Characterization and identification of a novel guanine nucleotide binding site on the bovine mitochondrial ribosome

MISSING IMAGE

Material Information

Title:
Characterization and identification of a novel guanine nucleotide binding site on the bovine mitochondrial ribosome
Physical Description:
xvi, 179 leaves : ill. ; 29 cm.
Language:
English
Creator:
Anders, John Claude, 1951-
Publication Date:

Subjects

Subjects / Keywords:
Research   ( mesh )
Guanine Nucleotides   ( mesh )
Guanosine Triphosphate -- analogs & derivatives   ( mesh )
Guanosine Triphosphate -- chemistry   ( mesh )
Mitochondria   ( mesh )
Ribosomes   ( mesh )
Binding Sites   ( mesh )
GTP-Binding Proteins   ( mesh )
Cattle   ( mesh )
Department of Biochemistry and Molecular Biology thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Biochemistry and Molecular Biology -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1990.
Bibliography:
Bibliography: leaves 167-178.
Statement of Responsibility:
by John Claude Anders.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001581949
oclc - 24594102
notis - AHK5867
sobekcm - AA00006098_00001
System ID:
AA00006098:00001

Full Text









CHARACTERIZATION AND IDENTIFICATION OF A NOVEL GUANINE
NUCLEOTIDE BINDING SITE ON THE BOVINE MITOCHONDRIAL RIBOSOME










by
JOHN CLAUDE ANDERS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1990









ACKNOWLEDGEMENTS


This work was made possible first by the love and support
of my wife, Valerie, and children, Jane, Benjamin, and
Katherine. Secondly, thanks are given to the U.S. Army which
provided the financial support for my graduate training. I
must thank Dr. Thomas W. O'Brien for his mentorship as my
major professor and the helpful guidance of the members of my
graduate studies committee. A special relationship often
develops between a student and their major professor and this
has been the case for me. Special thanks go to Dr. Nancy D.
Denslow who was instrumental in the initial discovery of GTP
binding to mitochondrial 28S ribosomes and also provided free
access to her work pertinent to this study and very helpful
insight and guidance. I must thank all the members of the
O'Brien laboratory for their patience in answering my
innumerable questions and especially to Mr. Michael Bryant who
often lead me through new techniques. Thanks go to fellow
graduate student Scott E. Fiesler, who collaborated with me in
the study of [a32P]ATP binding to mitochondrial ribosomes
presented in this study. Finally, thanks go to Michael Bryant
again, and to Mr. Bennie Parten of the protein core facility
in assistance with the chemical sequencing of mitochondrial
ribosomal proteins and peptides.









TABLE OF CONTENTS


ACKNOWLEDGEMENTS ...................................... ii
LIST OF TABLES ........................................ v
LIST OF FIGURES....... ............................ vi
ABBREVIATIONS......................................... ix
ABSTRACT.............................................. xv
CHAPTERS
1 INTRODUCTION..................................... 1
An Overview of GTP-Binding Proteins................ 1
Protein Synthesis.................. ................ 3
Conserved Amino Acid Sequences Among GTP-Binding
Proteins........... ...................... 6
Proteins Synthesis in Mitochondria...... ........ 12
Proposal .................................... .. 16
2 NONCOVALENT BINDING OF MITOCHONDRIAL RIBOSOMES BY
NTPs........ ....................................... 17
Introduction ...................................... 17
Materials and Methods............................. 19
Results......................................... 25
Discussion...................................... 33
3 PHOTOAFFINITY LABELING OF 28S RIBOSOMES WITH
8-AZIDO GTP ........................................ 37
Introduction......................... ............ 37
Materials and Methods............................. 43
Results....................................... .. 49
Discussion....................................... 72
4 PHQTOAFFINITY LABELING OF 28S RIBOSOMES WITH
[accP]GTP....................................... 79
Introduction ..................................... 79
Materials and Methods............................ 80
Results ........................................... 81
Discussion ......................................... 107


iii









5 ISOLATION IDENTIFICATION, AND PARTIAL AMINO ACID
SEQUENCE OF S5 AND OTHER RIBOSOMAL PROTEINS........ 112
Introduction............................. ..... 112
Materials and Methods ............................ 115
Results....................................... 128
Discussion ............................. ......... 159
6 CONCLUSIONS AND FUTURE DIRECTIONS................. 167
REFERENCES............................................. 167
BIOGRAPHICAL SKETCH................................... 179










LIST OF TABLES


Table PMe
1-1 Protein and rRNA Content of Ribosomes.......... 14
2-1 Noncovalent Binding of Various [a32P]
Nucleotides to 28S Ribosomes................... 27
2-2 Competition for [a32P]GTP Binding to 28S
Ribosomes by Various Nucleotides............... 29
2-3 Noncovalent Binding of GTP and 8-Azido GTP to
28S Ribosomes ................................. 33
3-1 The Effect of Substituting KCL for NH4C1, and
the Lowering of the MgCl2 Concentration in the
ReAction Buffer on the Noncovalent Binding of
[a P]GTP to 28S Ribosomes.................... 56
5-1 Summary of Peak Identification for RP-HPLC of
28S Ribosomal Proteins........................ 144
5-2 Summary of Ribosomal Proteins and Peptides
Submitted for Gas Phase Amino Acid Sequence
Analysis....................................... 155
5-3 Summary of Amino Acid Sequence for Several
Bovine Mitochondrial Ribosomal Proteins and
CNBR Peptides of S5........................... 156
5-4 Comparison of Amino Acid Sequence Found in S5
Corresponding to the E Site Found in Other GTP-
Binding Proteins Involved in Protein Synthesis. 162







LIST OF FIGURES


Figure gage
1-1 Comparison of amino acid sequences from con-
served binding regions, A,C,E, and G in GTP-
binding proteins ................................ 8
2-1 Noncovalent binding of [a3P]GTP, GMP, ATP,
UTP, and CTP to 28S bovine mitochondrial
ribosomes.......... ........................... 26
2-2 Binding inhibition of [a32P]GTP to bovine
mitochondrial ribosomes by various
nucleotides ................................. 28
2-3 Noncovalent binding of [a32P]8-Azido GTP and
[8- H]GTP to 28S bovine mitochondrial ribosomes 31
2-4 Binding inhibition of [a32'P]8-Azido GTP to
bovine mitochondrial ribosomes by GTP.......... 32
3-1 Generation of an aryl nitrene following uv
irradiation of 8-Azido GTP.................... 40
3-2 Photoaffinity labeling of 28S ribosomes with
[a, P]8-Azido GTP............................. 51
3-3 Two dimensional PAGE analysis .of photoaffinity
labeled 28S ribosomes with [a P]8-Azido GTP.... 53
3-4 Optimization of the intensity and duration of uv
irradiation for photoaffinity labeling of 28S
ribosomes with [a P]8-Azido GTP................ 58
3-5 Continued optimization of the intensity and
duration of uv irradiation for phAtoaffinity
labeling of 28S ribosomes with [y P]8-Azido GTP 60
3-6 The effect of decreasing concentrations of MgC12
on.photoaffinity labeling of 28S ribosomes with
[y P]8-Azido GTP.............................. 63
3-7 Two dimensional PAGE analysis of photoaffinity
labeled 28S ribosomes, with increasing
concentrations of [y P]8-Azido GTP............. 66
4-1 Optimization of the intensity and duration of uv
irradiation for photoaffinity labeling of 28S
ribosomes with [a P]GTP................ ....... 82








4-2 Specificity of the photoindured radiolabeling of
28S ribosomal proteins by [a P]GTP in the
presence of other nucleotides.................. 85
4-3 Determination of the binding affinity of [a32P]
GTP for the specifically labeled 38 kDa ribosomal
protein by photoaffinity labeling of 28S
ribosomes..................................... 90
4-4 Two dimensional PAGE analysis of photoaffinity
labeled 28S ribosomes with [a P]GTP........... 94
4-5 Determination of the requirement for uv
irradiation to rpdiolabel 28S ribosomes with
[a P]ATP and [a P]GTP......................... 100
4-6 Time course fur the uv independent radiolabeling
of S4 with [a P] and [y P]ATP................ 104
5-1 Reverse phase high pressure liquid chromatogram
of 1.6 nmol of acetic acid extracted 28S
ribosomes resolved on a wide pore (300 A) silica
bonded butyl column ........................... 130
5-2 Two-dimensional PAGE analysis of 28S ribosomal
proteins eluted from a RP-HPLC butyl column
coelectrophoresed with a minimal amount of acetic
acid extracted 28S ribosomes .................. 134
5-3 Immunoblot of 28S bovine ribosomal proteins with
an anti S5 monospecific rabbit sera........... 146
5-4 Immunoblot of 28S bovine ribosomal proteins with
an anti S5 monospecific rabbit sera........... 147
5-5 Immunoblot analysis of 28S bovine ribosomal
proteins separated by RP-HPLC with an anti S5
monospecific rabbit sera..................... 150
5-6 Immunoblot analysis of S5 protein isolated by
RP-HPLC............. ......................... 152
5-7 Electrophoretic separation of cyanogen bromide
peptides of S5................................ 154
5-8 Elucidation of a putative oxidized tryptophan
residue found in the 5 PTH amino acid residue
of the 13 kDa CNBR peptide of S5............... 158


vii








5-9 Comparison of putative conserved regions found in
the amino acid sequence of CNBR peptides of S5 to
homologous regions found in other GTP-binding
proteins... ......... ...... .................... 161


viii











A. ....... .....
AA ............
Ac............
ACS...........
ADP...........
AR............
ATP...........
ATZ...........
AU............
8N3GTP........
BCIP..........

bp............
BME ...........
OC........... .


C4............

Ci............
cm............
cm2...........

CNBR........
cpm ...........
CTP ...........
cGMP..........

cytoribosome..


ABBREVIATIONS

Angstrom, meter x 10-10
amino acid
acetate
American Chemical Society
adenosine 5'-diphosphate
autoradiogram
adenosine 5'-triphosphate
anilinothiazolinone
absorbance unit
8-Azido guanosine 5'-triphosphate
5-bromo-4-chloro-3-indolyl phosphate

base pairs
B-mercaptoethanol
degrees centigrade
percent cross inker (bisacrylamide)
butyl

curie, 2.2 x 1012 dpm
centimeter
centimeter squared

cyanogen bromide
counts per minute
cytosine 5'-triphosphate
3'-5'-cyclic guanosine monophosphate

cytoplasmic ribosome








d............
Da............
DEAE..........
DEPC..........
dGTP..........
DPTU..........
DTT...........
E. co i .......
EDTA..........
elF...........
EF-Tu.........
EF-Ts.........
EF-G..........
EF-1..........
f............
Fc............

g.............
GDP ...........
gm ...........
GMP ...........
GMPPCP........
GMPPNP........
GTP...........
h............
3H ............

HPLC..........


day
dal tons
diethylaminoethyl
diethylpyrocarbonate
2'-deoxyguanosine triphosphate
N,N'-diphenylthiourea
dithiothreitol
Escherichia coli
ethylenediaminetetraacetic acid
eucaryotic initiation factor
bacterial elongation factor-Tu
bacterial elongation factor-Ts
bacterial elongation factor-G
eucaryotic elongation factor-1
fraction
constant fragment of IgG molecule
force of gravity
guanosine 5'-diphosphate
grams
guanosine 5'-monophosphate
guanosine 5'-[s,y-methylene]triphosphate
guanosine 5'-[s,y-imido]triphosphate
guanosine 5'-triphosphate
hour
tritium
high pressure liquid chromatography








IF-1..........

IF-2 ..........

IF-3 .........

IgG ...........

Imm...........

ITP...........

Kd............

kDa ...........

Ki ............

m.............

M............

mA............

Mr............


ig............
Ig ............
Pi ............

pM ............

pW ............
mg............

min...........

mitoribosome..

ml ............

mM.... .......

mmol..........

mRNA..........

mt............


bacterial initiation factor-1

bacterial initiation factor-2

bacterial initiation factor-3

immunoglobulin G

immune

inosine 5'-triphosphate

binding dissociation constant

kilodaltons

binding inhibition constant

milli

molar, moles per liter

milliamphere

molecular mass

micro

microgram

microliter

micromolar

microwatt

milligram

minute

mitochondrial ribosome

milliliter

millimolar

millimole

messenger ribonucleic acid

mitochondria








MWM........... molecular weight markers
n............. number
NADP.......... nicotinamide adenine dinucleotide phosphate
ng............ nanogram, grams x 10.9
nm............ nanometer, meter x 10-9
nmol.......... nanomole, moles x 10-9
D. ............ one-dimensional
NBT........... nitro blue tetrazolium
NC.......... no competition
NEM........... n-ethylmaleimide
ND............ not detected
NTP........... nucleotide triphosphate
Org........... organic
P............. peak
2P........... phosphate
PAGE.......... polyacrylamide gel electrophoresis
pg.......... picograms, grams x 10-12
Pi............ phosphate
PI............ preimmune
PITC.......... phenylisothiocyanate
pmol.......... picomole, moles x 10-12
POPOP......... 1,4-Bis(5-phenyloxazol-2-yl)benzene
PPO........... 2,5-diphenyloxazole
Prep.......... preparation
PTH........... phenylthiohydantoin
PVDF.. ..... poly(vinylidene difluoride) membrane


xii








Rib .. .......
Res ..........
RF............
RP-HPLC.......
RNA...........
rRNA..........
S............
S............ .
S5............
SD............
sec...........
SE............
SDS. ..........
SW............
%T..... ......
TBS...........
TBSA..........
TEA...........
TEMED.........
TFA ...........
TMA...........
TP............
tRNA..........
Tris-HCL......
Triton X-100..
2D ............


ribosome
residue
release factor
reverse phase-HPLC
ribonucleic acid
ribosomal ribonucleic acid
stained gels (used in figures)
Svedberg
small subunit ribosomal protein #5 (or 1 to 33)
standard deviation
seconds
standard error
sodium dodecyl sulfate
sperm whale
percent total monomer
tris buffered saline
tris buffered saline with sodium azide
triethanolamine

N,N,N',N'-Tetramethylethylenediamine
trifluoroacetic acid
trimethylamine
total protein
transfer ribonucleic acid
tris-hydroxymethylaminomethane-hydrochloride
Octylphenoxypolyethoxyethanol
two-dimensional


xiii








UTP .......... uridine 5'-triphosphate
uv........... ultraviolet
V............. volt
X............. any amino acid


xiv













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

CHARACTERIZATION AND IDENTIFICATION OF A NOVEL GUANINE
NUCLEOTIDE BINDING SITE ON THE BOVINE MITOCHONDRIAL RIBOSOME
by
John Claude Anders
December, 1990
Chairman: Dr. T. W. O'Brien
Major Department: Biochemistry and Molecular Biology
Bovine mitochondrial ribosomes possess a single, high
affinity binding site for GTP and GDP on the small subunit.
This is in contrast to procaryotic and eucaryotic cytoplasmic
ribosomes which do not bind GTP directly but do so to soluble
protein factors which cycle on and off the ribosome. We used
a photoaffinity analog of GTP, 8-Azido GTP and well as GTP to
identify ribosome components associated with the GTP binding
site. 8-Azido GTP binding to the small subunit was saturable,
with an apparent dissociation constant (Kd) of 1.92 pM,
approximately 126 times higher than that for GTP. Photolysis
of both the photoreactive analog and GTP resulted in the
specific labeling of a single ribosomal protein, S5.
Photolabeling of S5 by both [a32P]GTP, and its 8-azido analog,
was blocked competitively by the addition of excess GTP,







demonstrating the labeling was site specific. Partial amino
acid sequence and analysis of this protein revealed two areas
of amino acid sequence having significant homology to
conserved sequences involved in GTP binding in other GTP-
binding proteins. This protein with a guanine nucleotide
binding site may participate in initiation complex formation
on mammalian mitochondrial ribosomes. These mitochondrial
ribosomes apparently employ a different mechanism for
initiation of protein synthesis than bacterial or eucaryotic
cytoplasmic ribosomes which are unable to interact directly
with GTP.


