. Cell 68: 109-118

Hu Y-F, Luscher B, Admon A, Mermod N, Tijan R (1990)
Transcription factor AP-4 contains multiple dimerization
domains that regulate dimer specificity. Genes Dev 4: 1741-1752

Jefferson RA (1987) Assaying chimeric genes in plants: the GUS gene
fusion system. Plant Mol Biol Reporter 5: 387-405

77

Kalderon D, Richardson WD, Markham AF, Smith AE (1984a)
Sequence requirements for nuclear location of simian virus 40
large T antigen. Nature 311: 33-38

Kalderon D, Roberts BL, Richardson WD, Smith AE (1984b) A short
amino acid sequence able to specify nuclear location. Cell 39:
499-509

Klein, T.M., Wolf, E.D., Wu, R., and Sanford, J.C. (1987). High-
velocity microprojectiles for delivering nucleic acids into living
cells. Nature 327: 70-73.

Kohl NE, Legouy E, DePhino RA, Nisen PD, Smith RK Gee CE, Alt
FW (1986) Human N-m_y_c is closely related in organization and
nucleotide sequence to C-myg. Nature 319: 73-77

Kunkel, T. A., Roberts, J. D., and Zakour, R. A. (1987). Rapid and
efficient site-specific mutagenesis without phenotypic selection.
Methods Enzymol 154: 367-382.

Lanford RE, Butel JS (1984) Construction and characterization of an
SV40 mutant defective in nuclear transport of T antigen. Cell 37 :
801-813

Ludwig SR, Habera LF, Dellaporta SL, Wessler SR (1989) Lo, a
member of the maize R gene family responsible for tissue-specific
anthocyanin production, encodes a protein similar to
transcriptional activators and contains the myc-homology region.

Proc Natl Acad Sci USA 86: 7092-7096

Ludwig SR, Wessler SR (1990) Maize R gene family tissue-specific
helix-loop-helix proteins. Cell 62: 849-851

Malim MH, Biihnlein S, Hauber J, Cullen BR (1989) Functional
dissection of the HIV-rev trans-activator- derivation of a trans-
dominant repressor of rev function. Cell 58: 205-214

Murashige T, Skoog F (1962) A revised medium for rapid growth and
bio-assays with tobacco tissue culture. Physiol Plant 15: 473-497

78

Newmeyer DD, Finlay DR, Forbes DJ (1986) In Vitro transport of a
fluorescent nuclear protein and exclusion of non-nuclear
proteins. J Cell Biol 103(#1, ptl): 2091-2102

Nigg EA, Baeuerle PA, Luhrmann R (1991) Nuclear import-
export: in search of signals and mechanisms. Cell 66: 15-22

Paine PL, Moore LC, Horowitz SB (1975) Nuclear envelope
permeability. Nature 254: 109-114

Perrot GH, Cone KC (1989) Nucleotide sequence of the maize R-_S
gene. Nucl Acid Res 17: 8003

Picard D, Yamamoto KR (1987) Two signals mediate hormone-

dependent nuclear localization of the glucocorticoid receptor.
EMBO J 6:3333-3340

Radicella PJ, Turks D, Chandler VL (1991) Cloning and nucleotide
sequence of a cDNA encoding B—Peru, a regulatory protein of the
anthocyanin pathway of maize. Plant Mol. Biol. 17: 127-130

Raikhel NV (1992) Transport of proteins to the nucleus. Plant Phys,
100: 1627-1632

Robbins J, Dilworth SM, Laskey RA, Dingwall C (1991) Two
interdependent basic domains in nucleoplasmin nuclear targeting

sequence: Identification of a class of bipartite nuclear targeting
sequence. Cell 64: 615-623

Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). Molecular
Cloning: A Laboratory Manual, Ed 2 Cold Spring Harbor, Cold
Spring Harbor NY

Standiford DM, Richter JD (1992) Analysis of a developmentally
regulated nuclear localization signal in Xenopus. J. Cell Biol. vol
118 5: 991-1002

Styles DE, Ceska O, Seah K (1973) Developmental differences in action
of R and B alleles in maize. Can J Genet Cytol 15: 59-72

79

Tapscott SJ, Davis RL, Thayer MJ, Cheng P, Weintraub H, Lassar
AB (1988) MyoDI: A nuclear phosphoprotein requiring a Myc
homology region to convert fibroblasts to myoblasts. Science 242:
405-411

Varagona MJ, Schmidt RJ, Raikhel NV (1991) Monocot regulatory
protein Opaque-2 is localized in the nucleus of maize endosperm
and transformed tobacco plants. Plant Cell 3: 105-113

