Citation
Proteins of the mammalian mitochondrial ribosome

Material Information

Title:
Proteins of the mammalian mitochondrial ribosome
Creator:
Matthews, David Earl, 1948-
Language:
English
Physical Description:
xi, 155 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Buoyancy ( jstor )
Cattle ( jstor )
Gels ( jstor )
Mitochondria ( jstor )
Molecular weight ( jstor )
Rats ( jstor )
Ribosomal proteins ( jstor )
Ribosomes ( jstor )
RNA ( jstor )
Species ( jstor )
Biochemistry and Molecular Biology thesis Ph. D
Dissertations, Academic -- Biochemistry and Molecular Biology -- UF
Mitochondria ( lcsh ) ( lcsh )
Proteins ( lcsh ) ( lcsh )
Ribosomes ( lcsh ) ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 142-154.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by David Earl Matthews.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
025798407 ( AlephBibNum )
03329360 ( OCLC )
AAV1742 ( NOTIS )

Downloads

This item has the following downloads:


Full Text












PROTEINS OF THE M~ttiALIAN MITOCHONDRIAL RTBOSOME


by

DAVTD EARL MATTHEW~S














A DISSERTATTON PRESENTED TO0 THE GRADUATE COUNCIL OF
THE UINIVERSITY OF FLORIDA
IN PARTIAL FULFLLLMLENT OF1 THE REQ!UTREMIENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY








UN'IVERSTTY OF FLORTDA


















ACKNOWLEDGEMENTS


The author gratefully acknowledges the financial support

awarded him in the form of fellowships from the National Science

Foundation and the Graduate Council.


Thanks are also due to Mark Critoph, Warren Clark, and Mark

Moffitt for their technical assistance in these experiments; and to


Nancy Denslow, Robert Hessler, and Mary Conde for helpful discussions.

The senior research project of Dean Kane contributed much to the

studies on the isoelectric points of mitochondrial ribosomal proteins.

The encouragement and assistance of Patty Matthews in the preparation

of this dissertation were invaluable.


The author also wishes to express his gratitude for the helpful-

ness of the faculty of the Department of Biochemistry, and of his


supervisory committee in particular; their good will, advice, and

scientific insights have made the past four years of graduate study

a most rewarding experience. Most of all, the profound influence of

Dr. Thomas W. O'Brien in shaping this research, as well as the author's

scientific training, goals, and attitudes, is humbly and gratefully

acknowledged.








11





















TABLE OF CONTENTS




Acknowledgements ............................................ ii
List of Tables ............................................. V
List of Figures ................... ................... ........... vi
Abbreviations Used ...........................................vi
Abstract ................................................ ix
Introduction .. .. .. . .. .. .. .. .. .. 1

I. Structural properties of non-micochondrial
ribosomes . .. . .. . . . .. .. 3

II. Structural properties of mitochondrial ribosomes ..... 7
A. Mitochondrial ribosomes of protests, fungi

and plants ....................................... 7
B. Mitochondrial ribosomes of animals ............... 12
III. Phylogenetic relationships in ribosome structure......, 17
IV. Summary ........................................... 21
Materials and Methods ......................................... 23
I. Materials ....*...................................., 23
II. Preparation of ribosomes ............................. 25
A. Bovine mitochondrial ribosomes ................... 25
1. Method A **................................... 26
2. Method B . . . . . . . . 27
3. Method C .. .. .. . .. .. .. 27

B. Rat mitochondrial ribosomes....................... 28
C. Bovine cytoplasmic ribosomes...................... 28
Dl. E. coli ribosomes ................................ 29
III. Extraction of ribosomal proteins ..................... 29
IV. Radioactive labelling of ribosomal proteins .......... 30
A. Radioidination ..................30

B. Reductive methylation ............................ 31
V. Twro-dimensional electrophoresis ...................... 31
VI. Isoelectric focusing ................................. 34
VII. Assays of ribosome function .......................... 35
VIII. Buoyant density determinations ....................... 36
IX. Quantitative measurements ............................ 37
Results ......................,,,,............. ........n 38
I. Preparation and characterization of
mitochondrial ribasomes .............................. 39
II. Electrophoretic analysis of mitochondrial
ribosomal proteins.................................. 47






111











A. Criteria for the identification of
ribosomal proteins .................. .................. 52
B. Identification of ribosomal proteins .................. 57
1. Large subunit ..................................... 57
a. Reproducibility ............................... 61
b. Resistance to salt-washing .................... 66
Low-salt treatment ............................ 68

High-salt treatment ..... ...................... 72
Extreme-salt treatment ............. ........... 73
c. Summary of the large-subunit proteins ......... 78
2. Small subunit ..................................... 79
3. Comparison of large and small submnits ........... 91
C. Experimental evaluation of possible artifacts ......... 96
III. Comparison with proteins of non-mitochondrial
ribosomes ............................................ 100
IV. Molecular weights of mitochondrial ribosomal
proteins .............................. ..........~......... 110
V. Comparison with proteins of rat mitochondrial
ribosomes ............................................ 119
Discussion ................................................. 128
Bibliography ................................................-... 142
Biographical Sketch .................. .... .........~................. 155




















LIST OF TABLES


I. Properties of cytoplasmic ribosomes of eukaryotes ........... 4


II. Properties of Moneran and chloroplast ribosomes ..........., 6


III. Properties of mitochondrial ribosomes ....................... 8


IV. Compositions of buffers used in preparation of
ribosomes .... .......................................... 24


V. Functional activity of mitochondrial ribosomes
prepared with or without DEAE-cellulose ................... 48


VI. Appearance of largie-subunit proteins in separate
experiments .. .......................................... 58


VII. Appearance of small-subunit proteins in separate
experiments ........................... ................ .... 82


VIII. Molecular weights of mitochondrial large subunit
proteins ................. ............................1


IX. Molec~ular weights of mitochondrial small-subunit
proteins ..............................................1


X. Calculation of the protein content of bovine
mitochondrial ribosomes ... ................................115



















LISTI OF' FIGURES


1. Phylogenetic relationships in ribosome structure ............ 20


2. Preparation of mitochondrial ribosomes by Method A ........... 41


3. Preparation of mitochondrial r-ibosomes by Method B ........... 43


4. Preparation of mitochiondrial ribosomes by Method C ........... 45


5. Elecilrophoretric pattern of large-subunit proteins
fi.o,, bovine mitochondrial ribosomes ........................ 50


6. Electrophoretic pattern of small-subunit proteins
from bovine mitochondrial ribosomes ........................ Sl


7. Effects of treatment with buffers of various tonic
composition on the peptidyl transferase activity
and buoyant density of havine mitochondrial large
subunits .............................................. 55


8. Schematic diagram of bovine mitochondrial large-
subunit proteins .......................................... 60


9. Electrophoretic pattern of prote.ins from large
subunits prepared in Buffer C (Experiment 2) ............... 63


10. Electrophoretic pattern of proteins from large
subunits prepared in Buffer C (Experiment 4) ............... 64


11. Electrophoretic pattern of proteins from large
subunits prepared in Buffer D (Experiment 6) ............... 65


12. Effect of high-salt treatment on thet protein content
of E. coli ribosomes ....................................... 67


13. Electrophoreftic patterns of proteins from low-sailt
treated large and small subunits .......................... 69


14. Electrophoretic pattern of proteins from large
subunits prepared in Biuffer C (Experiment 3) ............... 70










15, Electrophoretic pattern of proteins from large
subunits prepared in Buffer F ............................... 75


16. Electrophioietic pattern of proteins from large
subunits prepared in Buffer G, ............................... 76


17. Electrophoretic pattern of proteins from small
subunits prepared in Buffer C (Experiment 2) ................ 80


18. Electrophoretic pattern of proteins from small
submnits prepared in Buffer C (Experiment 3) ................ 81


19. Schematic diagram of bovine mitochondrial small-
subunit proteins .......................................... 84


20. Electrophoretic pattern of proteins from small
subunits prepared in Buffer E (Experiment 9) ................ 87


21. Electrophoretic pattern of proteins from small
subunits prepared in Buffer G, ............................... 89


22. Relative electrophoretic positions of bovine
mitochondrial large-subuinit and small-subunit
proteins ................... ..................... ........... 92


23. Relative eLectrophoretic positions of bovine
mitoribosomal and cytoribosomal large-subunit
proteins .............................................. 102


24. Relative electrophoretic positions of mitoribosomal
and cytoribosomal small-subunit proteins ................... 103


25. Isoelectric focusing of mitoribosomal proteins ............... 107


26. The pH gradient formed during isoelectric focusing
of mitoribosomal proteins .................................. 109


27. Electrophoretic pattern of proteins from rat mito-
chondlrial large subunits ................................... 120


28. Electrophoretic pattern of proteins from rat mito-
chondrial small subunits ................................... 121


29. Schematic diagram of rat mitochondrial large-
subunit proteins.................. .......................... 123


30. Schematic diagram of rat mitochondrial small-
subunit proteins .......................................... 124


















ABBREVIATIONS USED


ATP
bisacrylamide


cytoribosome

DNA
EDTA



GTP
Teu
mitoribosome

MigAc2
ML
IS
mRNA
phe
poly U
POPOP
PPO
RNA
rRNA

SDS
TEA
TEMED
Tris
tRNA
B-ME


adenosine-5'-triphosphate
N, N' -methylenebisac:rylamide
cytosine
cytoplasmic ribosome
daltons
deoxyribonucleic acid
(ethylenedinitrilo)tetraacetic acid
gravity
guanine
guanosine-5'-triphosphate
leucine
mitochondrial ribosome
magnesium acetate
mitoribosomal large-subunit protein
mitoribosomal small-subunit protein
messenger RNA
phenylalanine
polyuridylic acid
3, 4-bis[2-(5-phenyloxazolyl)]benzene
2, 5-diphenyloxazole
ribonucleic acid
ribosomal RNA
Svedberg unit
sodium dodecal sulfate
triethanolamine
N, N, N', N'-tetramethyl -ethylenediamine
tris(bydroxymethy1)aminomethane
transfer RNA
B-mercaptoethano1















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



PROTEINS OF THIE IIAMMALZLAN
M1ITOCHONDRIAL RIBOSOME


By

David Earl Matthews


December, 1976

Chaiirman: Thomas W. O'Bricn
Major Department: Biochemistry

For some years it was widely believed that all ribosomes were

either of the 70S type found in bacteria or of the 80S type occurring

in the cytoplasm of eukaryotic cells. With the discovery and subsequent

characterization of ribosomes in the mitochondria of eukaryotes, this

simple generalization had to be rejected. The mitochondrial ribosomes

of various organisms display a great diversity of structural properties,

even though they all show considerable functional homology with the 70S

prokaryotic ribosome. Animal species possess mitochondrial ribosomes

with particularly interesting physical-chemical properties, by virtue

of their high protein content relative to the quantity of RNA they

contain.


The present research~ is an investigation into the nature of the

individual proteins which maki up the large complement of total protein










found in mammalian mitochondrial ribosomes. Two-dimensional electro-


phoresis in polyacrylamide gels was used to separate, identify, and

characterize the proteins of mitochondrial ribosomes from bovine liver.

It was found that these ribosomles contain a relatively large number of

proteins: 52 are present in the large subunit and 41 in t-he small sub-

unit. Several kinds of criteria and experimental evidence argue that

these proteins are true ribosomal components rather than contaminants

in the riho~some preparations. However, only 81 electrophoretically

distinguishable proteins are present in a mixture of large and small

submnits; the possibility that some of the large-subunit proteins are

identical to proteins found in the small subunit is discussed.

The molecular weights of these proteins are similar to those of

bovine cytoplasmic ribosomal proteins, and considerably larger than

those of Escherichia coli ribosomal proteins. Comparisons of electro-

phoretic properties show that the mitochondrial proteins are more acidic

than those of mammalian cytoplasmic or most bacterial ribosomes. The

sum of the molecular weights of the 93 mitochondrial ribosomal proteins

is somewhat greater than predicted by physical-chemical measurements of

the total mass of protein in this ribosome; to account for this diiffer-

ence it is suggested that a proportion of the ribosomes as they are iso-

lated may be lacking some of the ribosomal proteins.

Other experiments demonstrated that exposure of the ribosomes to

certain ionic conditions could remove a small number of the proteins,

with a corresponding loss of the peptidyl transferase activity of the










ribosomes. One or more of the proteins removed may therefore be

involved in this particular ribosomal function. Finally, comparisons

between the proteins of bovine and rat mitochondrial ribosomes revealed

a large number of differences in their electrophoretic properties. The

level of divergence in these properties is greater than that reported

in similar comparisons of cytoplasmic ribosomal proteins. This obser-

vation parallels the great phylogenetic diversity observed in the over-

all structural properties of mitochondrial ribosomi!es fr~om more distantly

related organisms, and leads to the conclusion that these ribosomes have

diverged more widely during evolutionary history than have their extra-

mitochondrial counterparts located only a few Angstroms; away.

The large number and low isoelectric points of the proteins of

mammalian mitochondrial ribosomes are discussed with reference to their


implications for the structural organization of these particles. Some

mechanisms which may account for the rapid evolutionary divergence of

mitochondrial ribosomes are proposed.

















INTRODUCTION

Mitochondria are multifunctional organelles found in all eukary-

otic cells, their primary function being the aerobic production of ATP.

In addition to their important role in cellular energy metabolism,

mitochondria make an essential contribution to their own biogenesis.

Several of the multi-subunit enzymes of oxidative phosphorylation con-

tain one or more subunits synthesized within the mitochondrion (Schatz

and Mason, 1974). For this purpose mitochondria possess a large com-

plement of biosynthetic enzymes and other macromolecules, distinct from

their analogs in the nucleus and extramitochondrial cytoplasm. These

components of the mitochondrial biogenetic system include DNA, DNA and

RNA polymerases, messenger RNA, ribosomes, translation factors, transfer

RNA species and aminoacyl-tRNA synthetases. Mitochondrial DNA codes

for the ribosomal RNA and at least some of the tRNA and mRNA species

found in mitochondria, and these macromolecules are indispensable in the

biogenesis of functionally active mitochondria. On the other hand most

of the (equally essential) protein components of the mitochondrial bio-

genetic system --- ribosomal proteins, factors, and enzymes --- appear

to be coded by nuclear DNA and synthesized on cytoplasmic ribosomes.

Thus the mitochondrial and nuclear-cytoplasmic macromolecule-synthesizing

systems must cooperate as intimately in the production of the mitochon-










drial biogenetic system itself as they do in the synthesis of the


enzymes of oxidat-ive phosphorylation.

The ribosomes in a given organism's mitochondria are generally

distinguishable from its cytoplasmic ribosomers on the basis of sev-

eral functional or physical-chemical criteria. Indeed, especially in

their functional properties, mitochondrial ribosomes have been found

to be more like bacterial ribosomes than cytoplasmic ones. Some

structural similarities between mitochondrial and Moneran ribosomes


were noted early (Kiiatzel and Noll, 1967), and complemented reports

of other biochemical homologies between organelles and prakaryotes

that had already aroused considerable interest in the question of the

evolutionary origin of mitochondria and chloroplasts. More recent


comparisons of the structural parameters of bacterial and mitochondrial

ribosomes have shown some similarities but also a surprising number of

differences, both between the two groups and among mitochondrial ribo-

somes from different species. Thus, it is true that the mitochondrial

ribosomes of most organisms studied to date sediment more slowly than

the corresponding cytoplasmic ribosomes, and that some of them have

sedimentation coefficients close to that of prokaryotic ribosomes (705).

However, mitoribasomes from various species range in sedimentation

rates from 55S to 80S, a much wider variation than is found among

bacterial, eukaryotic~-cytoplasmic, or chloroplast ribosomes obtained

from different organisms. Indeed, when all ribosomal attributes are

considered, it seems that: only in mitochondria do so many different

kinds of ribasomes occur.





I. STRUCTURAL PROPERTIES OF NON-MIITOCHIONDRIAL RIBOSOMES

Before proceeding to the description of the physical and chemical

characteristics of mitochondrial ribosomes, it will be useful to sum-

marize the characteristics of the other kinds of ribosomes that exist.

This task is simplified by the fact that the characteristics of non-

mitochondrial ribosomes are not as divergent as might be thought. They

all fit reasonably well into two large categories within which the mem-

bers seem to share more similarities than differences. Table 1 shows

the relative homogeneity of the properties of cytoplasmic ribosomes,

whatever eukaryotic organism they are obtained from---protists, fungi,

plants or animals. Despite small differences among the cytoplasmic

ribosomes from these four taxonomic kingdoms, all of them appear to be

members of a single structural class typified by a sedimentation co-

efficient of 80S, a buoyant density of 1.57 g/cc, and rRNA molecules

of 0.7 and 1.3 million daltons containing 50 percent G + C.

The second category of non-mitochandrial ribosomes is also rela-

tively uniform in physical and chemical properties. Described in

Table II are several prokaryotic ribosomes---from E. coli, a mycoplasm,

and a blue-green alga---and those of various chloroplasts. The pro-

karyotic particles can all be adequately described by the values 70S,

1.64 g/cc, (0.56 + 1.10) x 106 daltons, and 50% for the structural

parameters tabulated. Chloroplast ribosomes are similar in all respects

except that some of them appear to be significantly lower in buoyant

density. Perhaps it should be mentioned that this homnogeneity of



















HM
LA try


U NN

















17 LA *


+ +







-






















CO
3NO
m CO O
Ir v 0


NN r
-1-
J -4


aaW


0 LD c a o




o a H





oe ca 1
04a rio
B B'di
,C ol 0 m
0 *HUM 0
rlH DNr
U O r- Oa I
Co g ue
Y7u gO
OC < ,C

rdm00
,C U Oii
"00 "
,CD0



H O ,C O





Cl r a *4 4
00 Hi O


Cro m r

,CQMO rd r(


0 N
r a to


vl I
rl to


3
~f~D m~
VIV)VI ~V]
~om ~o
mmm hao


2 2
me a
co a
06 0
RO OO
> "O
,C > 0 a
0 0 O ,C
0 U
WHOU
0 -C 0 <
H 0 CO

















0\5









as

lri O J Af r
~D~i~ ON
o~o ~ e *
.o .
-r -o m
LO N 0 ON~~m
*~ ON 01 0 0 Ht




H *Hmm AM

m Ci H 0 tr
Hut FNNOaD




& o c: c o ca o 2 c
G-a C O cel Nr i











u *~ D *r
2 1 01 Im.r





2 3 m r| -a 5 o a
m i 0l'
C ) 2 M c H a


*rlm c N M Fi
C ON:ct 0



















rI Y1


O

ilO L t







Co Ca O f3
I1 r- D





10 LO N 40


Ln
LD
..


~LO N N


ON N





co

H m
mt to
o LD


m H
00 r
0 *H ,

*H 0


01




m HO

O M





O*HCOO
H .0 0 m
a0 C) H *H
0 C) 0 c
*Hr N o O O


SE
'4 c
o ;
UO~ M

.0 CO 0


m04 ee~ ao

M 0 1 I CA .r


m m u





gross structural properties almost certainly conceals a great deal of

diversity in fine structure. Differences in such properties as ribo-

somal protein electrophoretic mobilities and immunologic identities are

the rule even for prokaryotic ribosomes related to each other more

closely than are the entries of Table II (Geisser et al., 1973;

Wittmann et al., 1970).

II. STRUCTURAL. PROPERTIES OF MIITOCHONDRIAL RIBOSOMES

In contrast to the rather simple classification scheme possible for

ribosomes from all other sources, mitochondrial ribosomes do not seem to

fall into one or even a few structural categories. Reference to Ta-

ble III shows the degree of diversity found in mitochondrial ribosomes

from different species, Elhny of the individual ribosome species in this

table are as distinct in physical and chemical properties from each

other as the 80S (Table I) and 70S (Table II) classes are. Furthermore,

among the protests and fungi no two genera have yet been shown to con-

tain similar mitoribosomes, so it seems likely that many more structur-

ally different ribosomes will be found as other species are investi-

gated. On the other hand, considerable homology is seen among the

mitoribosomes of several species of higher animals, from locust to man.

A. Mitochondrial Ribosomes of Protists, Fungi, and Plants

The best-characterized protist mitochondrial ribosomes are those

of Euglena gracilis and Tetrahymena pyriformis. The Euglena mitoribo-

some is unusual in that it is the only mitochondrial ribosome yet

described that shows a large degree of structural homology with the





J LA

m Ca m m


.t a _7
EN c en a
m~hm
2f`4J4


oa
+ +



- o


N N




+ +
eco

0 O


com






000


NNm
~nmo


m in
N~JaJ


CO
~r O
vvl m
rl o
r- co
rr
rn
m


c.0 CD
-O


tn CO
NN


E 4
o 4
LO 3
O 3


OO


OI 0 0
r **3
MQ H
"O "


UU -
illm8
,C C *


CLi 0>u


< 82 5 a t



H H
OZ






























0 ~ Mf O
OI~~ m mm



92 * 0 m rl 0 - 0

~~~ h 0r OD 01 rl 0
e a0 r r O1O~ C
~~~ E C 3 c> 0 u 0
O O O O to r*v] rd N O ar


0 N m a 0 a > a 0
N NJ N N N N N m al m a





Mi O O O c < r 0 O 4~
* aN e c r ul c *r O E a
O m m r 0 a c or r N 0-u u
'r H m O a 0 c A N *r






|cy O < 0 0 0| *

0~ C 0 j 0 a Q U

OI I 0 0 O O1 00
r( 0 * M W 0 m
,"1 > OH *H
M O N 0 T-H 5 0 0 m
"H *H i m m A D .C *Hl
0 L 0 O O










ribosomes of prokaryotes. The fact that only one of the many mitori-

bosomes shown in Table III is similar to bacterial ribosomes in its

physical-chemical properties is quite remarkable, in view of the

extensive functional homologies between mitochondrial and bacterial

ribosomes that have been found in every case tested. And even in this

instance the structural similarity is by no means complete. Although

the mitoribn~some of Euglena is similar to the E~. coli ribosome in its

sedimentation coefficient, the molecular weights of its rRN1A, and its

high buoyant density, the base composition of its rRNA (27% C + C) is

very different.

Tetrahymena mitoribosomes are quite dissimilar from those of

Euglena, and indeed from any other ribosome yet studied. Although the

Tetrahymena mitoribosome shares with that of Euglenay an exceptionally

low rRNA GC content, it more nearly resembles cytoplasmic ribosomes

withi respect to its sedimentation coefficient. And its buoyant density

and rRNA molecular weight values are smaller than those of any other

ribosomie except the miitoribusomes of animals.

The mitochondrial ribosomes of various fungi also show a rather

large diversity in sedimentation coefficients and buoyant densities.

Unfortunately, a nearly equal diversity in these parameters has been

reported by different research groups studying the same fungal species.

Thus the reported sedimentation coefficients range from 72S to 80S for

Sacchar-omyces mitoribosomes, and from 73S5 to 80S for those of Neuro-

spora. Although the buoyant density of a 74S form of the Saccharamyces










mitoriboscme was found to be 1.64 g/cc (Grivell et al., 1971), the

composition of the 80S version (Miorimota and Halvorson, 1971) corre-

sponds to a buoyant density of 1.56 g/ce. There is reasonably good

evidence that the 80S ribosomes described in these species really are

mitochondrial ribosomes rather than cytoplasmic contaminants (O'Brien

and Matthews, 1976). It therefore seems most likely that the contra-

dictory reports are due to artifactual alterations of the structure

of these ribosomes by som. of the preparative procedures used. How-

ever it is not yet clear which of the results are artifactual, so it

is difficult to compare the sedimentation coefficients and buoyant

densities of fungal mitoribosomes with those of other kinds of ribo-

somes at present.

There is much less controversy about the properties of the rRNA

of fungal mitoribosomes. Indeed, the mo~lecular weights and GC contents

of these molecules are quite similar in all the fungal species tested.

The CC contents are rather low, nearly as low as those found in pro-

tist mitoribosomal RNA. The molecular weights, on the other hand, are

quite high: these molecules are of about the same size as the rRNAs

found in cytoplasmic ribosomes.

The properties of mitochondrial ribosomes from higher plants are

not firmly established. The best-documented results are those shown

in Table III. It may be seen that the physical-chemical character-

istics of these ribosomes are quite similar to those of cytoplasmic

ribosomes in general. In fact, they differ only slightly or not at










all from those of the cytoplasmic ribosomes found in these same

organisms (Leaver and Harmey, 1973; Pring and Thornbury, 1975; Pring

and O'Brien, 19;1-, private communication). Such results raise the

possibility that these ribosomes are actually of cytoplasmic rather

than mitochondrial origin. This interpretation is supported by the

observation that these ribosomes are inhibited by anisomycin and not

by chloramnpbenicol, like cytoplasmic ribosomes and unlike any other

known mitochtondrial ribosomes (Pring, Denslow, and O'Brien, 1975,

private commlnunication).

There haive been isolated reports of plant mitoribosomes with

properties different from those described above. A 70S particle was

obtained from mung bean mitochondria (Vasconcelus and Bogorad, 1971),

and a 66S mitoribosome from maize (Wilson et al., 196,8). Mitochondrial

RNA from Virginia creeper was found to be relatively small, 0.42 x 106

and 0.84 x 106 in molecular weight. However, the functional properties

of these ribosomes have not been described, and the uncertainty about

the nature of mitochondrial ribosomien in higher plants remains unresolved.

B. Mitochondrial Ribasomes of Animals

In contrast to the fascinating and perplexing diversity of mito-

chondrial ribosomes from protist, fungal and plant species, Metazoan

mitoribosomes appear strikingly uniform in structure. Although most

of the animal species investigated have been mammalian, the mitoribo-

somes of the toad Xenopus have also been thoroughly characterized, and

enough data have been presented for several invertebrates to justify





a tentative conclusion that the littoribosomes of all multi-cellular

animals may be quite similar. Furthermore, this relatively homo-

geneous group of ribosomes is distinctly different in structural

properties from any other ribosomes yet described.

The first difference to be noticed was the low sedimentation

coefficient of animal mitoribosomes. Values within the range 54S to

61S have been obtained for these particles from rat (0'Brien and Kalf,

1967), HeLa cells (Perlman and Penman, 1970), rabbit, pig, cow (O'Brien,

1971), ha:nd~er (Coote et al., 1971), chicken (Rabbitts and Work, 1971),

toad (Swanson and Dawid, 1970), shark (0'Brien, 1972), and locust

(Kleinow et al., 1971). The early inference drawn from these low

sedimentation coefficients was that animal mitoribosomes were smaller

(lower in molecular weight) than other ribosomes (Borst and Crivell,

1971), and this inference was reinforced by the later observations of

unusually small rRNA, in these particles. An alternate possibility,

that the 55S particle is actually a subunit of the functional mono-

ribosome, has been excluded by the dissociation of the 55S structure

into two subunits and the demonstration t~hat eithler the 55S particle

or a mixture of both subunits is competent for poly U-dependent

phenylalanine incorporation (Leister and Dawid, 1974; O'Brien et al.,

1974).

But there is a third interpretation of the low sedimentation

coefficient of animal mitoribosomes, because in point of fact these

55S particles are not exceptionally small, either in molecular weight









or in physical dimensions. The particle weight of thle bovine mito-

ribosome as determined by high-speed equilibrium centrifugation is

2.8 mill-ion daltons (HIamilton and O'Brien, 1974), slightly greater

than the value reported for the ribosome of E. coli (Hill et al.,

1969). De Vries and Kroon (1974) have presented evidence that rat

mitoribosomes are even larger in volume than the E. coli particles,

though smaller than rat cytoribosomes. These investigators electro-

phoresed ribosomes into gels composed of a gradient of polyacrylamide

concentration until the particles could make no further progress

through the decreasing pore size of the gel matrix. The mitoribo-

somes penetrated farther into the gels than cytoribosomes but not as

far as bacterial ribosomes. The physical dimensions of ribosomes can

also be detennined by electron-microscopic measurements. The results

confirm the conclusion that mitoribosomes are smaller than cytoribo-

somes in rat (0'lHrien and KaLf, 1967; Aaij et al., 1972) and locust

(Kleinow et al., 1974), but direct comparisons of mitochondrial and

bacterial ribosomes (under the same conditions of fixation and stain-

ing) have not been performed.

How can two particles of the same molecular weight sediment

respectively at 55S and 70S? The significant difference appears to be

the much lower buoyant density of the animal mitochondrial ribosome.

Buoyant density values fromt 1.40 to 1.46 g/cc have been found for

mitoribosomes of HleLa cells (Perlman and Penman, 1970; Wengler et al.,

1972), rat (Sacchi et al., 1973; de Vries and Kroon, 1974), cow










(Hamilton and O'Brien, 1974), and toad (Leister and Dawid, 1974). The

buoyant density of a ribosomre can be used to calculate thle relative

proportions of RNA and protein in the particle (Hamilton, 1971); from

the relationship given in the notes to Table I an RNA content of about

30 percent can be calculated for animal nitoribosomes, as contrasted

with 63 percent for bacterial ribosomes (Tissibres et al., 1959). From

the RNA content and the sum of the molecular weights of the rRNA mole-

cules, the particle weight of the ribosome may be calculated. By this

meanis molecular weight estimates equal to or greater than those for

bacterial ribosomes were obtained for mitoribosomes from rat (Sacchi

et al., 1973; de Vries and Kroon, 1974), cow (O'Brien et al., 1974), and

toad (Leister and Dawid, 1974) even before the molecular weight was

determined directly by sedimentation equilibrium (Hamilton and O'Brien,

1974).

Doubts have been raised about the validity of buoyant density

values as measures of the protein content of ribosomes (MlcConkey, 1974).