xvi













CHAPTER 1
INTRODUCTION
An Overview of GTP-Bindina Proteins
G-proteins are a central component of the cellular
machinery driving olfaction, vision, control of cell prolif-
eration, cellular regulation, and protein synthesis (for
reviews see Stryer, 1986; Gilman, 1987; and Allende, 1988).
Their basic function in the first four instances is to mediate
signal transduction between the cell surface and the cyto-
plasm. This process begins when a ligand binds to the extra-
cellular portion of a receptor protein. Upon ligand binding,
a conformational change in the receptor allows binding of a
GDP binding protein to the cytoplasmic portion of the same
receptor. This results in the exchange of GDP for GTP on the
membrane bound G-protein. This exchange of guanine nucleo-
tides is often mediated by protein exchange factors which are
an important site for regulation of these processes. This
exchange results in conformational changes in the G-protein
which results in dissociation of two subunits of the G-protein
heterodimer, commonly called G8,, leaving an activated G,-GTP
protein. Activated G,-GTP now binds another membrane bound
acceptor protein such as cGMP phosphodiesterase, adenylate








2
cyclase, or phospholipase C, which results in catalytic
production of a second messenger molecule cGMP, cAMP, or
inositol triphosphate and diacylglycerol, respectively. These
soluble, second messenger molecules complete the signalling
event by binding to tissue specific intracellular receptors.
In the case of second messengers cGMP and cAMP, they modulate
conductance of various ions through the plasma membrane.
Inositol triphosphate and diacylglycerol bind to the
endoplasmic reticulum, resulting in release of intracellular
calcium, and to protein kinase C (PKC), forming an active PKC,
respectively.
Another feature of GTP-binding proteins is an intrinsic
GTPase activity which results in hydrolysis of GTP to GDP.
Following activation of the acceptor molecule, the GTP-binding
protein turns itself off by hydrolyzing the bound GTP to GDP
resulting in an altered conformation for the protein. This
results in a reduced affinity and dissociation from the
acceptor protein thereby inactivating the signaling event.
In protein synthesis, GTP-binding proteins function in
a different manner. There is no signaling across a cell
membrane upon ligand binding and no membrane bound G-protein
is involved, as described above. However, the G-proteins in
this system (initiation and elongation factors) do bind GTP.
Binding of GTP results in a conformational change in the
factor that allows it to bind to the ribosome and produce a
competent initiation or elongation complex leading to protein







3
synthesis. Hydrolysis of GTP, bound to these factors,
provides the energy to drive both initiation and elongation in
protein synthesis.
Protein Synthesis
G-Proteins are a key component in protein synthesis in
both procaryotic and eucaryotic systems. The components of
protein synthesis are analogous in both systems, differing
primarily in the increased complexity and numbers of accessary
proteins in the eucaryotic system. These accessary proteins
are involved in regulation of translation in the eucaryotic
system (for review see Moldave, 1985). Many features of
procaryotic ribosomes resemble that of eucaryotic ribosomes
therefore description of protein synthesis, will be largely
based upon the accumulated knowledge of translation in
procaryotic systems.
Translation of information encoded in messenger RNA
(mRNA) begins with the formation of a competent ribosomal
initiation complex. In E. coli, a GTP-binding initiation
factor, IF-2, binds GTP and initiator fmet-tRNA on the small
subunit of the ribosome forming a GTP-ternary complex. The
order of binding is random and IF-2 does not necessarily bind
initiator methionyl t-RNA prior to binding the ribosome
(Gualaerzi, 1977, and 1986). In conjunction with two other
protein initiation factors, IF-1 and IF-3, mRNA binds to
complete the initiation complex. Upon formation of the
complete initiation complex on the small subunit, the







4
intrinsic GTPase activity of IF-2 is activated and bound GTP
is hydrolyzed resulting in the release of GDP-bound IF-2, IF-
1, and IF-3 from the small subunit. The large subunit now
binds to the small subunit with the initiator tRNA held in
proper orientation with respect to the initiator codon
(usually AUG) of the mRNA to form a monosome to start
polypeptide elongation. IF-1 and IF-3 (which are not GTP-
binding proteins) are known to increase the binding affinity
of the GTP-ternary complex (Wintermeyer and Gualerzi, 1983)
and IF-3 also inhibits formation of monosomes (Noll and Noll,
1972).
Eucaryotic initiation factor, eIF-2, functions in a
manner nearly identical to IF-2. The major functional
difference is how GDP on the inactive initiation factor
exchange for GTP. The exchange of GDP for GTP occurs
spontaneously in procaryotes due to relatively low binding
affinities for both nucleotides. In eucaryotes the binding
affinity for GDP is typically several fold greater than for
GTP. An example of this is seen in Xenopus leavis oocytes
(Carvallo and Allende,1987) where GDP binds 50 times better
than GTP (Kd= 70 and 3800 nM for GDP and GTP, respectively).
The exchange of tightly bound GDP for GTP, on EIF-2, requires
a guanine exchange factor (GEF, Panniers and Henshaw, 1983)
which regulates initiation of protein synthesis in eucaryotes
(Ochoa, 1983).







5
Three elongation factors are required for protein
elongation, EF-Tu/Ts, EF-G in procaryotes and EF-1/1B, EF-2 in
eucaryotes. Two of these are GTP-binding proteins, EF-Tu, and
EF-G (EF-1, and EF-2 in eucaryotes) while EF-Ts (EF-1B in
eucaryotes) is an exchange factor. Both procaryotic and
eucaryotic elongation factors function in the same manner.
When GTP is bound to EF-Tu (or EF-1) a conformational change
occurs allowing binding of aminoacyl-tRNA to the factor (Crane
and Miller, 1974). Upon binding to the monosome and
positioning of the aminoacyl-tRNA in the P-site, the intrinsic
GTPase activity induces hydrolysis of GTP, and the GDP-bound
EF-Tu leaves the ribosome. The exchange factor, EF-Ts,
catalyses the exchange of GDP for GTP to reactivate the EF-Tu.
Elongation factor EF-Tu, bound with GTP can now bind the
ribosome. This induces the hydrolysis of GTP which provides
the energy to transfer of the aminoacyl-tRNA to the A site.
Once in the A site and the codon-anticodon pairing is correct,
the peptide chain is transferred from the P site (peptidyl
transferase reaction) to the amino group of the newly arrived
amino-acyl-tRNA. This generates a peptidyl-tRNA. Another
GTP-binding protein, EF-G catalyzes the translocation of the
peptidyl-tRNA from A to the P-site. The energy to drive
translocation is provided by the hydrolysis of GTP bound to
EF-G.
Termination of protein synthesis in procaryotes is not as
well worked out as are the initiation or elongation phases.







6
There are three release factors, RF-1, RF-2, and RF-3. RF-1
and RF-2 bind termination codons, UAA or UAG, and UGA,
respectively in mRNA (Tate and Caskey, 1974). The binding of
RF-1 or RF-2 to the ribosome stimulates the binding of the
third release factor, RF-3. This factor binds GTP and it
promotes the dissociation of release factors from the
ribosome, completing the synthesis of the nascent polypeptide.
A eucaryotic RF has been described from reticulocyte
preparations (Konecki et al., 1977) which also binds GTP and
has a GTPase activity.
The common elements seen in all the GTP-binding
initiation, elongation, and termination factors are that they
incur conformational changes upon binding GTP, that allow them
to mediate the binding of aminoacylated tRNAs or other protein
factors to the ribosome. Once completed they hydrolyze GTP to
GDP to provide the energy to release the factor or drive
protein synthesis.
Conserved Amino Acid Sequences Among GTP-Binding Proteins
Proteins that bind GTP and GDP comprise a diverse group.
The group includes the polypeptide initiation and elongation
factors, tubulin, transducin, adenylate cyclase receptor
proteins, and a family of ras proteins. The ras genes encode
a family of closely related proteins of approximately 21 kDa
and were initially identified in human lung and bladder
sarcomas (Scolnic et al., 1979; Shih et al., 1980; Papageorge
et al., 1982).







7
Halliday (1984) described four regions of homology among
various G-proteins whose amino acid sequences were known at
the time. The proteins were bacterial elongation factors (EF-
G and EF-Tu of E. coli), human c-H-rMsl gene product (c-H-
ras), and yeast c-rasS' gene product (figure 1-1). Van Meurs
et al. (1987) expanded this comparison to include the a-
subunits of both Ta rod and Ta cone transducin and the a-
subunits of GTP-binding proteins that modulate adenylate
cyclase, Gi and Gs. All sequence information for these
proteins were obtained from bovine sources. Recently,
several other notable G-proteins have been sequenced, a
bacterial initiation factor (IF2) and even a yeast
mitochondrial elongation factor (EF-Tumt, see figure 1-1 for
references). The same four regions of conserved sequence were
found in all these GTP-binding proteins from bacterial, yeast,
and mammalian sources and are shown in figure 1-1.
Two additional conserved sites for the mono-ADP
ribosylation by bacterial toxins were found in the a-subunits
of transducin and GTP-binding proteins that modulate adenylate
cyclase (for review see Ueda and Hayaishi, 1985). ADP-
ribosylation is a post-translational modification that
utilizes NAD as the donor of the ADP-ribose group and leads to
a decreased rate of GTP hydrolysis. This impaired intrinsic
GTPase activity leaves the GTP-binding protein in a
permanently activated (GTP-bound) state (Cassel and Selinger,
1977).

















HH H > >
H HH > H

N* 00 00 *0 00 0c #o 0 0
rVNo-W -4rrr-


** 00 #0 00 go 0* @0 ;4 ** 00 **

NoaM 4N -I v f-I

W M.M W P1


S1 W cl 10oy3 a 1
00 00 00


IE> >L LIN iEJ


com o r in Mvi c ciin


HHH Hto e




0 0 0 0 O 00O
SQ 0C) O



WH M 1 H 0 -1 0) L
Co oouoo gr H e





Q 0 O

U a-^ D OQ
Se v e I I I
0j (n I, n :$LL- LL-L L LH
00 0 W.~(1- U ILW W


> to -0>COCO -- -, 3 Q"

c ar' CO u Q ).II o
O c eLt-- .Ew .L; GJ, O 0

E a c o c .E
5-C t ZaI .o, CO 4
Uxa c E**- $ .
- O 0 t-
woo 0,-,.C qW.- W..O


-o r- !-


", o e = r.. :3
r 00o a C3
"4E .4 t0o ,
a a_ -0 V E 4 3
'0 r0 4- (1) 4 .

C- a mm a o 0 .






C4 cO-rf o Eo 3 .


a ve
0.-J s- .*-l .- -









SE"-?4-'..C 0Lzo -
-4 J ) .C


v -" .- t *e-i 0-- O "
S,--r-.- a-, S o, I




rC C- ..) to0o..
M r-- 1 W.-MC L)O (D 01 0%-






Li.-.'a. C ) U4 ( 4-' e
CO I C (0 CO 4--





00- C .4Jr- U'*~ 0 r m= -




u-0 (D u M t MO: Us^-^*-4 u~41 to C







9
Cholera toxin is a ADP-ribosyltransferase that modifies
the a-subunits of Gs (Cassel and Pfeuffer, 1978) and
transducin (Abood et al., 1982). The ADP-ribosylation site is
arginine specific, located at Arg'74 in a-Transducin within the
conserved peptide, Ser-Arg-Val-Lys.
The second enterotoxin, pertussis toxin, modifies the a-
subunits of Gi(Bokoch et al., 1983 and 1984), Go (Hurley et
al., 1984), and also transducin (West et al.,1985). This
second ADP-ribosylation site is located within three residues
of the COOH-terminus of a-transducin at Cys347 (S-glycoside,
Hurley et al., 1984). This site resides within a conserved
peptide, Asp-Cys-Gly-Leu.
Various means have been employed to identify the
nucleotide binding sites of most of these proteins. The first
to be well resolved was elongation factor Tu (EF-Tu) from
Escherichia coli, by means of X-ray crystallographic studies
(la Cour et al., 1985; Jurnak, 1985). De Vos et al. (1988)
determined the crystal of cellular ras p21 to the same
resolution as EF-Tu, 2.7 A. Essential features of both
structures are similar, only the structure of EF-Tu will be
discussed below.
Three domains of E.coli EF-Tu were found and the best
resolved was the GTP binding domain located at the NH2
terminal portion of the protein, comprising the first 200
residues. The GTP/GDP ligand site was found to be located at
the COOH terminal end of a sheet of four parallel B strands.