Varagona MJ, Schmidt RJ, Raikhel NV (1992) Nuclear localization
signal(s) required for nuclear targeting of the maize regulatory
protein, Opaque-2. Plant Cell 4: 1213-1227

Voronova A, Baltimore D (1990) Mutations that disrupt DNA binding
and dimer formation in the E47 helix-loop-helix protein map to
distinct domains. Proc Natl Acad Sci USA 87: 4722-4726

Wagner P, Knuz J, Koller A, Hall MN (1990) Active transport of
proteins into the nucleus. FEBS 275 1,2: 1-5

Zhao L, Padmanabhan R (1988) Nuclear transport of adenovirus DNA
polymerase is facilitated by interaction with preterminal protein.
Cell 55: 1005-1015

CHAPTER 3

CHARACTERIZATION OF THE

QARBOXY-TERMINAL NUCLEAR LOCALIZATION SEQUENCE
OF THE MAIZE R PROTEIN

80

81
INTRODUCTION

Eukaryotes contain nuclei which organize the genomic DNA and
separate the processes of transcription and translation. The separation
of transcription and translation by the nuclear envelope necessitates the
transport of transcription factors and other regulatory proteins from
their site of synthesis in the cytoplasm to the nucleoplasm. Therefore,
proteins to be transported to the nucleus must contain a targeting
signal, termed a nuclear localization sequence (NLS), recognizable to
the nuclear transport machinery. Translocation of proteins across the
nuclear envelope occurs at nuclear pore complexes which are aqueous
channels connecting the cytoplasm to the nucleoplasm. Though
diffusion of small molecules does occur (Paine et. al., 1975), the import
of most proteins to the nucleus is both energy and NLS dependent (for
review, see Forbes, 1992). NLSs vary from 7 to 40 amino acids in
length with no consensus sequence though they are enriched in basic
amino acids. Unlike targeting signals to other organelles, NLSs are
found in various locations within different nuclear proteins, are not
cleaved and can contain multiple targeting signals (for review, see

Garcia-Bustos et. al., 1991).

82

Based on amino acid composition and size, NLSs can be
separated into three groups: SV40 large T antigen-like (Kalderon et.
al., 1984a,b; Lanford and Butel, 1984), bipartite (nucleoplasmin;
Dingwall and Laskey, 1991) and Mata2-like (Hall et. al., 1984). SV40-
like signals are characterized as being 7-20 amino acids in length and
enriched in basic amino acids, whereas bipartite signals contain two
basic amino acid enriched regions separated by 10-30 amino acids.
Mata2-like NLSs contain several hydrophobic amino acids and a single
basic amino acid enriched region.

The Mata2 protein contains two NLSs, however, the MataZ-like
type of NLS is based on the targeting signal located in the amino
terminus of the protein (Hall et. al., 1984). The polypeptide motif
KIPIK in the amino terminal Mata2 NLS is conserved in several other
yeast nuclear proteins and is, therefore, considered to be a type of
NLSs, Mata2-like. However, there are no defined features for the
Mata2 NLS because it has not been thoroughly analyzed and no other
MataZ-like NLS has been identified. In addition, unlike other NLSs
which can function as targeting signals in animal, plant and fungal
systems, the yeast Mata2 NLS (amino terminal signal) does not

function as a targeting signal in animal cell lines (Chelsky et. al., 1989;

83
Lanford et. al., 1990). NLSs of other proteins, such as SV40 large T

antigen, have been shown to function in animal, yeast and plant cells
(Lanford and Butel, 1984; Nelson and Silver, 1989; Varagona and
Raikhel, 1994; van der Krol and Chua, 1991). It is unknown why the
Mata2 NLS does not function in animal systems.

Several NLSs have been identified in plants, similar to the animal
and yeast NLSs (see Raikhel, 1992 for review). Recently, we identified
three NLSs in the maize R protein, a transcriptional activator in the
anthocyanin biosynthesis pathway (Shieh et. al., 1993). The NLS -
located in the amino terminus of the R protein (NLS-A) is unusual in
its amino acid composition because it does not contain the basic amino
acid lysine. All eukaryotic NLSs identified contain the basic amino
acid lysine and the absence of lysines in NLSs is only found in some
viral NLSs. The NLS located in the middle of the R protein (NLS-M)
is adjacent to the helix-loop-helix motif and is a SV40-like NLS. The
carboxyl terminal NLS (NLS-C) of the R protein is similar to the
Mata2-like NLS because it contains several hydrophobic amino acids
and the basic amino acids are arranged in a similar pattern. Unlike
the well studied SV40-like NLSs (NLS-M) and bipartite NLSs, NLSs