Others have suggested that the low buoyant densities of animal and

Tetrahymena mitoribosomes might be due to membrane fragments adhering

to these particles (specifically, due to membrane lipids, which are

assumed to be absent in the calculation of protein content from buoyant

density) (Borst and Grivell, 1971). Several lines of evidence indicate

that these factors do not represent significant objections to the

description of animal mitoribosomes presented above. Determinations

of the protein content by either chemical analysis or ultra-violet





absorption spectra of the ribosomes are in agreement with estimates

from buoyant density, yielding values of 70-80 percent protein in rat

(0'Brien and Kalf, 1967) and toad (Leister and Dawid, 1974). No phos-

pholipids were detectable in rat mitoribosomes (de Vries andl Kroon,

1974). The molecular weight estimate for bovine mitoribosomes baised

on their buoyant density and the size of their rR~NA agrees quite closely

with the molecular weight determined by sedimentation equilibrium

(Hamilton aind O'Brien, 1974). Finally, similar values for the total pro-

tein content of Xenopus mitoribosomes have been obtained dircctiv by

summing the molecular weights of the individoll ribosomal proteins

(Leister and Dawid, 1974).

Besides their unusual sedimentation behavior and high protein

content, another unusual characteristic of` animal mitoribosomes is the

small size of their rRNA molecules. Values of 0.35 and 0.54 million

daltons have been obtained for HeLa mitoribosomal RNA by electronmicro-

scopic length measurements (Robberson etal., 1971), and these numbers

are in good agreement with those found by other methods for rat (Sacchi

et al., 1973), toad (Dawid and Chase, 1972; Leister and Dawid, 1974),

shrimp (Schmitt et al., 1974), and locust (Kleinow, 1974). Thus animal

mitoribosomes contain scarcely more than half as much RNA as any non-

mitochondrial ribosome known. To a first approximation, these particles

may be pictured as E. coli ribosomes modified by converting half of the

RNA into an equal mass of protein.





The base composition of the rRNA~ of animal mitoribosomes is

distinctly higher in guanine and cytosine than that of protist or fungal

mitoribosomnes, though still lower than that of animal cytoplasmic ribo-

somes. G + C contents of 40-47 percent have beeni found for mitochan-

drial rRNA of HeLa (Vesco and Penman, 1969), rat (Bartoov et al., 1970),

toad (Dawid and Chase, 1972), and shrimp (Schmitt et al., 1974). The

degree of methylation is also lower for animal mitoribosomal RNAs than

for either cytoplasmic or prokaryotic rRN~s (Dubin, 19)74).

Although mitochondrial ribosomes from all animal species examined

thus far appear quite similar in their physical and chemical properties,

it may be anticipated that they will differ in their detailed structure.

In fact, detectable non-homology has already been found between the base

sequences of mitoribosomal RNA from two species of toad, Xenopus laevis

and X. mulleri (Dawid, 1972). Electrophoretic differences have also been

demonstrated in several of the mitoribosoma~l proteins of these two spe-

cies (Leister and Dawid, 1975).

TIT. PHYLOGENETIC RELATIONSHIPS IN RIBOSOM(E STRUCTURE

The comparisons presented above permit some conclusions about the

variation in structural properties among mitochondrial ribosomes of

different organisms, and the differences between the ribosomes of mito-

chandria and those of prokaryotes, eukaryotic cytoplasm, and chloroplasts.

The most obvious generalization is that mitochondrial ribosomes show

more diversity in all of their structural properties than do any other

kinds of ribosomes. Cytoplasmic ribosomes of all four eukaryotic kingdoms





are relatively similar with respect to sedimentation coefficient,

buoyant density, rRNA size, and guanine plus cytosine content. Pro-

karyotic ribosomes display even more uniformity in these characteris-

tics, while chloroplast ribosomes, whether obtained from protests or

from higher plants, show remarkable homologies not only among them-

selves but to a large extent between themselves and prokaryotic ribo-

somes. Mitochondrial ribosomes, on the other hand, can vary in sedi-

mentation coefficient even within a kingdom: such differences do exist

between the protests Eugilena and TetrIrahyea and almost certainly

exist among the fungi. Similarly, differences in the buoyant densi-

ties of mitoribosomes are found within both the protist and the fungal

groups. Mitoribosomal RNA molecular weights and G + C contents, in con-

trast, appear to vary significantly between kingdoms but not within them.

The correlation between these last two structural properties of

mitochondrial ribosomes and the taxonomic kingdoms in which they are

found is illustrated in Figure 1. Also plotted in Figure 1 are the data

for chloroplast and prokaryotic ribosomes, which cluster together as

expected, and for four kingdoms of cytoplasmic ribosomes. The latter

group also forms a cluster, with the exception of the cytoribosome from

Tetrahymena. Both of these groups are dissimilar from any of the three

mitoribosome kingdoms plotted, and these are in turn distinct from each

other. Thus the two parameters, rRNA size and G + C content, discrim-

inate ribosomes along the lines of their phylogeny and intracellular

location,





Figure 1. Phylogenetic relationships in ribosome structure.
$)Mitochondrial Ribosomes
Fungi


S Sacchuomyces

A Aspergillus
C Candida utilis
N Neurospora

Protists
Eu Eugle~na

T Tetrahymena

An imals


H1 HetLa

R Rat
iProkaryotic Ribosomes
E Eschericiia
coli
M ElERoplasma
hominis
1Chloroplast Ribosomes
Eu Euglena

Sp Spinach
C)Cytup'Lasmic Ribosomes
Fungii
S Saccharomyces

A Aspergillus
C Candida utilis
N Neurospora

Protists
Eu Euglena

T Tetrahymenn

An imnals
X XYenopus
H HeLa

R\ Rat
Plants
P Pea


Reijnders et al., 1973; Mlorimouto and
Halvorson, 1971.
Vermai et al., 1970; Edelman et al., 1970.
Vignais et al., 1972.
Neupert et al., 1969; K~ntzel and Nol1,
19(,7.

Krawiec and Eisenstade, 1970; calculated
from Avadhuani and Buetow, 1972.
Recijnders et al., 1973; Chi and Suyama,
1970.

Leister and Dawid, 1974; Dawid and Chase,
1972.
Robberson et al., 1971; Vsco and Penman,
1969.
Sacchi et al., 1973; Bartoov ct al., 1970.

Kurland, 1960; Morimoto and Halvorson,
1971.
Johnson and Horowitz, 1971.


Rawson and Stutz, 1969; calculated from
Avadhoani and Buetow, 1972.
Hartley and Ellis, 1973; Lyttleton, 1962.


Reijnders et ;1l., 1973; Morimato and
Halvorson, 1971.
Verma et al., 1970; Edelman et al., 1970.
Vignais et al., 1972.
Neulpert et al., 1969; KUntzel and Noll,
1967.

Krawiec and Eisenstadt, 1970; calculated
from Avadhani and Buetow, 1972.
Reijnders et al., 19)73; Chi and Suyamia,
1970.

Loeniing ct al., 1969; Dawid et al., 1970.
Darnell, 1968; calculated from Vesco and
Penman, 1969.
Reijnders et at., 1973; Kirby, 1965.

Loening et al., 1969; Bonner and Varner,
1965.














_I __


_I~


I' '
O 30 40 50 60


, OC la-t



O Eu
~~NooS


2.5


i/S N1
-A C *



/ Eu
-I *
T Ti


2.0


1.5




1.0


tXH H
\ ** R


0.5


Content roless percent)


RNA G+C









Sedimentation coefficients and buoyant densities, on the other

hand, do not group mitoribosomes from the same kingdom together; nor

do they differentiate them from the various non-mitochondrial ribo-

somes. This observation suggests that these two structural proper-

ties have been less conserved than rRNA size and G + C content in the


course of ribosome evolution. It is noteworthy that buoyant density

is the only one of these four structural characteristics which

discriminatces chloroplast from prokaryotic ribosomes (Stutz and

Boschetti, 1976). However, some of the intrakingdom variability


reported for sedimentation coefficients and buoyant densities may be

due to the possibly greater sensitivity of these parameters to differ-

ing conditions used in the preparation of the mitoribosomes.

IV. SUMMAZRY

Perhaps the most remarkable structural property of mammalian

mitochondrial ribosomes is the very high ratio of protein to RNA found

in these particles. On the basis of various kinds of physical-chemical

measurements it appears that this ratio is approximately 2:1, or 1.8

x 106 daltons of protein to 0.9 x 106 daltons of RNA. These ribosomes

thus contain about twice as much protein as E. coli ribosomes, and

about half as much RNA.


Such an unusual composition poses questions about the molecular

architecture of these particles, and about the functional roles played


by this large quantity of protein in the process of protein synthesis.

Is the difference in protein content between the ribosomes of E. coli





and those of mammalian mitochondria due to a difference in the number

of the ribosomal proteins or to a difference in their size? Do the

different kinds of molecular interactions which must be involved in

maintaining the structural integrity of such a protein-rich ribosome

correlate with any identifiable differences in the properties of the

proteins? on the other hand, might the high protein content found in

this ribosome represent nothing more significant than the presence of

large quanlities of contaminating non-ribosomal proteins? A portion

of the present research is directed toward these questions.

Also of interest is the phylogenetic diversity in structural

properties which is found in comparisons of mitochondrial ribosomes

from distantly related organisms. Only small differences in the over-

all physical and chemical properties of mitoribosomes are seen when

species within the animal kingdom are compared. But if a high degree

of evolutionary divergence is indeed the rule for mitochondrial ribo-

somes in general, it should be possible to detect significant differ-

ences between the mitoribosomes of more closely related species at a

sufficiently detailed level of analysis. The rule further predicts

that such differences will be greater between mitochondrial ribosomes

than between the cytoplasmic ribosomes of the same species. The

experiments presented below include a test of these predictions.





RATERILrS AND METHODS

T. MIATERIALS

Sucrose (density gradient grade, ribonuclease free) was obtained

from Schwarz/M~ann. Urea (reagentt-grade) was purchased from JT. T. Baker

or frcm Manllinckradt. Stock solutions of 10 M urea were prepared,

filtered, and stored at room temperature for no muree than 24 br before

use; if the conductivity of the stock solution was greater than 50 pmbo,

it was deianized by stirring with AG:501-X8 resin (Bio-Rad). Acrylamide

and N, N'-methylene-bisacrylamide (Eastman) were recrystallized from

chloroform and acetone, respectively; stocks solutions were deionized

with Rexyn-300 (Fisher) and stored at room temperature in the dark.

Ampholine was obtained from LKB. Bovine serum albumin and human

y-globulin were obtained from~ Nutritional Biochemicals Corporation, and

egg albumin, equine myoglobin and egg-white lysozyme were from Sigma

Chemical Company. Puromycin dihydrochloride was from Nutritional Bio-

chemicals Corporation. Nal251 (carrier-free, in 0.1 N NaOH) and [14C]

formaldehyde (44 mCi/nmmal) were purchased from New England Nuclear;

and [4,5_3H}]L-leucine (55 Ci/mmol), 18_3H] G:TP (12 Ci/mmal, tetrasodium

salt), and [3H]LI-phenylalanine (7 Ci/mmol) from Schwarz/Mann.




























~

I--








H




o
r-

;r:
a W

cn
W
'3


~I


N
i3
CI
W
L3
*-(




o
ul

~ N
N
ri U












1L:

C

o -I

o


m

4

r- m
a

a
to
rl
m
rl H




o





I


"

L~I

ID
N
rl
N U
C~
o
eo
r "


o


ri
ri
u
r(
u


r

o
o
~o


u ~





Sa



CO t






OO ,


moumi


FrC C H O M M E ROO





II. PREPARATION OF RIBOSOMIES

A. Bovine Mitochondrial Ribosomes

In the course of this research the procedure for the preparation

of mitochondrial ribosomes from bovine liver was modified in several


ways to improve the yield and purity of the ribosomes obtained. The

major changes are discussed under Methods A, B, and C below. All of

the preparative methods shared the following common features. Livers

of freshly killed animals were obtained from a slaughterhouse and

transported to the laboratory on ice. All subsequent procedures were

performed in the cold unless otherwise noted. Four to 8 kg (fresh

weight) of liver were passed through a meat grinder, diluted with 4

volumes of Buffer K or Buffer L, strained through a coarse-mesh cloth

screen, and homogenized. In the earlier preparations homogenization was

performed with a Potter-Elvehjem homogenizer modified so that the ground

tissue could be pumped through it continuously. Increased cell breakage

and increased final yields of mitochondria were obtained with the use of

a high-frequency dispersion device (a Tekmar Company Super Dispax,

Model SD-45K) .

Unbroken cells and nuclei were removed by pumping the homogenate

at 880 ml/min through a Vernitron CFR-2 continuous-flow rotor rotating

at 11,00)0 rpml in a Vernitron LCA-2 centrifuge. Mitochondria were

harvested fromr the supernatant by continuous-flow centrifugation in a

Beckman JCF-Z rotor at 18,000 rpm and a flow rate of 440 ml/min. After

the mitochondria had been washed as described below, they were









resuspended to a concentration of 20 mg protein/ml in buffer and lysed

by the addition of non-ionic detergent. The lysate was clarified by

centrifugation in a Beckmaun Type 35 rotor at 28,000 rpm for 10 min.

The supernatant (after treatment with DEAE-cellulose, in the case of

Method C below) was centrifuged in a Beckman Type 35 rotor at 35,000 rpm

for 12 br. The ribosome pellets obtained at this step are referred to

as "crude ribosomes" below.

1. Method Ai

The bulffer used for th~e homogenization of the liver and the prepa-

ration of mitochondria was Buffer K. The mitochondria were washed three

times by resuspending in Buffer K and centrifuging in a Beckman JA-10

or Sorvall GS-3 rotor at 8,000 rpm for 10 min. Washed mitochondria

were suspended in Buffer A containing 34% sucrose (ribonuclease-free) ,

50 ug/ml heparin and 200 ug/ml oligonucleotides (prepared by partial

base hydrolysis of yeast tRNA according to Spencer and Poole (1965)).

Triton X-100 and sodium deoxycholate were added to final concentrations

of 2% and 0.5% respectively, and crude ribosomes were prepared ircom the

mitochondrial lysate as described above.

The ribosome pellets were resuspended in Buffer A containing

50 plg/ml heparin and 550 pg/ml puromycin, and incubated at 370C for

5 min. The ribosomes werr then purified by sedimentation into a linear

10-30% sucrose density gradient made up in Buffer A, in a Beckman SW27

rotor. After centrifugation the gradient was pumped through the flo~w

cell of a Gilford Model 2400 spectrophotometer, and its absorbance at

260 nm was recorded. One-ml fractions were collected.





2. Metthod B

Buffer L was used for the preparation and washing of the mito-

chouldria. After two washes the mitochondria were resuspended to a

concentration of 10 mg protein/ml in Buffer L containing 50 pg/ml

digitonin. The suspension was stirred for 15 min, and the mitochondria

pelleted at 8,000 rpm for 10 min. The mitochondria were washed once

more with Buffer L, resuspended to 20 mg protein/ml in Buffer H, and

lysed by the addition of Triton X-100 to a concentration of 1.6%. After

clarification as described above, the lysate was layered onto 20 ml of

Buffer H containing 34% sucrose and 1.6% Triton X-100, and centrifuged


to prepare crude ribosomes. Purification of the ribosomes by sucrose-

density-gradient centrifugation was as described above, except that

Bulffer H was used instead of Buffer A.

3. Method C


To thie clarified mitochondrial lysate prepared by Method B was

added a moist cake of DEAE-cellulose equilibrated in Buffer H containing

1.6% Triton X-100. The quantity of DEAE-cellulose used was about 1 g

(dry weight) per 50 ml of lysate. The slurry was stirred for 30 min,

placed in a large Buchner funnel, and filtered just until all excess

liquid was removed. Ten ml of Buffer H/Triton per gram DEAE-cellulose

was added and filtered off as before. The filtrates were discarded

and the cake of DEAE-cellulose was stirred for 30 min in Buffer I

(10 ml/g) to elute the ribosomes. The slurry was filtered and rinsed

with 2.5 ml/g of Buffer I. The ribosomes in the filtrate were then





centrifugied through a layer of Buffer H/1.6% Triton/34% sucrose and

purified as in Method B. By this procedure about 10 mg of ribosomes

cruld be obtained from 6 kg of liver.

B. Rat Mitochondrial Ribosomes

The livers of 20 to 50 young (80-100 g) female Sprague-Dawley

rats were homogenized in 4 volumes of Buffer L with a Potter-Elvehjem

homogenizer. Rapidly sedimenting material was removed by centrifuga-

tion at 3,000 rpm for 10 min in a Beckmran JA-10 rotor, and mitochon-

dria were obtained from the supernatant by centrifugation at 8,000 rpm

for 10 min in the same rotor. The mitochondria were washed as de-

scribed for bovine mitochondria above (Method B) except that the

concentration of digitanin used was 17 pg/ml. The washed mitochondria

were resuspended to a concentration of 5 mg protein/ml in Buffer H

and lysed by the addition of Triton X-100 to 1%. Riibosomes were then

prepared and purified as in Nethod B above. The yield was about 50 ipg

of ribosomes per rat liver.

C. Bovine Cytoplasmic Ribosomes

Cytoplasmic ribosomes were prepared from a microsomal fraction of

bovine liver. The liver was homogenized in 4 volumes of 25 mM KCL,

5 rmM MgC12, 0.34 M sucrose, 5 mMI R-mercaptoethanol, 10 mM Tris, pH 7.5.

Mitochondria and larger particles were removed at 8,000 rpm for 10 min

in a Beckman JA-10 rotor, andi microsomes were pelleted from the supe~r-

notant at 9,000 rpm for 45 min in the same rotor. The pellet was

suspended ini suffer M and Triton X-100 was added to 2%. Ribosomes









were then prepared as described above except that Buffer M was used

instead of Buffer A or H.

D). E, coli Ribosomes

Escherichia coli K-12, strain 1200F- end A 1100 rns A Su- or

strain Hlfr DIO RNase were grown in nutrient broth at 370C. The

cells were harvested by centrifugation at 6,000 rpm for 10 min in a

Beck~man JA-10O rotor, suspended in a small volume of cold Buffer N, and

ruptured by sonication. After centrifugation in a Beckman Type 65

rotor at 15,000 rpm for 10 min, the supernatant was centrifuged again

at 60,000 rpm for 2 br in the same rotor to pellet the ribosomnes.

These were then washed once by resuspending and recentrifuging in

either Buffer N (10w-salt ribo~somes) or Buffer 0 (salt-washed ribo-

somes).

III. EXTRAlCTION OF RIBOSOMAL. PROTEINS

Proteins were prepared for electrophoresis by a modification of

the method described by Letister and Dawid (1974). For each gel sample,

a pellet of ribosomes containing 200-300 pg of protein was suspended

in 40 pl1 of 0.1 MI KC1, 10 nmI MgC12. 40 pl of 10 M urea, 4 M L~iC1,

HC1, pH 3.5 was added and the mixture was stirred at 50C for 12 br.

The RNA-containing precipitate was removed by cerntrifugatlion in a

Beckman Type 65 rotor at 50,000 rpm for 1 hr, and re-extracted by stir-

ring with 80 Fil of 6 M urea, 3 M L~iC1, HC1, pH 3.5 for 2 hr, The

supernatants from the two extractions were combined, and disulfide

bonds were reduced by the addition of 40 pl of 7.5 M urea, 0.25 MI










dithiothreitol, 0.5 Mi EDTA, I M Tris, pH 8.8 and incubation at 370C

for 1 hr. The proteins were thecn dialyzed against Samp~le Buffer

(8 M urea, 60 r'M potassium acetate, 0.01% aminoethanethio1, pH 6.7).

IV. RADIOACTIVE LABELLING OF RIBOSOM1AL PROTEINS

A. Radioiodination

Chemical labelling of ribosomal proteins with Nal2I and Chlor-

amine T was performed according to Leister and Dawid (1974). A 15 pl

aliquot (20 to 30 lpg protein) of the ribosomal protein extract described

above was set aside before the disulfide reduction step. The pH was

adjusted with an equal volume of 6 N urea, 3 M LiC1, 100 mMr Tris,

pH 7.5. 100 pCi of Nal25 (100 Ci/mrmol) was added, followed immediately

by 30 pg of Chloramine T. The reaction (total volume = 55 pl) proceeded

at room temperature for 10 minutes, and was stopped with 1 pl of 1 M

B-mercaptoethanol. Disulfide bonds were then reduced as described above,

and the radioactive protein was separated from unreacted 12I by chroma-

tography on Sephadex G-25 equilibrated in Sample Buffer.

Control experiments were performed to determine whether the label-

lingi reaction affected the electrophoretic properties of the proteins.

Radioiodinated ribosomal proteins were added back to a large non-radio-

active sample of the same proteins, and the mixture was subjected to

electrophoresis, staining and autoradiography as described below. No

alteration in the electrophoretic mobilities of the proteins could be

detected. However some of the stained proteins were not detectably

labelled, and vice versa. Radioactive spots which did not correspond





to proteins regularly seen by staining were ignored. The electro-

phoretic positions of the proteins that could not be radiolabelled

were determined by interpolation between the nearest spots that were

labelled.

B. Reductive Methylation

A pellet of ribosomes containing 5-10 pg of protein was suspended

in 2 pl of 4 M guanidine hydrochloride, 100 mM sodium borate, 10 mM~


MIgC12, 20 mMl KC1, 6 ml&1 B-MIE, pHi 8.5. Thirty nmol of [14C] formalde-

hyde (44 Ci/mol) in 1 pl of water were added, and the mixture was

incubated on ice for 30 sec. One 1Jl of 30 nmol/pl sodium borohydride

was added, and this addition was repeated after one minute. The

reaction mixture was left standing in the cold for 5 hr. A resuspended

pellet of ribosomes containing 200-300 pg of protein in 40 pl of 0.1 Mi

KC1, 10 rmM MlgCl2 was then added, and the proteins were extracted for

electrophoresis as described above.

V. TW~O-DIMIENSIONAL ELECTROPHIORESIS

The procedure was modified from that of Leister and Dawid (1974).

Thec first dimension was essentially the discontinuous-buffer electro-

phoretic system of Reisfeld et al. (1962), providing electrophoretic

stacking of the protein sample, The composition of the first-dimension

separation gel (1.5 mm in diameter, 11 cm long) was 7.5% acrylamide,

0.023% bisacrylamide, 0.063% TEMIED, 8 M1 urea, 60 mM potassium acetate,

pH 4.3, It was polymerized by adding ammonium persulfate to a concen-

tration of 0.1%, and pre-electrophoresed before use for 12 br at 0.2 ma





per gel, in a tank buffer of 0.01% aminoethanethial, 8 MI urea, 60 rki

potassium acetate, pH L.3. The stacking gel (2 cm long) was 3.86%

acrylamide, 0.14% bisacrylamide, 0.05% TEMED, 8 M urea, 60 mM

potassium acetate, pH 6.7, polymerized by the addition of ammonium

persulfate to 0.02% and riboflavin to 0.001%. The tank buffer was

0.01% aminoethanethio1, 35 mHl B-alanine acetate, pH 5.0.

The dialyzed protein samples (about 200 pl) were loaded onto the

gels and electrophoresed at 0.1 ma per gel. When the tracking dye

(0.001% Pyronin Y in 200 pl of Sample Buffer) in a parallel gel tube

had reached the top of the separation gel, the current was increased

to 0.2 ma per gel. Electrophoresis was continued until the tracking

dye reached oble bottom of the gel. The gels were extruded from the

glass tubes with a 6-inch, 22 gauge needle through which water was

flowing under pressure.

The second-dimension gel slabs were 18 cm high, 20 em wide, and

1.5 mm thick, Each gol was poured between a pair of glass plates

separated by two 1.5 nmm-thick plexiglass strips, sealed at the bottom

and sides with a piece of silicane-rubber tubing, and held together

with spring clamps. After the gel was polymerized, the silicone tubing

was removed. The gel composition was 9.65% acrylamide, 0.35% bisacryl-

amide, 0.1% TEM1ED, 5 M urea, 0.5% SDS, 0.1 M1 sodium phosphate, pH 7.2,

polymerizedl with ammonium persulfate (0.025%).

The extruded first-dimnension gels were laid on top of the second-

dimension gel slabs and overlaid with 0.05% mercaptoacetic acid, 5 M





urea, 1% SDS. At each end of the first-dimension gel was placed a

small piece of agarose (1.5 mm in diameter, 5 mm long) containing

several malrker proteins. The composition of the marker mixture was

0.5 mg/ml bovine serum albumin, I mg/ml human y-gilobulin, 0.4 mg/ml

ovalbumin, 0.33 mg/ml equine myoglobin, 0.4 mg/ml egg-white lyso-

zyme, 0.33% SDS, 0.33% B-MIE, 0.016% mercaptoacetic acid, 2% agarose.

Electrophoresis was at 30 ma/gel with a tank buffer of 0.024% mer-

captoacetic acid, 0.5% SDS, 0.1 Ml sodium phosphate, pH 7.2, until

the dye marker (Bromopheno1 Blue) reached the bottom of the gel. The

total time for electrophoresis in both dimensions was about 24 br.

After electrophoresis the gels were soaked in the following

solutions with continuous agitation: 25% isopropanol, 10% acetic

acid (18 br); 0.25% Coomassie Brilliant Blue R, 50% ethanol, 7.5%

acetic acid (6 hr); and 5% ethanol., 10% acetic acid (several changes,

until the gels were adequately destinedd. For convenient storage

and ior autoradiography, the gels were soaked in 3% glycerol for 30

min and then dried onto Wh~atman #3MM filter under vacuum on a steam

bath, by the method of Maizel (1971).

For autoradiagraphiy the dri -d gel was placed in contact with

Kodak RP~14 medical X-ray film in an X-ray film cassette. Two small

holes had previously been drilled through the cassette. After the ge1

and film were loaded and the cassette was closed, a needle was inserted

through these holes to mark the ali!:nment of the gel with the film,





Gels containing reductively methylated proteins were prepared for

autoradiography as described by Bonnrer and Laskey (1974). They were

shaken in two changes of dimelthyl sulfoxide, for 30 min each time,

th-en in 4 volumes of a 20% solution of 2,5-diphenyloxazole in dimethyl-

sulfoxide for 3 br, and in water for I br. Th-e gels were then dried

onto filter paper as described above. The film used for autoradio-

graphy was Kodak RP/R-54 medical X-ray film, pre-exposed with a flash

from a photographic strobe light according to Laskey and Mlills (1975).

Exposure at -700C for 5 weeks was sufficient for gels containing 10,000

cpm of reductively methylated protein.

VI. ISOELECTRIC FOCUSING

A previously published procedure (Czempiel et al., 1976; Klose,

1975) was followed closely. The gel composition was 4.8% acrylamide,

0.2% bisacrylamide, 0.065% TEMED, 1.0% Ampholine (pH 3.5 10), 5%

sucrose, 8 Mf urea, polymerized by the addition of ammonium persulfate

to 0.019%. The gels were 0.4 cm in diameter and 7 cm long. The upper

tank buffer was 5% phosphoric acid and the lower tank buffer was 5%

ethylenediamine, with- the anode in the upper tank. The protein sample

for each gel was dissolved in 20X of 8 MI urea and then mixed with 30X

of Sephadex G-200 (superfine) swollen in 20% sucrose, S M urea, 10%

B-ME, 1% Ampholine (pH 3.5 10).

The details of the procedure listed above differed from those of

the published procedure only in three minor respects. In the latter

procedure the ammonium persulfate concentration was sllightly greater





(0.023%), the gel was longer (8 cm), and the upper tank buffer contained

urea at a concentration of 3 M. Besides these insignificant differences,

there was one further modification which was significant. Czempiel et

al. (1976) used an Ortec 4100 Pulsed Constant Power Suipply, which has

power output characteristics different from those of the constant-voltage

power supply uised in the present experiments. The published procedure

involved a program of increasing voltage (50V for 1 br, 100V for 1 br,

150V for 1 hlr, 200V for 2 hr, 300V for 2 min, and 400V for 2 min). In

the present experiments several parallel samples were electrofocused

according to this schedule but for different lengths of time. The first

sample was run for a total of 3 br (the first 3 br of the schedule),

the second sample for the whole 5 hr of the schedule, and the third

sample for the whole 5 br plus an additional 2 hr at 200V.

After electrofocusing, the gels were extruded and subjected to a

second dimension of electrophoresis in SDS as described in Section V.

VII. ASSAYS OF RIBOSOMIE FUNCTIONS


Peptidyl transferase activity was assayed by the modified fragment

reaction (Denslow andl O'Brien, 1974; de Vries et al., 1971). The

reaction mixture contained 0.1 mg ribosomes, 83 nMY (10,000 cpm) N-acetyl-


[3H] leucyl-tRNA, 0.66 mMi puromycin, 267 mai Ke~l, 13.3 mM~ MgAc2' 33%

ethanol, 33 ml1 Tris-HC1, pH 7.5 in a total volume of 0.15 ml. After

incubation at 250C for 10 min, KOH was added to 0.6 HL and the mixture

warmed to 40"C for 3 min. The N-acetyl-[3H]1eucyl-puromycin synthesized

in che reaction was extracted into 1.5 ml of ethyl acetate, and





radioactivity was determined by liquid scintillation counting in Triton

X-100/toluene (l:1) containing 0.5% PPO and 0.05% POPOP.

[3HJ GTP binding was measured by the Millipore filter assay of

Bodley et al. (1970). Fifty pll of reaction mixture containing 15 ug of

ribosomes, 42 pmol (0.5 pCi) of [8- H] CTP, 10 mM NH C1, 20 mM MgAc,,

5 mM B-ME, 10 mM Tris-HC1, pH 7.4 was incubated at 00C for 5 min and

then filtered thru a Millipore filter. The filter was washed with 10 net

NH C1, 10 mM~ MgAc?, 10 1,M Tris-HC1, pH 7.4, andl the ribosome-bound

radioactive y determined by liquid scintillation counting.