10
The guanine nucleotide was in the anti-conformation in respect
to the sugar and found in an unusual location, at the outer
edge of the domain, rather than within a hydrophobic pocket.
The guanine ring is partially buried in a cavity and appears
to be fixed in position by the side chains of four key
residues, Asn135-Lys-Cys-Asp. These residues represent a
highly conserved consensus sequence, of Asn-Lys-X-Asp, found
in region G (figure 1-1) of all GTP-binding proteins. The
ribose was exposed to the solvent. Asn135 and Asp38 lie near
the guanine and can bind the keto and amino substituents of
the guanine ring, respectively. Residues in both the E and G
region reside in two B-strands that interact with one side of
the guanine base (Jurnak, 1985, and la Cour et al., 1985).
Mutagenesis of Lys1, in conserved region G of bacterial
EF-Tu to an Arg or a Glu reduces the ability of the protein to
bind guanine nucleotides by about 20 or 100 fold, respectively
(Hwang et al., 1989). In the ras p21 protein, mutagenesis of
the corresponding Asp16 to Lys or Tyr in this same conserved
region abolishes GTP binding activity in this protein (Shih et
al., 1988).
In regions A and C (figure 1-1) are residues that
interact with the pyrophosphate group. Crystallographic
analysis by Jurnak (1985) and la Cour (1985) revealed that Mg2*
is located close to the B-phosphate of GDP, forming a salt
bridge to Asp80 of EF-Tu. This residue was found in the
conserved peptide, Leu-Asp-Thr-Ala-Gly, in region C. In the A







11
region, the sequence Gly-X4-Gly-Lys-Ser/Thr is found in all
GTP-binding proteins examined so far. Lysine24' in bacterial
EF-Tu, is at the NH2-terminus of a a-helix which has been
postulated to form a positive dipole that partially
neutralizes the negative charge of one the phosphates (Hol et
al.,1978, Jurnak, 1985). Additionally, Ohmi et al. (1988)
affinity labeled the corresponding Lys,1 in the ras oncogene
product p21 with guanosine diphospho- and triphospho-
pryridoxals. These observation suggest that this residue in
the conserved sequence is located in the guanine nucleotide
binding site, close to the B- or y-phosphate group of the
nucleotide.
Site directed mutagenesis in a family of closely related
genes, ras, which encode a group of homologous 21 Kda
proteins, has yielded additional information about guanine
interactions with these GTP-binding proteins. In region A the
conserved region Gly"1-X-Gly-Lys-Ser/Thr in ras p21 forms a
loop that interacts with the pyrophosphate group as described
above. Shih et al. (1988) mutated individually the Gly
residues at codons 10 and 15, to Val in ras p21 and found the
GTP binding was reduced by 33 and 1000 fold, respectively.
In Harvey and Kirsten sarcoma viruses, oncogenic ras
genes are found that contain a point mutation in this same
conserved region, resulting in the replacement of Gly12 for Val
(Taparowsky et al., 1982, Tabin et al., 1982, and Reddy et
al., 1982). Other transforming point mutations have been







12
identified and are discussed fully in the review by Barbacid
(1987).
Site directed mutagenesis of Gly12 in the cellular ras
proto-oncogene resulting in the substitution of any other
amino acid except proline for this residue conferred
transforming properties on the ras p21 gene product. The
authors postulated that a-helical structure in this region was
required for transformation (Seeburg et al., 1984). Insight
into the molecular events disturbed by this point mutation was
brought forth by McGrath et al. (1984). They found that this
single point mutation, resulting in the substitution of Gly12
by a Val, not only reduced GTP binding activity but also
impaired the intrinsic GTPase activity of ras p21 protein by
at least 10-fold. The transformation of the protoncogene
product, ras p21, was attributed to this lowered rate of
GTPase activity. Kinetic analysis of the hydrolysis of GTP by
p21 ras revealed that the mutant (Asp12) protein had a lowered
rate of hydrolysis and GDP release, in comparison to that of
the proto-oncogene (Val12), consistent with the reduced GTPase
activity found in this mutant (Neal et al., 1988).
Protein Synthesis in Mitochondria
The mechanism of protein synthesis in mammalian
mitochondria is poorly understood. Mammalian mitochondrial
ribosomes have been generally characterized as procaryotic in
nature due to some shared antibiotic sensitivities (Denslow
and O'Brien, 1974) and that both utilize f-Met-TRNA for







13
polypeptide initiation. However, a number of differences
exist. Susceptibility to lincosamies and macrolides is much
lower than in procaryotes suggesting some components of the
peptidyl transferase center are altered (Denslow and O'Brien,
1978). Additionally, mammalian mitochondrial ribosomes are
not sensitive to the antibiotic kirromycin, which stabilizes
the binding of the EF-Tu-GTP-aminoacyl-tRNA complex to the
bacterial ribosome, thereby inhibiting polypeptide elongation
(Schwartzbach and Spremulli, 1989).
In regard to their fine structure and physicochemical
properties, mitochondrial ribosomes differ considerably from
other kinds of ribosomes. The mass of the mitoribosome is
approximately the same as the procaryotic ribosome, but the
buoyant density is much less. This is due to the fact that
mitoribosomes contain only half the amount of rRNA and twice
the protein as bacterial ribosomes (Matthews, et al., 1982).
The number of ribosomal proteins is also much greater in the
mitoribosome (table 1-1). In addition, mitoribosomal proteins
do not appear to have any obvious homologue in the bacterial
or cytoribosome as identified by electrophoretic or immuno-
logic cross-reactivity (Pietromonaco et al., 1984 and Matthews
et al., 1982). Based upon this evidence and functional
studies described below, O'Brien and Matthews (1976) have
assigned mitochondrial translation systems to four separate
classes (Protista, fungi, plants, and animals).







14
Table 1-1
Protein and rRNA Content of Ribosomes


Source Ribosome Number of rRNA
(S units) Proteins (S units)

Bovine 28S" 33 12S
Liver 39Sb 52 16S
Mitochondria 55Sc 85
Procaryotic 30S8 21 16S
E. coli 50Sb 32 23S, 5S
70Sc 53
Mammalian 40S" 33 18S
Eucaryotic 60Sb 50 28S, 5.8S,5S
Cytoplasmic 80Sc 83

"Ribosome small subunit
Ribosome large subunit
cMonosome

The mechanism for binding mRNA in mitoribosomes appears
to differ from that of procaryotic and eucaryotic cytoplasmic
ribosomes. In procaryotes a leader sequence 5' to the
initiation codon (AUG) binds to a complementary sequence on
16S rRNA, called the Shine-Dalgarno (1974) sequence (Steitz
and Jakes, 1975). This mRNA binding site is absent in
mitochondrial mRNAs which often begin with the initiation
codon or have very short leader sequences which have no
complement on 12S rRNA (Clayton, 1984). Eucaryotic mRNAs also
lack any complementary binding to rRNA. However, they do have
a modified 5'-terminal modification, the m7G(5')ppp cap which
functions as a recognition signal (Shatkin, 1976) for a cap







15
binding protein. Once the cap binding protein binds the
modified mRNA it scans the 5'-untranslated mRNA leader for the
proper initiation codon. Though the mechanism for initiation
codon recognition in the mitoribosome is unknown, Denslow et
al. (1989) did discover that the major interaction of the mRNA
on the mammalian mitoribosome, covered a 30-nucleotide
segment. Also, the binding of authentic mitochondrial mRNAs
required homologous initiation factors.
Heterologous translocases and initiation factors from
several sources were tested to determine if they would
function with mammalian mitoribosomes. With few exceptions,
homologous factors were required to support in vitro poly
(U)-dependent phenylalanine polymerization in mammalian
mitoribosomes. The only heterologous factor that has been
shown to support poly (U)-directed systems in animal
mitoribosomes was EF-Tu from a variety of sources (E. coli, B.
subtilis, and a protist Euglena gracilis, Eberly and
Spremulli, 1985). Denslow et al. (1988) also showed that
bacterial IF-3 could bind to small subunits of mitoribosomes
and inhibit formation of the monosome as it does in E. coi .
Bacterial EF-G (Denslow and O'Brien, 1979) and other trans-
locases from gram-positive and gram-negative procaryotes,
protist, plants chloroplastt) and eucaryotic cytoplasmic
origins did not support in vitro phenylalanine polymerization
(Eberly et al., 1985). This was surprising in view of the
fact that a postribosomal supernatant from E. coli would







16
function in a poly (U)-directed system in Neurospora crassa
mitochondrial ribosomes (Grandi et al., 1971). This suggested
that mammalian mitochondrial ribosomes may differ signifi-
cantly from other types of mitochondrial ribosomes in their
interactions with soluble initiation and elongation factors.
The first reported isolation of any elongation factors
from mammalian mitochondria was EF-Tu/Ts complex from bovine
liver by Schwartzbach and Spremulli (1989).
Proposal
GTP binding proteins are an essential component in pro-
tein synthetic machinery in procaryotic and eucaryotic cyto-
plasmic systems. Though much is known about protein synthesis
in procaryotes and much more is being discovered about the
eucaryotic cytoplasmic system, little is known about protein
synthesis in mitochondrial systems. In a search for a puta-
tive mitochondrial elongation factor, O'Brien et al. (1990)
discovered a unique GTP binding activity on the small subunit
of the mammalian mitochondrial ribosome. This GTP-binding did
not occur on soluble initiation or elongation factors, as is
the case for other ribosomal systems, but on the surface of
the mitochondrial ribosome, suggesting a novel mechanism for
initiation of protein synthesis. The objective of this
research was to identify the site of GTP binding on the small
subunit and further characterize this binding activity. Once
identified, the GTP-binding ribosomal protein was purified
and partial amino acid sequence was determined in an effort to
find any shared homologies to other GTP-binding proteins.














CHAPTER 2
NONCOVALENT BINDING OF MITOCHONDRIAL RIBOSOMES BY NTPs
Introduction
Guanosine-5'-triphosphate plays a key role in protein
synthesis in both procaryotic and eucaryotic systems. In both
systems GTP binds to soluble protein factors in a cyclic
manner to promote the initiation or elongation phases of
protein synthesis (Moldave, 1985, and Allende, 1988).
Elucidation of the translation process in mammalian
mitochondria has been difficult. The substitution of
heterologous translation factors from a variety of procaryotic
and eucaryotic sources generally fail to support poly (U)-
directed phenylalanine synthesis with mammal i an mitoribosomes.
Only EF-Tu from a variety of sources was capable of
functioning in mammalian mitoribosomal translation (Eberly et
al., 1985) as discussed above.
In an effort to isolate a putative mitochondrial EF-G
from bovine liver O'Brien et al. (1990) serendipitously found
that [8-3H]GTP binds to salt washed small subunits of
mammalian mitochondrial ribosomes. This is surprising due to
the fact that GTP-binding to activity in both procaryotic and
eucaryotic cytoplasmic translation systems occurs only on







18
soluble protein factors that cycle on and off the ribosome in
protein synthesis. This binding activity was resistant to
high salt washes (300 to 500 mM KCL) and independent of any
exogenous factors or fusidic which is known to stabilize EFG-
GTP complex on bacterial ribosomes. Factor independent
binding of GTP to procaryotic (E. coli) and eucaryotic
cytoplasmic (Bovine) ribosomes has never been reported and was
not observed when tested in this laboratory (Sonnega et al.
personal communication).
Although no function has as yet been associated with this
activity [8-3H]GTP was found to bind with high affinity (Kd=
15.3 nM) and in a stoichiometric manner to salt washed 28S
ribosomes (O'Brien et al., 1990). The large subunit (39S) and
intact monosomes (55S) did not bind GTP. The binding site was
specific for GTP and GDP and no other nucleotide tested could
compete with [8-3H]GTP binding to the small subunit. Binding
affinity for the 28S ribosome was equally high for GDP (Kd=18
nM).
Nonhydrolyzable analogs of GTP, GMPPNP and GMPPCP, also
bound to the small subunit, but with reduced affinity,
suggesting that binding site is sensitive to alterations in
the structure of the triphosphate side chain. Additionally,
[y32P]GTP was not hydrolyzed following binding to small
subunits. These findings suggest that salt washed ribosomes
do not posses an intrinsic GTP hydrolytic activity. This of







19
course did not rule out the possibility that GTPase activity
may require exogenous factors, as yet identified.
The following study will expand on the above information
in an effort to determine the specificity of this GTP-binding
activity. GTP-binding will be determined using standard
nitrocellulose filter binding assays and under various salt
conditions to determine the optimal ionic conditions for
binding.
Materials and Methods
Reagents
Guanosine[8-3H] 5'-Triphosphate was purchased as a
ammonium salt in a 50% ethanol solution at a specific activity
of 12.6 Ci/mmol and <99% purity (determined by manufacturer)
from Amersham Corporation (Arlington Heights, IL). Adenosine,
cytidine, guanosine, and uridine 5'[a32P] Triphosphate were
purchased as triethylammonium salts in aqueous solution at a
specific activity of 410 Ci/mmole from Amersham Corporation.
[a32p] Guanosine monophosphate was purchased also as a
triethylammonium salt from New England Nuclear (Boston MA) at
a specific activity of approximately 3000 Ci/mmol and of at
least 90% purity. 8-Azidoguanosine 5'[a32P] Triphosphate, was
purchased as a triethylammonium salt at a specific activity of
4.6 Ci/mmole from ICN Radiochemicals (Irvine, CA). All
unlabeled nucleotide-5'-triphosphates, guanosine di- and
monophosphates were Type I reagents, purchased as sodium salts
from Sigma Chemical Company. The NTPs were isolated from







20
Equine muscle free of vanadate ion and all other nucleotides
isolated from Yeast.
Mitochondrial Ribosome Preparation Solutions
Solution Formulation
A 0.34 M sucrose, 1 mM EDTA, 5 mM Tris-HCL, pH 7.5
B 0.26 M sucrose, 40 mM KC1 15 mM MgC1,, 0.8 mM
EDTA, 5 mM 2-mercaptoethanol (BME), 50 AM
spermidine, 50 pM spermine, 14 mM Tris- HC1, pH 7.5
C 100 mM KC1, 20 mM MgCl2, 0.01% Triton X-100, 5 mM
BME, 20 mM triethanolamine (TEA), pH 7.5
D 40 mM KC1, 15 mM MgCl2, 0.05 mM EDTA 5mM BME 1.6%
Triton X-100, 0.005% diethylpyrocarbonate (DEPC),
10 mM Tris-HC1, pH 7.5
E 300 mM KC1, 40 mM MgCl2, 5 mM BME, 1.6% Triton X-
100, 0.005% DEPC, 10mM, Tris-HCl, pH 7.5
F 300 mM KC1 5mM MgCl2, 5 mM BME, 10 mM Tris-HCl, pH
7.5
G 0.34 M sucrose, 25 mM KC1, 5 mM MgC1,, 5 mM BME,
10 mM Tris-HC1, pH 7.5
Binding Assay Solutions
Reaction Buffer A: 100 mM KC1, 5 mM MgC1,, 10 mM Tris-HC1,
pH 7.4, and 1 mM BME
Reaction Buffer B: 10 mM NH4C1, 20 mM MgC12, 10 mM Tris-HC1,
pH 7.4, and 1 mM BME
Reaction Buffer C: 100 mM KC1, 20 mM MgCl2, 10 mM Tris-HC1,
pH 7.4, and 1 mM BME









B-Scintillation Solution
Liquid Scintillation cocktail: 0.5% PPO and 0.05% (w/v) POPOP
in toluene
Equipment
Beckman LS8000 B-liquid scintillation counter
Millipore 30 place vacuum manifold
Napco Vacuum oven, model# 5831
Preparation of Mitochondrial Ribosomes
Bovine mitochondrial ribosomes were prepared from bovine
liver as described previously (Matthews et al., 1982).
Mitochondria prepared from liver homogenates in buffer A were
treated with 100 pg/ml digitonin to remove microsomal
membranes and then stored at -70C in buffer B until needed.
To isolate the ribosomes, thawed mitochondria were lysed in
1.6% Triton X-100 and clarified by centrifugation at 10,000
rpm for 45 min in a Beckman JA-10 rotor at 40C. Bovine
mitoribosomes were concentrated by absorption to DEAE
cellulose in buffer D and eluted with high salt (buffer E).
The crude mitoribosomes eluted from the DEAE cellulose were
layered on a 34% sucrose cushion (20ml) prepared in buffer C
and centrifuged at 35,000 rpm for 24 h in a Beckman type 35
rotor. The pellets were resuspended in buffer C, treated with
1 mM puromycin to remove nascent polypeptides, and centrifuged
in 10-30% sucrose density gradients prepared in buffer F to
yield 28S and 39S subribosomal particles. Dissociation of
subribosomal particles in high salt buffer (buffer F with 300