which are similar to the Mata2-like NLSs have not been studied in

84

detail. Therefore, we choose to identify the essential amino acids
within NLS-C. To accomplish this, mutations in NLS-C were designed
to examine the role of the hydrophobic amino acids, basic amino acids,
NLS-C’s similarity to the yeast Mata2 NLS, and to determine if amino
acid context is important. The ability of the mutated NLSs to redirect
GUS to the nucleus was then assayed. Several of these mutations were
unable to redirect GUS activity to the nucleus. In addition, because
NLS-C is similar to the Mata2 NLS, we have made a Mata2 NLS f3-
glucuronidase (GUS) fusion construct and transiently expressed the
construct in onion epidermal cells to ascertain if the Mata2 NLS can

redirect the reporter protein to the nucleus.

MATERIALS AND NIETHODS

Materials

White onions were purchased locally, stored at 4° C in the dark
and used within two weeks. Oligonucleotides were synthesized by the
MSU Macromolecular Facility and enzymes for molecular

manipulations were obtained from Boehringer Mannheim Biochemicals

85
(Indianapolis, IN) and New England Biolabs (Beverly, MA). The

materials for the helium biolistic gun transformation system (Dupont,
Wilmington, DE) were from Bio-Rad (Richmond, CA).
Constructs

All standard recombinant DNA protocols were obtained from
Sambrook et. al. (1989). The protocol for site-directed mutagenesis
was performed as described by Kunkel et. al. (1987). After
mutagenesis constructs were sequenced to verify their integrity, and
completed constructs were subcloned into the expression vector pMF6
(Goff et. al., 1990).

The NLS-C:GUS construct was described in Shieh et. al., 1993.
These oligonucleotides were designed to mutagenize the NLS-C:GUS
construct to create the new sequence;

Mag; NLS:GUS TGCCGTCGTG CCCTGGATCG
ATTCTAGAAT GAACAAGATC CCGATCAAGG ACCTGCTGAA
CCCGCAGAGT GGGTACGGTC AG

NLS-C/MataZ hybrid:GUS CGAGGCTCTT CGCAAGATCC

 

CGATCAAGCG GAGTGGGTAC
Reverse NLS-C: US GCCGTCGTGC CCTGTATCGA

TCATATGCGG _AAGGGGATAG CTAAACGCCT TGCTGAGAGC

86
ATCATGAGTG GGTACGGTCA G

Minus Basic NLS-C:GUS GCCGTCGTGC CCTGTATCGA
TCATATGATC AGCGAGGCTC TGCGCCAGGC TATAGGGCAG
CGGAGTGGGT ACG

Minus Hydrophobic NLS-C:GUS GCCGTCGTGC CCTGTATCGA
TCATATGACC AGCGAGGCTC AGCGCAAAGC TACCGGGAAG

CGGAGTGGG.

Transformation of Onion Cells

Onion epidermal layers and plasmids were prepared as described
in Shieh et. al. (1993). The preparation of the gold particles and the
transformation conditions are the same as described in Shieh et. al.

(1993) except the plasmid DNA concentration was increased from 2.5

to 5.0 fig and the duration of sonication of the gold particles was

decreased (cup horn probe, 60% power, 10 s).

Histochemical Analysis
The colorimetric assay for B-glucuronidase (GUS) activity is the

same as described in Shieh et. al. (1993) except for the X-gluc buffer

87

and substrate incubation temperature. The onion cell incubation which
was increased to 37°C for 24 hours and the X-gluc buffer was altered
to improve the viability of the onion cells (2mM X-gluc, 20 mm NaPO4
pH 7.0, 0.05 mM K Ferrocyanide, 0.05 mM K Ferricyanide and 0.01%
Triton X100). As previous, the intracellular localization of the blue
precipitate was determined using a Zeiss Axiophot microscope with
Nomarski optics. The subcellular localization of each fusion protein
was determined from five to ten separate transformations. The

minimum number of cells analyzed for each construct was twenty.

RESULTS

Previously, it was demonstrated that the carboxyl terminus of the
maize R protein encoded a nuclear localization sequence (NLS C, 13
a.a.) which was able to redirect GUS protein from the cytoplasm to the
nucleus in onion epidermal cells (Shieh et. al. 1993). Based on amino
acid preference for helix formation, computer predicted secondary
structure and plotting on a helical wheel, NLS-C is predicted to form

an amphipathic alpha helix. To identify the amino acids which are

88

important for NLS-C to function, several constructs were made to
study the function of the hydrophobic amino acids, basic amino acids,
context of the amino acids and structural similarity between NLS-C
and the yeast Mata2 NLS.