Poly ii-dependent polyphenylalantne synthesis was assayed according

to Hosoktawa et al. (1966). The composition of the 0.25 ml reaction

mixture was 0.64 mg/ml poly U, 5.4 pM~ [3ll]phenylalanine (1.82 Ci/mmol),

25 mM tyrosine, 50 mM of each of the other 18 amino acids, 0.5 mg/ml

CRNA, 32 1PM GTP, I mEL ATP, 5 mM phosphoenolpyruvate, 0.1 mg/ml pyruvat-e

kinase, 1 mg/ml E. coli factors, 0.4 mg/ml ribosomes, 50 mM KC1, 20 mFt


MgAc7, 6 mM~ B-ME, 1 mMl dithiothreital, 10 n~I Tris-HC1, pH 7.8. Aliquots

of 50 pl were withdrawn at 5 min intervals, and radioactivity insolub~e

in hot trichloroacetic acid was determined by the method of Mans and

Novelli (1960).

VIII. BUOYANT DENSITY DETERM~INATIONIS

Ribosomes were dialyzed in 50 mML KC1, 5 mM MgC12, 20 mMl TEA,

pH 7.5, and then fixed by the addition of formaldehyde to a concentra-

tion of 5%. They were analyzed by equilibrium centrifugation in gradients

of CsC1 containing 50 mMi KC1, 5 mMI MgC1, 20 mM TEA, pH 7.5, 0.3%





formaldehyde in a Beckmnan SWJ39 rotor, as described by Brunk and Leick

(1969). The gradients were pumped through the flow cell of an Isco

UA2-2 absorbance monitor, and 0.25 ml fractions were collected. The

refractive index of each fraction was measured with a Bausch and Lomb

refractometer, and used to calculate the concentration of CsC1 and the

buoyant density in the fractions.

IX. QUANTTTATIVE MLEASUREMIENTS

Quantities of mitochondria were estimated by absorbance at 550 am.

Slnmples of mitochondria of known 5501 nm absorbance were analyzed by

the method of Lowry et al. (1951), to establish a calibration curve

relating the absorbance at this wavelength to the concentration of

mitochondrial protein in mg/m1.

Ribosome quantities were determined by absorbance at 260 nm, using

extinction coefficients (El% ) of 110 for mitochondrial ribosomes,
1 em

135 for cytoplasmic ribosomes, aind 160 for E. coli ribosomes.





RESUI.1S


The purposes of this research was to characterize the proteins of

mammalian mitochondrial ribosomes, especially in comparison with the

proteins of other kinds of ribosomes. The major characteristics to be

investigated were the number of ribosomal proteins, their molecular

weights and their electrophoretic properties. For this purpose Lt was

necessary to establish and apply a number of criteria to distinguish

the ribosomal proteins from the contaminating non-ribosomal proteins

which might be present in the mitochondrial ribosome preparation.

The results will be presented in five sections. First, the means

of preparing: mitochondrial ribosomes in adequate quantity and purity

for these experiments will be described. The two-dimensional electro-

phoretic patterns of bovine mitochondrial ribosomal proteins will be

presented and analyzed in terms of criteria for thle identification of

the proteins which are most likely to be true components of the ribo-

some in viva; the effect of the ribosome preparation procedure (partic-

ularly the effect of highi salt treatment) on the electrophoretic pattern

of ribosomal proteins will be evaluated, with a view to the possibility

of various kinds of artifacts. The electrophoretic properties of the

proteins will be compared with those of nrmammlian cytoplasmic andi

bacterial ribosomal proteins. The number and sizes of the mitochondrial














































































~_


ribosomal proteins will be used to estimate the total mass of protein

in the ribosome. Finally, the proteins of bovine mitochondrial ribo-

somes will be compared with those of other mammalian species to assess

the level of evolutionary divergence in these proteins.

I. PREPARATION AND CHARACTERIZATION OF MITOCHONDRIAL RIBOSOMES

At the time this research was begun, the method used for the prep-

aration of mitochondrial ribosomes (Method A in MATERIALS AND METHODS)

was not can able of producing adequate quantities of material for the

electrophoretic studies to be described below. Three separate prepa-

rations were required to produce enough ribosomes for a single electra-


phoretic analysis, each preparation being a week-long plalcedure.

Moreover, the crude mitochondrial ribosome preparation was heavily

contaminated with cytoplasmic ribosomles, as evidenced by the large

amount of material sedimenting at 80S when the ribosomes were subjected

to sucrose density gradient centrifugation (Figure 2A). The presence

of such quantities of 80S ribosomes suggested that even the 55S mito-

chondrial ribosome region of the gradient contained significant amounts


of cytoplasmic ribosomal subunits, as well as possibly other unknown

structures with similar sedimentation coefficients, For this reason

the intact 55S ribosomes were not suitable as samples for electro-

phoretic analysis.

Rather, a further purification step was performed by pooling the

material in thle 55S peak, treating it with a higher concentration of

KCL and a lower concentration of MgC12 to dissociate the mitochondrial





Figure 2. Preparation of mitochondrial ribosomes by Method A.
(A) Crude mitochondrial ribosomes were prepared from bovine
liver according to Method A in MIATERIALS AND METHODS. Thle ribosomes
were suspended in 2 ml. of Buffer A and layered onto a linear 10-30%
sucrose density gradient made up in Buffer A. After centrifugat ion
in a Beckman SW27 rotor at 27,000 rpm for 5 hours, the gradlient was
pumped through thle flow cell of~ a spectrophotometer and its absor-
bance at 260 nm was recorded. One-ml fractions were collected. The
direction of saJlmrentation is from left to right.
(B) Fractions corresponding to the 55S absorbance peak of the
sucrose density graidient shown in Figure 2A were pooled and centri-
fuged in a Beckman Typs 65 rotor at 65,000 rpm for 3 hours. The
pellet of mitochondrial ribosomes was suspended in 2 ml of Buffer E
and layered onto a sucrose density gradient made up in the same buf-
fer. Centrifugation was at 20,000 rpm for 13.5 hours.





55S


1.5 A
E
C
O
cD
1.0
O



8 0.5






E
C
O
<0
S0.2




C0 0.1
O
u)


28S 39S


Fraction Nlumber









ribosomes to their subunits, and re-ci~icrifuging on a second sucrose

density gradient. The only contaminants in the preparation of submnits

obtained in this manner, besides those that might be bound directly

to the subunits, would be those which sedimented near 55S under the

ionic conditions of the first centrifugation and in the 25S to 45S

region under the second conditions. The kinds of possible contaminants

which might behave in this manner seem intuitively to be few, and in

particular do not include cytopiasmic ribosomal subunits. In fact,

the discrete peaks of mitochondrial small and large ri!,,somal subunits

were found to be by far the major species visible on the absorbance

profile of the second centrifugation (Figure 2B), suggesting that the

preparation was indeed reasonably free of co-sedimenting material.

Later improvements in the procedure increased both the yield and

the purity of the mitochondrial ribosomes. The yield was approximately

tripled by the use of a more efficient homogenizer for the disruption

of the cell:;. Treatment of Lhe mitochondrial preparation with digitonin

(Method B) dramatically reduced the quantity of cytoplasmic ribosomes

in the crude mitochondrial ribosome preparation (Figure 3). This

detergent has; been used previously to remove cytoplasmic ribosomal RNA

(Malkin, 1971) and cytoplasmic ribosomes (de Vries and van der Koogh-

Schauiring, 1973) from rat liver mitochondrial preparations. In addi-

tion, treatment under these conditions has been reported to solubilize

latent lysosomal enzymes (Schnaitman and Greenawalt, 1968; Lowenstein

et al., 1970), thereby diminishing the possibiL ity of degradation of

the mitochondrial ribosomes during the preparation.





2.0



E 55S 80S



O

CV

0.










5 10 15 20 25

Fraction Nurnber



Figure 3. Preparation of mitochondrial ribosomes by Mlethod B.
Crude bovine mitochondrial ribosomes obtained by Method B
(MIATERIALS AND METHODS) were suspended in B~uffer H and analyzed
by sucrose density gradient centrifugation in this buffer as des-
cribed in Figure 2A, except that centrifugation was for 4 hours
at 27,000 rpm. (No differences have been observed between Buffer H
and Buffer A with respect to their effects on the mitochondrial
ribosomes.)





At this point the maximum yield of micochonidrial ribosomes from

a single preparation was limited by the volume of the largest prepar-

ative ultracentrifuge rotors available. After preparation and lysis

of the mitochondria, thre best meanns available for concentrating the

ribosomes for purification on sucrose density gradients was to

centrifuge them to a pellet at about 100,000xg. The volume of an

average yield of mitochondria (about 35g protein), suspended in buffer

to a concentration low enough for efficient detergent lysis, is far

greater than the capacity of three Beckman Type 35 rotors. Therefore,

a method was developed to preconcencrate the ribosomes before harvest-

ing them by centrifugation (Method C).

Thie mitochondrial lysate was stirred with DEAE-cellulose as de-

scribed in M1ATER1ALS AND NETHODS to adsorb out the mitochondrial ribo-

somes, and then the ribosomes were eluted by stirring with a smaller

volume of buffer at a higher ionic strength. This procedure is very

rapid andt effects a fourfold concentration of the ribosomes, thereby

quadrupling the yield of each- preparation. Furthermore, any remaining

traces of cytoplahlmic ribosomes, as we~ll as a considerable proportion

of the other contaminants of the mitochondrial ribosome preparation,

are removed by this means.

An assessment of the degree of purification achiieved by this method

is shown in Figurea 4. A sample of purified, digitonin-treated mito-

chondria was split into two equal aliquots. Ribosomes were prepared

from the first aliquot without DEAE-cellulose treatment and analyzed by





A5 ?" 6s B Js BSp
20- 20-




10 IC






o IO 2 30 0 10 20 30
Frachion Number Frochon Number
55S 8SOD s s







C j D 0" 9
10t 20 30 1020






gation inBufe H
(A iooe rm14g(rti)o oielvrmtcodi





(C) .4 g o Nmitchndi an 5mlofmcrosoe in mBuff


wa iltre 4red anteflrateo dicrdd Mitochondrial ribosomes b ehdC

were pepared fore nlzdb sucrose density gradient anlss codngtoiu

peae Method B. The 55S asrac peak contains 6.1 A26 unitsofibsms

(D) CTheonce-elute DEAE-mellls from (C) weas stirredf woith
Buffer J oeut h ctpasi iosomes. The 80S peak contains 56A6 ~t frbsms

4.6A26 unitse ord 82% of lthe dinputed ytpamitchna ribosomes.





sucrose density gradient centrifuga:.tion, giving the absorbance profile

shown in Figure 4A. A preparation of bovine cytoplasmic ribosomes

(having the sucrose density gradient profile shown in F~igure 4B) was

added to the second aliquot of mitrochrondria, to provider a more strin-

gent test of the method. The mitochondria were then lysed and the

rib~s~omes prepared by the DEAE-cellulose procedure described above. As

seen in Fiaure AC, the product consists primarily of mitochondrial

ribosomes and subunits, much of the slowly-sedimenting material and all

detectable 80S cytoplasmic ribosomes having been removed. Probably

because of the larger proportion of rRNAZ in cytopLasmic ribosomes and

their consequent higher density of negative charges, these particles

remain bound to the DEA2E-cellulose under the condiitions used for elution

of the mitochondrial ribosomes, as shown by re-elution of the DEAE-

cellulose with a buffer of higher ionic strength (Figure 41D).

As may be seen in Figure 4, the yield of mitochondrial ribosomes

per gram of mnitochiondria obtained with the DEAE-celluilose procedure is

as great as or slightly greater than that obtained without this step.

Since this method permits the processing of a four-fold greater quantity

of mitochondria, and gives a cleaner preparation of rihosomes as well,

it is clearly the method of choice unless it adversely affects the

structural or functional integrity of the ribosomes. Of particular con-

cern is the possibility that some of the more acidic ribosomal proteins

might bind more strongly to the DEAE-cellulose than to th~e ribosome

itself and thus might be stripped off when the ribosomes are eluted.





The most acidic proteins of E. coli ribosomes, L7 and L12, have been

found to be relatively loosely bound to the ribosomes (Hamel et al.,

1972),

To answer this question, the ribosomes were characterized by a

number of functional and physical tests. The peptidy1 transferase and

CTP-binding activities of the ribosome were essentially unaltered by

treatment with DEAE-cellulose (Table V). DEAE-cellulose-prepared ribo-

somes were also found to be active in the translation of poly U (400

pmoles phe incorporated/mg rRNA/15 min). The sediment stion coefficients

and buoyant densities of micochandrial ribosomes and the submnits de-

rived from them were essentially the same as those obtained from previous

preparations. The effect of this procedure on the two-dimensional

electrophoretic pattern of the ribosomal proteins will be evaluated in

Section IIC below.

II. ELECTROPHORETIC ANALYSIS OF MIITOCHONDRIAL, RIBOSOMAL PROTEINS

The very high proportion of protein which physical-chemical mleasure-

ments have shown to be present in bovine mitochondrial ribosomes (see

INTRODUCTION) predicts that a rather large number of individual proteins

are to be found in these ribosomes. If the 1.8 x 106 d. of total protein

per particle were made up of proteins similar in size to E. coli ribo-

somal proteins (average molecular weight about 17,000), for example,

each ribosome would contain more than a hundred protein molecules.

For this reason it seemed unlikely that electrophoresis in one

dimension would separate all the proteins sufficiently for thre purposes





Tabl. V. Fuinctional activity of mitochondrial ribosomes
prepaiced with or without DEAE-cellulose.


3H-Leu-Puromye in
Synthesis
(cpm/nnole ribs/10')

33,600
36,200


3H-GTP Binding
(males GTP/mole ribs~)

1.12
1.09


Preparation

No DEAE
DEAE





of this study. Several two-dimenisional ele~ctrophoretic systems have

been devised to provide adequate resolution for the analysis of ribo-

sonal proteins. The system used for most of the experiments to be

described below was that of L~eister and Dawid (1974), employing a


separation partially on the basis of charge at pHi 4.3 in the first

dimension and on the basis of size in the second (SDS) dimension. For


the purposes of this research, this electrophoretic system has several

advantages over thre one that i:, most widely used for ribosomal protein

studies at present (Kaltschmidt and Wittmann, 1970). It permits a

direct determination of the molecular weights of the proteins. It is

more sensitive, requiring only about 5 ug of each protein. And it has

been used for the analysis of Xenopus mitochondrial ribosomal proteins

(Leister and Dawid, 1974), making it possible to compare these published

results with those obtained in the present experiments on mammalian

mitochondrial ribosomes.

The results of one of the early electrophoretic analyses (shown

in Figures 5 and b) amply confirmed the expectations that a large number

of proteins would be found in mitochondrial ribosomes. 40 to 50 reason-


ably intense and distinct spots may be discerned in the photograph of

the electrophoretic pattern of proteins from each subunit. Somewrhat

larger numbers could be seen on the original gels, since some of the

resolution of closely adljoining spots has been last in the photographs.

It is evident in these figures that the description "reasonably

intense" involves a rather arbitrary judgment. In fact, particularly





YY
r)






I


Figure 5. Electrophoretic pattern of large-subunit proteins from bovine
mitochondrial ribosomes.
Mitochondrial ribosomes were dissociated to subunits by sucrose
density gradient centrifugation in Buffer E. Gradient fractions corre-
sponding to the large (39S) subunits were pooled and centrifuged in a
Beckman Type 65 rotor at 65,000 rpm for 5.5 hours. The proteins were
extracted from the ribosome pellet and analyzed by two-dimensional
polyacrylamide gel electrophoresis. The first dimension (left to right)
was run in urea at pH 4.3, and the second dimension (top to bottom) in
SDS.
Just before the second dimension was begun, pieces of agarose gel
containing a mixture of several proteins (bovine serum albumin, human
y-globulin heavy and light chains, ovalbumin, myoglobin, and lysozyme)
were placed on top of the gel slab, immediately adjacent to each end
of the first-dimension gel. These proteins may be seen at the left side
of the photograph.










-i


SIll~


Figure 6. Electrophoretic pattern of small-subunit proteins from bovine
mitochondrial ribosomes.
Small (28S) subunits were prepared by sucrose density gradient
centrifugation in Buffer E, and their proteins were subjected to two-
dimensional electrophoresis as described in Figure 5.


llk ~F~.

~"










in the small-subunit pattern, there is a large range of spot inten-

sities with a more or less continuous variation from the most intense

tc thie very faintest. This result was not what hiad been anticipated.

Since nearly all ribosomal proteins probably are present in one copy


per ribosome, at least in vivo (Hardy, 1975), it had been expected

that the gels would show a discrete group of ribosomal protein spots

of roughly similar staining intensities. In addition there might be

some other spots of varying intensity, corresponding to contaminating

proteins. The results obtained indicated that there were considerable

differences in the relative quantities of the ribosomal proteins pres-

ent in the sample, making it imprlossible to distinguish them from con-

taminating proteins on the basis of their staining intensity. Some

possible explanations for this phenomenon will be considered below

(Section IV).

A. Criteria for the Identification of Ribosomal Proteins

For this reason it seemed necessary to seek other criteria for the

purpose of identifying which of the spots appearing on the gels actually

represented ribosomal proteins. Several possible criteria were con-

sidered, including reproducibility of occurrence, subunit-specific local-

ization, and resistance to removal from the ribosomes by high salt

treatment. It seemed reasonable toi expect that the ribosomal proteins

would be fcund reproducibly in separate preparations of the ribosomes,

whereas the non-ribosomal contaminants might be more variable in occur-

rence. In fact, it was found thiat many of the proteins see~n in individual










gels such as those of Figures 5 and 6, particularly some of the

fainter ones, could be disregarded on the basis that they were not

seen in the majority of the electrophoretic analyses. This rule was

thus found to be a useful one, and the application of it will be

described below.

The idea that some contamination might arise from the non-specific

binding of some proteins to the ribosomes suggested that contaminants

of this type might be found nearly equally in both ribosomal submnits.

However, thle existence of some plausible mechanisms by which ribosomal

proteins might also show an apparent lack of subunit-specific local-

ization (Section IIB3) made the application of this criterion difficult,

and it was used only in conjunction with other kinds of evidence.

Yet another way to discriminate the riboso~mal and non-ribosomal

proteins, and the most unambiguous way, would be to purify all the pro-

teins and then reconstitute the ribosome, withholding each protein in

turn, to establish which proteins were essential for the generation of

a structurally and functionally normal ribosome. Unfortunately, this


approach is not practical in the case of ribosomes such as those of

mammalian mitochondria, which can be obtained only in relatively small

quantities. However the observations (01snes, 1971; K~urland, 1966;

Hardy and Kurland, 1966) that contaminating proteins are in general

more loosely bound to rib:osomes than are the ribosomal proteins them-

selves, suggested that an analogous approach might be feasible.

Specifically, it hlas been found that treatment of ribosomes with









moderately high concentrations (about 0.5 11` of monovalent salts dis-

lodges adventitiously bound proteins from the particles, while affecting

the binding of the ribosomal proteins to a lesser extent. Of course,

higher salt concentrations do remove ribosomal proteins (Delaunay et acl.,

1974; Gesteland and Stachelin, 1967) and the optimum concentration range,

that which removes the largest quantity of contaminants and the smallest

amount of ribosomal protein, may differ from one type of ribosome to

another.

The effect of increasing salt concentration on the content of

individual proteins in the bovine mitochiondrial ribosome was assessed by

centrifuging: the ribosomes throu,:h various high-salt buffers and then

analyzing the proteins remaining in the treated particles by two-dimen-

sional electrophoresis. To establish the point at which these treatments

began to remove ribosomal proteins, the peptidyl transferase activity of

the ribosomes prepared under each condition was assayed. The electro-

phoretic protein patterns and specific activities obtained were then

correlated, and any proceins whiich were found to be removed from the

ribosome by a given treatment, without diminishing its functional integ-

rity, were tentatively regarded as non-ribosomal contaminants. Since

the peptridyl transferase activity is a property of the large subunit

and is not dependent on the presence of the small subunit, this criter-

ion was directly applicable only to the large-subunit proteins.

Figure 7 shows the effect of treatment with buffers containing

increasing ratios of KCI concentration to MgC12 concentration on the





55









Ou










5 10 50 100 I,000 10,000


K Cl/MgClp



Figure 7. Effects of treatment with buffers of various ionic compo-
sitions on the peptidyl transferase activity and buoyant density of
bovine mitochondrial large subunits.
Large subunits were prepared by sucrose density gradient centri-
fugation in various buffers. Peptidyl transferase activity was
assayed as described in Materials and Methods, and expressed as a
percentage of the activity found in Buffer At treated subunits. The
buoyant densities of the formaldehyde-fixed subunits in CsC1 were
determined by isopyenic centrifugation.
The buffer conditions used are plotted along the horizontal axis
according to their ratio of KCl to MgC12 concentrations. The actual
concentrations of these two salts were:

Buffer KC1(M) IgCl,(mM)
A 0.1 20
C 0.5 10
D 0.3 5
E 0.5 5
F 1 5
G 1 1
The complete compositions of these buffers are given in Table IV.










activity and buoyant density of the large subunit. In this series,

buffets containing increasing concentrations of KCl and decreasing

concentrations of Mg~C12 were used to wash the ribosomes with increas-

ing stringency. For other ribosomes monovalent and divalent cations

have been found to produce opposing and competitive effects on the

structure of the particle, with respect to the association of the two

subunits with each other (Hamilton and Petermann, 1959; Zitomer and

Flaks, 19 2) and the binding of proteins to the ribosome (Spitnik-

Elson and Atsmnon, 1969; Stachelin et al., 1969). Such effects on

subunit association have also been found in mammalian mitochondrial

ribosomes (0'Brien, 1971), andi Fiigure 7 indicates that either increasing

KC1 or decreasing MgC17 can diminish both the functional activity and

the protein content (as measured by the increase in buoyant density)

of these ribosomes.

The upper curve of Figure 7 shows that the large subunit retained

most of its peptidyl transierase activity after treatment with either

Buffer C or Buffer D. The activity was significantly reduced by treat-

ment with Buffer E, and higher KC1/MlgCl2 I.tios yielded particles with

negligible activity. Therefore in the analysis of the electrophoretic

results presented below, proteins which were found to be absent from

large subunits treated with either Buffer C or Buffer D) are considered

to be dispenisable for this particular large-subunit function. These

proteins are regarded as possible non-ribosomal contaminants, and are

discussed separately.










B. Identification of the Ribosomal Proteins


1. ~L4arge Subui

In order to apply the criteria of reproducibility and resistance to

high-salt treatment, several electrophoretic analyses were performed on

large-subunit proteins taken from separate preparations of ribosomes,

after treatment with various high-salt buffers. The gels were scored for

the presence or absence of individual proteins, and the results are given

in Table VI. The electrophoretic positions of the proteins designated

in the table are shown in Figure 8. The 52 proteins which were seen

reproducibly and which were present both in large subunits treated with

Buffer C and in those treated with Buffer D are shown as numbered spots.

Lettered spots represent some of the proteins which did not satisfy

these criteria.


For the purposes of Table VI thle relatively subjective question of

the intensities of the protein spots was not considered, and a "+"

indicates only that the protein was clearly visible in the gel, whether

it was very intense or quite faint. Proteins which were near the limit

of visible detection, however, are marked with a "-"; some of these are

not visible in the photographs of thre gels. In a few cases ("nr") the

electrophoretic resolution in a given experiment was inadequate to

determine whether a protein was present or not. The three proteins

which move fastest in the first dimension of electrophoresis (ML45,

ML51, and MILS2) could have been run off the end of the first-dimension

gel in some experiments, and these cases are marked "ro".















++++++++++++++++f++++++++c+







++ 1+++++++++++++++++f+++++f


00 0
Li F 0


O *ma
*HE
m001
c0O
-Ho 0

UN



-r0 rd *







oua .c -







u ty



"o +




e l


7jOm
AUr


IIII++oo+++o++l II+++++++++++


f+l+ +++++++++++++++~++ +++++







f+l+++++++++++++++++++++++++







++ +++f++f++++++++++++++++++







HN M~ttlemmoloNmctric mosoHNOm tri
HHHHMHHMHHNNNNNNNN


*




a














I++f+++1+l+++f+++f+++++++++






+++++++1+f+++++++ofolf++1+






co
100++1++++++++++o++o++0++


I iso+o






coofo+1+






++o++c++


++oco+o I


++++++++++f+++++++8+o+o +B


olf++++1+l+l++i+++l++1+18






+++++++1++S++l+++1++++1+






I++++f~++1++1++1++++++f++++






II+++++oc+ol+++++o++o++cm


Io ololo






colo F+O






101000+o






lolo+c+o


(D L1 U a a~ u, LO C













+-










7 as 9O10
11*, *12 +13
S14
23.5- 16 rr


22 21 ()20
S26 #2427 ()25
*D 29 28 324 f 34
*72- 33 36 43
37 (938 3) 4
41 he 4*43~4
14.4- 46, 645
481) 4 49



514 42



Figure 8. Schematic diagram of bovine mitochondrial large-subunit
proteins.
The origin of electrophoresis is at the upper left corner of the
figure. The second-dimension electrophoretic positions of bovine serum
albumin (molecular weight 68,000), ovalbumin (44,000), human y-globulin
light chain (23,500), horse myoglobin (17,200), and egg-white lysozyme
(14,400) are indicated at the left of the figure.
Proteins which were reproducibly present in functionally active
large subunits are shown as numbered spots. Lettered spots designate
some of the other proteins which were seen in some of the electrophoretic
analyses .










a. Reproducibility

In order to discriminate ribosomal proteins from contaminants, an

arbitrary level of reproducibility of 50% was taken as the cut-off.

That is, a protein was considered to be non-ribosomal unless it was


present in over half of the ribosome preparations analyzed. In fact,

however, very few of the proteins in large-subunit preparations were

near this borderline. Most of thei proteins which were found in reason-

able quantity, and many of the fainit ones, in any one gul proved to be


quite reproducible. A small number of proteins in thre unreproducible

group were seen so rarely that it appears most unlikely that they

represent ribosomal proteins: the spot seen between ML16 and ML17 in

Figure 5, for example, was not found in any other experiment. Such

proteins are not listed in Table VI. Thie proteins designated by letters

in the table were seen more frequently, especially in ribosomes pre-


pared in certain buffers. Because the presence or absence of these

proteins was correlated with the ionic conditions with which thle ribo-

somes had been treated, they are considered below under "Resistance to

sa~l-washing".


Among the numbered proteins in Table VI many appeared in every

ribosome preparation, and w~it-h rCBsonable .staining intensity. Some

faint spots, like ML29 and ML41, were likewise very regular in occur-

rence although sometimes present only in trace quantities. Occasion-


ally a protein which was usually present in moderate qluantity would be

unaccountably lacking in a particular experiment: ML't7 for example,











can be clearly seen in experiments 2 (Figure 9), 4 (Figure 10), and

6 (F~igure 11) and is clearly absent from experiment 8 (Figure 5).

Similarly, ?nl1 and ;"L2 are always found to have about the same stain-

ing intensity seen in Figures 5, 9, and 10, except that in experiment

6 (Figure 11) they were both reduced to barely detectable traces.

Although such results are quite surprising, the overall reproducibil-

ity of these proteins was considered adequate to satisfy the criterion.

Predictably, the greatest variability was found among the faintest


spots, including the only two real borderline cases. EfLA8 was seen

with low but clearly visible intensity in three experiments (as in


Figure 9), as only a trace on three other occasions, and three times

not at all. MLA5 was found even less often, again was sometimes only a

trace when it was present, but was sometimes quite intense (Figure 9).

Overall, relatively little variability was observed in the

pattern of large-subunit proteins from one experiment to another. Mlost

of the proteins seeni on any one gel includingg the faintly-staining

ones) were regularly found in separate preparations of ribosomes, and

thus probably represent ribosomal proteins, according to this criter-

ion. Some possible explanations for those variations that were obtained

in these experiments will be discussed below in the analysis of the

small-subunit proteins, for which variability is a more serious prob-

lem.









il~i Rbt~~
'-r

ur-~

;~d



t

a)


4~t~:.~~Cr3)'k-r sc,
i


Figure 9. Electrophoretic pattern of proteins from large subunits
prepared in Buffer C (Experiment 2).





t


Figure 10. Electrophoretic pattern of proteins from large subunits
prepared in Buffer C (Experiment 4).


`~ ~b~


t





Y
i)



1191,

~ ii

r


ii



t 111,


t
;r rl)
i


Figure 11. Electrophoretic pattern of proteins from large subunits
prepared in Buffer D (Experiment 6) .










b. Resistance to salt-washing

The salt-washing procedure used in the preparation of mitochondrial

ribosomes was designed to remove two types of contaminants, those which

were bound directly to the ribosomes and those whlich were not bound,

but nonetheless accompanied the ribosomes through the early stages of


the preparation. As mentioned in Section IIA, treatment of ribosomes

with high co~ncentrationss of monovalent salts should preferentially

release noni-ribosomal matterial bound to them. Figure 12 shows the

effect of washing E. cali ribosomes with 1 M NH C1. It is enident that

the major difference between the two protein patterns is the presence

of many high-molecular-we ight proteins in the unwashed ribosomes. This


phenomenon is well-documented in the literature (Subramanian, 1974;

Hardy, 1975; Brouwer and Planta, 1975), and the proteins removed by

salt-washing are universally regarded as non-ribosomal for a number of

reasons, including the essentially undiminished functional activity of

the salt-washed ribosomes.

The fact that most high-salt buffers also dissociate mitochondrial

riboso~mes to their subunits permits the removal of non-ribosome-bound,

co-sedimenting contaminiants by the usre of two cycles of sucrose-density-


gradient centrifugation under different ionic conditions. As discussed

above (Section I) the isolation of the ribosomes first at low ionic


strength (0.1 M KC1) as intact 55S particles, and then a higher salt

concentrations as subunits should separate them from most non-ribosome-

bound contaminants.