22
mM KC1) removed any soluble initiation or elongation protein
factors. Subribosomal particles were recovered from gradient
fractions by centrifugation at 100,000 x g in a Beckman Ti 60
rotor for 15 h, resuspended in buffer G and stored at -700C
until needed.
Binding of GTP to 28S Ribosomes and Analysis by Millipore
Filter Assay
Binding activity was determined by means of a Millipore
filter assay. In this assay, ribosomal subunits with bound
radiolabeled NTP were absorbed to nitrocellulose and washed to
remove unbound NTP. The reaction mixture consisted of a fixed
concentration of ribosomal protein (0.1 pM) with varying
concentrations ( 0.05 to 5 pM) of radiolabeled NTP, in
reaction buffer (A, B, or C). The specific activity of the
radiolabeled NTP (specific activity was typically 410 Ci/mmol)
was reduced by dilution with the appropriate unlabeled NTP
just prior to use. The final volume was adjusted to 50 pl
with reaction buffer in a 1.5 ml conical plastic eppendorf
tube and incubated for 5 min on ice. The reaction was stopped
by the addition of 0.5 ml of ice cold reaction buffer and
immediately transferred to a wetted nitrocellulose filter.
The filter was in a Millipore 30 place manifold under vacuum
to remove the aqueous buffer. Another 0.5 ml volume of
reaction buffer was used to rinse out the reaction tube and
also passed through the nitrocellulose filter. The filters
with bound protein were immediately washed with an additional







23
8 ml of ice cold reaction buffer also under vacuum. The total
time for the wash steps was approximately 15 seconds.
The nitrocellulose filters were placed in empty 20 ml
scintillation vials and dried under vacuum and at 70C for
approximately 1 h. The dried filters were solubilized in 5 ml
of scintillation cocktail solution (0.5% PPO, and 0.05% w/v
POPOP in toluene). The filters were counted in a B-
scintillation spectrometer (Beckman, model #LS8000) at a
counting efficiency of 21% and 88% for 3H and 32P,
respectively. To determine the extent of nonspecific binding
of radiolabeled NTPs to nitrocellulose, control incubations
were done in an identical manner, but in the absence of
ribosomes. The amount of radiolabeled NTP that nonspecific-
ally bound to the filter was then subtracted from the corres-
ponding treatment. This nonspecific binding was typically <
1% of the experimental treatments. Additionally, total
amounts of radiolabeled NTPs (the same concentration and
specific activity of radiolabeled NTP used in the experiment)
were determined for each experiment. This was done by pipet-
ting an appropriate amount of radiolabeled NTP onto a wetted
nitrocellulose filter without washing, drying, then counting.
Binding data were plotted the aid of personal computer
using the Enzfitter Kinetics program by Leatherbarrow (1987).
Data was expressed as pmol of radiolabeled NTP bound per pmol
of ribosomes (R) versus the concentration of free NTP (C)







24
using an equation for ligand binding from Edsall and Wyman
(1958) as shown below:
R = n(C) / Kd + C
where: C= concentration of free ligand; Kd= dissociation
constant of ligand; and R= pmol ligand bound/ pmol ribosome
The binding data was linearly transformed by Scatchard
analysis (Scatchard, 1949) by the following equation:
R/C = -1/Kd (R) + n/Kd
where: R= pmol ligand bound / pmol ribosome; Kd= binding
dissociation constant n= number of binding sites; C=
concentration of free ligand
Binding inhibition studies were done by incubating 0.1 pM
28S ribosomes with increasing concentrations of (0.5 to 10 pM)
unlabeled NTP for 1 min in reaction buffer for 1 min on ice
under dim light. [a3P]GTP is then added at a saturating
concentration (0.1 pM) for 5 min under the same conditions.
The samples are then washed and counted as described above for
the amount of [a32P]GTP that is bound to the ribosome in the
presence of increasing concentrations of NTP.
Binding inhibition data were plotted as percent of
control (Y) versus inhibitor concentration with the aid of
personal computer using the Enzfitter Kinetics program by
Leatherbarrow (1987). A modified equation for binding inhibi-
tion (Cantor and Schimmel, 1980) was used and is shown below:
Y= 1-[ l-r(I/Ki)/ l+r[l+I/Ki] ]xl00
where: r= Kd/A=constant (Kd for GTP= 15.2 nM and A=
concentration of GTP); Y=l-vi/vo (vi/vo = percent inhibition).







25
Results
Binding of NTPs to 28S Ribosomes
O'Brien et al. (1990) found that [83H]GTP binds with high
affinity to 28S ribosomes and appears to be specific for GTP
and GDP only. To further investigate the nucleotide
specificity of binding for the small subunit, a number of NTPs
were tested for their ability to bind to ribosomes and also
their ability to compete for [ac32P]GTP binding to the 28S
ribosome. Millipore filter binding assays (figure 2-1 and
table 2-1) confirmed that [a32P]GTP did bind to 28S ribosomes
with high affinity (Kd of 14.3 nM) and in nearly unit
stoichiometry which agrees with the result (Kd = 15.3 nM) of
O'Brien et al. (1990). The only other [a32P]NTP tested that
was found to bind to this ribosome in a stoichiometric manner
was GMP. GMP binds with approximately 800 times lower
affinity to the 28S ribosome than GTP. [a32P]ATP, UTP, and CTP
did not bind to the 28S ribosome to any appreciable degree.
Binding inhibition studies were done to determine which
nucleotides could displace [a32P]GTP from the 28S ribosome.
Denslow et al. (personal communication) found that of the NTPs
(GDP, ATP, CTP, and UTP) tested only GDP was able to compete
for [a32P]GTP binding. In this experiment additional
nucleotides were tested to determine the structural
requirements for this GTP binding activity. In agreement with
the results of O'Brien et al. (1990), GDP did compete strongly
for [aC32']GTP binding to the ribosome with a resulting Ki of














0.8

o 0 GTP
S 0.6 GMP
m A ATP

. 00.4
S- A UTP
z --CTP
z 0.2 T
S............ ............
M A : ... -A- .. I-_
o.o I------
I I I I I I I I I I I
0.0 2.0 4.0 6.0 8.0 10.0
FREE NTP (X 106M)



Figure 2-1 Noncovalent binding of 1a32P]GTP, GMP, ATP, UTP,
and CTP to 28S bovine mitochondrial ribosomes. Binding of the
radiolabeled nucleotides to the ribosomes was determined by
the Millipore filter binding assay. The same ribosomal
preparation was used for all nucleotides and the assay was
done in reaction buffer A (100 mM KCL, 5 mM MgC2, 10 mM Tris-
HC, pH 7.4, and 1 mM BME) as described in Methods. The
initial specific activity of the [a P]NTPs was 410 Ci/mmol and
was diluted approximately 160 fold with unlabeled NTP to a
specific activity of 2.15, 1.67, 3.39,and 3.07 Ci/mmol for
GtP, ATP, CTP, and UTP, respectively just prior to use.
[a P]GMP was initially at a specific activity of 3000 Ci/mmol
and diluted with GMP to a specific activity of 10.2 Ci/mmol
just prior to use. The concentration of the ribosomes was 0.1
pM and the radiolabeled nucleotides varied from 0.05 to 10 pM.
Each data poipt represents the mean of duplicate samples
except for [a P]GTP where individual data points are shown
without averaging. Refer to table 2-1 for Scatchard analysis
of this data and to Methods for additional experimental
details.







27
9.2 nM (figure 2-2 and table 2-2) suggesting that GTP and GDP
bind to the same site on the 28S ribosome equivalently.
Additionally, ITP and GMP were able to compete to a lesser
extent, with binding inhibition constants (Ki) of 124 and 1140
nM, respectively.


Table 2-1
Noncovalent Binding of Various Fa32PI Nucleotides to 28S
Ribosomes


Nucleotide (Kd SE)8 Number Binding Sites
(nM) (n SE)

GTP 14.1 2.3 0.80 0.02
GMP 11,500 1,300 1.00 0.06
ATP 77 44 0.09 0.01
CTP ND ND
UTP 1,200 990 0.07 0.02

"Sptchard analysis of noncovalent binding of [a32P]NTP and
[a P]GMP to 28S ribosomes shown in figure 2-1. Data
represents binding dissociation values (Kd) and standard error
(SE) from a single binding experiment. ND: Not detected



The ionic conditions used in which these binding assays
were modified from those used in earlier work by O'Brien et
al. (1990) in order to determine optimal stringencies for the
study of this GTP binding activity. The original work by
O'Brien et al. (1990) used reaction buffer B (10 mM NH4CL, 20
mM Mg2CL2,, 10 mM Tris-HC1, pH 7.4, and 5 mM BME) which was
















100
\~~~ -" ------
6 80 A UTP
z IO CTP
8 ATP
L. 60 OdGTP
0
z 40 GMP
0o V ITP
S 20 GDP
W 20
a-

. I I .
0 2 4 6 8 10
INHIBITOR CONCENTRATION (X106M)





Figure 2-2 Binding inhibition of [aCP]GTP to bovine
mitochondrial ribosomes by various nucleotides. Displacement
of bound [a P]GTP (O.I1pM) from ribosomes (0.1 pM) by
increasing concentrations (0.5 to 10 kM) of competing
nucleotide was dane in reaction buffer A as described in
Methods. The [a P]GTP was diluted with GTP to a specific
activity of 21.6 and 14.8 Ci/mmol just prior to use for two
separate binding inhibition experiments (GDP, GMP,
dGTP,ATP,CTP, and GDP, UTP, ITP respectively}. The same
preparation of ribosomes and chemical lot of [a PJGTP was
used in both experiments. The binding of [a P]GTP to
ribosomes in the presence of competing unlabeled nucleotide
was compared to binding in the absence of the competing
unlabeled nucleotide and represented as percent of control.
The data are single data points and each curve represents a
single experiment except for GDP where each data point is a
mean of two experiments conducted on consecutive days. Refer
to table 2-2 for analysis of binding inhibition for each
nucleotide described above.











Table 2-2
Binding Inhibition for ra 3P1 GTP Binding to 28S Ribosomes by
Various Nucleotides
-----------------------------"'----------------""

Nucleotide Ki SE'
(nM)

GDP 9.2 1.0
ITP 124 19
GMP 1,140 270
dGTP NC
CTP NC
UTP NC
ATP NC

"B pding inhibition (Ki) of various nucleotides to compete for
a P]GTP binding to the 28S ribosome. Data represents values
from a single binding experiment and standard error, shown in
figure 2-2, for a single binding inhibition experiment.
NC: No binding inhibition


developed by Bodley (Bacca, et al., 1976) for the study of GTP
binding to bacterial EFG. Reaction buffer A (100 mM KC1, 5 mM
MgC12, 10 mM Tris-HC1, pH 7.4, and 1 mM BME) was developed
during the progress of the present study and was used in the
above binding and inhibition binding studies. The development
of these ionic conditions will be described later in this
paper. At this point it suffices to say that GTP binds
specifically to the 28S ribosome and quantitatively in the
same manner over a wide range of ionic conditions and
stringencies.







30
Both [8-3H GTP and rI3Pl8-Azido GTP Bind to the Same Site on
the 28S Ribosome
In order to determine if the 8-Azido analog of GTP would
be a suitable photoaffinity probe for labeling the GTP binding
site on the 28S ribosome, filter binding and binding
inhibition studies were performed. In an experiment directly
comparing binding of [ac32P]8-Azido GTP and [8-3H]GTP to 28S
ribosomes, the GTP analog did bind to the small subunit with
an apparent dissociation constant of 1.35 pM (figure 2-3).
The dissociation constant for [8-3H]GTP was 18.1 nM which is
in close agreement with that found by O'Brien et al. (Kd=15.3
nM, 1990) and both bound in unit stoichiometry.
A mean dissociation factor for [a32P]8-Azido GTP and
[a32P]GTP have been determined (table 2-3). These values were
1.92 0.84 pM and 15.2 8.64 nM, respectively. Both
compounds bound to the ribosome in a stoichiometric manner
(1.2 0.4 and 0.84 0.17 sites occupied per ribosome,
respectively). These data suggest that the GTP analog binds
to the 28S ribosome with an affinity approximately 125 times
less than GTP. These mean Kd values for 8-Azido GTP and GTP
will be used throughout the remainder of this paper.
GTP was found to compete with [a32P]8-Azido GTP for
binding to the 28S ribosome with an apparent Ki of 106 nM
(figure 2-4). This suggests that GTP and its photoreactive
analog bind to the same binding site on the 28S ribosome.
These data imply that [a32P]8-Azido GTP could be used as a
photoaffinity probe for the identification of the GTP binding
site on the 28S ribosome.
















1.2
ILJ
=5 1.0
0
(U
0
D 0.8



z 0.4 4 GTP
z O 8N3GTP
o 0.2
m
0.0

0.0 0.2 0.4 0.6 0.8

FREE NTP (X 106M)








figure 2-3 Noncovalent binding of [aP3]8-Azido GTP and [8-
H]GTP to 28S bovine mitochondrial ribosomes. Binding of the
radiolabeled nucleotides to the ribosomes was determined by
the Millipore filter binding assay as described in Methods.
The same ribosomal preparation was used for both nucleotides
and the assay was done in reaction buffer B (10 mM NH4CI, 20
mM MgC12, 10 mM Tris-HC1, pH 7.4, and 1 mM BME) as described
in Methods. The specific activity of [8-3H]GTP and [a P]8-
Azido GTP was 3.1 and 0.98 Ci/mmol, respectively. Each data
point represents the mean of duplicate samples. Scatchard
analysis for the binding of GTP and its 8-Azido analog to 28S
ribosomes revealed an apparent Kd of 18.1 and 1.35 pM,
respectively.

















0 1.0

S 0.8 -Ki = 106nM
-J
0


,. 0.4


0 1 2 3 4 5 6 7 8
I---

S 0.2-
0

Sti 0.0i a t o i i i -
g I I I I I | I I p I I I I P I I
0 1 2 3 4 5 6 7 8
INHIBITOR CONCENTRATION (X 106M)






Figure 2-4 Binding inhibition of [a3P] -Azido GTP to bovine
mitochondrial ribosomes by GTP Displacement of saturating
concentration (1.67 iM) of [a P]8-Azido GTP from 0.083 uM
ribosomes by increasing concentrations (0 to 16.7 pM) of GTP
was done in reaction buffer B as described in Methods. The
specific activity of [a P]8-Azido GTP was 0.98 Ci/mmol The
same preparation of ribosomes and lot of [a P]8-Azido GTP was
used in this experiment and the one described in figure 2-3.
The binding of [a P]8-Azido GTP to ribosomes in the presence
of competing unlabeled nucleotide was compared to binding in
the absence of the competing unlabeled nucleotide and
represented as percent of control. The data are mean values
from duplicate samples and each curve represnts a single
experiment. GTP was found to compete for [a P]8-Azido GTP
binding to 28S ribosomes with an apparent Ki of 106 nM.