To construct the GUS:modified NLS gene fusions, the NLS-
C:GUS construct was altered by site-directed mutagenesis. The
constructs were designed to encode the modified NLS-C at the amino
terminus of GUS as this orientation was previously shown to be optimal
for targeting of the gene fusion constructs to the nucleus (Shieh et. al.,
1993). The constructs were transformed into a monolayer of onion
epidermal cells by particle gun bombardment and subcellular
localization of the fusion proteins was determined with the
histochemical substrate X-gluc. Addition of NLS-C to GUS redirected
the GUS activity to the nucleus as indicated by the histochemical stain

X-gluc in the nucleus (Fig. 3.1 and 3.2A, Table 3.1).

Context of the Amino Acid Sequence Influences NLS Function
Since all NLSs contain basic amino acids it is hypothesized that
they are the sole determinant for a functional NLS (Boulikas. 1993).

To ascertain whether or not this hypothesis is correct, a construct was

EALR

w H

GIAK A

EALR

I B I I
H

a: a: N tn

Hi0 w H

:v :v i." iv
H

II!

EAQR

MISEALRRIP

MNKIPIKDLL

89

GKR
ESIM
GQR

GKR

IKR

NPQ

NLS - C

rev - C NLS - C

minus -basic NLS - C
minus - hydrophobic
NLS - C

NLS-cluataz

NLS diatom

figure 3.1. Amino acid sequences of the mutated NLS-C polypeptides.

90

Figure 3.2. Histochemical localization of GUS activity of each of the
mutated NLS-C GUS fusion constructs (A-G). The corresponding
locations of the nuclei are shown with the DNA specific dye DAPI (A’-
F’). A, NLS-C GUS; B, rev-NLS-C GUS; C, minus hydrophobic NLS-
C GUS; D, minus basic NLS-C GUS; E NLS-C/MataZ GUS; F, Mata2

GUS

91

 

 

 

92
Table 3.1

N_L_S Localization
NLS-C Nuclear
Rev-NLS-C Cytoplasmic
Minus-Basic Cytoplasmic
Minus-Hydrophobic NLS-C Nuclear/Cytoplasmic
NLS-C/Mata2 Nuclear
NLS-Mata2 Nuclear/ Cytoplasmic

 

 

 

Table 3.1 Summary of the histochemical analysis for the mutated

NLS-C:GUS fusion protein.

93

designed to reverse the amino acid sequence of NLS-C (Fig. 3.1, rev-
NLS-C). The reversed NLS-C retains the spacing of the basic amino
acids while changing the context in which the amino acids are
presented. The reverse NLS-C polypeptide was unable to redirect GUS
activity to the nucleus as indicated by the cytoplasmic localization of
the histochemical stain X-gluc (Fig. 3.2B, Table 3.1). The dark spot
in the area of the nucleus are the gold particles which carried the
reverse NLS-C:GUS DNA into the nucleus and does not represent GUS
activity. The micrograph in Figure 3.2B shows that GUS activity
remained in the area surrounding the nucleus; this may represent the
signal binding at the nuclear envelope without translocation across the

nuclear envelope.

The Hydrophobic Amino Acids are not Essential

The role of the hydrophobic amino acids in NLS-C was examined
by substituting the hydrophobic amino acids with more charged
molecules (Fig. 3.1, minus-hydrophobic NLS-C); this removes the
predicted amphipathic nature of the signal. The threonine for
isoleucine and glutamine for leucine substitutions were chosen to negate

the hydrophobicity while maintaining average surface volume and

94

normalized frequency of occurrence in an alpha-helix. When expressed
in onion epidermal cells, the histochemical stain X-gluc was partitioned
between the nucleus and cytoplasm (Fig. 3.2C, Table 3.1), indicating
that the minus-hydrophobic NLS:GUS protein was partitioned between
the nucleus and cytoplasm. Although the amount of partitioning
varied, distinct nuclear (or cytoplasmic) targeting was not observed

during any experiment with this construct.