9 *


QhQO
oma
WH~1
lr OR
Ma C
m1 *Ho
UIO O r
;IHa
OuR 0r
" -& ,C

mo*

U U
MI 0
ao UN

CMw
OHO OP
UO W~
MO
0 oA
*Ho '0*


c 40ir


~FOd
Um O -H
0 40
00000
U -3 F *
rd UN 00-

,ICO H
rll a0
HOOD~C
m > ,CC
n~ eHM
,i, ,
00 ,
*r- > cI)


,C Qr z C 41



* CH Z ,C


N OHO
H~ 0 C
* M *


,











Low-salt treatment. Thre effectiveness of this procedure can be

evaluated by comparing subunits prepariied in this way ("derived submnits")

with the subunits which are present in small amounts in the first, low-

salt sucrose density gradient ("native submnits"; see Figure 3).

Figure 13 shows the eleictrophoretic patterns obtained from native large

and small subunits, and Figure 14 shows the pattern of the derived large

subunit from thre same preparation of rihosomes. The most obvious

difference Es the presence of two high-molecu~lar-weight proteins in

large quanntity in the native subunits. Several other proteins of sim-

ilar and higher molecular weights are also present in native but not in

derived subunits; many of these, like th~e two major proteins, are found

equally in both the large and small submnits. Another native-subunit-

specific protein, seen to the left of ML30, is also fcund in somewhat

smaller quantity in thre small submit. Additionally, numerous proteins

of various molecular weights and staining intensities appear in the

native large subunit but not in either of the other two patterns.

It seems reasonable to conclude that most if not all of the proteins

found in native but not derived subunits represent non-ribosomal contamn-

in~ants. The similarity of these results to those obtained with E. coli

ribosomes (Figure 12) is clear, particularly withi respect to the high-

molecular-weight proteins. The lack of subunit-specificity of many of

these proteins is also incriminating. Most importantly, they are

essentially abssent from large subunits derived in Buffer C or Buffer D,

and such subunits are functionally active. By visual comparison of the

















C 60
=OM
.aua








rd

Jl0



30m


CIcu
rdi

*MDU
3000

10

its
Ba r
10
OW *~
0000





ORr

wU)

nBi


*~ ~


b,


sit


F 0 a


3l B





rt


Figure 14. Electrophoretic pattern of proteins from large subuniLs
prepared in Buffer C (Experiment 3) .
The irregular white line in the upper left is a crack in the gel.









relative staining intensities of these proteins in native and derived

subunits most of them appear to be reduced in quantity by 90% or more

in derived subunits. Yet the activity of Buffer C derived subunits

is 75% of that found in native subunits (shown as Buffer A subunits


in Figure 7). Thus none of these proteins can be essential for the


peptidyl transferase activity of the ribosomie.

The kinds of contaminants which have thius been identified are

various, although some loose generalizations can be maide about them.

The contamlinants may be present in a very wide range of staining in-

tensities, have molecular weights ranging from values similar to those

of the ribosomnal proteins to considerably larger values, and may be


present subunit-specifically or not. However, a major part of the

contamination seems to be represented by a group of relatively intense,


high-molecular-weightgbt non-subunit-spec ific proteins. It is of somne

interest to ask whether this contamination is due primarily to adherent


proteins or to protein-containing structures that are not bound to the

ribosomes but are large enough to sediment in the same region of the

sucrose density gradient. The buoyant density of the native submit

is significantly lower than that of the Buffer C or Buffer D derived

subunits (Figure 7), suggesting that at least some of the native-

subunit-specific protein must be bound directly to the particle under

these low-salt conditions. Further, inl two experiments the subunits

were derived directly from the crude ribosome pellet, rather than from

the 55S fractions of a sucrose density gradient. Such sulbunits showed










only small quantities of these contaminants (Figure 10). Mlost of

these proteins thus seem to be bound directly to the ribosome~s under

low-salt but not high-salt conditions, rather than components of

separate structures which also sediment in the 25S to 45S region of

the gradient.

High-salt treatment. In addition to these proteins which were

absent from all derived large subunit preparations, several proteins

were found to be removed when the ribosomes were dissociated under

some ionic conditions but not others. These proteins are the ones

designated by letters in Table VI. Thle buffer conditions of interest

are Buffer C and Buffer D: since large subunits prepared in either

of these conditions are functionally active (Figure 7), the absence

of a protein from either kind of preparation indicates that that pro-

tein is not required for the function. Thus MfLe and M~g are not

essential for peptidy1 transferase activity because Buffer D subunits

lack them but possess the activity. bfLb, d, e, f, and b are specif-

ically absent from Buffer C subunits. M~a is found in Buffer C

preparations, but only in mucch smaller quantity than in Buffer D sub-

units; since this large quantitative difference does not correlate with

the similarity between the two kinds of submnits in functional activity,

this protein too may be regarded as nonessential.

The fact that these proteins are dispensable for the peptidy3

transferase activity of the ribosome does not necessarily imiply that

they are contaminating non-ribosomal proteins. It only indicates that










they are not components of the active site for this function, and are

not structurally required to maintain the proper conformation of the

active site. Peptidyl transferase is only one of many reactions and

interactions involved in protein synthesis, Including translocation,

initiation, termination, and binding to factors and mRNA. And not all

of the ribosomal proteins are required for each function. For example,

the E. coli ribosomal protein L,7/12, which is needed for the inter-

action of the elongation factors EF-T and EF-G with the ribosome, can

be removed witlhout: significantly affecting the peptidyl transferase

activity (Hamel et al., 1972). Therefore, in the absence of data

demonstrating that both Buffer C and Buffer D ribosomes are active in

a more demanding assay (such as che translation of poly U or a natural

mRNA), it remains quite possible that all of the lettered spots in

Figure 8 do, represent ribosomal proteins.

Extreme-salt treatment. As an attempt to identify some proteins

which were required for peptidyl transferase activity (and whiich there-

fore must clearly be ribosomal proteins), the electrophoretic patterns

of large subunits prepared with Buffers E, F, and G were evaluated.

Since Buffer F and Buffer C subunits are essentially inactive, it seemed

likely that they would be mrissingi one or more proteins, and that among

these proteins at least one was required for the activity. One or sev-

eral of these proteins might also be found in diminished quantities in

Buffer E subunits, which have diminished activity.










For this purpose, two preparations of Buffer F subunits andli two

of Buffer G subunits were subjected to two-dimensional electropho-

resis. One of the Buffer F analyses is shown in Figure 15,and one of

the Buffer G gels in Figure 16,. Some differences were observed be-

tween the Buffer F preparations and between the Buffer G preparations,

but there were several proteins which were absent or significantly

diminished in staining intensity in both of the analyses in each pair.

Buffer F didl not completely remove any of the numbered large-subunit

proteins, but it did reduce five of them to trace-level staining inten-

sity. ML1,l 15, 16, and 18 were diminished to a tenth or less, and ML2

to less than hallf, of the quantity normally present in Buffer C or

Buffer D subunits. In subunits treated with Buffer C, small quantities

of ML16 and MLl8 were still present, but ML1 and ML15 were not detectable.

In addicion ML23, 29, 30, 33, 43, and 48 were absent and the quantity

of MiL47 was considerably reduced.


Interestingly, the amount of ML2 found in Buffer G subunits did not

appear to be significantly less than that in submnits treated with

Buffer C or Buffer D, even though the less stringent treatment with


1 M KC1 at 5 mM MgC12 (Buffer F) did remove som- of this protein. A

similar phenomenon was observed among the large-subunit proteins which

are designated by letters. Mng, which is absent from sub is its treated

with the moderate-salt Buffer D and is present only in trace quantities

after treatment with Buffer F, is not removed by Buffer G. Such results

are at adds with the expectation that the set of proteins removed by a











,~awrrsrclwj'.~LJd*kr91P~~ ~~e ~n it~4PCr~rrshirrr*L
~CI~"L~C



r
ut



-t





-3~
EQ1`


(JI


Figure 15. Electrophoretic pattern of proteins from large subunits
prepared in Buffer F.















1

2




r
O




i
I

9

Ib
1

F
1"

II


of proteins from large subunits


Figure 16. Electrophoretic pattern
prepared in Buffer G.










given ionic medium should be a subset of the proteins removed by more

stringent conditions. This expectation derives from the continuous

decrease in the total protein content of the subunit which is ob-


served as the KC1/MgC12 ratio in the medium is increased (Figure 7).

Apparently the generalization that increasing KC1/MgfC12 ratios remove

increasing amounts of protein applies to the total protein content (as

measured by the buoyant density) and to some of the individual p~ro-

teins, but "fL2 and MLg are exceptions. That a few such exceptions to

this rule might exist, due to specific effects of particular buffer

compositions on the binding of some of the proteins to the ribosome,

is not too surprising.

The protein complement found in Buffer E treated large subunits,

however, represents a more serious anomaly. None of the numbered pro-

teins were found to be removed or significantly diminished in quantity

by washing with this buffer, offering no explanation for the reduction

in functional activity produced by this treatment. Furthermore, even

the lettered proteins (which are all removed by either Buffer C or

Buffer D) were each found at least once in preparations of Buffer E

subunits (Table VI). Thus the buoyant density data (Figure 7) indicate

that these particles muist be lackingi some proteins present in Buffer C

or Buffer D subunits, but no missing proteins can be identified by

electrophoretic analysis. The reasons for this discrepancy are not

clear, although some considerations described below (Section TV) may be

relevant to this question.










As mentioned above, several proteins are specifically missing or

reduced in quantity in large subunits that have been treated with ionic

conditions sufficiently stringent to inactivate them. In particular,

the loss of ML1, 15, 16 and 18 correlates with the loss of functional

activity. In order to prove that the inactivity of the stripped sub-

unit is due to the loss of these proteins (and not, for example, due

to a confornmtional alteration of the subunit induced, by these salt

conditions) it is necessary to demonstrate that the activity can be

reconstituted by adding the proteins back. Such a partial reconst i u-

tion would also be useful as an assay to determine which of the missing

proteins is (are) required for thie activity. However, attempts to

reconstitute the subunit were unsuccessful.

c. Summary of the large-subunit proteins

Electrophoretic analysis of the large subunit of the boviine mito-

chondrial ribosome revealed 52 protein components which were repro-

ducibly present in functionally active large subunits and which there-

fore probably represent ribosomnal proteins. Eight additional proteins

were present in large subunits prepared only under certain ionic

conditions; these may also be ribosomal pro~teins, but they are not

required for the peptidyl transferase activity of the large subunit.

Treatment with very-high-salt buffers removes a small number of the 52

proteins and destroys the peptidyl transferase activity, suggesting

that one or more of the missing proteins is required for this function.









2. Small Sulbunict

As shown in Figure 13 above, native small submnits prepared in low-

salt buffers contain a stet of mainly high-mallecular-weight proteins

which are also0 found in the native large subunit. Although lesser

quantities of these proteins are sometimes found in derived small sub-

units treated with moderate salt concentrations (Figures 6 and 17),

they are usually much diminished relative to the quantities present in

the native subunit. Figure 18, obtained from the derived small subunit

from the same ribosome preparation whose native subunits are shown in

Figure 13, is completely lacking in these proteins. For some of the

same reasons given in the discussion of the large subunit, it seems

unlikely that these proteins represent functional components of the

small subunit either.

Table VII shows the results of several electrophoretic analyses of

high-salt-washed small siiubits. The proteins which were found regu-

larly in separate ribosome preparations and thus satisfy the criterion

of reproducibility are designated by numbers. Some of the proteins

which were seen less frequently are also tabulated and are designated

by letters. (To avoid confusion with the numerals I and 0, the letters

1 and o have been skipped.) The electrophoretic positions of these pro-

teins are given in Figure 19.

The overall level of reproducibility of the small-subunit proteins

was Less thian that of the large-subunit proteins. To some extent thi4

variability is due to the absence of many of the proteins from a partic-





Figure 17. Electrophoretic pattern of proteins from
prepared in Buffer C (Experiment 2)


small subunits


r*q


*J)


rli

',3


r





sull~


Figure 18. Electrophoretic pattern of proteins from small subunits
prepared in Buffer C (Experiment 3).


,,
r,


L
e





+t++++++++++++++C+++++1++++








++++++t+++++l++l+++++++++++








+ee++++c++++o++o++++++6++++,








oi lt++++++++o++o+o+ocol+o+c








o+++++++++++of++f++++++++++


IO



CI


0.::
























CO:







C O


WlcO .

~IU
~clla
uix
>iw
P31





U
a
x
q





I"