Table 2-3
Noncovalent Binding of GTP and 8-Azido GTP to 28S Ribosomes
---------------------""'''-----------'-----------

Nucleotide Reaction (Kd SD) Binding Sites
Bufferc (nM) (n SD)

[8_3H & a32P1GTPa A 15.2 8.64 0.84 0.17
[a32P]8-Azido GTPb B 1,920 a 840 1.20 0.40

"Data for the binding of [8-3H] and [a32PGTP to 28S ribosomes
are the mean and standard deviation of 7 experiments.
bData for the binding of [a32P18-Azido GTP to 28S ribosomes are
the mean and standard deviation of 3 experiments.
cFormulations for reaction buffers A and B are listed in
Materials


Discussion
There is one (stoichiometric) nucleotide binding site on
the 28S ribosome and it is specific for GTP and GDP.
Additionally, O'Brien et al. (1990) suggested that the binding
may occur on the interface of the ribosomes because [8-3H]GTP
did not bind to 55S monosomes. However it is not ruled out
that monosome formation induces a conformational change in the
GTP binding site, preventing binding of the nucleotide.
Only GTP and GDP bind the small subunit, and they do so
equally and with high affinity. Most GTP binding proteins
bind GDP more tightly than GTP, requiring an exchange factor
in order to regenerate the active GTP-bound protein. However







34
there are several instances where GTP-binding proteins do have
equal affinities for both GTP and GDP, such as the photo-
receptor G-protein of frogs (Robinson et al., 1986), E. coli
EF-G (Kaziro et al., 1978), and p21 ras protein (Shih et al.,
1988).
The structural requirements for nucleotide binding in the
GTP/GDP binding site on the 28S ribosome were elucidated from
binding and binding inhibition assays as described above. No
difference were seen in binding affinities for GTP vs GDP
suggesting that the loss of the y-phosphate does not perturb
binding. However the loss of the B-phosphate as seen in GMP
does decrease binding affinity by at least 100-fold (table 2-
2). Additionally, O'Brien et al. (1990) reported that
nonhydrolyzable analogs of GTP, GMPPNP and GMPPCP, do bind the
28S ribosome, but with much reduced affinity. This was
attributed to alteration of the P,-X-P angles in these analogs
perturbing the binding interactions with the pyrophosphate
group (Yount, 1971).
Substitutions on the guanine ring also appear to be
poorly tolerated (see figure 3-1 for ring numbering system).
The substitution of the purine by a pyrimidine ring as seen
for CTP and UTP completely abolishes the ability of the
nucleotide to bind to or compete for [a32P]GTP binding to the
28S ribosome (table 2-1 and 2-2, respectively). This suggest
that functional groups on the base are misaligned in the
single ring base, preventing interactions with the binding







35
site. Inosine triphosphate did compete partially with
[a32P]GTP binding to the small subunit suggesting that the loss
of the amino group at C2 of GTP can be tolerated to some
extent. However the loss of the same amino group and the
substitution of an amino for the keto group at C6 as in ATP,
abolished all binding activity, suggesting that the keto group
was required for binding. Both the keto and amino group of
the guanine base have been shown in crystallographic studies
to bind to Asn and Asp residues, respectively in the highly
conserved G region (figure 1-1) in bacterial EF-Tu and ras p21
(Jurnak, 1985, la Cour et al., 1985, and De Vos, et al.,
1988).
Alterations of the ribose moiety are not tolerated
either. dGTP did not displace [a32p]GTP from the binding site,
indicating the importance for the hydroxyl group at C2, on the
ribose. O'Brien et al. (1990) also reported that periodate
cleaved GTP did not bind the small subunit suggesting that any
distortion in the conformation of the ribose affects its
binding. Additionally, these investigators found that n-
ethylmaleimide (NEM), a sulfhydryl modifier had no effect on
GTP binding activity, suggesting that a cysteine residue is
not located near critical GTP-binding residues.
The addition of the azido (N3) moiety at C8 of the guanine
base reduces the affinity of this GTP analog for the small
subunit (Anders et al., 1989). This could be attributed to
stearic hindrance of the planar azido group and a change to







36
the syn conformation of the base preferred by 8-azido purines
(Colman, 1983). In crystallographic studies of EF-Tu and ras
p21 the GTP molecule resides in the GTP binding pocket in the
anti conformation (Jurnak, 1985, la Cour et al., 1985, and De
Vos et al., 1988) suggesting this conformational change could
contribute, at least in part, to the lowered affinity of this
GTP analog for the 28S ribosome.
This nucleotide binding activity on the ribosome, which
is highly specific for GTP and GDP, is consistent with other
GTP-binding proteins. GTP-binding proteins bind GTP in the
nM to low pM concentration range and GDP commonly in the low
nM range (see reviews by Stryer, 1986; Gilman, 1987; and
Allende, 1988). They also have a single nucleotide binding
site that is specific for GTP and GDP. In a comparison of
nucleotide specificity, number of binding sites, and binding
affinity, 28S ribosomes bind GTP and GDP in a similar manner
as reported for other GTP-binding proteins. The only possible
difference is that an intrinsic GTPase activity has not yet
been observed on the salt washed 28S ribosomes. This does not
rule out the possibility that an exogenous factorss, as yet
unidentified, may be required for GTP hydrolysis.














CHAPTER 3
PHOTOAFFINITY LABELING OF 28S RIBOSOMES WITH 8-AZIDO GTP
Introduction
Photoaffinitv Labeling of GTP-Bindina Proteins
Labeling receptor molecules with photoreactive analogs of
effector molecules has become a popular method in elucidating
the mechanisms by which their natural analogs regulate
biological phenomena. The chemical approach to receptor site
labeling is still used but suffers from the necessity to
contrive a molecule that can bind in sufficient concentration
to ensure preferential reaction with suitable residues that
can react with the modified ligand. Most of these reagents
are electrophiles which leads to selective incorporation that
may not reveal the residues in close proximity to the binding
site of the ligand.
Photoreactive ligands can be constructed with addition of
one or just a few photosensitive constituents available today.
The photoreactive moieties commonly used are either diazo- or
azido-derivatives of their natural ligands (i.e.,ATP, cAMP,
GTP, cGMP etc.) and which should bind their receptor, assuming
sufficient latitude existed in the ligand-receptor recognition







38
process. They remain unreactive until irradiated with light
by the investigator. Irradiation of the diazo- or azido-
purine analog provides the energy to form a reactive carbene
or nitrene radical, respectively, that may form a covalent
bond to residues at or near the binding site in a nonselective
manner. These reactive intermediates do display some
electrophilic character. However, a-carbonyl carbenes and
aryl nitrenes, which have been used extensively to identify
nucleotide binding proteins, display wide chemical reactivity
by being able to insert into C-H bonds and other heteroatomic
bonds (Cooperman, 1980).
The ability to control the labeling event and provide
relatively nonselective incorporation of modified ligands has
made this a popular technique for the identification of purine
binding proteins. With the development and commercial
production of 8-Azido analogs of ATP, cAMP, GTP, and cGMP by
the methods of Haley, (1976), and Czarnecki et al. (1979), the
use of these reagents has increased. 8-Azido GTP was used in
this study. Therefore, the following discussion will be
limited to the nitrene, and not carbenes, even though the two
are of the same isoelectronic specie (six valence electrons)
and their chemistries are expected to be qualitatively
similar. Some examples are; azido analogs of GTP have been
used as photoaffinity probes to study GTP-subunit interaction
in adenylate cyclase (Pfeuffer, 1977), tubulin (Potter and
Haley, 1983), transducin (Kohnken and Mc Connell, 1985), and







39
bacterial EF-G (Girshovich and Kurtskhalia, 1979). The number
of applications of these reagents is vast and the reader
should consult reviews by Bayley and Knowles, 1977; and
Chowdry and Westheimer, 1979).
The photoreactive azido compounds remain unreactive in
the absence of light, thus allowing the investigator to bind
the ligand to the binding site before activating the
photoreactive reagent. Once bound, the sample is irradiated
with uv light and a molecule of N2 is photoeliminated and a
reactive nitrene forms (figure 3-1). This nitrene is highly
reactive (Turro, 1980; Bayley and Knowles, 1977; and Chowdhry
and Westheimer, 1979) and it can insert into heteroatomic
bonds of residues at or near the binding site, in a
nonselective manner (Richards and Konigsberg, 1977). The
type of bond formed is dependent upon the nature of the
nitrene. The nitrene comes in two forms; a very reactive
singlet nitrene that prefers to insert into 0-H and N-H bonds,
and the less reactive triplet nitrene that prefers C-H bonds
(Bayley and Knowles, 1977). The nitrene atom in both species
is linear in shape and the orbital occupancy is two
filled(sp2)4 orbitals for the singlet, and one filled (sp2)2
and two p orbitals (px)' and (pz)' for the triplet nitrene
(Turro, 1980). The singlet nitrene can covalently attach
itself in a single step reaction such as the addition to a C=C
bond, nucleophilic attack of N-H or 0-H bonds, or direct
insertion into C-H bonds in the backbone or sidechains of





























HO

0 0 0
HO-P-O-P-O- P-O-CH,
OH OH OH H
H H
HO OH


AZIDE FORM


0



HN" H
HO-P-O--P-O-"P-OCH,
OH OH OH H, H

"HO OH


NITRENE FORM


Figure 3-1 Generation of an aryl nitrene following uv
irradiation of 8-Azido GTP. Formation of the reactive nitrene
is accompanied by the photo-elimination of a molecule of Nz.







41
residues in the polypeptide residing near the nitrene while in
the binding site. The triplet nitrene requires a coupled
reaction before a covalent bond can form. The first reaction
can be an abstraction of an electron from a C-H bond. This
generates a short lived radical ('CH,) that can couple with
reduced nitrene to form a useful covalent bond or form some
other product. The only other major reactions are
rearrangements, to which aryl nitrenes are not susceptible
(Bayley and Knowles, 1977). The 8-Azido purine analogs do not
form tetrazoles (an isomer that forms a third five membered
nitrogen ring attached at C8 to C9 of the purine ring) in
aqueous solvents to any extent (Haley, B., personal communi-
cation). These isomers of 8-Azido ATP or GTP are poor photo-
reactive reagents and would quench photoaffinity reactions.
The productive covalent attachment of a ligand to a
receptor by photololysis is a competition between the rates of
photoincorporation and rates of decomposition plus
decomposition of the reagent. The rate of photoincorporation
consists of the formation of the reactive nitrene which in
turn must react with surrounding amino acid residues to insert
into heteratomic linkages such as C=O, C-S, C-N, C-H, N-H or
water. The estimates for this total reaction rate in aryl
nitrenes are relatively long, approximately 104sec when
measured in a soft polystyrene matrix (Reiser et al., 1968).
This suggests that if binding affinities are strong enough, at
least in the millimolar range, then the ligand should reside







42
long enough in the binding site for productive photoinduced
covalent binding to the receptor (Chowdry and Westheimer,
1979; and Potter and Haley, 1983. Additionally, the stronger
the binding affinity of the ligand for its site, fewer
residues will be available for interaction with the nitrene
increasing the likelihood of covalently labeling a specific
residue at the binding site.
Other concerns in photoaffinity labeling, besides binding
affinity as discussed above, are the temperature of incubation
and photolysis, ionic strength, pH, protein concentration,
intensity of photolyzing light and its affect upon the
receptor molecule. In the vast majority of photolabeling
experiments described in the literature, the incubation and
photolysis were done near 00C and under dim red light to
prevent decomposition of the azido compound. This was done in
most instances to minimize the rates of exchange thereby
increasing the residency of the radiolabeled ligand in the
binding site and to dissipate heat from the irradiation by uv
light. The ionic strength and pH should be optimal to ensure
not only tight binding of ligand to the receptor but also that
photoincorporation is not diminished in some manner. In the
following study I will describe the photoincorporation of [a
and 32]8-Azido GTP in 28S ribosomes. I will show that the GTP
analog binds specifically to a single ribosomal protein and
will describe the development of optimal photolabeling
conditions for this reaction.







43
Materials and Methods
Reagents
Both [a32P]8-Azido GTP and [y32P]8-Azido GTP were obtained
as triethylammonium salt at specific activities of 6.7 and
10.2 Ci/mmole, respectively, from ICN radiochemicals (Irvine,
CA). GTP was obtained from Sigma Chemical Co. as a Type I
grade reagent.
Binding Assay Solutions
Reaction Buffer A: 100 mM KC1, 5 mM MgC12, 10 mM Tris-HC1,
pH 7.4, and 1 mM BME
Reaction Buffer B: 10 mM NH4C1, 20 mM MgCl2, 10 mM Tris-HC1,
pH 7.4, and 1 mM BME
Reaction Buffer C: 100 mM KC1, 20 mM MgC12, 10 mM Tris-HC1,
pH 7.4, and 1 mM BME
Polyacrvlamide Gel Solutions
Laemmli Sample Buffer: 2% SDS, 5% 2-BME, 10% glycerol, and
62.5 mM Tris-HC1, pH 6.8
1D-PAGE Tank Buffer: 24 mM Tris-HC1, pH 8.3, 92 mM Glycine,
0.2% SDS and 0.1 M sodium thioglycolate
2D-PAGE Sample Buffer: 9 M urea, 10 mM 2-BME, 60 mM
potassium acetate, pH 6.7, and 0.01% aminoethanethiol
2D-PAGE Prep. Buffer: 5 M urea, 2% SDS, and 10 mM
sodium phosphate, pH 7.2
2D-PAGE Pre-Run Buffer: 9 M urea, 57 mM potassium hydroxide,
and 340 mM HPLC grade acetic acid, pH 5.0







44
2D-PAGE Tank Buffer: 0.5% SDS, 0.1 M sodium phosphate, pH
7.2, 3 mM 2-BME and 0.1 mM thioglycolate.
rRNA Extraction Buffer: 15 mM sodium acetate, pH 5.0, 100 mM
sodium chloride, 5 mM disodium EDTA, and 2% SDS
rRNA Run Buffer: 0.5 M Tris-HC1, pH 8.3, 0.5 M boric acid,
and 10 mM EDTA.
Photolabeling of 28S Ribosomes with Radiolabeled 8-Azido-GTP
The conditions used for photoincorporation of
radiolabeled 8-Azido-GTP were nearly identical to those used
for the Millipore filter assay. The final reaction mixture
contained 28S ribosomes (0.5 and 1.0 VM for 1D and 2D-PAGE
analyses, respectively) and radiolabeled NTP at varying
concentrations (0.05 to 6 pM) and GTP (5 and 10 pM for 1D and
2D-PAGE analyses, respectively) if a binding site blocker was
desired. The order of addition was important. The ribosomes
and GTP (if appropriate) were mixed in the reaction buffer (A,
B, or C) and incubated on ice and in the dark for 5 min. The
radiolabeled NTP dissolved in reaction buffer was then added
to a final volume of 30 or 100 pl for 1D or 2D-PAGE analyses,
respectively and incubated on ice in the dark for an
additional 15 min.
Photolysis was performed with a hand held ultraviolet
lamp with a purple filter that allowed transmission at 254 or
377 nm. Irradiation with 254 nm light was done in all cases
and at varying distances directly above the samples resulting
in intensities of 400 to 800 pW/cm2 for photolabeling with







45
8-Azido-GTP or GTP, respectively. The intensity of the uv
light was determined with the use of Spectroline digital uv
intensity meter (model # 254X). In initial photolabeling
experiments with [a32P]8-Azido GTP, 2500 PW/cm2 was used.
Irradiation of samples was allowed to proceed for 5 to 10 min
typically, on ice, in the dark with only dim red light for the
8-Azido GTP and with minimal room light for GTP. Samples were
then prepared for either 1D or 2D PAGE analysis.
Polvacrvlamide Gel Electrophoresis
Photolabeled ribosomal samples were often resolved on 1D
SDS slab gels (1.5 X 160 X 200 mm) as described by Laemmli
(1970). Samples were loaded into sample wells of a
polyacrylamide SDS gel (12% or 15%T, 3.5%C, 0.375 M Tris-HC1,
pH 8.8, and 0.1% SDS; T is the total monomer concentration, C
is the crosslinker, bisacrylamide, in percentage T) with a
stacker gel (3%T, 3.1%C). Electrophoresis was performed in
1D-PAGE tank buffer (24 mM Tris-HC1, pH 8.3, 0.192 M Glycine,
and 0.2% SDS) at an applied constant current of 10 mA for
approximately 14 h at room temperature. Following
electrophoresis the gel was fixed in 50% methanol and allowed
to equilibrate overnight with repeated changes of the
methanol. The gel was then silver stained (Wray et al.,
1981)and refixed in 50% methanol for at least 4 h to ensure
complete shrinkage of the gel before autoradiography was
performed. The stained and fixed gel was then sandwiched
between two layers of wetted Bio-gel wrap and allowed to air