Basic Amino Acids are Essential

The single feature which all NLSS share is the presence of basic
amino acids which are considered essential. To determine if the basic
amino acids are essential for NLS-C’s function, the two lysines were
substituted with two polar uncharged glutamine residues (Fig. 3.1,
minus-basic NLS-C). This change alters the charge density of the
polypeptide while maintaining the average surface volume,
hydrophilicity and normalized frequency of occurrence in an alpha-
helix. This construct did not redirect GUS activity to the nucleus as
indicated by the histochemical stain X-gluc in the cytoplasm (Fig. 3.2D,
Table 3.1). Therefore, the lysines are essential for NLS-C’s function.

However, it is possible that another amino acid, such as an arginine

95

may substitute for the lysine.

Similarity of NLS-C to the NLS of Mata2

To determine if NLS-C and the yeast Mata2 NLS are similar, a
hybrid NLS was constructed substituting the consensus region of the
Mata2 NLS (KIPIK; Hall et. al., 1984) into NLS-C (Fig. 3.1, NLS-
C/MataZ). The two lysines of both NLSS are separated by three amino
acids indicating that the two signals are similar. If the hybrid NLS
functions as a targeting signal then the implication is that yeast and
plant NLSS are similar. The hybrid targeting signal, NLS-C/MataZ,
redirected GUS activity from the cytoplasm to the nucleus (Fig. 3.2E,
Table 3.1) and there was no notable difference between the localization

of either NLS-C/MataZ or NLS-C GUS fusion proteins.

Mata2 NLS in Plants

NLS-C and the yeast Mata2 NLS are similar in the basic amino
acid spacing and content of hydrophobic amino acids. Therefore,
based on the similarity of NLS-C to Mata2, we wanted to determine if

the Mata2 NLS can function in plant cells. When Mata2 NLS was

96

fused to GUS, the activity was localized to the nucleus and cytoplasm
as indicated by the histochemical stain X-gluc in the nucleus and
cytoplasm (Fig. 3.2F, Table 3.1). This indicated that Mata2 NLS

retains partial function in plants.

DISCUSSION

NLS-C of the maize R protein contains several hydrophobic
amino acids and the lysines are spaced in the same pattern as those in
the Mata2-like NLSS. Therefore, NLS-C is more similar to the Mata2-
like NLS than to the SV40-like NLS. As evidence to indicate that NLS-
C and Mata2 NLS are similar, a fusion construct of the two signals
was constructed. If the two targeting signals are similar, then regions
from each should be capable of complementing a similar region in the
other. Therefore, the five central amino acids of NLS-C (KAIGK)
were substituted with the KIPIK sequence of the Mata2 NLS (construct
NLS-C/MataZ, Fig. 3.1). The KIPIK polypeptide is conserved in
several yeast nuclear proteins and is therefore considered the core of

the Mata2 NLS (Hall et. al., 1984). The NLS-C/Mata2 hybrid NLS

A.-

 

' ‘h-‘h
t ._ II

97

when fused to GUS redirected GUS activity to the nucleus with no
notable difference between the NLS-C and NLS-C/Mata2 in their
ability to direct GUS activity to the nucleus (Fig. 3.2E, Table 3.1).
This indicates that the secondary structure in the basic region of NLS-
C and Mata2 are homologous and suggests that NLS-C is a Mata2-like
signal. Since the KIPIK substitution in NLS-C contains a proline
residue it may be that the predicted alpha-helix structure of NLS-C
would be disrupted. However, a proline within four amino acids of the
end of a helix can be tolerated in a helix and therefore it may be
tolerated within the NLS-C/Mataz NLS.

If NLS-C is a Mata2-like NLSS, then the Mata2 NLS should
function in plants. Therefore, the ability of the Mata2 NLS to redirect
GUS to the nucleus was assayed. When the Mata2 NLS was fused to
GUS, the GUS protein was located in both the nuclear and cytoplasmic
compartments (Fig. 3.2F, Table 3.1). This indicated that despite its
inability to function as an NLS in animal systems, the Mata2 NLS can
function in plant systems but it is not as strong a signal as NLS-C. In
addition, preliminary experiments have been performed to determine
if NLS-C functions as a nuclear targeting signal in yeast and animal

cells. When overexpressed in yeast, the NLS-C and Mata2 NLS

98
redirected GUS to the nucleus and, like the Mata2 NLS, NLS-C was

unable to redirect GUS to the nucleus in the Xenopus oocytes in vitro
transport system (preliminary data not shown). This substantiates the
hypothesis that NLS-C is a Mata2-like NLS.