Ix
w
,I

II,



x


a+++++l++++++ .4 44 4++


+++++o0++o+o ++1+1cl++++o


+d"++++f+++'+++++++++++++++


~Nm~~~~~~O~NmJ~~~~~O~Nm~~~~
~~~~~~~~~~N~~~N~NN


ct+c




- 1














+f++f++++++f+++







++++1+1++++f+1+







+ +++~+1+0++++







+oo++++:-i+++o+







++++fi+0+0++1+







+o0++ +++f++ +







++0++1+++++++1+







+++++ff+++++1+







+cec~+o+o+o++1+


oioo+ocoolioco







co+ocooaico 1+







coo++oo+oo co+







occoo++coococo







000++o0+o+i+o+







Iocoocooococo







++oco+ooootoco







++o+oococo+oco







concao+o++10+o


m~C~N~~~~~m~O~
NNmmmmmmm~mmu~


mo Un a,~c~ co C -I ~~ E d a





+ -D -

2 -D


++
68-

44-1 1 (




deL 711
12deY 13
23.5' 14)*. 1S***
e9 (17 j 21
1 M20 **22
fl) *23
24~ 25
26 92
I-i. 2 h 239 m.,
33~ 3k 34 35
369
I4 .4- - mr34*,
40~ nG 939

P 41



Figure 19). Schematic diagram of bovine mitochondrial small-subunit
proteins.
Proteins which were reproducibly observed in preparations of small
subunits are designated by numbers, and proteins seen less frequently
are marked with letters.










ular experiment. In Experiment 6 for example, the quantity of protein

in the sample was inadequate; as a result all of the spots obtained

were faint, so the absence of many of the spots which are normally

faint in adequately loaded gels is understandable. The absence of

6 of the numbered proteins from Experiment 3 (Figure 18), however, is

not so simply explained. On the other hand, there was a subset of the

experiments in which variability was not a problem at all: none of

the numbered proteins were lacking in Experiments 2, 8, or 9, and only

two were missing in Experiment 4.

The variability which was observed in the other experiments seemed

to be random with respect to which proteins were found to be missing.

That is, it was not due to a high level of variability in a few specific

proteins. Only two of the numbered proteins, MS13 and MS21, were absent

from as many as three of the nine experiments. In the case of these

two proteins (as well as MS1),, much of the lack of reproducibility was

due to their absence from subuinits prepared in bufifer D; thus9 there may

be a specific ionic effect on the binding of these proteins to the ribo-

some, as was seen for MLec and MLg above. A complementary effect was

found for MSa and MSb. These two proteins were always present in sub-

units prepared in Buffer C, but only once in any other preparation.

The general impression of unreproducibility in the small subunit

pattern derives partly from the relatively large number of spots which

appear in fewer than half of the analyses and thus are designated by

letters in Table VII. Some of thesP, like MiSc and MISn (Figure 6),










appeared in only one experiment and can be easily disregarded. Others

(MISa, b, f, k, and p) were present in four of the nine experiments and

thius represent borderline cases which are not clearly Excluded from the

ribosomal protein category.

Another factor which contributes to this impression of unrepro-

ducib~ility is the variation in staining intensity among the different

spots on a given gel. In most experiments (Figiure 17 is a conspicuous

exception) the electrophoretic pattern was found to consist of a minor-

ity of heavily-stained spots, usually including MlS4, 5, 6, 10, 17, 19,

24, 28, 31, 37, and 38, and a majority of fainter spots. Furthermore

the relative intensity of a given spot, particularly among the fainter


group, varied considerably among different grls. (Compa~re M1S13, 15,

and 16 in Figures 6 and 20, for example.) Some differences in the

stain-binding abilities of different proteins hcve been reported

(Fazekas de St. Croth et al., 1963; Bickle and Traut, 1971), but the

differences are not of the magnitude required to explain the present

results if all of the smlall-subunit proteins were present in equimolar


quantities in these gels. And in any case th~is phenomenon could not

account for the variations observed for a particular protein in


separate experiments. It is therefore necessary to conclude that the

small-subunit proteins are present in differing molar quantities in

these preparations. It follows that the individual subunits in each


preparation are heterogeneous withi respect to the set of proteins they

contain.









































10






Figure 20. Electrophoretic pattern of proteins from small subunits
prepared in Buiffer E (Experiment 9) .





This heterogeneity could be due to the non-specific binding of con-

taminating proteins to some of the subunits. However, in view of the

small number- of heavily-staining proteins found in these particles, it

seems unlikely that all of the fainter spots represent contaminants. A

more plausible hypothesis is that a proportion of the subunits are

lacking some of the ribosomal proteins, as a result of an in vitra

artifact. This hypothesis is supported by the similar results which

have been nL gained (Kaltschmidt and Wittmann, 1970) and explained

(Hardy, 1975) for E. coli ribosomal proteins. Some possible causes of

such an artifact will be considered below (Section IV); but whatever

the cause, it is probably responsible for the variability between

ribosome preparations as well as the heterogeneity within each prepa-

ration. Similar considerations also apply (to a lesser extent) to the

proteins of the large subunit, which show a lower level of variability

and non-uniformity of staining intensities.

The effect of extreme-salt treatment (Buffer G) on the protein

content of the small subunit was also investigated. One of thle gels

obtained is shown in Figure 21. The most notictarble result was a

general reduction in the quantity of the low-molecul ar-weight proteins,

those below MIS31. Several proteins (M~S13, 16, and 35) were completely

absent from all three preparations of Buffer G subunits, and MS27, 28,

33, 36, 37, and 41 were diminished to negligible levels. Some proteins,

on the other hand, were increased relative to the quantities present

in high-salt-treated (Buffers C and D) subunits: this was particularly





t


Figure 21. Electrophoretic pattern of proteins from small subunits
prepared in Buffer G.
The irregular spots seen above and to the right of M3S4 are
artifacts due to precipitation of thle stain on the surface of the gel.


Ir )


**




Full Text

PAGE 1

PROTEINS OF THE MAMMALIAN MITOCHONDRIAL RIBOSOME by DAVID EARL MATTHEWS A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1976

PAGE 2

ACKNOWLEDGEMENTS The author gratefully acknowledges the financial support awarded him in the form of fellowships from the National Science Foundation and the Graduate Council. Thanks are also due to Mark Critoph, Warren Clark, and Mark Moffitt for their technical assistance in these experiments; and to Nancy Denslow, Robert Hessler, and Mary Conde for helpful discussions. The senior research project of Dean Kane contributed much to the studies on the isoelectric points of mitochondrial ribosomal proteins. The encouragement and assistance of Patty Matthews in the preparation of this dissertation were invaluable. The author also wishes to express his gratitude for the helpfulness of the faculty of the Department of Biochemistry, and of his supervisory committee in particular; their good will, advice, and scientific insights have made the past four years of graduate study a most rewarding experience. Most of all, the profound influence of Dr. Thomas W. O'Brien in shaping this research, as well as the author's scientific training, goals, and attitudes, is humbly and gratefully acknowledged .

PAGE 3

TABLE OF CONTENTS Acknowledgements ii List of Tables v List of Figures vi Abbreviations Used viii Abstract ix Introduction ; 1 I. Structural properties of non-mitochondrial ribosomes 3 II. Structural properties of mitochondrial ribosomes 7 A. Mitochondrial ribosomes of protists, fungi and plants 7 B. Mitochondrial ribosomes of animals 12 III. Phylogenetic relationships in ribosome structure 17 IV. Summary 21 Materials and Methods 23 I. Materials 23 II. Preparation of ribosomes 25 A. Bovine mitochondrial ribosomes 25 1. Method A 26 2. Method B 27 3. Method C 27 B. Rat mitochondrial ribosomes 28 C. Bovine cytoplasmic ribosomes 28 D. E. coli ribosomes 29 III. Extraction of ribosomal proteins 29 IV. Radioactive labelling of ribosomal proteins 30 A. Radioidination 30 B. Reductive methylation 31 V. Two-dimensional electrophoresis 31 VI. Isoelectric focusing 34 VII. Assays of ribosome function 35 VIII. Buoyant density determinations 36 IX. Quantitative measurements 37 Results 38 I. Preparation and characterization of mitochondrial ribosomes 39 II. Electrophoretic analysis of mitochondrial ribosomal proteins 47

PAGE 4

A. Criteria for the identification of ribosomal proteins m m 52 B. Identification of ribosomal proteins 57 1. Large subunit . ., 57 a. Reproducibility 61 b. Resistance to salt-washing 66 Low-salt treatment 68 High-salt treatment 72 Extreme-salt treatment 73 c. Summary of the large-subunit proteins 78 2. Small subunit 79 3. Comparison of large and small subunits 91 C. Experimental evaluation of possible artifacts 96 III. Comparison with proteins of non-mitochondrial ribosomes _ 100 IV. Molecular weights of mitochondrial ribosomal proteins t m ;qq V. Comparison with proteins of rat mitochondrial ribosomes 219 Discussion 128 Bibliography , -. / 9 Biographical Sketch ^55

PAGE 5

LIST OF TABLES I. Properties of cytoplasmic ribosomes of eukaryotes 4 II. Properties of Moneran and chloroplast ribosomes 6 III. Properties of mitochondrial ribosomes 8 IV. Compositions of buffers used in preparation of ribosomes 24 V. Functional activity of mitochondrial ribosomes prepared with or without DEAE-cel lulose 48 VI. Appearance of large-subunit proteins in separate experiments „ 58 VII. Appearance of small-subunit proteins in separate experiments 82 VIII. Molecular weights of mitochondrial large subunit proteins Ill IX. Molecular weights of mitochondrial small-subunit proteins i \2 X. Calculation of the protein content of bovine mitochondrial ribosomes 115

PAGE 6

LIST OF FIGURES 1. Phylogenetic relationships in ribosome structure 20 2. Preparation of mitochondrial ribosomes by Method A 41 3. Preparation of mitochondrial ribosomes by Method B 43 4. Preparation of mitochondrial ribosomes by Method C 45 5. Elect rophoretic pattern of large-subunit proteins f i oui bovine mitochondrial ribosomes 50 6. Electrophoretic pattern of small-subunit proteins from bovine mitochondrial ribosomes 51 7. Effects of treatment with buffers of various ionic composition on the peptidyl transferase activity and buoyant density of bovine mitochondrial large subunits 55 8. Schematic diagram of bovine mitochondrial largesubunit proteins 60 9. Electrophoretic pattern of proteins from large subunits prepared in Buffer C (Experiment 2) 63 10. Electrophoretic pattern of proteins from large subunits prepared in Buffer C (Experiment 4) 64 11. Electrophoretic pattern of proteins from large subunits prepared in Buffer D (Experiment 6) 65 12. Effect of high-salt treatment on the protein content of E. coli ribosomes 67 13. Electrophoretic patterns of proteins from low-salt treated large and small subunits 69 14. Electrophoretic pattern of proteins from large subunits prepared in Buffer C (Experiment 3) 70

PAGE 7

15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

PAGE 8

ABBREVIATIONS USED ATP bisacrylamide C cytoribosome d DNA EDTA g G GTP leu mitoribosome MgAc 2 ML MS mRNA phe poly U POPOP PPO RNA rRNA S SDS TEA TEMED Trls tRNA e-ME adenosine-5 '-triphosphate N, N '-methylenebisacrylamide cytos Lne cytoplasmic ribosome daltons deoxyribonucleic acid (ethylenedinitrilo) tetraacetic acid gravity guanine guanosine-5 '-triphosphate leucine mitochondrial ribosome magnesium acetate mitoribosomal large-subunit protein mitoribosomal small-subunit protein messenger RNA phenylalanine polyuridylic acid 1 , 4-bis [2-(5-phenyloxazolyl) ] benzene 2, 5-diphenyloxazole ribonucleic acid ribosomal RNA Svedberg unit sodium dodec*'! sulfate triethanolamine N, N, M', N'-tetramethyl-ethylenediamine tris ( hy d r oxyme thyl) am inome thane transfer RNA (vtnercaptoethanol

PAGE 9

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PROTEINS OF THE MAMMALIAN MITOCHONDRIAL RIBOSOME By David Earl Matthews December, 1976 Chairman: Thomas W. O'Brien Major Department: Biochemistry For some years it was widely believed that all ribosomes were either of the 70S type found in bacteria or of the 80S type occurring in the cytoplasm of eukaryotic cells. With the discovery and subsequent characterization of ribosomes in the mitochondria of eukaryotes, this simple generalization had to be rejected. The mitochondrial ribosomes of various organisms display a great diversity of structural properties, even though they all show considerable functional homology with the 70S prokaryotic ribosome. Animal species possess mitochondrial ribosomes with particularly interesting physical-chemical properties, by virtue of their high protein content relative to the quantity of RNA they contain. The present research is an investigation into the nature of the individual proteins which make up the large complement of total protein

PAGE 10

found in mammalian mitochondrial ribosomes. Two-dimensional electrophoresis in polyacrylamide gels was used to separate, identify, and characterize the proteins of mitochondrial ribosomes from bovine liver. It was found that these ribosomes contain a relatively large number of proteins: 52 are present in the large subunit and 41 in the small subunit. Several kinds of criteria and experimental evidence argue that these proteins are true ribosomal components rather than contaminants in the ribosome preparations. However, only 81 electrophoretically distinguishable proteins are present in a mixture of large and small subunits; the possibility that some of the large-subunit proteins are identical to proteins found in the small subunit is discussed. The molecular weights of these proteins are similar to those of bovine cytoplasmic ribosomal proteins, and considerably larger than those of Escherichia coli ribosomal proteins. Comparisons of electrophoretic properties show that the mitochondrial proteins are more acidic than those of mammalian cytoplasmic or most bacterial ribosomes. The sum of the molecular weights of the 93 mitochondrial ribosomal proteins is somewhat greater than predicted by physical-chemical measurements of the total mass of protein in this ribosome; to account for this difference it is suggested that a proportion of the ribosomes as they are isolated may be lacking some of the ribosomal proteins. Other experiments demonstrated that exposure of the ribosomes to certain ionic conditions could remove a small number of the proteins, with a corresponding loss of the peptidyl transferase activity of the

PAGE 11

ribosomes. One or more of the proteins removed may therefore be involved in this particular ribosomal function. Finally, comparisons between the proteins of bovine and rat mitochondrial ribosomes revealed a large number of differences in their electrophoretic properties. The level of divergence in these properties is greater than that reported in similar comparisons of cytoplasmic ribosomal proteins. This observation parallels the great phylogenetic diversity observed in the overall structural properties of mitochondrial ribosomes from more distantly related organisms, and leads to the conclusion that these ribosomes have diverged more widely during evolutionary history than have their extrao mitochondrial counterparts located only a few Angstroms away. The large number and Low isoelectric points of the proteins of mammalian mitochondrial ribosomes are discussed with reference to their implications for the structural organization of these particles. Some mechanisms which may account for the rapid evolutionary divergence of mitochondrial ribosomes are proposed. XL

PAGE 12

INTRODUCTION Mitochondria are multifunctional organelles found in all eukaryotic cells, their primary function being the aerobic production of ATP. In addition to their important role in cellular energy metabolism, mitochondria make an essential contribution to their own biogenesis. Several of the multi-subunit enzymes of oxidative phosphorylation contain one or more subunits synthesized within the mitochondrion (Schatz and Mason, 1974). For this purpose mitochondria possess a large complement of biosynthetic enzymes and other macromolecules, distinct from their analogs in the nucleus and extramitochondrial cytoplasm. These components of the mitochondrial biogenetic system include DMA, DNA and RNA polymerases, messenger RNA, ribosomes, translation factors, transfer RNA species and aminoacyl-tRNA synthetases. Mitochondrial DNA codes for the ribosomal RNA and at least some of the tRNA and mRNA species found in mitochondria, and these macromolecules are indispensable in the biogenesis of functionally active mitochondria. On the other hand most of the (equally essential) protein components of the mitochondrial biogenetic system ribosomal proteins, factors, and enzymes appear to be coded by nuclear DNA and synthesized on cytoplasmic ribosomes. Thus the mitochondrial and nuclear-cytoplasmic macromolecult.--synthesizing systems must cooperate as intimately in the production of the mitochon-

PAGE 13

drial biogenetic system itself as they do in the synthesis of the enzymes of oxidative phosphorylation. The ribosomes in a given organism's mitochondria are generally distinguishable from its cytoplasmic ribosomes on the basis of several functional or physical-chemical criteria. Indeed, especially in their functional properties, mitochondrial ribosomes have been found to be more like bacterial ribosomes than cytoplasmic ones. Some structural similarities between mitochondrial and Moneran ribosomes were noted early (Kuntzel and Noll, 1967), and complemented reports of other biochemical homologies between organelles and prokaryotes that had already aroused considerable interest in the question of the evolutionary origin of mitochondria and chloroplasts. More recent comparisons of the structural parameters of bacterial and mitochondrial ribosomes have shown some similarities but also a surprising number of differences, both between the two groups and among mitochondrial ribosomes from different species. Thus, it is true that the mitochondrial ribosomes of most organisms studied to date sediment more slowly than the corresponding cytoplasmic ribosomes, and that some of them have sedimentation coefficients close to that of prokaryotic ribosomes (70S). However, mitoribosomes from various species range in sedimentation rates from 55S to 80S, a much wider variation than is found among bacterial, eukaryotic-cytoplasmic , or chloroplast ribosomes obtained from different organisms. Indeed, when all ribosomal attributes are considered, it seems that only in mitochondria do so many different kinds of ribosomes occur.

PAGE 14

I . STRUCTURAL PROPERTIES OF NON-MITOCHONDRJ AL RIBOSOMBS Before proceeding to the description of the physical and chemical characteristics of mitochondrial ribosomes, it will be useful to summarize the characteristics of the other kinds of ribosomes that exist. This task is simplified by the fact that the characteristics of nonmitochondrial ribosomes are not as divergent as might be thought. They all fit reasonably well into two large categories within which the members seem to share more similarities than differences. Table I shows the relative homogeneity of the properties of cytoplasmic ribosomes, whatever eukaryotic organism they are obtained from protists, fungi, plants or animals. Despite small differences among the cytoplasmic ribosomes from these four taxonomic kingdoms, all of them appear to be members of a single structural class typified by a sedimentation coefficient of SOS, a buoyant density of 1.57 g/cc, and rRNA molecules of 0.7 and 1.3 million daltons containing 50 percent G + C. The second category of non-mitochondrial ribosomes is also relatively uniform in physical and chemical properties. Described in Table II are several prokaryotic ribosomes from E. coli, a mycoplasm, and a blue-green alga and those of various chloroplasts. The prokaryotic particles can all be adequately described by the values 70S, 1.64 g/cc, (0.56 + 1.10) x 10 6 daltons, and 50% for the structural parameters tabulated. Chloroplast ribosomes are similar in all respects except that some of them appear to be significantly lower in buoyant density. Perhaps it should be mentioned that this homogeneity of

PAGE 15

H + 0) (-1 to < 01 4-> 4-1 O + + +

PAGE 16

H

PAGE 17

,3 iX> &o o H r-l H Z 3 2 + + + + m vo vo o 4-1 u 3 m oi m m -•:

PAGE 18

gross structural properties almost certainly conceals a great deal of diversity in fine structure. Differences in such properties as ribosomal protein electrophoretic mobilities and immunologic identities are the rule even for prokaryotic ribosomes related to each other more closely than are the entries of Table II (Geisser e_t aj_. , 1973; Wittmann et al., 1970). II. STRUCTURA L PROPE RTIES OF MITOCHONDRIAL RIBOSOMES In contrast to the rather simple classification scheme possible for ribosomes from all other sources, mitochondrial ribosomes do not seem to fall into one or even a few structural categories. Reference to Table III shows the degree of diversity found in mitochondrial ribosomes from different species. Many of the individual ribosome species in this table are as distinct in physical and chemical properties from each other as the 80S (Table I) and 70S (Table II) classes are. Furthermore, among the protists and fungi no two genera have yet been shown to contain similar mitoribosomes, so it seems likely that many more structurally different ribosomes will be found as other species are investigated. On the other hand, considerable homology is seen among the mitoribosomes of several species of higher animals, from locust to man. A. Mitochondrial Ribosomes of Protists, Fungi, and Plants The best-characterized protist mitochondrial ribosomes are those of Euglena gracilis and Tetrahymena pyriformis . The Euglena mitoribosome is unusual in that it is the only mitochondrial ribosome yet described that shows a large degree of structural homology with the

PAGE 19

c

PAGE 20

u vi ^

PAGE 21

10 ribosomes of prokaryotes. The fact that only one of the many mitoribosomes shown in Table III is similar to bacterial ribosomes in its physical-chemical properties is quite remarkable, in view of the extensive functional homologies between mitochondrial and bacterial ribosomes that have been found in every case tested. And even in this instance the structural similarity is by no means complete. Although the mitoribosome of Euglena is similar to the E. coli ribosome in its sedimentation coefficient, the molecular weights of its rRNA, and its high buoyant density, the base composition of its rRNA (27% G 4C) is very different. Tetrahymena mitoribosomes are quite dissimilar from those of Euglena , and indeed from any other ribosome yet studied. Although the Tetrahymena mitoribosome shares with that of Euglena an exceptionally low rRNA GC content, it more nearly resembles cytoplasmic ribosomes with respect to its sedimentation coefficient. And its buoyant density and rRNA molecular weight values are smaller than those of any other ribosome except the mitoribosomes of animals. The mitochondrial ribosomes of various fungi also show a rather large diversity in sedimentation coefficients and buoyant densities. Unfortunately, a nearly equal diversity in these parameters has been reported by different research groups studying the same fungal species. Thus the reported sedimentation coefficients range from 72S to 80S for Saccharomyces mitoribosomes, and from 73S to 80S for those of Neurospora. Although the buoyant density of a 74S form of the Saccharomyces

PAGE 22

1 1 raitoribosrme was found to be 1.64 g/cc (Grivell et_ aK , 1971), the composition of the 80S version (Morimoto and Halvorson, 1971) corresponds to a buoyant density of 1.56 g/cc. There is reasonably good evidence that the 80S ribosomes described in these species really are mitochondrial ribosomes rather than cytoplasmic contaminants (O'Brien and Matthews, 1976). It therefore seems most likely that the contradictory reports are due to artifactual alterations of the structure of these ribosomes by som. of the preparative procedures used. However it is not yet clear which of the results are artifactual, so it is difficult to compare the sedimentation coefficients and buoyant densities of fungal mitoribosomes with those of other kinds of ribosomes at present. There is much less controversy about the properties of the rRNA of fungal mitoribosomes. Indeed, the molecular weights and GC contents of these molecules are quite similar in all the fungal species tested. The GC contents are rather low, nearly as low as those found in protist mitoribosomal RNA. The molecular weights, on the other hand, are quite high: these molecules are of about the same size as the rRNAs found in cytoplasmic ribosomes. The properties of mitochondrial ribosomes from higher plants are not firmly established. The best-documented results are those shown in Table III. It may be seen that the physical-chemical characteristics of these ribosomes are quite similar to those of cytoplasmic ribosomes in general. In fact, they differ only slightly or not at

PAGE 23

12 all from those of the cytoplasmic ribosomes found in these same organisms (Leaver and Harmey, 1973; Pring and Thornbury, i975; Pring and O'Brien, 19/., private communication). Such results raise the possibility that these ribosomes are actually of cytoplasmic rather than mitochondrial origin. This interpretation is supported by the observation that these ribosomes are inhibited by anisomycin and not by chloramphenicol, like cytoplasmic ribosomes and unlike any other known mitochondrial ribosomes (Pring, Denslow, and O'Brien, 1975, private communication) . There have been isolated reports of plant mitoribosomes with properties different from those described above. A 70S particle was obtained from mung bean mitochondria (Vasconcelos and Bogorad, 1971), and a 66S mitoribosome from maize (Wilson e^ al . , 1968). Mitochondrial RNA from Virginia creeper was found to be relatively small, 0.42 x 10 and 0.84 x 10 in molecular weight. However, the functional properties of these ribosomes have not been described, and the uncertainty about the nature of mitochondrial ribosomes in higher plants remains unresolved. B . Mitochondrial Ribosomes of Animals In contrast to the fascinating and perplexing diversity of mitochondrial ribosomes from protist, fungal and plant species, Metazoan mitoribosomes appear strikingly uniform in structure. Although most of the animal species investigated have been mammalian, the mitoribosomes of the toad Xenopus have also been thoroughly characterized, and enough data have been presented for several invertebrates to justify

PAGE 24

13 a tentative conclusion that the mitoribosomes of all multi-cellular animals may be quite similar. Furthermore, this relatively homogeneous group of ribosomes is distinctly different in structural properties from any other ribosomes yet described. The first difference to be noticed was the low sedimentation coefficient of animal mitoribosomes. Values within the range 54S to 61S have been obtained for these particles from rat (O'Brien and Kalf, 1967), HeLa cells (Perlman and Penman, 1970), rabbit, pig, cow (O'Brien. 1971), hamster (Coote et al. , 1971), chicken (Rabbitts and Work, 1971), toad (Swanson and Dawid , 1970), shark (O'Brien, 1972), and locust (Kleinow e_t al . , 1971) . The early inference drawn from these low sedimentation coefficients was that animal mitoribosomes were smaller (lower in molecular weight) than other ribosomes (Borst and Grivell, 1971), and this inference was reinforced by the later observations of unusually small rRNA in these particles. An alternate possibility, that the 55S particle is actually a subunit of the functional monoribosome, has been excluded by the dissociation of the 55S structure into two subunits and the demonstration that either the 55S particle or a mixture of both subunits is competent for poly U-dependent phenylalanine incorporation (Leister and Dawid, 1974; O'Brien et al., 1974). But there is a third interpretation of the low sedimentation coefficient of animal mitoribosomes, because in point of fact these 55S particles are not exceptionally small, either in molecular weight

PAGE 25

14 or in physical dimensions. The particle weight of the bovine mitoribosoiriu as determined by high-speed equilibrium centrifugation is 2.8 million daltons (Hamilton and O'Brien, 1974), slightly greater than the value reported for the ribosome of E. coli (Hill et al., 1969). De Vries and Kroon (1974) have presented evidence that rat mitoribosomes are even larger in volume than the E. coli particles, though smaller than rat cytoribosomes. These investigators electrophoresed ribosomes into gels composed of a gradient of polyacrylamide concentration until the particles could make no further progress through the decreasing pore size of the gel matrix. The mitoribosomes penetrated farther into the gels than cytoribosomes but not as far as bacterial ribosomes. The physical dimensions of ribosomes can also be determined by electron-microscopic measurements. The results confirm the conclusion that mitoribosomes are smaller than cytoribosomes in rat (O'Brien and Kal.f, 1967; Aaij et al. , 1972) and locust (Kleinow e_t al_. , 1974), but direct comparisons of mitochondrial and bacterial ribosomes (under the same conditions of fixation and staining) have not been performed. How can two particles of the same molecular weight sediment respectively at 55S and 70S? The significant difference appears to be the much lower buoyant density of the animal mitochondrial ribosome. Buoyant density values from 1.40 to 1.46 g/cc have been found for mitoribosomes of HeLa cells (Perlman and Penman, 1970; Wengler et al., 1972), rat (Sacchi et al., 1973; de Vries and Kroon, 1974), cow

PAGE 26

L5 (Hamilton and O'Brien, 1974), and toad (Leister and Dawid, 1974). The buoyant density of a ribosome can be used to calculate the relative proportions of RNA and protein in the particle (Hamilton, 1971); from the relationship given in the notes to Table I an RNA content of about 30 percent can be calculated for animal mitoribosomes, as contrasted with 63 percent for bacterial ribosomes (Tissieres e_t_ a_l. , 1959) . From the RNA content and the sum of the molecular weights of the rRNA molecules, the particle weight of the ribosome may be calculated. By this means molecular weight estimates equal to or greater than those for bacterial ribosomes were obtained for mitoribosomes from rat (Sacchi et al . , 1973; de Vries and Kroon, 1974), cow (O'Brien et al . , 1974), and toad (Leister and Dawid, 1974) even before the molecular weight was determined directly by sedimentation equilibrium (Hamilton and O'Brien, 1974). Doubts have been raised about the validity of buoyant density values as measures of the protein content of ribosomes (McConkey, 1974). Others have suggested that the low buoyant densities of animal and Tetrahymena mitoribosomes might be due to membrane fragments adhering to these particles (specifically, due to membrane lipids, which are assumed to be absent in the calculation of protein content from buoyant density) (Borst and Grivell, 1971). Several lines of evidence indicate that these factors do not represent significant objections to the description of animal mitoribosomes presented above. Determinations of the protein content by either chemical analysis or ultra-violet

PAGE 27

16 absorption spectra of the ribosomes are in agreement with estimates from buoyant density, yielding values of 70-80 percent protein in rat (O'Brien and Kalf, 1967) and toad (Leister and Dawid , 1974). No phospholipids were detectable in rat mitoribosomes (de Vries and Kroon, 1974). The molecular weight estimate for bovine mitoribosomes based on their buoyant density and the size of their rRNA agrees quite closely with the molecular weight determined by sedimentation equilibrium (Hamilton and O'Brien, 1974). Finally, similar values for the total protein content of Xenopus mitoribosomes have been obtained direci ly by summing the molecular weights of the individu i] ribosomal proteins (Leister and Dawid, 1974). Besides their unusual sedimentation behavior and high protein content, another unusual characteristic of animal mitoribosomes is the small size of their rRNA molecules. Values of 0.35 and 0.54 million daltons have been obtained for HeLa mitoribosomal RNA by electronmicroscopic length measurements (Robberson e_t_ al_. , 1971), and these numbers are in good agreement with those found by other methods for rat (Sacchi et al., 1973), toad (Dawid and Chase, 1972; Leister and Dawid, 1974), shrimp (Schmitt et al., 1974), and locust (Rleinow, 1974). Thus animal mitoribosomes contain scarcely more than half as much RNA as any nonmitochondrial ribosome known. To a first approximation, these particles may be pictured as E. coli ribosomes modified by converting half of the RNA into an equal mass of protein.

PAGE 28

17 The base composition of the rRNA of animal mitoribosomes is distinctly higher in guanine and cytosine than that of protist or fungal mitoribosomes, though still lower than that of animal cytoplasmic ribosomes. G + C contents of 40-47 percent have been found for mitochondrial rRNA of HeLa (Vesco and Penman, 1969), rat (Bartoov et al . , 1970), toad (Dawid and Chase, 1972), and shrimp (Schmitt et al., 1974). The degree of methylation is also lower for animal mitoribosomal RNAs than for either cytoplasmic or prokaryot ic rRNAs (Dubin, 1974). Although mitochondrial ribosomes from all animal species examined thus far appear quite similar in their physical and chemical properties, it may be anticipated that they will differ in their detailed structure. In fact, detectable non-homology has already been found between the base sequences of mitoribosomal RNA from two species of toad, Xenopus laevis and X. mulleri (Dawid, 1972). Electrophoretic differences have also been demonstrated in several of the mitoribosomal. proteins of these two species (Leister and Dawid, 1975). III. PHYLOGENETIC RELATION SHIPS IN RIBOSOME STRUCTURE The comparisons presented above permit some conclusions about the variation in structural properties among mitochondrial ribosomes of different organisms, and the differences between the ribosomes of mitochondria and those of prokaryotes, eukaryotic cytoplasm, and chloroplasts. The most obvious generalization is that mitochondrial ribosomes show more diversity in all of their structural properties than do any other kinds of ribosomes. Cytoplasmic ribosomes of all four eukaryotic kingdoms

PAGE 29

18 are relatively similar with respect to sedimentation coefficient, buoyant density, rRNA size, and guanine plus cytosine content. Prokaryotic ribosomes display even more uniformity in these characteristics, while chloroplast ribosomes, whether obtained from protists or from higher plants, show remarkable homologies not only among themselves but to a large extent between themselves and prokaryotic ribosomes. Mitochondrial ribosomes, on the other hand, can vary in sedimentation coefficient even within a kingdom: such differences do exist between the protists Euglena and Tetrahymena , and almost certainly exist among the fungi. Similarly, differences in the buoyant densities of mitoribosomes are found within both the protist and the fungal groups. Mitoribosomal RNA molecular weights and G + C contents, in contrast, appear to vary significantly between kingdoms but not within them. The correlation between these last two structural properties of mitochondrial ribosomes and the. taxonomic kingdoms in which they are found is illustrated in Figure 1. Also plotted in Figure 1 are the data for chloroplast and prokaryotic ribosomes, which cluster together as expected, and for four kingdoms of cytoplasmic ribosomes. The latter group also forms a cluster, with the exception of the cytoribosome from Te trahymena . Both of these groups are dissimilar from any of the three mitoribosome kingdoms plotted, and these are in turn distinct from each other. Thus the two parameters, rRNA size and G + C content, discriminate ribosomes along the lines of their phylogeny and intracellular location.

PAGE 30

Figure 1. Phylogenetic relationships in ribosome structure. ©Mitochondrial Ribosomes Fungi S Saccha romyces Reijnders e_t_ _al. , 1973; Morimoto and Halvorson, 1971. A Asper gillus Verma et al. , 1970; Edelman et al. , 1970. C Candida utilis Vignais et al. , 1972. N Ne urospora Neupert e_t al . , 1969; Kiintzel and Noll, 1967. Protists Eu Euglena Krawiec and Eisenstadt, 1970; calculated from Avadhani and Buetow, 1972. T Tetrahymena Reijnders et al. , 1973; Chi and Suyama, 1970. Animals X Xenopus Leister and Dawid , 1974; Dawid and Chase, 1972. H HeLa Robberson et al. , 1971; Vesco and Penman, 1969. R Rat Sacchi et al. , 1973; Bartoov et al. , 1970. BProkaryotic Ribosomes E Escherichia Kurland, 1960; Morimoto and Halvorson, coli 1971. M Mycoplasm a Johnson and Horowitz, 1971. hominis Adtloroplast Ribosomes Eu Euglena Rawson and SLutz, 1969; calculated from Avadhani and Buetow, 1972. Sp Spinach Hartley and Ellis, 1973; Lyttleton, 1962. O Cy top lasmic Ribosomes Fungi S Saccharomyees Reijnders et al., 1973; Morimoto and Halvorson, 1971. A A spergillus Verma et al. , 1970; Edelman et al . , 1970. C Candida util is Vignais et al. , 1972. N Neurospora Neupert et al. , 1969; Kiintzel and Noll, 1967. Protists Eu Euglena Krawiec and Eisenstadt, 1970; calculated from Avadhani and Buetow, 1972. T Tetrahym ena Reijnders et al. , 1973; Chi and Suyama, 1970. Animals X Xenopus Loening et al. , 1969; Dawid et al. , 1970. H HeLa Darnell, 1968; calculated from Vesco and Penman, 1969. R Rat Reijnders et al. , 1973; Kirby, 1965. Plants P Pea Loening et al . , 1969; Bonner and Varner, 1965.

PAGE 31

20 CO O to c o ~o o c c o o < cr 2.5 2.0 0.5 20 30 40 50 60 RNA G + C Content (moles percent)

PAGE 32