46
dry while clamped in a plexiglas holding frame. The dried gel
was then autoradiographed at room temperature for an
appropriate period of time.
A modified 2D-PAGE system of Leister and Dawid (1975)
was used to determine which ribosomal protein was photolabeled
by either radiolabeled 8-Azido-GTP or GTP. Following
photolabeling as described above, ribosomal samples were
extracted with urea and lithium chloride as described below.
The first dimension was run in tube gels (3 X 110 mm) cast in
siliconized (Sigmacote, Sigma) glass tubing. The acrylamide
concentration of the separation tube gel was 4.6%T, 3.2%C,
with 9 M urea, and 60 mM potassium acetate, pH 4.3. The first
dimension was pre-run in 2D-PAGE pre-run buffer (9 M urea, 60
mM potassium acetate, pH 4.3) to remove the polymerizing
agent, ammonium persulfate (0.1%). A stacker gel (4%T and
3.5%C) was then cast (3 cm long) on top of the separator tube
gel. To each il of dialyzed sample, 0.25 jil of sample gel
(4%T, and 3.4%C) was added and then cast on top of the stacker
gel. The first dimension was then electrophoresed at a
constant amperage of 0.2 mA/gel toward the cathode in 2D-PAGE
tank buffer (35 mM B-alanine, pH 4.5, and 0.01%
aminoethanethiol) until the tracking dye, pyronine Y, had
entered the separation gel, then the current was boosted to
0.5 mA/gel. The polymerizing agent of both stacker and sample
gel was 0.1% riboflavin. When the tracker dye was within 1 cm
of the end of the tube gel, electrophoresis was stopped. The







47
tube gel was removed from the glass tube, then equilibrated in
2D-PAGE prep. buffer (5 M urea, 2% SDS, and 10 mM sodium
phosphate, pH 7.2) for 30 min while shaking gently at room
temperature.
For the second dimension a SDS slab (1.5 X 160 X 200 mm)
gel (10%T, and 3.5%C in 5.1 M urea, 0.5% SDS, 0.1 M sodium
phosphate, pH 7.2) was used. In preformed wells adjacent to
the first dimension tube gel, molecular weight markers (94,
67, 43, 30, 20.1, and 14.4 kDa) were run. The second
dimension was run at a constant current of 45 mA/gel toward
the anode in tank buffer (0.5%SDS, 0.1 M sodium phosphate, pH
7.2, and 3 mM BME).
The slab gels were fixed in 50% methanol, with several
changes, overnight. The gels were then silver stained (Wray
et al., 1981) and dried in a sandwich of Bio-gel wrap and
prepared for autoradiography.
Electrophoresis of Mitochondrial 12S rRNA
Photolabeling of mitochondrial 12S rRNA, as well as
ribosomal proteins with [a32P]8-Azido-GTP or [a32P]GTP, was
also examined. Following extraction of the photolabeled 28S
ribosomes by urea and LiCl, the rRNA pellet was stored at -
20'C until analyzed. The rRNA pellet was dissolved in an
extraction buffer (15 mM sodium acetate, pH 5.0, 100 mM sodium
chloride, 5 mM disodium EDTA, and 2% SDS) and extracted three
times with an equal volume (100 il) of equilibrated phenol
followed by centrifugation at high speed in a Beckman tabletop







48
microcentrifuge for 5 min at room temperature. Sodium acetate
was then added to a final concentration of 200 mM to the
pooled aqueous phases. To precipitate the rRNA, 2.5 volumes
of ice cold 95% ethanol was added and the mixture was kept at
-20C for at least 4 h. The sample was centrifuged for 10 min
in a microcentrifuge at high speed to pellet the rRNA. The
rRNA pellet was washed two more times with 70% ethanol
followed by a final wash with 95% ethanol. The sample was
dried under vacuum for approximately 15 min.
The rRNA pellet was redissolved in a 1:1 ratio of rRNA to
formamide loading buffer (80% formamide, 10 mM EDTA, pH 8, 1
mg/ml bromophenol blue, and 1 mg/ml xylene cyanol FF) prior
to 1D urea PAGE. Samples were resolved with molecular markers
of six synthetic poly(A)-tailed RNAs ranging in length from
0.24 to 9.49 kb (BRL, Gaithersburg, MD) on a one dimensional
urea PAGE (7.5%T, 3.5%C, 7 M urea, 0.5 M Tris-HC1, pH 8.3, 0.5
M boric acid, 10 mM EDTA, 0.05% TEMED, and 0.017% ammonium
persulfate) with a stacker gel (5%T, 3.5%C). Electrophoresis
was done in rRNA run buffer (0.5 M Tris-HCl, pH 8.3, 0.5 M
boric acid, and 10 mM EDTA) at a constant voltage of 500 V
toward the anode. The gel was then soaked in 0.2 M sodium
acetate, pH 5.0, for 30 min prior to staining with methylene
blue. The stained gel was washed with water, destined in 5%
acetic acid overnight, and then washed with water again just
prior to photography. The gel was dried overnight in a
sandwich of Bio-gel wrap and prepared for autoradiography.







49
Extraction of Ribosomal Proteins for 2D-PAGE Electrophoretic
Analysis
Ribosomes were extracted in urea-LiC1 to remove rRNA from
ribosomal proteins as described by Matthews et al. (1982)
prior to 2D-PAGE analysis. Thawed 28S ribosomes were adjusted
to 9 M urea, 3 M LiC1 with addition of solid urea and LiC1
with a resulting neutral pH. The samples were stirred
overnight at 4C followed by centrifugation at 50,000 rpm
(100,000 X g) for 60 min in a Beckman type 65 rotor to
precipitate the rRNA. The supernatant containing ribosomal
proteins was removed and the rRNA pellet was extracted once
again, by stirring for 2 h at 40C, with the above buffer.
After centrifugation the supernatants were pooled and dialyzed
against 2D-PAGE sample buffer (9 M urea, 0.01%
aminoethanethiol, 10 mM BME, and 60 mM potassium acetate, pH
6.7) overnight at room temperature in preparation for 2D-PAGE
analysis. The rRNA pellets were stored in the type 65
centrifuge tubes at -20C until needed.
Results
ra3Pp8-Azido GTP Photolabels a 38 kDa Ribosomal Protein in a
Specific Manner.
Filter binding studies suggest that a single and
specific binding site for GTP exists on the mitochondrial
ribosomal small subunit. To identify which ribosomal
protein(s) may be involved a photochemical approach was used.
28S ribosomes were incubated with increasing concentrations
(0.15 to 6.0 pM) of [a3'P]8-Azido GTP in reaction buffer B.







50
The samples were irradiated with uv (254 nm) light for 20 min
at 2500 PW/cm2 on ice (and in the dark). The photoaffinity
labeled ribosomes were resolved on a 12% polyacrylamide
Laemmli gel, silver stained and prepared for autoradiography
as described in Methods. Photoaffinity labeling of ribosomal
small subunits with [c32P]8-Azido GTP yielded a single
specifically radiolabeled 38 kDa ribosomal protein (figure 3-
2A). The addition of excess unlabeled GTP blocked
photolabeling of this protein, suggesting that the binding
site was also specific for GTP (figure 3-2 A). The amount of
radiolabeling increased as the concentration of the [aC32P]8-
Azido GTP increased but is approaching saturation at 6pM
[a32'P]8-Azido GTP. This is consistent with the apparent Kd of
1.92 pM (75% of the binding sites on 0.5pM 28S ribosomes
should be occupied at this Kd and concentration of 8-Azido GTP
used in this experiment) determined by filter binding assays
(table 2-3). Several ribosomal proteins with Mr > 40 kDa were
nonspecifically photolabeled at concentrations of [a32P]8-Azido
GTP above the Kd for 8-Azido GTP.
Unambiguous identification of the specifically photo-
labeled 40 kDa ribosomal protein was not possible by one
dimensional gel analysis due to the existence of other 28S
ribosomal proteins of similar molecular weight. Two dimen-
sional gel analysis of photoaffinity labeled small subunits
was done in order to do so. In figure 3-3 a single ribosomal
protein, S5, was specifically photolabeled with [a32p]8-Azido















S AR
94 -
67!
43e _
30 -

20.1 -

14.4 -



8N3 GTP(pM) 0.3 0.15 0.3 0.8 1.5 6.0
GTP(5spM) + + + + + +





FiLure 3-2 A Photoaffinity labeling of 28S ribosomes with
[a P]8-Azido GTP. Bovine 28S ribosomes (0.5 pM) were
irradiated in the presence of increasing concentrations (0.15
to 6 pM) of the 8-azido analog of GTP in reaction buffer B.
The samples were irradiated with uv light for 20 min at 2500
pW/cm ,pn ice under dim red light. The specific activity of
the [a P]8-Azido GTP was 1.35 Ci/mmol just prior to use. In
pane A a representative lane of silver stained (S) proteins
electrophoresed on a 12% SDS-polyacrylamide gel and the
corresponding autoradiogram (AR) are shown. In alternate
assays the GTP binding site was blocked by the addition (+) of
5uM GTP. The Mr markers were as follows: Phosphorylase b (94
ka), Bovine Serum Albumin (67 kDa), Ovalbumin (43 kDa),
Carbonic Anhydrase (30 kDa), Soybean Trypsin Inhibitor (20.1
kDa), and alpha-Lactalbumin (14.4 kDa). These same Mr markers
were used throughout this entire study. The arrow indicates
the presence of a 38 kDa protein specifically radiolabeled
with [a P]8-Azido GTP. See Methods for additional details.



















1 200


4 150

0.
W 100


O^ 50 -
50



0 1 2 3 4 5 6

8N3GTP CONCENTRATION (X 106M)





Figure 3-2 B Densitometric analysis of the autoradiogram for
the specifically radiolabeled 38 kDa ribosomal protein in
figure 3-2 A is shown. Data were the relative peak height
densitometric measurements of the 38 kDa baqd (panel A) in the
presence of increasing concentrations of [a P1]8-Azido GTP and
the absence of GTP.








A

94- "
67 -S45


*
43- -C0

30-

20.1 -


14.4-


B


-S4
S5











Figure 3-3 Two dimensional PAGE analysis of photoaffinity labeled
28S ribosomes with [a P]8-Azido GTP. Bovine M8S ribosomes (1 M)
were irradiated in the presence of 2 pM [a P]8-Azido GTP (45%
saturation) and in the presence (B) and absence (A) of 10 M GTP
The samples were irradiated with uv light for 10 min at 2500 pW/cm
on ice as described in Methods. The specific activity of the
[a P]8-Azido GTP was 1 Ci/mmol just prior to use. Silver stained
2D-PAGE gels are on the left and their corresponding autoradiograms
are on the right. Autoradiograms were exposed for 14 days. The
molecular weight markers used were the same ones described in
figure 3-2. The arrows indicate the presence of the nonspecific-
ally labeled (S4) and the specifically labeled (S5) proteins. See
Methods for additional experimental details.







54
GTP. The presence of excess amounts of unlabeled GTP com-
pletely blocked photoincorporation of [a32P]8-Azido GTP into
S5. Another ribosomal protein, S4, was also labeled
nonspecifically, since the presence of excess amounts of GTP
did not block photolabeling of S4.
Development of Optimal Photoaffinity Labeling Conditions for
8-Azido GTP
The extent of specific photolabeling of S5 by [a32P]8-
Azido GTP was small, especially in comparison to the much
greater nonspecific labeling of S4 (figure 3-3). Optimization
of photoirradiation and ionic conditions was undertaken to
improve the quality of the photoaffinity labeling of 28S
ribosomes by 8-Azido GTP. Optimal conditions were determined
by both filter binding assays and photoaffinity labeling
methods.
The development of optimal ionic conditions focused on
the modification of two components of the reaction buffer B as
developed by Bodley (Bacca, et al., 1976). Bodley developed
this reaction buffer for the study of GTP hydrolysis by EFG in
bacterial systems. One component, NH4C1, is known to be toxic
to peptidyl transferase activity (Denslow and O'Brien,
personal communication) on the large subunit of the
mitochondrial ribosome and its replacement with another
monovalent salt such as KC1 was studied.
In table 3-1 the effect of replacing 10 mM NH4C1 with KC1
on the binding of [8-3H]GTP to 28S ribosomes was examined by







55
filter binding assays. Increasing concentrations of [a32P]GTP
were incubated with 0.1 pM 28S ribosomes in reaction buffer B
(treatment #1 in table 3-1) as a control, and the test
reactions were in the same reaction buffer but with 10 mM
NH4C1 replaced with increasing concentrations (10 to 100 mM)
of KC1 and decreasing concentrations of MgCl2. The
substitution of NH4CI by KC1 or the reduction of MgCl2 from 20
to 5 mM did not alter the binding of [a32P]GTP to 28S
ribosomes. The most stringent condition, 100 mM KC1 and 20 mM
MgCl2, supports binding of [8-3H]GTP to the 28S ribosome in an
equivalent manner as controls. This reaction buffer C (100 mM
KC1, 20 mM MgC12, 10 mM Tris-HC1, pH 7.4, and 1 mM BME) and
was used in experiments discussed below.
Potter and Haley (1983) reported that in their
experience with a number of azido-purine analogs, photolysis
at 254 nm and at intensities of 180 to 800 PW/cm2 for 1 to 10
min caused little if any detectable damage in a variety of
enzymatic systems. In our initial experiments (figure 3-2 and
3-3) these guidelines were exceeded.
The intensity and duration of photoirradiation was
reduced to evaluate the minimal amount of uv irradiation
necessary to adequately photolabel 28S ribosomes with [a32P]8-
Azido GTP. Ribosomes (0.5 pM) were photolabeled in the
presence of 0.75 pM [a32P]8-Azido GTP, corresponding to
approximately 24% saturation of the GTP binding site
(calculation based upon a Kd of 1.92 pM) in reaction buffer B.