The amino acids which are important for a Mata2-like NLSS to
function have not been identified. Therefore, we analyzed NLS-C to
determine if the basic amino acids, charge density and hydrophobic
amino acids are important for its function. Basic amino acids are the
single feature common in all NLSS and their importance is best
indicated by the fact that, unlike other amino acids, a single alteration
of a basic residue in a targeting signal can completely negate its ability
to function as a targeting signal. The best studied of these mutations
is the SV40 large T antigen NLS (PmKKmKRKV) in which the lysine
at position 128 was replaced with a threonine (Kalderon et. al., 1984b).
Similar studies in plants were performed on bipartite NLSS where a
substitution of the basic residues affects the NLS’s targeting function
(V aragona et. al., 1994).

Therefore, two constructs were designed to determine if the
context of the amino acids or the basic amino acids are important in

NLS-C. It has been proposed that if four amino acids are basic in a

99

hexapeptide then that region will constitute a NLS (Boulikas, 1993) as
the basic charge is the major factor defining a polypeptide as a
targeting signal. To determine if the basic charge density of a NLS is
an essential factor for NLS functioning, the amino acid order in NLS-C
was reversed (Fig. 3.1, reverse NLS-C), thereby altering the context of
the NLS without changing the charge density or hydrophilicity of the
signal. When fused to the GUS protein, the reversed NLS-C was
unable to redirect GUS activity to the nucleus (Fig. 3.2B, Table 3.1)
and, therefore, the charge density does not determine NLS function.
Rather, the context of the amino acids was crucial for it to function as
a targeting signal. These findings are interesting because the seven
amino acids encompassing NLS-C (RKAIGKR) are virtually
palindromic (alanine and glycine are very similar in structure). This
inversion may have created a subtle change in the signal which negated
its function as a targeting signal. Alternatively, residues amino
terminal to the basic amino acids may have a strong influence on the
structure of the NLS, possibly to initiate the formation of the alpha-
helix. Similarly, Adam et. al. (1989) demonstrated that a reversed-
order SV40 large T antigen NLS will not compete against a wild type

SV40 large T antigen NLS for binding to a putative NLS receptor in

100

mammalian cells.

To determine if the basic amino acids are essential in NLS-C, the
two lysine residues were substituted with glutamic acids (minus-basic
NLS-C). Substitution of the two lysine residues negated the ability of
the targeting signal to redirect GUS activity to the nucleus (Fig. 3.2D,
Table 3.1). Therefore, the lysines were essential for NLS-C to function
and this correlates with numerous examples indicating that the basic
amino acids are essential in NLSS (Kalderon et. al., 1984; Varagona
and Raikhel, 1994).

Multiple hydrophobic residues are not frequently found in NLSS,
therefore, the role of the hydrophobic amino acids in NLS-C was
investigated. The predicted secondary structure for NLS-C is an
amphipathic alpha-helix which exposes the basic amino acids on one
side of the helix while the hydrophobic residues are hidden from the
surface of the protein. Substitution of the hydrophobic amino acids
with polar uncharged amino acids would eliminate the amphipathicity.
The minus-hydrophobic NLS-C partially redirected GUS activity to the
nucleus (Fig. 3.2C, Table 3.1). Therefore, the hydrophobic amino
acids are important but not essential for NLS-C to function. Unlike

other constructs tested in this and our previous study (Shieh et. al.

 

101

1993), the ratio of GUS activity in the nucleus and cytoplasm varied in
different experiments but always partitioned (data not shown). The
variability suggests that the minus-hydrophobic NLS does not form a
stable structure recognizable by the transport machinery.

In our study, it has been demonstrated that NLS-C is a Mata2-
like targeting signal because it contains several hydrophobic amino
acids, the basic amino acids are similarly spaced and the conserved
sequence KIPIK of the Mata2 can substitute for five amino acids in
NLS-C. It was determined that the hydrophobic amino acids are
important for NLS-C to function as a targeting signal. In addition,
because the Mata2 NLS from yeast was able to redirect GUS activity
to the plant cell nucleus, the nuclear transport mechanism in plant and
yeast systems are closely related.

Since several of the mutations delete the targeting function of
NLS-C, future experiments will use these constructs to identify
components of the nuclear transport machinery. Other mutated NLSS
tested in plants (i.e. SV40 large T antigen mutant and the mutated
bipartite signal of the maize 02 protein; Varagona and Raikhel, 1994)
were capable of redirecting some GUS activity to the nucleus.

Therefore, the minus-basic and reverse NLS-C constructs, which are

102

restricted to the cytoplasm, are stronger negative controls for
transport. In addition, the reverse NLS-C contains the same amino
acids and charge density as NLS-C while not functioning as a targeting
signal. The reverse-NLS-C is, therefore, a more appropriate control
to distinguish non-specific binding of NLSS to NLS-binding proteins

which are interacting based upon charge.