21 Sedimentation coefficients and buoyant densities, on the other hand, do not group mitoribosomes from the same kingdom together; nor do they differentiate them from the various non-mitochondrial ribosomes. This observation suggests that these two structural properties have been less conserved than rRNA size and G + C content in the course of ribosome evolution. It is noteworthy that buoyant density is the only one of these four structural characteristics which discriminates chloroplast from prokaryotic ribosomes (Stutz and Boschetti, 1976). However, some of the im rakingdom variability reported for sedimentation coefficients and buoyant densities may be due to the possibly greater sensitivity of these parameters to differing conditions used in the preparation of the mitoribosomes. IV. SUMMARY Perhaps the most remarkable structural property of mammalian mitochondrial ribosomes is the very high ratio of protein to RNA found in these particles. On the basis of various kinds of physical-chemical measurements it appears that this ratio is approximately 2:1, or 1.8 x 10 daltons of protein to 0.9 >: 10 6 daltons of RNA. These ribosomes thus contain about twice as much protein as E. coli ribosomes, and about half as much RNA. Such an unusual composition poses questions about the molecular architecture of these particles, and about the functional roles played by this large quantity of protein in the process of protein synthesis. Is the difference in protein content between the ribosomes of E. coli

PAGE 33

22 and those of mammalian mitochondria due to a difference in the number of the ribosomal proteins or to a difference in their size? Do the different kinds of molecular interactions which must be involved in maintaining the structural integrity of such a protein-rich ribosome correlate with any identifiable differences in the properties of the proteins? On the other hand, might the high protein content found in this ribosome represent nothing more significant than the presence of large quantities of contaminating non-ribosomal proteins? A portion of the present research is directed toward these questions. Also of interest is i iie phylogenetic diversity in structural properties which is found in comparisons of mitochondrial ribosomes from distantly related organisms. Only small differences in the overall physical and chemical properties of mitoribosomes are seen when species within the animal kingdom are compared. But if a high degree of evolutionary divergence is indeed the rule for mitochondrial ribosomes in general, it should be possible to detect significant differences between the mitoribosomes of more closely related species at a sufficiently detailed level of analysis. The rule further predicts that such differences will be greater between mitochondrial ribosomes than between the cytoplasmic ribosomes of the same species. The experiments presented below include a test of these predictions.

PAGE 34

MATERIALS AND METHODS I. MATERIALS Sucrose (density gradient grade, ribonuclease free) was obtained from Schwarz/Mann. Urea (reagent-grade) was purchased from J. T. Baker or frcm Mallinckrodt . Stock solutions of 10 M urea were prepared, filtered, and stored at room temperature for no more than 24 hr before use; if the conductivity of the stock solution was greater than 50 pmho, it was deionized by stirring with AC501-X8 resin (Bio-Rad) . Acrylamide and N, N'-methylene-bisacrylamide (Eastman) were recrystallized from chloroform and acetone, respectively; stock solutions were deionized with Rexyn-300 (Fisher) and stored at room temperature in the dark. Ampholine was obtained from LKB. Bovine serum albumin and human Y-globulin were obtained from Nutritional Biochemicals Corporation, and egg albumin, equine myoglobin and egg-white lysozyme were from Sigma Chemical Company. Puromycin dihydrochloride was from Nutritional Biochemicals Corporation. Na 125 I (carrier-free, in 0.1 N NaOH) and [ 1 ^ f C] formaldehyde (44 mCi/mmol) were purchased from New England Nuclear; and [4,53 H]L-leucine (55 Ci/mmol) , [83 H] GTP (12 Ci/mmol, tetrasodium salt), and [ J H] I. -phenylalanine (7 Ci/mmol) from Schwarz/Mann. 23

PAGE 35

<

PAGE 36

25 II. PREPARATION OF RIBOSOMES A. Bovine Mitochondrial Ribosomes In the course of this research the procedure for the preparation of mitochondrial ribosomes from bovine liver was modified in several ways to improve the yield and purity of the ribosomes obtained. The major changes are discussed under Methods A, B, and C below. All of the preparative methods shared the following common features. Livers of freshly killed animals were obtained from a slaughterhouse and transported to the laboratory on ice. All subsequent procedures were performed in the cold unless otherwise noted. Four to 8 kg (fresh weight) of liver were passed through a meat grinder, diluted with 4 volumes of Buffer K or Buffer L, strained through a coarse-mesh cloth screen, and homogenized. In the earlier preparations homogenization was performed with a Potter-Elvehjem homogenizer modified so that the ground tissue could be pumped through it continuously. Increased cell breakage and increased final yields of mitochondria were obtained with the use of a high-frequency dispersion device (a Tekmar Company Super Dispax, Model SD-4 5K) . Unbroken cells and nuclei were removed by pumping the homogenate at 880 ml/min through a Vernitron CFR-2 continuous-flow rotor rotating at 11,000 rpm in a Vernitron LCA-2 centrifuge. Mitochondria were harvested from the supernatant by continuous-flow centrifugation in a Beckman JCF-Z rotor at 18,000 rpm and a flow rate of 440 ml/min. After the mitochondria had been washed as described below, they were

PAGE 37

26 resuspended to a concentration of 20 mg protein/ml in buffer and lysed by the addition of non-ionic detergent. The lysate was clarified by centrifugation in a Beckman Type 35 rotor at 28,000 rpm for 10 min. The supernatant (after treatment with DEAE-cellulose, in the case of Method C below) was centrifuged in a Beckman Type 35 rotor at 35,000 rpm for 12 hr. The ribosome pellets obtained at this step are referred to as "crude ribosomes" below. 1. Method A The buffer used for the homogenization of the liver and the preparation of mitochondria was Buffer K. The mitochondria were washed three times by resuspending in Buffer K and centrifuging in a Beckman JA-10 or Sorvall GS-3 rotor at 8,000 rpm for 10 min. Washed mitochondria were suspended in Buffer A containing 34% sucrose (ribonuclease-f ree) , 50 ug/ml heparin and 200 ug/ml oligonucleotides (prepared by partial base hydrolysis of yeast tRNA according to Spencer and Poole (1965)). Triton X-100 and sodium deoxycholate were added to final concentrations of 2% and 0.5% respectively, and crude ribosomes were prepared from the mitochondrial lysate as described above. The ribosome pellets were resuspended in Buffer A containing 50 ug/ml heparin and 550 ug/ml puromycin, and incubated at 37°C for 5 min. The ribosomes were then purified by sedimentation into a linear 10-30% sucrose density gradient made up in Buffer A, in a Beckman SW27 rotor. After centrifugation the gradient was pumped through the flow cell of a Gilford Model 2400 spectrophotometer, and its absorbance at 260 nm was recorded. One-ml fractions were collected.

PAGE 38

27 2. Method B Buffer L was used for the preparation and washing of the mitochondria. After two washes the mitochondria were resuspended to a concentration of 10 mg protein/ml in Buffer L containing 50 pg/ml digitonin. The suspension was stirred for 15 min, and the mitochondria pelleted at 8,000 rpm for 10 min. The mitochondria were washed once more with Buffer L, resuspended to 20 mg protein/ml in Buffer H, and lysed by the addition of Triton X-100 to a concentration of 1.6%. After clarification as described above, the lysate was layered onto 20 ml of Buffer H containing 34% sucrose and 1.6% Triton X-100, and centrifuged to prepare crude ribosomes. Purification of the ribosomes by sucrosedensity-gradient centrifugation was as described above, except that Buffer H was used instead of Buffer A. 3. Method C To the clarified mitochondrial lysate prepared by Method B was added a moist cake of DEAE-cellulose equilibrated in Buffer H containing 1.6% Triton X-100. The quantity of DEAE-cellulose used was about 1 g (dry weight) per 50 ml of lysate. The slurry was stirred for 30 min, placed in a large Buchner funnel, and filtered just until all excess liquid was removed. Ten ml of Buffer H/Triton per gram DEAE-cellulose was added and filtered off as before. The filtrates were discarded and the cake of DEAE-cellulose was stirred for 30 min in Buffer I (10 ml/g) to elute the ribosomes. The slurry was filtered and rinsed with 2.5 ml/g of Buffer I. The ribosomes in the filtrate were then

PAGE 39

28 centrifuged through a layer of Buffer H/l.6% Triton/34% sucrose and purified as in Method B. By this procedure about 10 mg of ribosomes could be obtained from 6 kg of liver. B. Rat Mitochondrial Ribosomes The livers of 20 to 50 young (80-100 g) female Sprague-Dawley rats were homogenized in 4 volumes of Buffer L with a Potter-Elvehjem homogenizer. Rapidly sedimenting material was removed by centrifugation at 3,000 rpm for 10 rain in a Beckman JA-10 rotor, and mitochondria were obtained from the supernatant by centrifugation at 8,000 rpm for 10 min in the same rotor. The mitochondria were washed as described for bovine mitochondria above (Method B) except that the concentration of digitonin used was 17 ug/ml. The washed mitochondria were resuspended to a concentration of 5 mg protein/ml in Buffer H and lysed by the addition of Triton X-100 to 1%. Ribosomes were then prepared and purified as in Method B above. The yield was about 50 yg of ribosomes per rat liver. C. Bovine Cytoplasmic Ribosomes Cytoplasmic ribosomes were prepared from a microsomal fraction of bovine liver. The liver was homogenized in 4 volumes of 25 mM KC1 , 5 mM MgCl ? , 0.34 M sucrose, 5 mM g-mercaptoethanol, 10 mM Tris, pH 7.5. Mitochondria and larger particles were removed at 8,000 rpm for 10 min in a Beckman JA-10 rotor, and microsomes were pelleted from the supernatant at 9,000 rpm for 45 min in the same rotor. The pellet was suspended in Buffer M and Triton X-100 was added to 2%. Ribosomes

PAGE 40

29 were then prepared as described above except that Buffer M was used instead of Buffer A or H. D. E. coli Ribosome s Escherichia coli K-12, strain 1200F" end A 1100 rns A Su" or strain Hfr DIO RNase", were grown in nutrient broth at 37°C. The cells were harvested by centrif ugation at 6,000 rpm for 10 min in a Beckman .IA-10 rotor, suspended in a small volume of cold Buffer N, and ruptured by sonication. After centrifugation in a Beckman Type 65 rotor at ] 5 , 000 rpm for 10 min, the supernatant was centrif uged again at 60,000 rpm for 2 hr in the same rotor to pellet the ribosomes. These wer • then washed once by resuspending and recentrif uging in either Buffer N (low-salt ribosomes) or Buffer (salt-washed ribosomes) . III. EXTRACTION OF RIBOSOMAL PROTEINS Proteins were prepared for electrophoresis by a modification of the method described by Leister and Dawid (1974). For each gel sample, a pellet of ribosomes containing 200-300 ug of protein was suspended in 40 pi of 0.1 M KCl, 10 mM MgCl 2< 40 ul of 10 M urea, 4 M LiCl, HC1, pH 3.5 was added and the mixture was stirred at 5°C for 12 hr. The RNA-containing precipitate was removed by centrifugation in a Beckman Type 65 rotor at 50,000 rpm for 1 hr, and re-extracted by stirring with 80 pi of 6 M urea, 3 M LiCl, HC1, pH 3.5 for 2 hr. The supernatants from the two extractions were combined, and disulfide bonds were reduced by the addition of 40 ul of 7.5 M urea, 0.25 M

PAGE 42

31 to proteins regularly seen by staining were ignored. The electrophoretic positions of the proteins that could not be radiolabelled were determined by interpolation between the nearest spots that were labelled. B. Reductive Methylation A pellet of ribosomes containing 5-10 ug of protein was suspended in 2 ul of 4 M guanidine hydrochloride, 100 mM sodium borate, 10 mM MgCl 2 , 20 mM KC1, 6 mM 3-ME, pH 8.5. Thirty nmol of [ 14 C] formaldehyde (44 Ci/mol) in 1 ul of water were added, and the mixture was incubated on ice for 30 sec. One pi of 30 nmol/ul sodium borohydride was added, and this addition was repeated after one minute. The reaction mixture was left standing in the cold for 5 hr. A resuspended pellet of ribosomes containing 200-300 ug of protein in 40 ul of 0.1 M KC1, 10 mM MgCl„ was then added, and the proteins were extracted for electrophoresis as described above. V. TWO-DIMENSIONAL ELECTROPHORESIS The procedure was modified from that of Leister and Dawid (1974). The firs! dimension was essentially the discontinuous-buffer electrophoretic system of Reisfeld et_ al^. (1962), providing electrophoretic stacking of the protein sample. The composition of the first-dimension separation gel (1.5 mm in diameter, 11 cm long) was 7.5% acrylamide, 0.023% bisacrylamide, 0.063% TEMED, 8 M urea, 60 mM potassium acetate, pH 4.3. It was polymerized by adding ammonium persulfate to a concentration of 0.1%, and pre-electrophoresed before use for 12 hr at 0.2 ma

PAGE 43

32 per gel, in a tank buffer of 0.01% aminoethanethiol, 8 M urea, 60 mM potassium acetate, pH 4.3. The stacking gel (2 cm long) was 3.86% acrylamide, 0.14% bisacrylamide, 0.05% TEMED, 8 M urea, 60 mM potassium acetate, pH 6.7, polymerized by the addition of ammonium persulfate to 0.02% and riboflavin to 0.001%. The tank buffer was 0.01% aminoethanethiol, 35 mM g-alanine acetate, pH 5.0. The dialyzed protein samples (about 200 yl) were loaded onto the gels and . iectrophoresed at 0.1 ma per gel. When the tracking dye (0.001% Pyronin Y in 200 pi of Sample Buffer) in a parallel gel tube had reached the top of the separation gel, the current was increased to 0.2 ma per gel. Electrophoresis was continued until the tracking dye reached the bottom of the gel. The gels were extruded from the glass tubes with a 6-inch, 22 gauge needle through which water was flowing under pressure. The second-dimension gel slabs were 18 cm high, 20 cm wide, and 1.5 mm thick. Each gel was poured between a pair of glass plates separated by two 1.5 mm-thick plexiglass strips, sealed at the bottom and sides with a piece of silicone-rubber tubing, and held together with spring clamps. After the gel was polymerized, the silicone tubing was removed. The gel composition was 9.65% acrylamide, 0.35% bisacrylamide, 0.1% TEMED, 5 M urea, 0.5% SDS, 0.1 M sodium phosphate, pH 7.2, polymerized with ammonium persulfate (0.025%). The extruded first-dimension gels were laid on top of the seconddimension gel slabs and overlaid with 0.05% mercaptoacetic acid, 5 M

PAGE 44

33 urea, 1% SDS. At each end of the first -dimension gel was placed a small piece of agarose (1.5 mm in diameter, 5 mm long) containing several marker proteins. The composition of the marker mixture was 0.5 mg/ml bovine serum albumin, 1 mg/ml human y-globulin, 0.4 mg/ml ovalbumin, 0.33 mg/ml equine myoglobin, 0.4 mg/ml egg-white lysozyme, 0.33% SDS, 0.33% B-ME, 0.016% mercaptoacetic acid, 2% agarose. Electrophoresis was at 30 ma/gel with a tank buffer of 0.024% mercaptoacetic acid, 0.5% SDS, 0.1 M sodium phosphate, pH 7.2, until the dye marker (Bromophenol Blue) reached the bottom of the gel. The total time for electrophoresis in both dimensions was about 24 hr. After electrophoresis the gels were soaked in the following solutions with continuous agitation: 25% isopropanol, 10% acetic acid (18 hr); 0.25% Coomassie Brilliant Blue R, 50% ethanol, 7.5% acetic acid (6 hr) ; and 5% ethanol, 10% acetic acid (several changes, until the gels were adequately destained) . For convenient storage ami ! or autoradiography, the gels were soaked in 3% glycerol for 30 min and then dried onto Whatman //3MM filter under vacuum on a steam bath, by the method of Maizel (1971). For autoradiography the dri A gel was placed in contact with Kodak RP14 medical X-ray film in an X-ray film cassette. Two small holes had previously been drilled through the cassette. After the gel and film were loaded and the cassette was closed, a needle was inserted through these holes to mark the alignment of the gel with the film.

PAGE 45

34 Gels containing reductively methylated proteins were prepared for autoradiography as described by Bonner and Laskey (1974) . They were shaken in two changes of dimethyl sulfoxide, for 30 min each time, then in 4 volumes of a 20% solution of 2,5-diphenyloxazole in dimethylsulfoxide for 3 hr, and in water for 1 hr. The gels were then dried onto filter paper as described above. The film used for autoradiography was Kodak RP/R-54 medical X-ray film, pre-exposed with a flash from a photographic strobe light according to Laskey and Mills (1975). Exposure at -70°C for 5 weeks was sufficient for gels containing 10,000 cpm of reductively methylated protein. VI. ISOELECTRIC FOCUSING A previously published procedure (Czempiel et_ al_. , 1976; Klose, 1975) was followed closely. The gel composition was 4.8% acrylamide, 0.2% bisacrylamide, 0.065% TEMED, 1.0% Ampholine (pH 3.5 10), 5% sucrose, 8 M urea, polymerized by the addition of ammonium persulfate to 0.019%. The gels were 0.4 cm in diameter and 7 cm long. The upper tank buffer was 5% phosphoric acid and the lower tank buffer was 5% ethylenediamine, with the anode in the upper tank. The protein sample for each gel was dissolved in 20A of 8 M urea and then mixed with 30A of Sephadex G-200 (superfine) swollen in 20% sucrose, 8 M urea, 10% B-ME, 1% Ampholine (pH 3.5 10). The details of the procedure listed above differed from those of the published procedure only in three minor respects. In the latter procedure the ammonium persulfate concentration was slightly greater

PAGE 46

35 (0.023%), the gel was longer (8 cm), and the upper tank buffer contained urea at a concentration of 3 M. Besides these insignificant differences, there was one further modification which was significant. Czempiel e_t al . (1976) used an Ortec 4100 Pulsed Constant Power Supply, which has power output characteristics different from those of the constant-voltage power supply used in the present experiments. The published procedure involved a program of increasing voltage (50V for 1 hr, 100V for 1 hr, 150V for 1 hr, 200V for 2 hr, 300V for 2 rain, and 400V for 2 min). In the present experiments several parallel samples were electrofocused according to this schedule but for different lengths of time. The first sample was run for a total of 3 hr (the first 3 hr of the schedule), the second sample for the whole 5 hr of the schedule, and the third sample for the whole 5 hr plus an additional 2 hr at 200V. After electrof ocusing, the gels were extruded and subjected to a second dimension of electrophoresis in SDS as described in Section V. VII. ASSAYS OF RIBOSOME FUNCTIONS Peptidyl transferase activity was assayed by the modified fragment reaction (Denslow and O'Brien, 1974; de Vries et_ al_. , 1971). The reaction mixture contained 0.1 mg ribosomes, 83 hM (10,000 cpm) N-acetyl[ 3 H] leucyl-tRNA, 0.66 mM puromycin, 267 mM KC1, 13.3 mM MgAc ? , 33% ethanol, 33 mM Tris-HCl, pH 7.5 in a total volume of 0.15 ml. After incubation at 25°C for 10 min, KOH was added to 0.6 M and the mixture 3 warmed to 40°C for 3 min. The N-acetyl-[ H] leucyl-puromycin synthesized in the reaction was extracted into 1.5 ml of ethyl acetate, and

PAGE 47

36 radioactivity was determined by liquid scintillation counting in Triton X-100/toluene (1:1) containing 0.57. PPO and 0.05% POPOP. 3 [ H] GTP binding was measured by the Millipore filter assay of Bodley et_ al. (1970). Fifty ul of reaction mixture containing 15 yg of ribosomes, 42 pmol (0.5 uCi) of [83 H] GTP, 10 mM NH CI, 20 mM MgAc ? , 5 mM (2-ME, 10 mM Tris-HCl, pH 7.4 was incubated at 0°C for 5 min and then filtered thru a Millipore filter. The filter was washed with 10 mM NH CI, 10 mM MgAc 2 , 10 niM Tris-HCl, pH 7.4, and the ribosome-bound radioactivity determined by liquid scintillation counting. Poly li-dependent polypheny lalanine synthesis was assayed according to Hosokawa et al. (19o6) . The composition of the 0.25 ml reaction mixture was 0.64 mg/ml poly U, 5.4 uM [ II] phenylalanine (1.82 Ci/mmol) , 25 mM tyrosine, 50 mM of each of the other 18 amino acids, 0.5 mg/ml tRNA, 32 uM GTP, i mM ATP, 5 mM phosphoenolpyruvate, 0.1 mg/ml pyruvate kinase, 1 mg/ml E_. coli factors, 0.4 mg/ml ribosomes, 50 mM KC1, 20 mM MgAc , 6 mM g-ME, 1 mM dithiothreitol, 10 mM Tris-HCl, pH 7.8. Aliquots of 50 pi were withdrawn at 5 min intervals, and radioactivity insoluble in hot trichloroacetic acid was determined by the method of Mans and Novelli (1960). VIII. B UOYANT DENSITY DE T ERMINATIONS Ribosomes were dialyzed in 50 mM KC1, 5 mM MgCl„, 20 mM TEA, pH 7.5, and then fixed by the addition of formaldehyde to a concentration of 5%. They were analyzed by equilibrium centrifugation in gradients of CsCl containing 50 mM KC1, 5 mM MgCl, 20 mM TEA, pH 7.5, 0.3%

PAGE 48

37 formaldehyde in a Beckman SW39 rotor, as described by Brunk and Leick (1969). The gradients were pumped through the flow cell of an Isco UA-2 absorbance monitor, and 0.25 ml fractions were collected. The refractive index of each fraction was measured with a Bausch and Lomb refractometer, and used to calculate the concentration of CsCl and the buoyant density in the fractions. I X . QUANTITATIVE MEASUREMENT S Quantities of mitochondria were estimated by absorbance at 550 ran. Samples of mitochondria of known 550 nm absorbance were analyzed by the method of Lowry et al. (1951), to establish a calibration curve relating the absorbance at this wavelength to the concentration of mitochondrial protein in mg/ml. Ribosome quantities were determined by absorbance at 260 nm, usinj extinction coefficients (E, ) of 110 for mitochondrial ribosomes, 1 cm 135 for cytoplasmic ribosomes, and 160 for E. coli ribosomes.

PAGE 49

RESULTS The purpose of this research was to characterize the proteins of mammalian mitochondrial ribosomes, especially in comparison with the proteins of other kinds of ribosomes. The major characteristics to be investigated were the number of ribosomal proteins, their molecular weights and their electrophoretic properties. For this purpose it was necessary to establish and apply a number of criteria to distinguish the ribosomal proteins from the contaminating non-ribosomal proteins which might be present in the mitochondrial ribosome preparation. The results will be presented in five sections. First, the means of preparing mitochondrial ribosomes in adequate quantity and purity for these experiments will be described. The two-dimensional electrophoretic patterns of bovine mitochondrial ribosomal proteins will be presented and analyzed in terms of criteria for the identification of the proteins which are most likely to be true components of the ribosome in vivo; the effect of the ribosome preparation procedure (particularly the effect of high salt treatment) on the electrophoretic pattern of ribosomal proteins will be evaluated, with a view to the possibility of various kinds of artifacts. The electrophoretic properties of the proteins will be compared with ttiose of mammalian cytoplasmic and bacterial ribosomal proteins. The number and sizes of the mitochondrial 38

PAGE 50

39 ribosomal proteins will be used to estimate the total mass of protein in the ribosome. Finally, the proteins of bovine mitochondrial ribosomes will be compared with those of other mammalian species to assess the level of evolutionary divergence in these proteins. I . PREPARATION AND CHARACTE R IZATION OF MITOCHONDRIAL R1B0S0MES At the time this research was begun, the method used for the preparation of mitochondrial ribosomes (Method A in MATERIALS AND METHODS) was not capable of producing adequate quantities of material for the electrophoretic studies to be described below. Three separate preparations were required to produce enough ribosomes for a single electrophoretic analysis, each preparation being a week-long procedure. Moreover, the crude mitochondrial ribosome preparation was heavily contaminated with cytoplasmic ribosomes, as evidenced by the large amount of material sedimenting at 80S when the ribosomes were subjected to sucrose density gradient centrifugation (Figure 2A) . The presence of such quantities of 80S ribosomes suggested that even the 55S mitochondrial ribosome region of the gradient contained significant amounts of cytoplasmic ribosomal subunits, as well as possibly other unknown structures with similar sedimentation coefficients. For this reason the intact 55S ribosomes were not suitable as samples for electrophoretic analysis. Rather, a further purification step was performed by pooling the material in the 55S peak, treating it with a higher concentration of KC1 and a lower concentration of MgCl ? to dissociate the mitochondrial

PAGE 51

Figure 2. Preparation of mitochondrial ribosomes by Method A. (A) Crude mitochondrial ribosomes were prepared from bovine liver according to Method A in MATERIALS AND METHODS. The ribosomes were suspended in 2 ml. of Buffer A and layered onto a linear 10-30% sucrose density gradient made up in Buffer A. After centrifugalion in a Beckman SW27 rotor at 27,000 rpm for 5 hours, the gradient was pumped through the flow cell of a spectrophotometer and its absorbance at 260 nm was recorded. One-ml fractions were collected. The direction of sedimentation is from left to right. (B) Fractions corresponding to the 55S absorbance peak, of the sucrose density gradient shown in Figure 2A were pooled and centrifuged in a Beckman Type 65 rotor at 65,000 rpm for 3 hours. The pellet of mitociiondrial ribosomes was suspended in 2 ml of Buffer E and layered onto a sucrose density gradient made up in the same buffer. Centrifugation was at 20,000 rpm for 13.5 hours.

PAGE 52

41 10 20 Fraction Number 30

PAGE 53

42 ribosomes to their subunits, and re-ctntrifuging on a second sucrose density gradient. The only contaminants in the preparation of subunits obtained in this manner, besides those that might be bound directly to the subunits, would be those which sedimented near 55S under the ionic conditions of the first centrifugation and in the 25S to 45S region under the second conditions. The kinds of possible contaminants which might behave in this manner seem intuitively to be few, and in particular do not include cytoplasmic ribosomal subunits. In fact, the discrete peaks of mitochondrial small and large ribosomal subunits were found to be by far the major species visible on the absorbance profile of the second centrifugation (Figure 2B) , suggesting that the preparation was indeed reasonably free of co-sedimenting material. Later improvements in the procedure increased both the yield and the purity of the mitochondrial ribosomes. The yield was approximately tripled by the use of a more efficient homogenizer for the disruption of the cells. Treatment of the mitochondrial preparation with digitonin (Method B) dramatically reduced the quantity of cytoplasmic ribosomes in the crudemitochondrial ribosome preparation (Figure 3). This detergent has been used previously to remove cytoplasmic ribosomal RNA (Malkin, 1971) and cytoplasmic ribosomes (de Vries and van der KooghSchuuring, 1973) from rat liver mitochondrial preparations. In addition, treatment under these conditions has been reported to solubilize latent lysosomal enzymes (Schnaitman and Greenawalt, 1968; Lowenstein e_t a_l . , 1970), thereby diminishing the possibility of degradation of the mitochondrial ribosomes during the preparation.

PAGE 54

43 2.0 £ c O CD C\J 0) o c a .a k. o < 0.5 10 15 Fraction Number 20 Figure 3. Preparation of mitochondrial ribosomes by Method B. Crude bovine mitochondrial ribosomes obtained by Method B (MATERIALS AND METHODS) were suspended in Buffer H and analyzed by sucrose density gradient centrifugation in this buffer as described in Figure 2A, except that centrifugation was for 4 hours at 27,000 rpm. (No differences have been observed between Buffer H and Buffer A with respect to their effects on the mitochondrial ribosomes. )

PAGE 55

44 At this point the maximum yield of mitochondrial ribosomes from a single preparation was limited by the volume of the largest preparative ultracentrifuge rotors available. After preparation and lysis of the mitochondria, the best means available for concentrating the ribosomes for purification on sucrose density gradients was to centrifuge them to a pellet at about 100,000xg. The volume of an average yield of mitochondria (about 35g protein), suspended in buffer to a concentration low enough for efficient detergent lysis, is far greater than the capacity of three Beckman Type 35 rotors. Therefore, a method was developed to preconcentrate the ribosomes before harvesting them by centrifugation (Method C) . The mitochondrial lysate was stirred with DEAE-cellulose as described in MATERIALS AND METHODS to adsorb out the mitochondrial ribosomes, and then the ribosomes were eluted by stirring with a smaller volume of buffer at a higher ionic strength. This procedure is very rapid and effects a fourfold concentration of the ribosomes, thereby quadrupling the yield of each preparation. Furthermore, any remaining traces of cytoplasmic ribosomes, as well as a considerable proportion of the other contaminants of the mitochondrial ribosome preparation, are removed by this means. An assessment of the degree of purification achieved by this method is shown in Figure 4. A sample of purified, digitonin-treated mitochondria was split into two equal aliquots. Ribosomes were prepared from the first aliquot without DEAE-cellulose treatment and analyzed by

PAGE 56

45 10 20 Fraction Numbe 10 20 Fraction Number Figure 4. Preparation of mitochondrial ribosomes by Method C. All samples were analyzed by sucrose density gradient centrifugation in Buffer H. (A) Ribosomes from 1.4 g (protein) of bovine liver mitochondria prepared by Method B. The 55S absorbance peak contains 5.7 A„,_ units , ., ZoU of rxbosomes. (B) Cytoplasmic ribosomes from 5 ml of a preparation of bovine liver microsomes. The 80S peak contains 5.6 A , units of ribosomes. (C) 1.4 g of mitochondria and 5 ml of microsomes in Buffer H were mixed together, lysed with Triton X-100, and stirred with 2 g (dry weight) of DEAE-cellulose equilibrated in Buffer H. The slurry was filtered and the filtrate discarded. Mitochondrial ribosomes were then eluted by stirring the DEAE-cellulose with Buffer I, and were prepared for sucrose density gradient analysis according to Method B. The 55S peak contains 6.1 A„,„ units of ribosomes. (D) The once-eluted DEAE-cellulose from (C) was stirred with Buffer J to elute the cytoplasmic ribosomes. The 80S peak contains 4.6 A„, units, or 82% of the input cytoplasmic ribosomes.

PAGE 57

46 sucrose density gradient centrifuv.it ion, giving the absorbance profile shown in Figure 4A. A preparation of bovine cytoplasmic ribosomes (having the sucrose density gradient profile shown in Figure 4B) was added to the second aliquot of mitochondria, to provide a more stringent test of the method. The mitochondria were then lysed and the ribosomes prepared by the DEAE-cellulose procedure described above. As seen in Figure 4C, the product consists primarily of mitochondrial ribosomes and subunits, much of the slowly-sedimenting material and all detectable SOS cytoplasmic ribosomes having been removed. Probably because of the larger proportion of rRNA in cytoplasmic ribosomes and their consequent higher density of negative charges, these particles remain bound to the DEAE-cellulose under the conditions used for elution of the mitochondrial ribosomes, as shown by re-elut ion of the DEAEcellulose with a buffer of higher ionic strength (Figure 4D) . As may be seen in Figure 4, the yield of mitochondrial ribosomes per gram of mitochondria obtained with the DEAE-cellulose procedure is as great as or slightly greater than that obtained without this step. Since this method permits the processing of a four-fold greater quantity of mitochondria, and gives a cleaner preparation of ribosomes as well, it is clearly the method of choice unless it adversely affects the structural or functional integrity of the ribosomes. Of particular concern is thi j possibility that some of the more acidic ribosomal proteins might bind more strongly to the DEAE-cellulose than to the ribosome itself and thus might be stripped off when the ribosomes are eluted.

PAGE 58

47 The most acidic proteins of E. coli ribosomes, L7 and L12, have been found to be relatively loosely bound to the ribosomes (Hamel et al. , 1972). To answer this question, the ribosomes were characterized by a number of functional and physical tests. The peptidyl transferase and GTP-binding activities of the ribosome were essentially unaltered by treatment with DEAE-cellulose (Table V). DEAE-cellulose-prepared ribosomes were also found to be active in the translation of poly U (400 pmoles phe incorporated/mg rRNA/15 min) . The sediment :tion coefficients and buoyant densities of mitochondrial ribosomes and the subunits derived from them were essentially the same as those obtained from previous preparations. The effect of this procedure on the two-dimensional electrophoretic pattern of the ribosomal proteins will be evaluated in Section IIC below. II. ELECTROPHORETIC ANALYSIS OF MITOCHONDRIAL RIBOSOMAL PROTEINS The very high proportion of protein which physical-chemical measurements have shown to be present in bovine mitochondrial ribosomes (see INTRODUCTION) predicts that a rather large number of individual proteins are to be found in these ribosomes. If the 1.8 x 10 6 d. of total protein per particle were made up of proteins similar in size to _E. coli ribosomal proteins (average molecular weight about 17,000), for example, each ribosome would contain more than a hundred protein molecules. For this reason it seemed unlikely that electrophoresis in one dimension would separate all the proteins sufficiently for the purposes

PAGE 59

48 Tab! Preparation V. Functional activity of mitochondrial ribosomes prepared with or without DEAE-cellulose. •%-Leu-Puroiuyc in Synthesis (cpm/nmole ribs/10') J H-GTP Binding (moles GTP/mole ribs) No DEAE DEAE 33,600 36,200 1.12 1.09

PAGE 60

49 of this study. Several two-dimensional electrophoret ic systems have been devised to provide adequate resolution for the analysis of ribosomal proteins. The system used for most of the experiments to be described below was that of Leister and Dawid (1974), employing a separation partially on the basis of charge at pH 4.3 in the first dimension and on the basis of size in the second (SDS) dimension. For the purposes of this research, this electrophoretic system has several advantages over the one that is most widely used for ribosomal protein studies at present (Kaltschmidt and Wittmann, 1970). It permits a direct determination of the molecular weights of the proteins. It is more sensitive, requiring only about 5 ug of each protein. And it has been used for the analysis of Xenopus mitochondrial ribosomal proteins (Leister and Dawid, 1974), making it possible to compare these published results with those obtained in the present experiments on mammalian mitochondrial ribosomes. The results of one of the early electroohoretic analyses (shown in Figures 5 and 6) amp] v confirmed the expectation that a large number of proteins would be found in mitochondrial ribosomes. 40 to 50 reasonably intense and distinct spots may be discerned in the photograph of the electrophoretic pattern of proteins from each subunit. Somewhat larger numbers could be seen on the original gels, since some of the resolution of closely adjoining spots has been lost in the photographs. It is evident in these figures that the descripi 'on "reasonably intense" involves a rather arbitrary judgment. In fact, particularly

PAGE 61

50 • ^ .* Ifc* Figure 5. Electrophoretic pattern of large-subunit proteins from bovine mitochondrial ribosomes. Mitochondrial ribosomes were dissociated to subunits by sucrose density gradient centrifugation in Buffer E. Gradient fractions corresponding to the large (39S) subunits were pooled and centrifuged in a Beckman Type 65 rotor at 65,000 rpm for 5.5 hours. The proteins were extracted from the ribosome pellet and analyzed by two-dimensional polyacrylamide gel electrophoresis. The first dimension (left to right) was run in urea at pH 4.3, and the second dimension (top to bottom) in SDS. Just before the second dimension was begun, pieces of agarose gel containing a mixture of several proteins (bovine serum albumin, human y-globulin heavy and light chains, ovalbumin, myoglobin, and lysozyme) were placed on top of the gel slab, immediately adjacent to each end of the first-dimension gel. These proteins may be seen at the left side of the photograph.

PAGE 62

51 I * Figure 6. Electrophoretic pattern of small-subunit proteins from bovine mitochondrial ribosomes. Small (28S) subunits were prepared by sucrose density gradient centrifugation in Buffer E, and their proteins were subjected to twodimensional electrophoresis as described in Figure 5.

PAGE 63

52 in the small-subunit pattern, there is a large range of spot intensities with a more or less continuous variation from the most intense tc the very faintest. This result was not what had been anticipated. Since nearly all ribosomal proteins probably are present in one copy per ribosome, at least in vivo (Hardy, 1975), it had been expected that the gels would show a discrete group of ribosomal protein spots of roughly similar staining intensities. In addition there might be some other spots of varying intensity, corresponding to contaminating proteins. The results obtained indicated that there were considerable differences in the relative quantities of the ribosomal proteins present in the sample, making it impossible to distinguish them from contaminating proteins on the basis of their staining intensity. Some possible explanations for this phenomenon will be considered below (Section IV) . A. Criteria for the Identification of Ribosomal Proteins For this reason it seemed necessary to seek other criteria for the purpose of identifying which of the spots appearing on the gels actually represented ribosomal proteins. Several possible criteria were considered, including reproducibility of occurrence, subunit-specif ic localization, and resistance to removal from the ribosomes by high salt treatment. It seemed reasonable to expect that the ribosomal proteins would be fcund reproducibly in separate preparations of the ribosomes, whereas the non-ribosomal contaminants might be more variable in occurrence. In fact, it was found that many of the proteins seen in individual

PAGE 64