Table 3-1
The Effect of Substituting KC1 for NH,CL and the Lowering of
the MaCl Concentration in .he Reaction Buffer on the
Noncovalent Bindina of [a"Pl GTP to 28S Ribosomes
Treatment NH C1 M Cll KC1 a Kd SE n SE
# (mM (mM ( mM) (nM)
1 (B) 10 20 75 3.8 2.2 0.91 0.08
2 20 10 75 4.0 1.0 0.99 0.05
3 20 20 85 3.1 0.8 0.92 0.04
4 20 40 105 1.7 1.0 0.96 0.11
5 20 60 125 8.3 3.4 0.91 0.07
6 (C) 20 100 165 5.8 2.5 0.87 0.06
7 -10 100 135 2.5 0.9 0.92 0.06
8 (A) 5 100 120 2.7 0.8 1.00 0.06

Scatchard analysis for the binding of [a32p]GTP to 28S
ribosomes under various ionic conditions by the Millipore
filter assay. The data are binding affinities (Kd) and number
of binding sites (n) for a single binding curve. The
concentration of the ribosomes was 0.1 VM and th a32P]GTP
0.02 to 1 pM. The specific activity of the [a P GTP was
diluted to 25.4 Ci/mmol with GTP just prior to use. Reaction
buffers A, B, and C (for complete formulations see Methods)
are anotated in parentheses, beside their respective
treatments In all reaction buffers 1mM BME was used. No
differences were observed with the substitution of NH4Cl in
reaction buffer B (treatment #1) for increasing concentrations
of KC1 (treatments 2 to 6) or decreasing concentrations of
MgCl2 (treatments 7 & 8).
ap = total ionic strength


In figure 3-4 A the reduction of 254 nm uv irradiation from
2500 to 800 pW/cm2 for 10 min reduced the photolabeling of the
38 kDa ribosomal protein (no excess GTP present) by only 47%,
as determined by densitometric analysis (figure 3-4 B). The
amount of labeling was still adequate for analysis. However,
irradiation for 2 min reduced photolabeling of the same 38 kDa
protein by 75 to 90% at 2500 and 800 pW/cm2, respectively.







57
This provided minimal detectable labeling at 2500 PW/cm2 and
virtually no labeling at the lower level of uv irradiation
even after a 2 week exposure of the autoradiogram. The
pattern of photolabeling remained the same as seen in figure
3-2 A. Photolabeling of the 38 kDa protein by [a32P]8-Azido
GTP was blocked in the presence of excess GTP and not in the
case of the 43 kDa ribosomal protein.
In an continuing effort to find the minimal exposure of
ribosomes with uv light for adequate photolabeling [y32P]8-
Azido GTP was used. This isotope was used because [a32P]8-
Azido GTP was no longer available. [y32P]8-Azido GTP was also
available in higher purity (>98%), more than twice the
specific activity (10.2 Ci/mmole),and at lower cost. Optimal
photoirradiation conditions for the photolabeling of 28S
ribosomes with [y32P]8-Azido GTP were determined in reaction
buffer C (100 mM KC1, 20 mM MgC12, 10 mM Tris-HC1, pH 7.4, and
1 mM BME). Substoichiometric concentrations of [y32P]8-Azido
GTP (0.43 pM), corresponding to approximately 18% binding,
were incubated with 28S ribosomes (0.5 pM) and irradiated at
300 and 800 iW/cm2 for 4 and 8 min on ice in the dark. The
minimal uv (254 nm) irradiation intensity and duration found
for adequate labeling were 300 PW/cm2 for 10 min (figure 3-5
A and B). These conditions were utilized in subsequent
studies involving this analog of GTP.












S AR

94-
67-
43-

30- :.. :!...

"...i "' .
20.1-


14.4-


Ribosomes -+ + + + + + +
GTP (51M) --+ + +-+
Irrad. time (min) 10 2 10 2

Intensity (pW/cm2) 2500 800




Figure 3-4 A Optimization of the intensity and duration of uv
irradiation for photoaffinity labeling of 28S ribosomes with
[ P]8-Azido GTP. 28S ribosomes (0.5 pM) were incubated with
.75 pM [a P]8-Azido GTP (24% saturation) in reaction buffer
B. The specific activity of the [a P8-Azido GTP was 1.0
Ci/mmol. In panel A is a representative lane of silver
stained ribosomal proteins on the left and the corresponding
autoradiogram on the right. The AR was exposed for 14 days
before development. The presence (+) and absence (-) of
ribosomes and unlabeled GTP as well as the intensity and
duration of uv (254 nm) irradiation are noted below the
autoradiogram. The arrow indicates the position of the
specifically radiolabeled 38 kDa protein. For more
experimental details see Methods.















[ac32]8-Azido GTP


MI 38 kDa
ICl 43 kDa


in


*~f~ a ----


- + +
2 min 10 min
800AW / cm2


m + +


- + -+
2 min 10 min
2500/W / cm2


Figure 3-4 B Densitometric analysis of the autoradiogram in
figure 3-4 A is shown. Densitometric analysis was done for
the specifically radiolabeled ribosomal protein (38 kDa) and
the nonspecifically labeled protein (43 kDa) ribosomal protein
at varying intensities and duration of uv irradiation. Note
that in the presence (+) of 5 pM GTP the 38 kDa protein was
not radiolabeled. For additional experimental details see
Methods.


w~ n


r"l












AR


94 -
67 -
43--


30--



20.1-


S. _


o..


-+-+-+-+

4 10 4 10

300 800


Figure 3-5 A Continued optimization of the intensity and
duration of uv irradiation for photoaffinity labeling of 28S
ribosomes with [y P]8-Azido GTP. 28S ribosomes (0.5 pM) were
incubated with 0.5 1M [y P]8-Azido GTP (18% saturation) in
reaction buffer C. The specific activity of the [y P]8-Azido
GTP was 8.4 Ci/mmol. In panel A is a representative lane of
silver stained (S) ribosomal proteins on the left and the
corresponding autoradiogram (AR) on the right. The AR was
exposed for 3 days before development. The presence (+) and
absence (-) of unlabeled GTP as well as the intensity and
duration of uv (254 nm) irradiation are noted below the
autoradiogram.


4-


I .


GTP (5pM)

Irrod.time (min)

Intensity (LW/cm2)


J
I


I







61






[y32P]8-Azido GTP


eM 38 kDa
-CD 43 kDa












- + + + +


4 min


10 min


4 min


10 min


300pW / cm2


800uW / cm2


Figure 3-5 B Densitometric analysis of the autoradiogram in
figure 3-5 A is shown. Densitometric analysis was done for
the specifically labeled ribosomal protein (38 kDa) and the
nonspecifically labeled protein (43 kDa) ribosomal protein at
varying intensities and duration of uv irradiation. The
presence (+) and absence (-) of 5 VM GTP is indicated below
the figure. In the presence of excess amounts of GTP (+) the
38 kDa protein was not radiolabeled. For additional
experimental details see Methods.


120







62
No significant differences were seen in the photolabeling
pattern of the two isotopes ( compare figures 3-4 A and 3-5 A)
though the intensity of radiolabeling was greater for the
[y32p] isotope due to its higher specific activity. The
[y32P]8-Azido GTP was utilized in the remainder of 8-Azido GTP
photoaffinity experiments described in this paper.
The last component of the reaction buffer to be modified
was the divalent cation, MgC12. MgC12 has been shown by
Denslow et al. (1988) to promote adherence of 28 and 39S
subunits to one another due presumably to salt bridging
effects at higher concentrations of the divalent cation. To
reduce the possibility of nonspecific binding due to salt
bridging of the nucleotide, MgC12 concentration was reduced
from 20 to 5 mM and photolabeling was examined under these
conditions. The same incubation conditions were used as
described in the above experiment and uv photoirradiation was
done at 300 PW/cm2 for 10 min. The 38 kDa ribosomal protein
was still specifically photolabeled (figure 3-6 A) by the GTP
analog, [y32P]8-Azido GTP, and a 45% increase in this labeling
was observed by densitometric analysis as the concentration of
MgCl2 dropped (figure 3-6 B) to 5 mM. Since the lowest
concentration of MgC1, provided a modest increase in
photolabeling of the specifically labeled 38 kDa ribosomal
protein it was adopted for subsequent studies. The amount
oflabeling on the nonspecifically labeled 43 kDa ribosomal
protein was unchanged in the absence of excess GTP but did
increase 47% in the presence of 5 mM GTP.












S

94-- U
67-
43--

30--


0.1 -

44---


GTP(5/M)

MgCL2(mM)


AR
Q o


9 4


Figure 3-6 A The effect of decreasing concentration~ of MgCl2
on photoaffinity labeling of 28S ribosomes with [y 'P]8-Azido
GTP. Bovine 28S ribosomes (0.5 iM) were incubated with 0.5 pM
[y P]8-Azido GTP (18% saturation) and irradiated with uv light
for 10 min at 300 pW/cm on ice. The specific activity of the
[y P]8-Azido GTP was 6.3 Ci/mmol. The concentration of the
MgCl1 component of the reaction buffer was decreased from 20
mM (reaction buffer C) to 5 mM (reaction buffer A). In panel
A a representative silver stained lane of 28S ribosomal
proteins and molecular weight markers are shown with the
corresponding autoradiogram on the right. The concentration
of the MgC1, component of the reaction buffer and the presence
(+) and absence (-) of 5 pM unlabeled GTP is shown below the
autoradiogram. The autoradiogram was exposed for 3.5 days
before developing. The arrow indicates the position of the 38
kDa specifically radiolabeled protein.


2

I


- +- + +

20 10 5


* *.




























40


20


MgCl2 (X 106M)








Figure 3-6 B Densitometric analysis of the autoradiogram in
figure 3-6 A. The amount of radiolabeling of the 38 and 43
kDa ribosomal proteins was plotted versus the concentration of
MgCl in the reaction buffer. Note that no radiolabeling of
the 38 kDa protein was detected in the presence of 5 pM TP.


* S5 -GTP
* S4 -GTP
0 S4 +GTP)



I I I







65
The three previous experiments suggest that the optimal
ionic composition of the reaction buffer for the binding of
GTP and its 8-Azido analog to 28S ribosomes was 100 mM KC1, 5
mM MgCl2, 10 mM Tris-HCl, pH 7.4, and 1 mM BME and will be
referred to as reaction buffer A in this paper. The optimal
uv photolabeling conditions were 300 pW/cm2 for 10 min for 8-
Azido GTP.
With these new conditions another attempt was made to
analyze photolabeled ribosomal proteins by two dimensional
PAGE. In figure 3-7 A to F, 28S ribosomes (1 pM) were
photoaffinity labeled with increasing concentrations of
[y32P]8-Azido GTP (0.3, 1, and 3 pM) in the presence and
absence of excess (10 pM) GTP. These amounts of [y32P]8-Azido
GTP correspond to 9, 20, and 60% saturation of the GTP binding
site on the ribosome (based on a Kd of 1.92 pM). Small
subunit ribosomes were resolved by 2D-PAGE and radiolabeled
proteins were detected by autoradiography.
At all concentrations tested, S5 was specifically
radiolabeled by [y32P]8-Azido GTP (figure 3-7 A to F) and the
radiolabeling increased directly proportional to the
concentration of [y32P]8-Azido GTP (figure 3-7 G). The
presence of GTP completely blocked photoincorporation of the
GTP analog into S5. S4 was also radiolabeled, but in a
nonspecific manner, since excess amounts of GTP failed to
block photoincorporation of [y32P]8-Azido GTP. These results
support the tentative finding seen in the initial attempt to























Figure 3-7 Two dimensional PAGE analysis of photoaffinity
labeled 28S ribosomes with increasing concentrations of
[y P]8-Azido GTP. Bovine 28S ribosomes (I jM) were irradiated
with uv light for 10 min at 300 IW/CJ on ice in the presence
of increasing concentrations of [y P]8-Azido GTP (0.3 to 3
IM). Results in the presence (B, & F) and absence (A, C,
& E) of 10 pM GTP is also shown. The s-amples were irradiated
with uv light for 10 min at 300 pW/cm on ice and under dim
red light. The specific activity of the [y P]8-Azido GTP was
2.9 Ci7mmol just prior to use. Silver stained gels are on the
left and their corresponding autoradiograms on the right.
Autoradiograms were exposed for 7 days. The Mr markers used
were the same ones described in figure 3-2. See Methods for
additional experimental details.






67
A


94 -
67 -
43 -

30 -

20.1 -


14.4 -


4
SS5


B

i *


S4


Figure 3-7 A & B Two dimensional PAGE analysis of 2photoaffinity
labeled 28S ribosomes with 0.3 pM (9% saturation) [y P]8-Azido GTP
(panels A & B). The presence of 10 pM GTP (panel B) blocked
radiolabeling of S5.


84
0S5














S5


/s4
"60


S4


Figure 3-7 C & D Two dimensional PAGE analysis ofphotoaffinity
labeled 28S ribosomes with 1.0 iM (28% saturation) [y P]8-Azido GTP
(anels C & D). Note, that in the presence of 10 pM GTP (panel D)
was not radiolabeled.


94-
67-
43-

30-

20.1-


14.4 -












94-
67-
43-

30-

20.1 -

14.4-


SI
/ 4


399
S13'


S4


;4
LS5


60
(i
L,


Figure 3-7 E & F Two dimensional PAGE analysis of 3photoaffinity
labeled 28S ribosomes with 3.0 IM (56% saturation) [y P]8-Azido GTP
(anels E & F). Note, that in the presence of 10 M GTP (panel F)
S5 was not radiolabeled.




















150



100



50


1 2 3


[y32p]8N3GTP CONCENTRATION (X106M)






Figure 3-7 G Densitometric analysis of the GTP specific and
nonspecific radiolabeling of ribosomal proteins S5 and S4
respectively. In panel G densitometric analysis of the
autoradiograms in panels A to F for the specific labeling of
S5 and nonspecific labeling of S4 is shown.












9.5
7.5-
4.4-..
2.47
1.4




0.24-


S AR











r


GTP(IO1M) + +



Figure 3-7 H Analysis of rRNA extracted from 28S ribosomes
irradiated in the presence of [y P]8-Azido GTP. This is an
analysis of the rRNA extracted from photoaffinity labeled 28S
ribosomes shown in figure 3-8, panels C & D ( the result in
all other panels was the same, no radiolabeling of rRNA). The
rRNA was extracted following uv irradiation by 9M urea and 3M
LiCl and electrophoresed by 1D urea PAGE (7.5% PAGE) as
described in Methods. The left panel is a methylene blue
stained gel (S) and its corresponding autoradiogram (AR) on
the right. The autoradiogram was exposed for 41 days. The
molecular weight markers are six synthetic poly(A)-tailed RNAs
ranging in length from 0.24 to 9.49 kb (Bethesda Research
Labs). No rRNA fragments were radiolabeled.