103

REFERENCES

Adam SA, Lobi TJ, Mitchell MA, Gerace L (1989) Identification
of specific binding proteins for a nuclear location sequence.
Nature 337: 276-279

Boulikas T (1993) Nuclear localization signals. CRC Crit. Rev.
Euk. Gene Expr. 3(3): 193-227

Chelsky D, Ralph R, Jonak G (1989) Sequence requirements for
synthetic peptide-mediated translocation to the nucleus. Mol.
Cell. Biol. 9: 2487-2492

Dingwall C, Laskey RA (1991) Nuclear Targeting sequences- a
consensus? TIBS 16: 478-481

Forbes D (1992) Structure and function of the nuclear pore
complex. Ann. Rev. Cell Biol. 8: 495-527

Garcia-Bustos J, Heitman J, Hall MN (1991) Nuclear protein
localization. Biochim Biophys Acta 1071: 83-101

Goff SA, Klein TM, Roth BA, Fromm ME, Cone KC, Radicella
JP, Chandler VL (1990) Transactivation of anthocyanin
biosynthetic genes following transfer of R regulatory genes

into maize tissue. EMBO J 9: 2517-2801

Hall MN, Hereford L, Herskowitz I (1984) Targeting of E.coli B-
galactosidase to the nucleus in yeast. Cell 36: 1057-1065

Kalderon D, Richardson WD, Markham AF, Smith AE (1984a)
Sequence requirements for nuclear location of simian virus 40
large T antigen. Nature 311: 33-38

Kalderon D, Roberts BL, Richardson WD, Smith AE (1984b) A short
amino acid sequence able to specify nuclear location. Cell 39:
499-509

104

Kunkel, T. A., Roberts, J. D., and Zakour, R. A. (1987). Rapid and
efficient site-specific mutagenesis without phenotypic selection.
Methods Enzymol 154: 367- 382.

Lanford RE, Butel JS (1984) Construction and characterization of an
SV40 mutant defective in nuclear transport of T antigen. Cell 37 :
801-813

Lanford RE, Feldherr CM, White RG, Dunham RG, Kanda P
(1990) Comparison of diverse transport signals in synthetic
peptide-induced nuclear transport. Exp. Cell Res. 186: 32-
38

Nelson M, Silver P (1989) Context affects nuclear protein
localization in Saccharomyces cerevisiae. Molec. and Cell.
Biol. 9: 384.389

Paine PL, Moore LC, Horowitz SB (1975) Nuclear envelope
permeability. Nature 254: 109-114

Raikhel NV (1992) Transport of proteins to the nucleus. Plant Phys.
100: 1627-1632

Sambrook, J., Fritsch, E- F., and Maniatis, T. (1989). Molecular
Cloning: A Laboratory Manual, Ed 2 Cold Spring Harbor, Cold
Spring Harbor NY

Shieh MW, Wessler SR, Raikhel NV (1993) Nuclear targeting of
the maize R protein requires two nuclear localization
sequences. Plant Physiol. 101: 353-361

van der Krol AR, Chua N-H (1991) The basic domain of plant
B-ZIP proteins facilitates import of a reporter protein into
plant nuclei. Plant Cell 3:667-675

Varagona MJ, Raikhel NV (1994) The basic domain in the bZIP
regulatory protein Opaque2 serves two independent
functions: DNA binding and nuclear localization. Plant J.
5: 207-214

CHAPTER4

FUTURE RESEARCH PROSPECTIVES

105

106

Towards a consensus NLS

The structural elements which are fundamental for a NLS are
unknown and will require either identification of additional NLSS or
more detailed information on the NLSS identified if a consensus is to
ever to be discovered. Site-specific mutations of SV40 large T antigen
NLS (Kalderon et. al.,1984), bipartite opaque 2 NLS (Varagona and
Raikhel, 1994) and the MataZ-like maize NLS-C (chapter 3) have all
indicated that the basic amino acids are essential for NLS function.
However, the amino acid spacing of the basic residues is not the sole
factor defining a NLS because rev-NLS-C does not redirect GUS to the
nucleus. Analysis of the amino acid sequence of the NLSS identified
(Boulikas, 1993) and the mutational NLS studies do not reveal a
common structural feature which could denote a consensus sequence.
Rather, the variation in NLS amino acid composition and length
suggests that there are structural features which are not obvious from
the sequence. In addition, NLS function is dependent on the context
in which it is presented. As an example, the NLSS (A, M and C) of the
maize R protein redirected more GUS protein to the nucleus when
fused to the amino- rather than the carboxy-terminus of GUS, see