53 gels such as those of Figures 5 and 6, particularly some of the fainter ones, could be disregarded on the basis that they were not seen in the majority of the electrophoretic analyses. This rule was thus found to be a useful one, and the application of it will be described below. The idea that some contamination might arise from the non-specific binding of some proteins to the ribosomes suggested that contaminants of this type might be found nearly equally in both ribosomal subunits. However, the existence of some plausible mechanisms by which ribosomal proteins might also show an apparent lack of subunit-specif ic localization (Section IIB3) made the application of this criterion difficult, and it was used only in conjunction with other kinds of evidence. Yet another way to discriminate the ribosomal and non-ribosoma] proteins, and the most unambiguous way, would be to purify all the proteins and then reconstitute the riboscme, withholding each protein in turn, to establish which proteins were essential for the generation of a structurally and functionally normal ribosome. Unfortunately, this approach is not practical in the case of ribosomes such as those of mammalian mitochondria, which can be obtained only in relatively small quantities. However the observations (Olsnes, 1971; Kurland, 1966; Hardy and Kurland, 1966) that contaminating proteins are in general more loosely bound to ribosomes than are the ribosomal proteins themselves, suggested that an analogous approach might be feasible. Specifically, it has been found that treatment of ribosomes with

PAGE 65

54 moderately high concentrations (about 0.5 M) of monovalent salts dislodges adventitiously bound proteins from the particles, while affecting the binding of the ribosomal proteins to a lesser extent. Of course, higher salt concentrations do remove ribosomal proteins (Delaunay et a.l . , 1974; Gesteland and Staehelin, 1967) and the optimum concentration range, that which removes the largest quantity of contaminants and the smallest amount of ribosomal protein, may differ from one type of ribosome to another. The effect of increasing salt concentration on the content of individual proteins in the bovine mitochondrial ribosome was assessed by centrifuging the ribosomes through various high-salt buffers and then analyzing the proteins remaining in the treated particles by two-dimensional electrophoresis. To establish the point at which these treatments began to remove ribosomal proteins, the peptidyl transferase activity of the ribosomes prepared under each condition was assayed. The electrophoretic protein patterns and specific activities obtained were then correlated, and any proteins which were found to be removed from the ribosome by a given treatment, without diminishing its functional integrity, were tentatively regarded as non-ribosomal contaminants. Since the peptidyl transferase activity is a property of the large subunit and is not dependent on the presence of the small subunit, this criterion was directly applicable only to the large-subunit proteins. Figure 7 shows the effect of treatment with buffers containing increasing ratios of KC1 concentration to MgClconcentration on the

PAGE 66

55 100 ""1 ' I ' Ml| 1 1 I I I I IT "I I — I I I I ! ™ % "S 50 1.50 145 1.40 t I I I | \\ I H 1 — I I I I I l| CD E ^r^-^-i r-n*y* -+ 1 — i i i i i+ 1 + ' ' ' U 1 1 ' I I I I ll l i i M | 5 10 50 100 ,000 I I I I I J 10,000 KCI/MgCI 2 Figure 7. Effects of treatment with buffers of various ionic compositions on the peptidyl transferase activity and buoyant density of bovine mitochondrial large subunits. Large subunits were prepared by sucrose density gradient centrifugation in various buffers. Peptidyl transferase activity was assayed as described in Materials and Methods , and expressed as a percentage of the activity found in Buffer A treated subunits. The buoyant densities of the formaldehyde-fixed subunits in CsCl were determined by isopycnic centrifugation. The buffer conditions used are plotted along the horizontal axis accordxng to their ratio of KC1 to MgClo concentrations. The actual concentrations of these two salts were: Buffer A C 1J E F G KC1(M)

PAGE 67

56 activity and buoyant density of the large subunit. In this series, buffers containing increasing concentration:, of KC1 and decreasing concentrations of MgCl~ were used to wash the ribosomes with increasing stringency. For other ribosomes monovalent and divalent cations have been found to produce opposing and competitive effects on the structure of the particle, with respect to the association of the two subunits with each other (Hamilton and Petermann, 19 59; Zitomer and Flaks, 1 L < ' I) and the binding of proteins to the ribosorr.e (SpitnikElson and Atsmon, 1969; Staehelin et al ., 1969). Such effects on subunit association have also been found in mammalian mitochondrial ribosomes (O'Brien, 1971), and Figure 7 indicates that either increasing KC1 or decreasing MgCl ? can diminish both the functional activity and the protein content (as measured by the increase in buoyant density) of these ribosomes. The upper curve of Figure 7 shows that the large subunit retained most of its peptidyl transferase act ivity after treatment with either Buffer C or Buffer D. The activity was significantly reduced by treatment with Buffer E, and higher KCl/MgCl 2 ratios yielded particles with negligible activity. Therefore in the analysis of the electrophoretic results presented below, proteins which were found to be absent from large subunits treated with either Buffer C or Buffer D are considered to be dispensable for this particular large-subunit function. These proteins are regarded as possible non-r ibosomal contaminants, and are discussed separately.

PAGE 68

57 B. Identification of the Ribosomal Proteins 1. Large Subunit In order to apply the criteria of reproducibility and resistance to high-salt treatment, several electrophoretic analyses were performed on large-subunit proteins taken from separate preparations of ribosomes, after treatment with various high-salt buffers. The gels were scored for the presence or absence of individual proteins, and the results are given in Table VT . The electrophoretic positions of the proteins designated in the tablt! are shown in Figure 8. The 52 proteins which were seen reproducibly and which were present both in large subunits treated with Buffer C and in those treated with Buffer D are shown as numbered spots. Lettered spots represent some of the proteins which did not satisfy these criteria. For the purposes of Table VI the relatively subjective question of the intensities of the protein spots was not considered, and a "+" indicates only that the protein was clearly visible in the gel, whether it was very intense or quite faint. Proteins which were near the limit of visible detection, however, are marked with a "-"; some of these are not visible in the photographs of the gels. In a few cases ("nr") the electrophoretic resolution in a given experiment was inadequate to determine whether a protein was present or not. The three proteins which move fastest in the first dimension of electrophoresis (ML45, ML51, and ML52) could have been run off the end of the first-dimension gel in some experiments, and these cases are marked "ro".

PAGE 69

58 en

PAGE 70

59 i i + + + + + i ++ i + + + + + + + + + + + + + + I I i i i o + o + + + + + + + I + + + + + I + + + + + + I + O O + I O + I + I CC++I+ + + + + + + + + + + + + + +HW + +O++ C+ + + + + + + + + OO + + + + + + + + O + + | + + i + + O O O + O I + + + + + + + + + + + + + + + + + M++0 I + u u + +0 + + +0 + I + + + + + I + + I I + + + + + + + I I O I O I O I o + + + + + + + I + + + + + I + + I 4+ + + + I + O O I o o c + o I ++++++ I ++ I ++ I +++++++++++ I O I o o o + o i i + + + + + oc + o i + + + + + + + + + cc; I O I o + ODt^OHMnsfu-ivOi^aDi^OHMrOsrin CO J3 O T3
PAGE 71

60 2-D 684 423.517.2. 14.4-D — > 4

PAGE 72

61 a . Reproducibilit y In order to discriminate ribosomal proteins from contaminants, an arbitrary level of reproducibility of 50% was taken as the cut-off. That is, a protein was considered to be non-ribosomal. unless it was present in over half of the ribosome preparations analyzed. In fact, however, very few of the proteins in large-subunit preparations were near this borderline. Most of the proteins which were found in reasonable quantity, and many of the faint ones, in any one geJ proved to be quite reproducible. A small number of proteins in the unreproducible group were seen so rarely that it appears most unlikely that they represent ribosomal proteins: the spot seen between ML16 and ML17 in Figure 5, for example, was not found in any other experiment. Such proteins are not listed in Table VI. The proteins designated by letters in the table were seen more frequently, especially in ribosomes prepared in certain buffers. Because the presence or absence of these proteins was correlated with the ionic conditions with which the ribosomes had been treated, they are considered below under "Resistance to salt-washing" . Among the numbered proteins in Table VI , many appeared in every ribosome preparation, and with reasonable staining intensity. Some faint spots, like ML29 and ML41, were likewise very regular in occurrence although sometimes present only in trace quantities. Occasionally a protein which was usually present in moderate quantity would be unaccountably lacking in a particular experiment: MLt7, for example,

PAGE 73

62 can be clearly seen in experiments 2 (Figure 9), 4 (Figure 10), and 6 (Figure 11) and is clearly absent from experiment 8 (Figure 5) . Similarly, ML1 and ML2 are always found to have about the same staining intensity seen in Figures 5, 9, and 10, except that in experiment 6 (Figure 11) they were both reduced to barely detectable traces. Although such results are quite surprising, the overall reproducibility of these proteins was considered adequate to satisfy the criterion. Predictably, the greatest variability was found among the faintest spots, including the only two real borderline cases. ML48 was seen with low but clearly visible intensity in three experiments (as in Figure 9), as only a trace on three other occasions, and three times not at all. ML45 was found even less often, again was sometimes only a trace when it was present, but was sometimes quite intense (Figure 9). Overall, relatively little variability was observed in the pattern of large-subunit proteins from one experiment to anottier. Most of the proteins seen on any one gel (including the faintly-staining ones) were regularly found in separate preparations of ribosomes, and thus probably represent ribosomal proteins, according to this criterion. Some possible explanations for those variations that were obtained in these experiments will be discussed below in the analysis of the small-subunit proteins, for which variability is a more serious prob1 em.

PAGE 74

63 ~% ' t Figure 9. Electrophoretic pattern of proteins from large subunits prepared in Buffer C (Experiment 2).

PAGE 75

64 ». I Figure 10. Electrophoretic pattern of proteins from large subunits prepared in Buffer C (Experiment 4).

PAGE 76

65 Figure 11. Electrophoretic pattern of proteins from large subunits prepared in Buffer D (Experiment 6).

PAGE 77

66 b . Re sistance to salt-washing The salt-washing procedure used in the preparation of mitochondrial ribosomes was designed to remove two types of contaminants, those which were bound directly to the ribosomes and those which were not bound, but nonetheless accompanied the ribosomes through the early stages of the preparation. As mentioned in Section IIA, treatment of ribosomes with high c 'iicentrations of monovalent salts should preferentially release non-ribosomal material bound to them. Figure 12 shows the effect of washing E. coli ribosomes with 1 M NH CI . It is evident that the major difference between the two protein patterns is the presence of many high-molecular-weight proteins in the unwashed ribosomes. This phenomenon is well-documented in the literature (Subramanian, 1974; Hardy, 1975; Brouwer and Planta, 1975), and the proteins removed by salt-washing are universally regarded as non-ribosomal for a number of reasons, including the essentially undiminished functional activity of the salt-washed ribosomes. The fact that most high-salt buffers also dissociate mitochondrial ribosomes to their subunits permits the removal of non-ribosome-bound , co-sedimenting contaminants by the use of two cycles of sucrose-densitygradient centrifugat ion under different ionic conditions. As discussed above (Section I) the isolation of the ribosomes first at low ionic strength (0.1 M KCl) as intact 55S particles, and then a higher salt concentrations as subunits should separate them from most non-ribosomebound contaminants.

PAGE 78

67 • • •tt .; 3 w| cu d a) 4-4 -H ,£ U (DO) O H 4-1 O O O CU O • !-i vO 4-1 > 0,-^s 4-1 U H CO 0) O n-i (!) M-| 1 i i .

PAGE 79

68 Low-salt treatment . The effectiveness of this procedure can be evaluated by comparing subunits prepared in this way ("derived subunits") with the subunits which are present in small amounts in the first, lowsalt sucrose density gradient ("native subunits"; see Figure 3). Figure 13 shows the electrophoret ic patterns obtained from native large and small subunits, and Figure 14 shows the pattern of the derived large subunit from the same preparation of rihosomes. The most obvious difference is the presence of two high-molecular-weight proteins in large quantity in the native subunits. Several other proteins of similar and higher molecular weights are also present in native but not in derived subunits; many of these, like the two major proteins, are found equally in both the large and small subunits. Another native-subunitspecific protein, seen to the left of ML30, is also fcund in somewhat smaller quantity in the small subunit. Additionally, numerous proteins of various molecular weights and staining intensities appear in the native large subunit but not in either of the other two patterns. It seems reasonable to conclude that most if not all of the proteins found in native but not derived subunits represent non-r ibosomal contaminants. The similarity of these results to those obtained with E. coli rihosomes (Figure 12) is clear, particularly with respect to the highmolecular-weight proteins. The lack of subunit-specif icity of many of these proteins is also incriminating. Most importantly, they are essentially absent from large subunits derived in Buffer C or Buffer D, and such subunits are functionally active. By visual comparison of the

PAGE 80

69 CI | «w # si d ffl * • t t • a s: o en •H (J U -H u o a en (j — h cn u u a a u -j C Vj cj II -H [J HI — I 0) Q, —I

PAGE 81

70 • Figure 14. Electrophoretic pattern of proteins from large subunils prepared in Buffer C (Experiment 3) . The irregular white line in the upper left is a crack in the gel.

PAGE 82

71 relative staining intensities of these proteins in native and derived subunits most of them appear to be reduced in quantity by 90% or more in derived subunits. Yet the activity of Buffer C derived subunits is 75% of that found in native subunits (shown as Buffer A subunits in Figure 7). Thus none of these proteins can be essential for the peptidyl transferase activity of the ribosome. The kinds of contaminants which have thus been identified are various, although some loose generalizations can be made about them. The contaminants may be present in a very wide range of staining intensities, have molecular weights ranging from values similar to those of the ribosomal proteins to considerably larger values, and may be present subunit-specifically or not. However, a major part of the contamination seems to be represented by a group of relatively intense, high-molecular-weight, non-subuni t-spec i f ic proteins. It is of some interest to ask whether this contamination is due primarily to adherent proteins or to protein-containing structures that are not bound to the ribosomes but are large enough to sediment in the same region of the sucrose density gradient. The buoyant density of the native subunit is significantly lower than that of the Buffer C or Buffer D derived subunits (Figure 7), suggesting that at least some of the nativesubunit-specific protein must be bound directly to the particle under these low-salt conditions. Further, in two experiments the subunits were derived directly from the crude ribosome pellet, rather than from the 55S fractions of a sucrose density gradient. Such subunits showed

PAGE 83

72 only small quantities of these contaminants (Figure 10). Most of these proteins thus seem to be bound directly to the ribosomes under low-salt but not high-salt conditions, rather than components of separate structures which also sediment in the 25S to 45S region of the gradient. High-salt treatment . In addition to these proteins which were absent from all derived large subunit preparations, several proteins were found to be removed when the ribosomes were dissociated under some ionic conditions but not others. These proteins are the ones designated by letters in Table VI. The buffer conditions of interest are Buffer C and Buffer D: since large subunits prepared in either of these conditions are functionally active (Figure 7), the absence of a protein from either kind of preparation indicates that that protein is not required for the function. Thus MLc and MLg are not essential for peptidyl transferase activity because Buffer D subunits lack them but possess the activity. MLb, d, e, f, and h are specifically absent from Buffer C subunits. MLa is found in Buffer C preparations, but only in much smaller quantity than in Buffer D subunits; since this large quantitative difference does not correlate with the similarity between the two kinds of subunits in functional activity, this protein too may be regarded as nonessential. The fact that these proteins are dispensable for the peptidyl transferase activity of the ribosome does not necessarily imply that they are contaminating non-ribosomal proteins. It only indicates that

PAGE 84

73 they are not components of the active site for this function, and are not structurally required to maintain the proper conformation of the active site. Peptidyl transferase is only one of many reactions and interactions involved in protein synthesis, including translocation, initiation, termination, and binding to factors and raRNA. And not all of the ribosomal proteins are required for each function. For example, the E. coli ribosomal protein L7/12, which is needed for the interaction of the elongation factors EF-T and EF-G with the ribosome, can be removed without significantly affecting the peptidyl transferase activity (Hamel et al . , 1972). Therefore, in the absence of data demonstrating that both Buffer C and Buffer D ribosomes are active in a more demanding assay (such as the translation of poly U or a natural mRNA), it remains quite possible that all of the lettered spots in Figure 8 do represent ribosomal proteins. Extreme-salt treatment . As an attempt to identify some proteins which were required for peptidyl transferase activity (and which therefore must clearly be ribosomal proteins), the electrophoret ic patterns of large subunits prepared with Buffers E, F, and G were evaluated. Since Buffer F and Buffer G subunits are essentially inactive, it seemed likely that they would be missing one or more proteins, and that among these proteins at least one was required for the activity. One or several of these proteins might also be found in diminished quantities in Buffer E subunits, which have diminished activity.

PAGE 85

74 For this purpose, two preparations of Buffer F subunits and two of Buffer G subunits were subjected to two-dimensional electrophoresis. One of the Buffer F analyses is shown in Figure 15, and one of the Buffer G gels in Figure 16. Some differences were observed between the Buffer F preparations and between the Buffer G preparations, but there were several proteins which were absent or significantly diminished in staining intensity in both of the analyses in each pair. Buffer F did not completely remove any of the numbered large-subunit proteins, but it did reduce five of them to trace-level staining intensity. ML1, 15, 16, and 18 were diminished to a tenth or less, and ML2 to less than half, of the quantity normally present in Buffer C or Buffer D subunits. In subunits treated with Buffer G, small quantities of ML16 and ML18 were still present, but ML1 and ML15 were not detectable In addiLion ML23, 29, 30, 33, 43, and 48 were absent and the quantity of ML47 was considerably reduced. Interestingly, the amount of ML2 found in Buffer G subunits did not appear to be significantly less than that in subunits treated with Buffer C or Buffer D, even though the less stringent treatment with 1 M KC1 at 5 mM MgCl 2 (Buffer F) did remove sonu of this protein. A similar phenomenon was observed among the large-subunit proteins which are designated by letters. MLg, which is absent from sub. , its treated with the moderate-salt Buffer D and is present only in trace quantities after treatment with Buffer F, is not removed by Buffer G. Such results are at odds with the expectation that the set of proteins removed by a

PAGE 86

75 mmmmmm mn.miM-w* 4 * * * *£ % Figure 15. Electrophoretic pattern of proteins from large subunits prepared in Buffer F.

PAGE 87

76 Figure 16. Electrophoretic pattern of proteins from large subunits prepared in Buffer G.

PAGE 88

77 given ionic medium should be a subset of the proteins removed by more stringent conditions. This expectation derives from the continuous decrease in the total protein content of the subunit which is observed as the KCl/MgCl 2 ratio in the medium is increased (Figure 7). Apparently the generalization that increasing KCl/MgCl ratios remove increasing amounts of protein applies to the total protein content (as measured by the buoyant density) and to some of the individual proteins, bur ML 2 and MLg are exceptions. That a few such exceptions to this rule might exist, due to specific effects of particular buffer compositions on the binding of some of the proteins to the ribosome, is not too surprising. The protein complement found in Buffer E treated large subunits, however, represents a more serious anomaly. None of the numbered proteins were found to be removed or significantly diminished in quantity by washing with this buffer, offering no explanation for the reduction in functional activity produced by this treatment. Furthermore, even the lettered proteins (which are all removed by either Buffer C or Buffer D) were each found at least once in preparations of Buffer E subunits (Table VI). Thus the buoyant density data (Figure 7) indicate that these particles must be lacking some proteins present in Buffer C or Buffer D subunits, but no missing proteins can be identified by electrophoretic analysis. The reasons for this discrepancy are not clear, although some considerations described below (Section IV) may be relevant to this question.

PAGE 89

78 As mentioned above, several proteins are specifically missing or reduced in quantity in large subunits that have been treated with ionic conditions sufficiently stringent to inactivate them. in particular, the loss of MLl, 15, 16 and 18 correlates with the loss of functional activity. In order to prove that the inactivity of the stripped subunit is due to the loss of these proteins (and not, for example, due to a conformational alteration of the subunit induced by these salt conditions) it is necessary to demonstrate that the activity can be reconstituted by adding the proteins back. Such a partial reconst itution would also be useful as an assay to determine which of the missing proteins is (are) required for the activity. However, attempts to reconstitute the subunit were unsuccessful, c. Summary of the large-subunit proteins Electrophoretic analysis of the large subunit of the bovine mitochondrial ribosome revealed 52 protein components which were reproducibly present in functionally active large subunits and which therefore probably represent ribosomal proteins. Eight additional proteins were present in large subunits prepared only under certain ionic conditions; these may also be ribosomal proteins, but they are not required for the peptidyl transferase activity of the large subunit. Treatment with very-high-salt buffers removes a small number of the 52 proteins and destroys the peptidyl transferase activity, suggesting that one or more of the missing proteins is required for this function.

PAGE 90

79 2. Small S ubunit As shown in Figure 13 above, native small subunits prepared in lowsalt buffers contain a set of mainly high-molecular-weight proteins which are also found in the native large subunit. Although lesser quantities of these proteins are sometimes found in derived small subunits treated with moderate salt concentrations (Figures 6 and 17), they are usually much diminished relative to the quantities present in the native subunit. Figure 18, obtained from the derived small subunit from the same ribosome preparation whose native subunits are shown in Figure 13, is completely lacking in these proteins. For some of the same reasons given in the discussion of the large subunit, it seems unlikely that these proteins represent functional components of the small subunit either. Table VII shows the results of several electrophoretic analyses of high-salt-washed small subunits. The proteins which were found regularly in separate ribosome preparations and thus satisfy the criterion of reproducibility are designated by numbers. Some of the proteins which were seen less frequently are also tabulated and are designated by letters. (To avoid confusion with the numerals 1 and 0, the letters 1 and o have been skipped.) The electrophoretic positions of these proteins are given in Figure 19. The overall level of reproducibility of the small-subunit proteins was Less than that of the large-subunit proteins. To some extent this variability is due to the absence of many of the proteins from a partic-

PAGE 91

Figure 17. Electrophoretic pattern of proteins from small subunits prepared in Buffer C (Experiment 2) .

PAGE 92

HI % Figure 18. Electrophoretic pattern of proteins from small subunits prepared in Buffer C (Experiment 3) .

PAGE 93

82 CT\

PAGE 94

83 + + + + + i + + + + + + i + I OOO + OOOO I ooo + + + +I + I+ + + + +I + OO + OOOOOI I O O I + + C+ + + + I + O + + O + + 000++00 + 00 coo + + o o + 4. + + I + + + 0+ OOOOO + +OOOOOOO + + + + +IO + O + +I |+ OOC+ + 00+0 + + O + + O O + + I + + + + + + 1 + I I Ooooooooooo I + + 1 + + + + + + 1 + ++OOO + OOOO + OOO + + + + + + + + + + + + 1 + OOOOOO + OOO C + + + + + 1 + ooroo + o + + 10 + 00 cOONOHNrn^-irnor-oOffiOH MNnnmmrtcocnn(
PAGE 95

84 2-D 68 44 + I-D23.5t.d14 41 o

PAGE 96

85 ular experiment. In Experiment 6 for example, the quantity of protein in the sample was inadequate; as a result all of the spots obtained were faint, so the absence of many of the spots which are normally faint in adequately loaded gels is understandable. The absence of 6 of the numbered proteins from Experiment 3 (Figure 18), however, is not so simply explained. On the other hand, there was a subset of the experiments in which variability was not a problem at all: none of the numbered proteins were lacking in Experiments 2, 8, or 9, and only two were missing in Experiment 4. The variability which was observed in the other experiments seemed to be random with respect to which proteins were found to be missing. That is, it was not due to a high level of variability in a few specific proteins. Only two of the numbered proteins, MS13 and MS21, were absent from as many as three of the nine experiments. In the case of these two proteins (as well as MSI), much of the lack of reproducibility was due to their absence from subunits prepared in buffer D; thus there may be a specific ionic effect on the binding of these proteins to the riboscme, as was seen for MLc and MLg above. A complementary effect was found for MSa and MSb. These two proteins were always present in subunits prepared in Buffer C, but only once in any other preparation. The general impression of unreproducibility in the small subunit pattern derives partly from the relatively large number of spots which appear in fewer than half of the analyses and thus are designated by letters in Table VII. Some of these, like MSc and MSn (Figure 6),

PAGE 97

86 appeared in only one experiment and can be easily disregarded. Others (MSa, b, f, k, and p) were present in four of the nine experiments and thus represent borderline cases which are not clearly excluded from the ribosomal protein category. Another factor which contributes to this impression of unreproducihility is the variation in staining intensity among Lhe different spots on a given gel. In most experiments (Figure 17 is a conspicuous exception) the electrophoret ic pattern was found to consist of a minority of heavily-stained spots, usually including MS4, 5, 6, 10, 17, 19, 24, 28, 31, 37, and 38, and a majority of fainter spots. Furthermore the relativ,intensity of a given spot, particularly among the fainter group, varied considerably among different gels. (Compare MS13, 15, and 1ft in Figures 6 and 20, for example.) Some differences in the stain-binding abilities of different proteins hcve been reported (Fazekas de St. Groth et ad. , 1963; Bickle and Traut, 1971), but the differences are not of the magnitude required to explain the present results if all of the sma 1 1-subunit proteins were present in equimolar quantities in these gels. And in any case this phenomenon could not account for the variations observed for a particular protein in separate experiments. It is therefore necessary to conclude that the small-subunit proteins are present in differing molar quantities in these preparations. It follows that the individual subunits in each preparation are heterogeneous with respect to the set of proteins they

PAGE 98

87 1

PAGE 99

This heterogeneity could be due to the non-specific binding of contaminating proteins to some of the subunits. However, in view of the small number of heavily-staining proteins found in these particles, it seems unlikely that all of the fainter spots represent contaminants. A more plausible hypothesis is that a proportion of the subunits are lacking some of the ribosomal proteins, as a result of an in vitro artifact. This hypothesis is supported by the similar results which have been ub ained (Kaltschmidt and Wittmann, 1970) and explained (Hardy, 1975) for E. coli ribosomal proteins. Some possible causes of such an artifact will be considered below (Section IV); but whatever the cause, it is probably responsible for the variability between ribosome preparations as well as the heterogeneity within each preparation. Similar considerations also apply (to a lesser extent) to the proteins of the large subunit, which show a lower level of variability and non-uniformity of staining intensities. The effect of extreme-salt t reatment (Buffer G) on the protein content of the small subunit was also investigated. One of the gels obtained is shown in Figure 21. The most noticeable result was a general reduction in the quantity of the low-molecul ir-weight. proteins, those below MS31. Several proteins (MS13, 16, and 35) were completely absent from all three preparations of Buffer G subunits, and MS27, 28, 33, 36, 37, and 41 were diminished to negligible levels. Some proteins, on the other hand, were increased relative to the quantities present in high-salt-treated (Buffers C and D) subunits: this was particularly

PAGE 100

89 • Figure 21. Electrophoretic pattern of proteins from small subunits prepared in Buffer G. The irregular spots seen above and to the right of MS4 are artifacts due to precipitation of the stain on the surface of the gel

PAGE 101

90 true of MS8, 12, and 14. In addition, some of the lettered, unreproducible proteins were found reproducibly in Buffer G subunits. MSd, e, f, and m were each found in all three preparations. It was also observed that several of the high-molecular-weight proteins present in low-salt but not high-salt subunits are present in considerable quantity in these extreme-salt subunits. This rather confusing set of observations is further evidence for a conclusion reached earlier in the discussion of the large subunit. The general effect of treating these rlbosomes with buffers of increasing KCl/MgCl 2 composition is to reduce both the total amount of protein as measured by the buoyant density of the particles and the number of proteins as detected by gel electrophoresis. But some of the individual proteins may show the opposite response to such treatments, being preferentially removed by buffers of lower KCl/MgCL 2 ratios. As was the case with the large subunit, no proteins were found to be removed or significantly diminished by treatment with Buffer E (Table VII), even though the buoyant density of the Buffer E subunit is somewhat higher than that of Buffer C or Buffer D subunits. Thus buoyant density measurements appear to be a more sensitive means of detecting the loss of protein as a function of KCl/MgCl 2 than the electrophoretic analyses are. This lack of sensitivity may be due lo the high background level of variability between electrophoretic experiments, especially in the case of the small subunit. However, it also seems likely that these treatments are acting non-selectively to some extent, removing small quantities of many of the proteins

PAGE 102

91 (and thus increasing the heterogeneity of the subunits with respect to their protein content) rather than larger quantities of only a few proteins. Any such non-selective effects would not be detected by comparing the relative staining intensities of the proteins in the gel. 3 . Comparison of La rge and Small Subunits In order to determine whether the proteins of the large subunit were all distinct from those of the small subunit, the two electrophoretic patterns were compared with each other. To obtain a sufficiently accurate relative positioning of the patterns, samples of largeand small-subunit proteins were mixed together and eleclrophoresed on the same gel: one sample was a trace quantity of protein which had been radioiodinated as described in MATERIALS AND METHODS, and the other sample was a larger (stainable) quantity of protein. After electrophoresis, staining, and autoradiography, the patterns of stained and radioactive proteins were compared. Control experiments in which a small aliquot of a protein sample was radioiodinated and then mixed with a large aliquot of the same sample before electrophoresis showed that the labelling reaction had no detectable effect on the electrophoret ic mobilities of the proteins. As shown in Eigure 22, it was found that many of the small-subunit proteins overlapped at least partially with proteins of the large subunit, and there were twelve pairs of proteins which were electrophoretically indistinguishable. These were MSI and ML 2 , MS4 and ML3,

PAGE 103

92 l-D > 2-D 68 44 23.5 17.214.4w % o^>' 9 © O «0» to o » Figure 22. Relative electrophoretic positions of bovine mitochondrial large-subunit and small-subunit proteins. Filled spots, large-subunit proteins; open spots, small-subunit proteins; hatched spots, overlapping large-subunit and small-subunit proteins.

PAGE 104

93 MS7 and ML5, MS11 and ML12, MS15 and Mil 7, MS23 and ML20, MS25 and ML25, MS26 and ML27, MS30 and ML32, MS31 and ML31, MS32 and ML35, and MS40 and ML.46. Only the proteins which have been designated by numbers in the discussion above are shown in Figure 22. In addition the electrophoretic positions of MSk and ML38 overlap, as do those of MS 3 7 and MLh. There are several possible explanations for the relatively large number oi electrophoi etically identical proteins in the two subunits. First, it might be that two copies of some proteins are present in the ribosome, one in each subunit, even though such a double role for a single protein has not been demonstrated in any other ribosome to date. Alternatively it may be that the two proteins in each pair are nonidentical but are not resolved by the electrophoretic system; in view of the generally crowded appearance of Figure 22 it would not be surprising if some proteins occupied the same electrophoretic position by coincidence. A third possibility is that some of these proteins are contaminants similar to the non-subunit-specif ic proteins found in native subunits (Figure 13), except that they have molecular weights similar to those of the ribosomal proteins and remain bound to the ribosomes even after washing with high-salt buffers. Finally, it is possible that each of the protein pairs represents a single ribosomal protein which is present at the level of one copy per intact ribosome, but fails to bind exclusively to one subunit or the other when the ribosome is dissociated. Such a partitioning of

PAGE 105

94 individual proteins between the dissociated subunits has been observed in other ribosomes. The E. coli small-subunit protein S20, which coelectrophoreses with the large-subunit protein L26, has been shown to be identical to it by immunochemical and genetic studies (Tischendorf et al. , 1974; Wittmann et al. , 1974). Significant quantities of S5 have been found in the large subunit as well (Tischendorf et al., 1974). Likewise, three proteins of rat cytoplasmic ribosomes appear to be shared between the two subunits upon dissociation (Sherton and Wool, 1974). If this phenomenon is responsible for some of the overlapping protein spots observed, several related effects might also be seen. First, this hypothesis requires that the total quantity of protein present in each pair represent no more than one copy per ribosome. This requirement could be satisfied by ilmost all of the protein pairs (MS31/ML31 is the exception), since in each case one or both members of the pair are seen in quantities significantly smaller than the average of the proteins present on the gel. It might also be expected that even though these proteins do not bind exclusively to either of the subunits, they do bind somewhat more strongly to the subunit with which they are associated in the intact ribosome and thus are found preferentially in this subunit upon electrophoretic analysis. In fact, in six of the protein pairs, one protein of each pair does appear in much greater quantity than the corresponding protein in the other subunit. The overlapping proteins which are present in relatively

PAGE 106