72
identify the GTP binding protein by 2D-PAGE analysis (figure
3-3), that S5 was specifically labeled by 8-Azido GTP. The
development of more stringent ionic and milder photoirradia-
tion conditions did not eliminate nonspecific labeling of S4.
S4 was the most intensely labeled protein which increased
proportionally with the concentration of [y32P]8-Azido GTP.
S4 appears to be reciprocally labeled in the presence of
excess amounts of GTP (figure 3-7 G). The amount of S4
labeling increases in the presence of 10 pM GTP, even though
the radionucleotide was diluted as much as 33 fold. This
suggests this site may be a lower affinity site for 8-Azido
GTP. No radiolabeling of 12S rRNA was detected (figure 3-7
H), suggesting no direct involvement of rRNA in binding of 8-
Azido GTP to 28S ribosomes.
Discussion
Photoaffinity labeling of 28S ribosomes with 8-Azido GTP
revealed that a single 38 kDa ribosomal protein bound the GTP
analog in a specific manner (figure 3-2 to 3-7). Unlabeled
GTP could displace the radiolabeling of this protein
suggesting the site was specific for GTP. A 43 kDa protein
was also radiolabeled but the labeling was nonspecific because
excess amounts of GTP failed to block binding. The amount of
radiolabeling on the 38 kDa protein increased with increasing
concentrations of [aP32P]8-Azido GTP and was saturable (figure
3-2 B) in a fashion consistent with the binding of this
compound to 28S ribosomes as determined by filter binding







73
assays (figure 2-3). The photolabeling reaction was done
under the same ionic conditions (reaction buffer B) as were
used in filter binding assays to assess [aP32P8-Azido GTP
binding to 28S ribosomes.
The identity of this labeled protein was examined by 2-
dimensional PAGE. Two proteins retained the label, one S4,
was the predominately radiolabeled protein but the labeling
was nonspecific because it was not diminished in the presence
of excess amounts of GTP. The second protein, S5, was
specifically radiolabeled by [a32P]8-Azido GTP. Excess
concentrations of GTP blocked binding to this protein
suggesting that it is the specific binding site for GTP.
Comparing these results to those of the 1-dimensional
analysis (figure 3-2), the 43 kDa nonspecifically labeled
protein correlates to S4 in both Mr and binding specificity,
and the 38 kDa protein would be S5 for the same reasons. The
only disparity was that the radiolabeling of S5 was less than
that seen on S4 in the two-dimensional PAGE analysis (figure
3-7). The disparity likely resides in differences in sample
preparation. In two-dimensional PAGE analysis, photolabeled
28S ribosomes are extracted under acidic conditions (pH 2.5)
to remove rRNA prior to separation in the first dimension in
urea also under acidic conditions (pH 4.3). Temple et al.
(1966) and Maliarik and Goldstein (1988) have observed that
some covalent bonds formed by the aryl purinee) nitrene
adducts are acid labile. In previous attempts to resolve







74
photolabeled ribosomes with the azido analog no radiolabeled
proteins were detected. Increasing the pH of the extraction
step to neutrality, only one radiolabeled protein, S4, was
found and it was not blocked in the presence of GTP (data not
shown). However when the pH of the first dimensional
separation was raised from 4.3 to 5.0 we were able to observe
not only the nonspecific labeling of S4 but also specific
labeling on S5 (figure 3-3 and 3-7). The separation of these
basic ribosomal proteins in the first dimension in urea could
not be done at any higher pH without risking alteration of the
relative mobilities of the proteins in a defined system
(Matthews et al.,1982). Never-the-less, under the elevated pH
conditions during sample preparation and analysis, it is clear
that S5 is the protein binding GTP in a specific manner.
Concerns about the survivability of the covalent bound formed
by the nitrene adduct, and photolabeling under more
physiological ionic conditions lead to studies to determine
the minimum photoirradiation required and optimal ionic
conditions for this reaction.
The minimum fluency (iW/cm2) and duration of irradiation
to achieve labeling of ribosomal proteins using 254 nm light
with 8-Azido GTP, were 300 PW/cm2 and 10 min, respectively
(figure 3-5). This was determined over the span of two
experiments in which both the [a32P] and [y32P]8-Azido GTP were
used, with no differences seen in the pattern of radiolabeled
proteins. O'Brien et al. (personal communication) found that







75
binding of [y32P]GTP to 28S ribosomes, under conditions similar
to those used here, resulted in no hydrolysis of the [y32P]
label even after several hours of incubation with the
ribosome. This indicates that under these binding conditions
[y32P]8-Azido GTP can be used with no hydrolysis of the [y32p]
to obscure the results of photoaffinity labeling.
The length of time required (10 min) for adequate levels
of radiolabeling are somewhat longer than reported by others
(typically 2 to 6 min at fluency < 1000 PW/cm2 on ice in
various buffer biological solutions, see Potter and Haley,
1983). However, irradiation for 4 min at 300 W/cm2 was
minimally acceptable in 1-dimensional systems, but due to
greater concentrations of protein (1 versus 0.5pM ribosomes)
it was not judged adequate for 2-dimensional analysis.
Several enzymatic systems which were subjected to
photoaffinity labeling saw no appreciable loss in function
when irradiated with 254 nm light at low temperatures for as
long as 10 min if the fluency remained below 800 W/cm2 (Potter
and Haley, 1983). These guidelines were used, since no
functional assay exist for the 28S ribosome which could be
used to monitor uv induced damage.
Geahlen and Haley (1977) reported that the half-life of
the 8-azido GMP (160 pM at neutral pH) was approximately 20
sec in protein free buffer solution at approximately 1000
PW/cm2 with 254 nm light. Since only minute quantities of the
8-azido GTP were available, it was not possible to examine the







76
uv degradation of the 8-azido to N2 and 8-NH2-GTP by
spectrophotometric means as done by Geahlen and Haley (1977).
Also, no suitable thin layer chromatography or other method
could be found that could discriminate between 8-azido and the
8-NH2 GTP though it was determined that the gamma phosphate
was not lost following irradiation under the conditions used
in this study (data not shown).
Efforts to develop more physiologically relevant ionic
conditions for GTP binding to 28S ribosomes focused upon two
components of the reaction buffer, MgCl2, and NH4C1. In the
initial binding experiments of this study the ionic
composition of the reaction buffer (B) was adopted from the
work of Bodley (Bacca, et al., 1976) who developed it to study
GTP binding to bacterial EF-G. The concentration of the
divalent cation, MgC1, (20 mM), was reduced to 5 mM to
alleviate concerns about subunit dimerization which could
alter GTP binding to the ribosome. O'Brien (1971) showed that
the mitochondrial monoribosomes are stabilized at the higher
concentration of MgCl2 (20 mM MgCl2 in 100 mM KC1) or greater.
This occurs because divalent cations form salt bridges which
stabilize rRNA interactions between subunits which results in
subunit association or aggregation. At 5 mM MgCl2,
mitoribosomes subunits exist as discrete particles. GTP-Mg2
binds as a complex to GTP-binding proteins. The Mg+2 ion binds
tightly and nM concentrations of the divalent cation commonly
suffice to fulfill this requirement (Gilman, 1987). Though







77
only modest increases in the photoincorporation of [y32P8-
Azido GTP in 28S ribosomes were observed at 5 mM MgC1,
(figure 3-6), this reduced concentration of the divalent
cation was adopted.
The replacement of the monovalent cation NH4' by K' was
examined because NH4C1 is known to be toxic to the peptidyl
transferase activity of the mitoribosome large subunit
(Denslow and O'Brien, personal communication). Though we are
studying a nucleotide binding activity on the small subunit it
was deemed prudent to replace this monovalent cation in the
event NH4C1 is also toxic to some function of the small
subunit, and in the event that future studies of peptidyl
transferase activity might be undertaken in conjunction with
this GTP-binding activity. Additionally, we wanted to
increase the ionic strength to determine if GTP-binding would
be affected by increasing the stringency to more
physiologically relevant conditions. This was accomplished by
replacing NH4C1 with KC1 and testing the effect of higher
concentrations of KC1, equivalent to physiological
concentrations (70 to 100 mM). GTP-binding activity was
essentially unaltered at 100 mM KC1 when compared to binding
activity at 10 mM NH4C1 (table 3-1). This indicated that GTP
can bind the 28S ribosome even at moderate salt concentrations
suggesting the binding is not predominately electrostatic in
nature. Having established new ionic conditions (reaction
buffer A) photoaffinity labeling of 28S ribosomes by







78
[y32P]8-Azido GTP was examined again. Photoaffinity labeling
with increasing concentrations of 8-Azido GTP to a fixed
concentra-tion of the ribosome, resulted in radiolabeling of
the same proteins seen previously. S4 was nonspecifically
radiolabeled and S5 was specifically radiolabeled by [y32P]8-
Azido GTP. At very high saturating concentrations of the GTP
analog, other ribosomal proteins were also labeled in a
nonspecific manner (figure 3-7 E). At all concentrations of
[y32P]8-Azido GTP used, S5 radiolabeling was blocked in the
presence of excess GTP and the radiolabeling of S5 increased
with increased concentrations of the [y32P]8-Azido GTP.
Nonspecific radio-labeling of S4 also increased in a manner
comparable to that seen for S5 but interestingly S4
radiolabeling increased in the presence of excess GTP and
appeared to be saturable (figure 3-7 G). This suggested that
S4 bound some of the [y32P]8-Azido GTP displaced from the
binding site by GTP (10 pM) even though the specific activity
of the radiolabeled GTP analog was decreased by 33 to 3.3
fold. This suggests that S4 binds GTP in a manner different
than for S5, and is likely to have a lower affinity binding
site.














CHAPTER 4
PHOTOAFFINITY LABELING OF 28S RIBOSOMES BY [a32P]GTP
Introduction
Direct photoaffinity labeling of phosphofructokinase
(Ferguson, and MacInnes, 1980) and tubulin (Nath, et al.,
1985; Hesse, et al., 1987;and Linse and Mandelkow, 1988) with
radiolabeled cAMP and GTP/GDP, respectively, encouraged us to
explore the use of the [a32P]GTP to photolabel the 28S ribo-
some. The chemical structure at the site of covalent attach-
ment for these purines is not known, but the fact that maximal
photolabeling of phosphofructokinase occurred at a wavelength
equivalent to its absorption maxima suggests that the adenine
portion of the nucleotide is the light sensitive chromophore
(Ferguson and MacInnes, 1980). The use of the natural ligand
takes advantage of the increased binding affinity thereby
circumventing difficulties encountered due to lowered binding
affinities of the 8-Azido GTP analog for the 28S ribosome
(table 2-3). The disadvantages are that the mechanism of
photoactivation is unknown and therefore we do not know if
covalent bond formation is as nonselective a process as was
the case for the nitrene. Another disadvantage is that the
minimal quantum yield of product would be lower than for the







80
nitrene. In the case of cAMP, photolabeling of phosphofructo-
kinase was reported to be approximately 0.01% (Ferguson and
Maclnnes, 1980) several hundred times less than for the
nitrene. However that difficulty can be overcome to some
degree because [ 32P] GTP can be obtained at specific activities
as much as 300 to 600 times greater than [a32P]- or [y32]8-
Azido GTP, respectively.
In this study we report successful photoaffinity labeling
of 28S ribosomes with [a32P]GTP. The photolabeling of ribo-
somal proteins was equivalent to that seen for [a32P]- and
[y32P]8-Azido GTP, confirming that S5 is the specific site for
photoincorporation of GTP.
Materials and Methods
Reagents
[a32P]GTP, and ATP and [y32P]ATP in triethylammonium salt
were purchased from Amersham at specific activities of 410 and
>3000 Ci/mmole, respectively. All unlabeled nucleotides were
purchased from Sigma Chemical Co. as Type I reagents as
described above.
Solutions
The same solutions as described above, in 8-Azido GTP
photoaffinity labeling experiments, were used in these
studies.









Methods
The same methods as described above, in 8-Azido GTP
photoaffinity labeling experiments, were used in these
studies.
Results
Photoaffinitv Labeling of 28S Ribosomes with a32P1GTP
In an effort to substantiate the above findings
photolabeling of small subunits with [a32P]GTP was attempted.
Hesse et al. (1987) successfully photolabeled B-tubulin with
[a32P]GTP suggesting the natural ligand could be used to
photoaffinity label the GTP binding ribosomal protein.
First, photolabeling conditions were varied to optimize
photolabeling. 28S Ribosomes (0.5 pM) were photoaffinity
labeled in presence of 0.5 VM [a32P]GTP, which corresponds to
80% saturation (calculation based on a Kd of 15.2 nM) of the
GTP binding site. Ribosomes were photolabeled for 2 and 10
min at increasing intensities (800, 1500, 2000, and 2500
PW/cm2) of uv light (254 nm) on ice and under dim room light.
The pattern of radiolabeling of 28S ribosomes by [a32P]GTP
(figure 4-1) was similar to that for [a32P]8-Azido GTP (figure
3-4) and [y32P]8-Azido GTP (figure 3-5). The 38 kDa ribosomal
protein was specifically labeled by GTP because excess amounts
of GTP blocked labeling of that protein. The 43 kDa ribosomal
protein was still nonspecifically labeled in the presence of
excess amounts of GTP as seen with the 8-Azido analog.











AR S
__ ---- -

67
43

30


I -- 20.1

14.4




Ribosomes ++++++++++ + + + ++-

G TP(5PM) -+-+ -+-+ + -+--
Irrod. time (min) 2 10 2 2 2 5 10 200

Intensity (mW/cma) 0.8 1.5 2.0- 2.5--- 0


Figure 4-1 A Optimization of the intensity and duration of uv
irradiation for photoaffinity labeling of 28S ribosomes with
[caP GTP. 28S ribosomes (0.5 pM) were incubated with 0.5 IM
[a32P 8-Azido GTP (80% saturation) in reaction buffer B. The
samples were irradiated with uv light fqr 5 min at 800 pW/cm
on ice, prior to electrophoresis. [a P]GTP was diluted 16
fold with GTP to a specific activity of 32 Ci/mmol just prior
to use. In panel A is a representative lane of silver stained
(S) ribosomal proteins on the right and the corresponding
autoradiogram (AR) on the left. The AR was exposed for 24 h
before development. The presence and absence of ribosomes and
unlabeled GTP as well as the intensity and duration of uv (254
nm) irradiation are noted below the autoradiogram. The arrow
indicates the radiolabeled 38 kDa protein that is specifically
labeled by GTP.

















[a32p] GTP


M 38 kDa
IF- 43 kDa


~j1


Ji~n~i


In P


n


-+ -+ -+ -+ -+ -+ -+ -+
2 10 2 2 2 5 10 20


2.5mW


0.8mW 1.5mW 2.0mW


Figure 4-1 B Densitometric analysis of the autoradiogram
shown in figure 4-1 A. Densitometric analysis was done for
the specifically labeled ribosomal protein (38 kDa) and the
nonspecifically labeled protein (43 kDa) ribosomal protein at
varying intensities and duration of uv irradiation. In the
presence (+) of 5 VM GTP, the 38 kDa protein was not
radiolabeled. The duration and intensity of uv irradiation is
indicated below the figure. For additional experimental
details see Methods.


'~ w"' '~L"







84
Photolabeling of the 38 kDa protein by [a32P]GTP was more
dependent on the duration of irradiation than the intensity of
the uv light (figure 4-1 B). No differences were seen by
densitometric analysis in the amount of photolabeling of the
38 kDa protein when the intensity of uv irradiation was
increased from 800 to 2500 pW/cm2 for 2 min. However the
amount of photolabeled 38 kDa protein was dramatically
increased nearly 10 fold when the duration of irradiation was
lengthened to 10 min at 2500 PW/cm2. The amount of
photolabeling at 800 and 2500 PW/cm2 was in fact nearly
identical for irradiation at 10 min. This data suggests that
irradiation for 5 min at 800 W//cm2 would produce equivalent
amounts of radiolabeling to that seen for 5 min at 2500 pW/cm2
(figure 4-1 B). The amount of radiolabeling in that case was
sufficient for photoaffinity labeling of 28S ribosomes so uv
(254 nm) irradiation at 800 PW/cm2 for 5 min was adopted for
subsequent photolabeling experiments with [a32P]GTP.
To determine the extent of binding specificity of the 28S
ribosome several nucleotide tri- and diphosphates were tested
for their ability to block photoaffinity labeling of 28S
ribosomes by [a32P]GTP. Of the nucleotides tested, only GTP
and GDP, and to a much lesser extent, GMP, were able to block
photolabeling of the 38 kDa ribosomal protein by [a32P]GTP
(figure 4-2 A). These findings agree with the results of
filter binding assays (figure 2-2) and photoaffinity labeling
experiments with 8-Azido GTP, that there is a single GTP