chapter 2 (also see chapter 1; SV40 NLS in pyruvate kinase). The

107

large variation in NLS amino acid length and context in which they are
presented distinguishes them from other organelle targeting signals
with consensus structures. Chloroplast transit peptides, mitochondrial
signals and signal peptides (secretory) have consensus sequences and
are typically found as the amino-terminal residues and are not as
variable in size and number as NLSS (Boulikas, 1993). Therefore, it
is my assertion that there is not a single consensus sequence for all
NLSS. Rather, NLS function is based upon common structural
features. Therefore, to identify the NLS consensus structure it will be
necessary to derive the structure from X-ray crystallographic data of
the NLSS in their native proteins.
The nuclear transport machinery in plants

Having identified the NLSS of the maize R protein, the next
objective is to use the NLSS to identify other components of the nuclear
transport machinery in plants. Along with the Opaque2 bipartite NLS
(V aragona et. al., 1992) we have plant NLSS similar to the three types
of NLSS, the SV40-like (R-NLS-M), bipartite (opaque2) and MataZ-like
(R-NLS-C). The plant bipartite NLS from opaque2 has already been
shown to have a higher affinity for binding to tobacco and maize nuclei

than the animal SV40 large T antigen NLS (Hick and Raikhel, 1993).

108
Similar experiments have been performed with R-NLS-C and binding

to the nucleus is NLS specific with similar characteristics to those of
the bipartite opaque2 NLS (Smith S, Hick GR and Raikhel NV;
unpublished results). Chemical-crosslinking of radiolabeled opaque2
bipartite NLS to proteins extracted from the nuclear envelope has
labeled three putative NLS-binding proteins which have binding
characteristics similar to those of the isolated nuclei (Hick and Raikhel,
unpublished). These putative NLS-binding proteins will be purified so
that they can be studied in detail.

To determine if the putative NLS-binding proteins are part of the
nuclear transport machinery, it will be necessary to develop an in vitro
nuclear transport system. Reconstituting transport will allow the
identification of individual components of the transport machinery.
Also, the energy and cellular requirements for transport are unknown
and there may be cytosolic NLS-binding proteins required for
transport. The mutated NLS, rev-NLS-C, will be a useful control in
the characterization of both an in vitro transport system and NLS-
binding proteins because it maintains the charge density and amino
acid content of NLS-C without functioning as a NLS in onion

epidermal cells.

109

Understanding the molecular mechanism of nuclear transport is
essential if we are to learn how cellular and developmental processes
are regulated. Nuclear transport is typically thought to be constitutive,
such that all translated nuclear proteins are immediately transported
into the nucleus. However, some nuclear proteins are retained in the
cytoplasm until activated for transport. For example, the
glucocorticoid steroid hormone receptor is retained in the cytoplasm
until it binds hormone and is then transported into the nucleus to
initiate transcription. Regulation of nuclear import is of major
importance in cellular differentiation. Nuclear localization of the Rel
related proteins dorsal and NF-Kappa B is the determinant for dorsal-
ventral axis formation and immunoglobulin synthesis. Both of these
changes are key steps in the final determination of cell type. Since
plant cells are fixed in place by the cell wall, they will be a model
system to study regulated nuclear transport during cellular
differentiation. In addition, due to the totipotency of plant cells, there
must be additional levels of regulation unique to plants. Therefore, the
study of nuclear transport in plants will give additional insights into

the complexity of cellular regulation and differentiation.

1 10
REFERENCES

Boulikas T (1993) Nuclear localization signals. CRC Crit. Rev.
Euk. Gene Expr. 3(3): 193-227

Hick GR, Raikhel NV (1993) Specific binding of nuclear
localization sequences to plant nuclei. Plant Cell 5:983-994

Kalderon D, Richardson WD, Markham AF, Smith AE (1984a)
Sequence requirements for nuclear location of simian virus 40
large T antigen. Nature 311: 33-38

Varagona MJ, Schmidt RJ, Raikhel NV (1992) Nuclear localization
signal(s) required for nuclear targeting of the maize regulatory
protein, Opaque-2. Plant Cell 4: 1213-1227

Varagona MJ, Raikhel NV (1994) The basic domain in the bZIP
regulatory protein Opaque2 serves two independent

functions: DNA binding and nuclear localization. Plant J.
5: 207-214