95 larger quantity on one subunit are ML2, 5, 17, 25, and 27, which are found in greater amounts than the corresponding smalJ -subunit proteins, and MS4, which occurs in greater amounts than ML3. Since the failure of such proteins to segregate exclusively to one subunit or the other upon dissociation probably represents an artifact of the method used to effect the dissociation, it seems likely that differences in the ionic composition of the buffer used for this purpose could alter the extent of the protein partitioning. In fact it was reported (Sherton and Wool, 1974) that the subunit localization of one of the three partitioning proteins in rat cytoplasmic ribosomes was determined by the ratio of potassium to magnesium ions in the dissociating buffer. In the present results such a quantitative shift of a protein from presence only in one subunit to presence only in the other, due to differences in the ionic medium, was not observed. However MSI was found to be specifically absent from small subunits prepared in Buffer D, and MS23 and 25 were significantly reduced in quantity in this ionic condition as well. Clearly, this phenomenon could contribute to the heterogeneity in protein content observed in these subunits, although it could not completely account for it. In any case more direct evidence to test this hypothesis will not be easy to obtain. Whether the proteins in each pair are in fact identical can perhaps be determined by chemical or immunochemical analysis of the isolated proteins. If they are identical, however, the question whether they are ribosomal proteins

PAGE 107

96 or non-ribosomal contaminants still rests primarily on the criteria discussed above. Some additional evidence that they are ribosomal proteins would be obtained if they could be shown to be local ed at the interface between the two subunits in the intact ribosome; if these proteins are responsible for the binding of the subunits to each oth. i , their partitioning behavior is more easily understood. The E. col i S20/L26 pair has been found to be a subunit-interface protein (Morrison et al. , 1973). ( . . Experiment al Evaluation of Possible Artifacts A total of 93 proteins (the numbered spots in Figures 8 and 19) were found reproducibly in high-salt-treated mitochondrial ribosome subunits. Eighty-one of these proteins were distinguishable from each other on the basis of their electrophoretic mobilities. Although neither of these numbers is greater than had been expected on the basis of measurements of the total protein content of this ribosome, they are considerably in excess of the numbers of proteins which have been reported for any other kind of ribosome. For this reason it was necessary to consider the possibility of various artifacts which might generate an unrealistically ! uge number of apparent ribosomal proteins. One such possibility, the presence of adventitiously bound contaminants in the ribosome preparation, has been discussed above. Another possibility is that some single proteins may be represented by more than one spot on a gel. This could occur if a protein had been partially modified as the result of either a normal in vivo process or

PAGE 108

97 an artifact of the isolation or electrophoretic procedure. Thus E. coli L7 is actually identical to L12, except that it has been acetylated at the N-terminus in vivo (Terhorst et al,, 1972); similarly, the rat liver cytoplasmic ribosomal protein S6 sometimes appears as multiple electrophoretic species due to in vivo phosphorylation (Gressner and Wool, 1974). Two electrophoretically distinct forms, thought to represent different states of oxidation, have been observed for each of the E. c oli proteins Sll, S12, and S17 (Wittmann, 1974). Carbamylation by evanate ions formed spontaneously in urea solutions can alter the electrophoretic mobility of proteins (Gerding et al . , 1971). In the electrophoretic system used in the present experiments, these kinds of chemical modification.-; could result in a small alteration of a protein's mobility in the first electrophoretic dimension (in urea at pH 4.5), but would not affect the migration in the second dimension (in SDS). Some groups of spots which do show this electrophoretic pattern include ML42, 43, and 44; MS30 and il ; and MS37 and 38. Extra spots might also be produced by proteolytic degradation of some of the proteins to discrete polypeptide products. Alternatively, some of the proteins may be covalently crosslinked to give specific aggregates. Disulfide crosslinks have been reported to form during two-dimensional electrophoresis (Kaltschmidt and Wittmann, 1970). Peroxidation of lipids in the crude ribosome preparation could also lead to protein crosslinking (Tappel, 1973).

PAGE 109

98 Some precautions against these possible artifacts were taken in the experiments presented above. BHT was added to all solutions used in the ribosome preparation to prevent peroxidation-induced crosslinks. Disulfide bonds were reduced with dithiothreitol, and the proteins were maintained in the reduced state by the inclusion of reducing agents throughout the electrophoretic procedure. In addition, the electrophoretic results themselves argue against the likelihood of some of these artifacts. Carbamyl.! i ion, for example, wouln be expected to affect most or all of the proteins to a similar extent; thus the presence of many proteins which do not show the peculiar electrophoretic pattern represented by ML42, 43, and 44 argues that such patterns are not produced by a non-specific chemical modification. To obtain more evidence about the possibilities of proteolytic, crosslinking, or other chemical alterations, as well as the nonspecific binding of contaminants to the ribosomes, the effect of the ribosome isolation procedure on the proteins of purified E. coli ribosomes was assessed. To a preparation of bovine liver mitochondria was added a small quantity of salt-washed E. coli ribosomes, equal to the quantity of mitochondrial ribosomes the mitochondria were estimated to contain. The mitochondria were then lysed with Triton X-100 and ribosomes pelleted at 100,000 x g as usual. After sucrose density gradient centri f ugat ion in Ruffer H, the ribosomes in the 70S (E. coli ribosome) peak were dissociated to subunits by a second centrifugation in a sucrose density gradient in Buffer E. The

PAGE 110

99 proteins extracted from these subunits were then analyzed by twodimensional electrophoresis. The results suggested that the types of artifacts mentioned above are not a problem in these experiments. The electrophoretic pattern of the proteins of the _E. coli 50S subunit showed no significant difference from that obtained from ribosomes subjected to the same reisolation procedure in the absence of mitochondria, or from freshly prepared ribosomes. The 30S subunits were apparently slightly contaminated with mitochondrial large (39S) subunits, since traces of all of the most intensely staining 39S proteins were found on the gel of 30S proteins; with this exception, the 30S subunit protein pattern was likewise unaltered in the number and electrophoretic mobilities of the proteins. Thus incubation with components of the mitochondrial lysate under the conditions of our normal preparation procedure does not cause any increase in the apparent number of ribosomal proteins of exogenouslv added ribosomes. This result argues especially against the possibilities that the number of proteins found in mitochondrial ribosomes is overestimated due to proteolysis, crosslinking, or the adsorption of non-ribosoma.l proteins. Another possible artifact of the ribosome isolation procedure was mentioned in Section I above: the binding of the ribosomes to DEAEcellulose, or the subsequent elution, during the preparation of ribosomes by Method C might dislodge some of the riboscmal proteins. This possibility seems unlikely in view of the undiminished functional

PAGE 111

100 activity of ribosomes prepared in this way. However, the electrophoretic results presented above include analyses of subunits prepared both with and without the use of DEAE-cellulose, so a comparison of the protein patterns produced by these two methods is easily made. In Tables VI and VII the samples for Experiments 2, 3, and 4 were DEAEcellulose prepared ribosomes. Comparing these three experiments with tie other experiments tabulated shows no proteins which are clearly removed by treatment with DEAE-cellulose, although some possibly significant effects can be seen. They most likely involve proteins MS9 and MS14, which are both absent from two of the DEAE-cellulose preparations whereas they are otherwise quite reproducible. Somewhat weaker correlations are observed for MS30, MS33, and ML21. Thus it remains possible that a few proteins, particularly in the small subunit, are specifically removed by treatment with DEAE-cellulose. It may also be that MS9 and MS14 are unnecessary for the translation of poly II; however, a more direct comparison between the activities of ribosome preparations known to contain and to lack these proteins will be necessary to establish this point. III. COMPARIS ON WITH PROTEINS OF N0N-MIT0CH0NDRIA1 , RIBOSOMES Yet another way in which the number of proteins in mitochondrial ribosomes could be over-estimated by the results of these experiments would be the possible contamination of the preparations by significant quantities of cytoplasmic ribosomes. Although the more purified preparations described in Section I contain no detectable cytoplasmic

PAGE 112

101 ribosomes, it seemed worthwhile to demonstrate directly the absence of cytoplasmic ribosomal proteins from the electrophoretic patterns of mitochondrial ribosomes. It was expected that the results would also provide a comparative basis for the description of the properties of mitochondrial ribosomal proteins. The relative electrophoretic positions of the two sets of ribosomal proteins were determined by co-electrophoresing radioiodinated and stainable protein samples as described in Section I1B3. The results for the large subunits of the two ribosomes are shown in Figure 23, and for the small subunits in Figure 24. With the exception of one protein which comigrates with MS25, ail of the mitoribosomal proteins are electrophoretically distinct from those of the corresponding subunit of the cytoplasmic ribosome. Indeed most of the cytoribosomal proteins are found in a different region of the gel from that which contains most of the mitoribosomal proteins. Comparing proteins of similar molecular weights (similar positions in the second electrophoretic dimension), it is evident that most of the cytoribosomal proteins migrate more rapidly in the first dimension than do the mitoribosomal proteins. This tendency, which is most evident in the comparison of the large subunits, implies that the cytoplasmic ribosomal proteins bear a greater positive charge at pH 4.3 than the mitoribosomal proteins. It also suggests that the isoelectric points of many of the cytoribosomal proteins may be higher than those of the miLoribosomal proteins.

PAGE 113

102 + l-D Figure 23. Relative electrophoretic positions of bovine mitoribosomal and cytoribosomal large-subunit proteins. Filled spots, mitoribosomal proteins; open spots, cytoribosomal proteins.

PAGE 114

103 + l-D > 17214.4Figure 24. Relative electrophoretic positions of mitoribosomal and cytonbosomal small-subunit proteins. Filled spots, mitoribosomal proteins; open spots, cytoribosomal proteins; hatched spot, overlapping mitoribosomal and cytoribosomal proteins ,

PAGE 115

104 Comparing the mobilities of the two sets of proteins in the second electrophoretic dimension shows that the distributions of molecular weights of the mitoribosomal and cytoribosomal proteins are similar. The large-subunit proteins of the mitochondrial ribosome are slightly smaller on the average than those of the cytoplasmic ribosome, and the reverse is true for the small-subunit prcteins. The major difference in electrophoretic characteristics between mitochondrial ana bacterial ribosomal proteins lies in the much greater second-dimension mobility of the latter. As seen in Figure 12, over half of the ribosomal proteins of E. col i have molecular weights less than tit -i l of myoglobin (17,200), whereas relatively few mitoribosomal proteins are this small. A more detailed description of the molecular weight properties of the mitoribosomal proteins is given in Section IV. The suggestion above that mitoribosomal proteins may have isoelectric points significantly lower than those of other kinds of ribosomal proteins has some support in the literature. The relatively slow first-dimension migration of mitoribosomal proteins can be seen in the comparison of the large subunits of Xenopus laevis ribosomes (Leister and Dawid, 1974, Figure /). About half of the proteins of rat mitochondrial ribosomes are anionic at pH 8.6 (van den Bogert and de Vries, 1976), whereas nearly all eukaryotic cytoplasmic and bacterial ribosomal proteins are cations at this pH. Indeed, while most ribosomal proteins have isoelectric points that are too high to permit analysis by isoelectric focusing, it lias been reported that rat mitoribosomal proteins

PAGE 116

105 can be analyzed by this technique (Czempiel et al., 1976). The isoelectric points of most of the proteins were reported to lie below a pi of 8.5, with none above a pi of 10. Since the two-dimensional electrof ocusing/electrophoresis system described in the latter report should be capable of much greater resolution than the system used in the experiments described above, an attempt was made to reproduce these results. The methods given for the eleclrofocusing procedure (Czempiel et al., 1976; Klose, 1975) were followed as closely as possible; the only significant exception was the use of a constant-voltage power source instead of a pulsedpower source. The electrof ocusing was performed on parallel samples for various periods of time in order to determine the length of time required to focus the proteins. This question is of considerable importance in isoelectric focusing experiments. Short periods of time may be adequate for the carrier ampholytes to reach their equilibrium positions and establish the pH gradient, but not for the proteins, which are retarded by the gel due to their higher molecular weights, to reach their equilibrium positions. Thus false estimates of the isoelectric points of the proteins would be obtained. On the other hand after long times the pH gradient begins to break down, probably due to the migration of the carrier ampholytes out of the gel (Chrambach et al., 1973). The results of this experiment are shown in Figure 25. Alter 3 hours of electrofocusing the pattern of protein spots looked similar to

PAGE 117

Figure 25. Isoelectric focusing of mitoribosomal proteins. Proteins from bovine mitochondrial ribosomes prepared in Buffer A were electrofocused as described in MATERIALS AND METHODS for 3 hr (upper photograph) , 5 hr (middle photograph) , or 7 hr (lower photograph) . The electrofocusing gels were then extruded and placed at the top of SDS-gel slabs for electrophoresis in the second dimension. The isoelectric focusing dimension is horizontal in the photographs, and the SDS-electrophoresis dimension is from top to bottom. The point of application of the protein sample is at the upper left-hand corner of each photograph.

PAGE 118

107 ." ^ •

PAGE 119

108 that shown by Czempiel et al. (1976). It may be seen that the spots tend to be arranged diagonally, especially on the right-hand side of the pattern, rather than randomly. This diagonal tendency, which is seen to a greater extent in the standard two-dimensional system used in Section II, is an indication that the proteins are being differentially retarded cording to their molecular weights as they move through the first-dimension gel. Elect rofoeusing/SDS-electrophoresis patterns obtained after the proteins have reached their equilibrium positions in the first dimension show an apparently random distribution of the spots (O'i'arrell, 1975). The pattern observed after 5 hours of electrofocusing confirms the conclusion that most of the proteins were not focused at the 3-hour time point. Many of the proteins continued to move until they were caught at the basic end of the gel. Furthermore several proteins of relatively high molecular weight, seen in the upper right-hand corner of this pattern, also moved to the basic end of the gel after 7 hours. In the 7-hour pattern, those proteins which remained in the body of the first-dimension gel showed the randomly-dispersed appearance expected of focused proteins. It may be noted (Figure 26) that the pH gradient was already formed at 3 hours and remained stable through the 7-hour time point, although it had deteriorated by 10 hours. The results thus seem to contradict those of Czempiel et al. (1976) and to suggest that the very low isoelectric points reported by these authors are in error due to the failure of the proteins to reach their

PAGE 120

109 o o o in o o

PAGE 121

no isoelectric positions in the gel. It also appears that considerably fewer than half of the mitoribosomal proteins have isoelectric points below 8.5, in contrast to the results of van den Bogert and de Vries (1976). Although the latter results were obtained by electrophoresis at pH 8.6 rather than by isoelectric focusing, it would be surprising if these methodological differences were the source of the discrepancy. Some preliminary electrophoretic experiments, however, appear to confirm that about half of the large-subunit proteins are anions at pH 8.6 (Robert Hessler, 1976, private communication). Thus the question of the actual isoelectric points of the mitoribosomal proteins remains somewhat uncertain. But even the electrof ocusing experiments presented above show that a relatively large proportion of these proteins are isoelectric below pH 8.5. All of the evidence therefore agrees that these proteins are significantly more acidic than the proteins of E. c oli or mammalian cytoplasmic ribosomes, over half of which have isoelectric points above 10 (Kaltschmidt , 1971; Huynh-van-Tan et al_. , 1974). IV. MOLECULAR WEIG H TS OF MITOCHON D RIAL RI B OSOMAL PROTEINS The molecular weights of the mitoribosomal proteins were determined from their mobilities in the second electrophoretic dimension and are given in Table VIII (large-subunit proteins) and Table IX (small-subunit proteins). The molecular weights of the large-subunit proteins range frcm 11,100 to 45,500 (average: 21,200); the small subunit contains proteins from 12,300 to 43,000, averaging 22,600, For comparison, the average molecular weights of E. coli large-subunit and small-subunit

PAGE 122

Ill Table VIII. Molecular weights of mitochondrial large-subunit proteins. Protein 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 MW x 10 45.5 42.5 38.5 32 31.5 31.5 31 29.2 28.1 27.6 26.3 26.0 26.0 25.7 25.0 23.6 23.5 22.2 21.6 20.6 20.4 19.9 19.5 19.2 19.0 18.8 -3 Protein 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 MW x 10 18.3 18.2 18.2 18.0 17.8 17.6 17.4 17.1 16.8 17.0 16.5 16.2 15.8 15.9 15.3 15.2 15.2 15.4 14.7 14.0 13.6 13.4 13.5 13.3 11.2 11.1

PAGE 123

112 Table iX. Molecular weights of mitochondrial small-subunit proteins. Protein 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 MW x 10 43 40 38 38 33.5 32 32 29 28.1 26.7 25.6 24.8 24.1 23.7 23.5 23.3 22.0 21.8 21.6 21.1 Protein 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 MW x 10 21.2 21.0 20.3 19.7 18.9 18.3 18.3 17.8 17.6 17.5 17.5 16.9 16.3 15.9 15, 15. 1.4. 14.7 14.0 13.7 12.3

PAGE 124

113 proteins are 16,300 and 19,000 (Dzionara et al. , 1970). This rather large difference in the molecular weight distributions of mitochondrial and bacterial ribosomal proteins can be easily observed by visual comparison of their electrophoretic patterns (Figures 5 and 6 vs. Figure 12) . Reported average molecular \jeights of mammalian cytoplasmic ribosomal proteins range from values similar to those of bovine mitoribosomal proteins (Martini an 1 Gould, 1971; Creusot et al. , 1975) to significantly larger values (Howard et al., 1975; Lin and Wool, 1974). The data in Tables VIII and IX, with an assumption, permit a calculation of the total content ol protein in the mitochondrial ribosome. If one copy of each of these proteins is present per ribosome, the sum of their molecular weights should equal the total protein content in daltons. This sum is 2.03 x 10 . The sum of the molecul ar weights of the E. coli ribosomal proteins, for comparison, is only 0.94 x 10 (Dzionara et_ al_. , 1970) . This two-fold difference is due partially to the larger size of the mitoribosomal proteins, but mainly to their greater number. In fact, mitochondrial ribosomes appear to contain at least as much protein as SOS mammalian cytoplasmic ribosomes. As mentioned above, there is some uncertainty about the molecular weights of the cytoribosomal proteins; but because of the smaller number of proteins in these ribosomes, even the larger values reported for their molecular weights lead to estimates for the total protein content of 1.73 x 10 6 to 2.05 x 10 6 daltons (Howard et al., 1975; Lin and Wool, 1974), about the same as that of bovine mitor ibosomes.

PAGE 125

114 This relatively large protein content had been expected on the basis of physical-chemical measurements on the mitochondrial ribosomes. Since these measurements provided quantitative predictions, it was of interest to determine how closely the different kinds of estimates for the total protein content agreed. The protein content can be calculated from the buoyant density, p, of the ribosome in CsCl and the molecular weight of the ribosomal RNA, Mp NA , by the equations % RNA = 309.8 I 96 6 (Hamilton and O'Brien, 1974) P M and Proteii, content = J*M x u,,, ... M D . 1A . The results of such a cal%RNA RNA dilation, using a value of 1.43 g/ml for the buoyant density of either subunit and 0.54 x 10 and 0.35 x 10 for the molecular weights of the large and small ribosomal RM.\'s (Robberson et al., 1971), are given in the first line of Table X. A second, independent measurement of the protein content can be obtained by determining the molecular weight of the ribosome by highspeed equilibrium centrifuge tion, and subtracting M 1 . The second line of Table X is based on the values 1 . 65 x 10 6 and 1.10 x 10 6 for the molecular weights of the large and small subunits, respectively (Hamilton and O'Brien, 1974). The results obtained from the electrophoretic analysis are given in the last line of the table. The sum of the molecular weights of the 52 proteins in the large subunit is 1.10 x 10 , a value which agrees well with the protein content as calculated from the buoyant density of this subunit or from its particle weight. However, the total of the 41

PAGE 126

115 Table X. Calculation of the protein content of bovine mitochondrial ribosomes. — f, Data used in calculation Protein content x 10 (daltons) Large subunit Small subunit Total Buoyant density of subunit, 1.10 0.71 1.81 molecular weight of rRNA Molecular weight of subunit, 1.11 0.75 1.86 molecular weight of rRNA Molecular weights of 1.10 0.93 2.03 individual proteins

PAGE 127

116 small-subunit proteins, 0.93 x 10 daltons, is somewhat greater than the protein content determined by other methods. One interpretation of this discrepancy is that it indicates an overestimate of the number of proteins in the small subunit by about 20%. An argument in favor of this interpretation can be found in the results presented in Section IIB3. Several of the proteins in the large and small subunits were found to be electrophoretically indistinguishable from each other, possibly because each pair represents a single protein which partitions between the two subunits when the ribosome is dissociated. Five of these proteins are present in much greater quantity in the large subunit than in the small, suggesting that they are actually large-subunit proteins and thus should not be included in the small-subunit list. If MSI, 7, 15, 25, and 26 were disregarded for this reason, the total of the small-subunit protein molecular weights would be 0.79 x 10 daltons rather ihan 0.93 x 10 6 daltons, agreeing more closely with the total protein content determined by physical-chemical measurements (Table X). A more convim hig interpretation, however, is that the original assumption (that each ribosome contains one copy of each of the proteins) is false. If, on the average, the proteins are present in less than unit quantity in each ribosome, the total protein content would be overestimated by simply adding up the proteins' molecular weights. It is clear in the electrophoretic patterns shown above that the proteins are not all present in equal quantities, particularly in the

PAGE 128

1 17 small subunit, but there i: no direct evidence whether the average spot intensity in these patterns represents one copy per ribosome or less than one. However the results which have been obtained with other ribosomes indicate that ribosomal proteins are more likely to be deficient than to be found in greater than unit quantities when ribosomes are prepared according to the methods used in the present experiments . Although E. coli ribosomes prepared under very gentle mechanical and ionic conditions appear to contain unit amounts of nearly all of the ribosomal proteins, the more usual procedures (particularly the salt-wa.sln'ng steps used to remove non-ribosomal proteins) remove varying quantities of many of the proteins (Hardy, 1975). As a result, several ribosomal proteins appeared on two-dimensional gels as spots that stained only faintly (KaJ tschmidt and Wittmann, 1970), quantitative analyses showed that many of the proteins were present in less than unit stoichiometry (Voynow and Kurland, 1971; Weber, 1972), and the sum of the molecular weights of the E_. coli small-subunit proteins was found to exceed the measured total protein content of the subunit by 25% (Dzionara et al. , 1970). Tt seems likely that this partial removal of ribosomal proteins by salt treatment is responsible for the finding that the mitochondrial ribosomes in each preparation appear to be heterogeneous with respect to their content of the individual ribosomal proteins, and for the variations observed among separate ribosome preparations. Such effects

PAGE 129

118 may also explain the observation that none of the proteins are found to be removed by treatment with Buffer E, even though the buoyant densities of the subunits are increased by this treatment (Section IIB) . The action of Buffer E may he relatively nonspecific, removing some proteins from all of the ribosomes but not always the same set of proteins. In this case none of the proteins would be found to be absent in the el ec trophoretic analyses, and even reductions in the relative quantities of some of the proteins might be undetectable. Unfortunately no method has yet been developed for preparing ribosomes free of contaminating proteins but containing stoichiometric quantities of all the ribosomal proteins. But if this were done, the results presented above indicau that the measured protein content of mitochondrial ribosomes would be as given in the bottom line of Table X. Since f unc t Lonal ribosomes in vivo probably do contain unit copies of nearly all the proteins (Hardy, 1975), the comparison shown in Table X corresponds to some extent to the difference between the mitochondrial ribosome jin vivo and irt vitro . Thus one useful result of the electrophoretic experiments is that they provide information about the nature and degree of the differences between the ribosomes as isolated and as they probably exist in vivo. The table, and the electrophoretic patterns themselves, permit the conclusion that the isolated ribosomes are lacking considerable quantities of many of the proteins; further, it appears that most of the difference is localized in the small subunit, whereas the large subunit is more nearly intact.

PAGE 130

119 V. COMPARISON WITH PROTEINS OF RA T MITOCHONDRIAL RIBOS OMES The proteins of eukaryotic cytoplasmic ribosomes show a relatively low degree of evolutionary divergence. No differences were found among mammals, birds, or reptiles in the two-dimensional electrophoretic pattern of their cytor ibosomal proteins by Delaunay _et al. ( L973) . Other laboratories (Martini and Gould, 1975; Kuter and Rodgers, 1974) have been able to detect some electrophoretic differences among these species, but in all cases the number of such differences has been quite small. Two proteins were found to differ in comparisons of rat, mouse, and hamster cytor ibosomal proteins, and one further protein was altered in HeLa cell ; (Kuter and Rodgers, 1974). However, the mitochondrial ribosomes of various distantly related organisms show greater differences in their physical-chemical properties than do the cytoplasmic ribosomes of the same organisms (Figure 1). This diversity at the level of the overall structure of the ribosome suggested that differences might be found between more closely related organisms at the level of the electrophoretic properties of their ribosomal proteins. To test this possibility the mitoribosomal proteins of another mammal, the rat, were analysed and compared with the proteins found in bovine mitoribosomes. The electrophoretic results from one of the two preparations of rat mitoribosomes which were analyzed as described above for bovine mitoribosomes are shown in Figures 27 (large subunit) and 28 (small subunit). The large-subunit pattern was similar in overall appearance to that obtained from bovine mitoribosomes. In

PAGE 131

120 . Figure 27. Electrophoretic pattern of proteins from rat mitochondrial large subunits.

PAGE 132

121 ^ * Figure 28. Electrophoretic pattern of proteins from rat mitochondrial small subunits.

PAGE 133

122 comparisons of such complex patterns, it is easiest to recognize correspondences between proteins that have unusual electrophoretic properties. Thus a very acidic protein (that is, a protein which migrates more slowly in the first dimension than other proteins of the same molecular weight) _ is seen in the rat pattern at a position near that of the bovine protein MI. 30, and a faintly-stained protein corresponding to ML33 is present as well. Spots corresponding to the very bash protein ML13 and the low-molecular-weight proteins ML49 and 50 can also be identified. Some obvious differences between the two sets of proteins are also evident, such as the very basic rat protein just to the right of the spot corresponding to ML20, which was never found in bovine mitoribosomes. A detailed comparison shows many other differences. The small-subunit pattern showed no obvious similarity to that of bovine small-subunit proteins, and no individual rat proteins clearly corresponded to any bovine proteins. In order to compare the rat and bovine patterns more accurately, samples of proteins from both sources were mixed together and coelectrophoresed. The bovine sample was a large (stainable) quantity of protein and the rat sample was a trace quantity (from a third preparation of ribosomes) radioac tivelv labelled by reductive methylation as described in MATERIALS AND METHODS. The results are shown in Figures 29 and 30. In these figures only those proteins which were found in at least two of the three rat mitoribosome preparations are given; most of these appeared in all three. Rat proteins which have

PAGE 134

123 + I-D " 2-D J — > + V 684 423.517214.4m

PAGE 135

124 + |-D2-D -^ 684423.514 4 (6)< 12* 14 24« 36%> ^34 J 38 X39) '41 Figure 30. Schematic diagram of rat mitochondrial small-subunit proteins.

PAGE 136

L25 electrophoretic mobilities identical to some of the bovine proteins are marked with the number of the bovine protein, as given in Figure 8 and Figure 19. In addition some of the rat proteins were found in positions very close to those of some bovine proteins and are designated by parenthesized numbers. Some of these near-correspondences occur among the proteins with unusual electrophoretic properties which appeared to represent clear similarities between rat and bovine largesubunit proteins. In these cases (ML13, 30, 33, 49, and 50) it seems very likely that the rat protein is closely related to that of the cow even though some difference is detectable. Bovine mitoribosomal proteins are clearly more similar to the mitoribosomal proteins of rat than they are to the proteins of bovine cytoplasmic or bacterial ribosomes. Both kinds of mitoribosomal proteins have similar molecular weight and charge properties, which are different from those of cytoplasmic or E. coli ribosomal proteins (Section III) . It seems likely that the numbers of proteins in both mitoribosomes are also about the same. Forty-eight proteins are seen in the rat large subunit as compared with 51 for the bovine large subunit. The small number of proteins found in rat small subunits (24) is rather puzzling. However, the quantity of protein used in these experiments may have been too small to permit the detection of faintly staining proteins. The similar physical-chemical properties of rat and bovine mitoribosomal subunits (O'Brien, 1971; O'Brien et al., 1974; de Vries and Kroon, 1974) indicate that the protein contents of the two should be nearly the same.

PAGE 137

126 At a more detailed level of analysis, the electrophoretic positions of the individual proteins show more differences than similarities between these two mammalian species. A convenient means of quantitating the degree of similarity between the electrophoretic patterns of two sets of riboscmal proteins has been suggested by Delaunay and Schapira (1974), using the formula degree of similarity, P = (a + b) n %(a + b) where a is the number of proteins in one of the ribosomes, b is the number in the other ribosome, and n is the total number of electrophoretically different proteins found in a mixture of the two samples. If all of the proteins are electrophoretically distinct from each other, P = 0. When no differences at all are found, as reported by Delaunay et al. (1973) for mammals, birds, and reptiles, a value of P = 1 is obtained. The three protein differences seen between rat and HeLa cytoribosomal proteins (Kuter and Rodgers, 1974) correspond to P = 0.95. When this formula is applied to the proteins of mitochondrial large subunits, in which only 13 of the 48 rat proteins are electrophoretically identical to bovine proteins, it is found that P = 0.26. A similar calculation for the small subunits, in which there are only 8 electrophoretic identities, gives a value of 0.25. The formula slightly underestimates the actual degree of similarity between the small subunits, since even if all 24 of the rat proteins detected were identical to cow proteins, a value of P = 1 would not be obtained.

PAGE 138

127 However if a correction is made for the apparent failure of these experiments to detect all of the rat small-subunit proteins, the value of P is still only 0.33. These results are consistent with the finding of at least seven mitochondrial large-subunit proteins which differ even between two species of the frog genus Xenopus (P = 0.83) (Leister and Dawid, 1975). The evidence thus indicates that the degree of evolutionary divergence among mitoribosomal proteins is much higher than among the cytoribosomal proteins of the same organisms. This conclusion extends to more closelyrelated species the generalization (see INTRODUCTION) that most of the physical and chemical properties of ribosorn-:s appear to diverge more rapidly during evolution for the ribosomes of mitochondria than for their extramitochondrial counterparts. Some speculations about the molecular and genetic basis of this higher evolutionary rate will be discussed below.

PAGE 139

DISCUSSION The results presented above are relevant to questions about the structure and the evolution of mammalian mitochondrial ribosomes. It is found that the large complement of protein observed in these ribosomes by physical-chemical measurements can be accounted for by the number and sizes of the individual ribosomal proteins. The molecular weights of these proteins are not exceptionally high; rather, it is the large number of proteins present in these ribosomes which is primarily responsible for their high protein content. The proteins are found to be somewhat more acidic than those of bacterial or eukaryotic cytoplasmic ribosomes, a property which may be related to differences in the kinds of intermolecular bonding interactions which prevail in such a protein-rich structure. Finally, the electrophoretic properties of mammalian mitoribosomal proteins, like many of the other structural properties of mitochondrial ribosomes in general, show a greater degree of phylogenetic variation than is found among cytoplasmic ribosomes. Ninety-three proteins are found in bovine mitochondrial ribosomes, as compared with about 70 in mammalian cytoplasmic ribosomes (Wool and Stoffler, 1974) and 53 in E. coli ribosomes (Wittmann, 1974; Pettersson et al., 1976). Similar relatively large numbers of ribosomal proteins have been reported for the mitochondrial ribosomes of other vertebrate species. Leister and Dawid (1974) found 84 mitoribosomal proteins in 128

PAGE 140

129 the frog Xenopus laevis . Recently the proteins of rat mitochondrial ribosomes were analyzed, and numbers ranging from 70 (van den Bogert and de Vries, 1976) to 107 (Czempiel et al., 1976) were reported. The large number of proteins found in bovine mitochondrial ribosomes does not seem to be attributable to contamination of the ribosomes by non-ribosomal proteins, on the basis of a variety of evidence. These proteins are reproducibly found in separate preparations of ribosomes, and they are not removed by treatments that have been found to remove con! nminating proteins from other kinds of ribosomes. Indeed, only a few of them are removed by even more stringent treatments that result in complete inactivation of the peptidyl transferase function of this ribosome. Furthermore, when heterologous ribosomes are added to the mitochondria and then re-purified by the same method used for the preparation of mitochondrial ribosomes, they are not contaminated by adsorbed proteins. There is a possibility that the present results overestimate the number of bovine mitoribosomal proteins to some extent, due to a partitioning of some proteins between the two subunits. Although the total number of proteins found in analyses of the large and small subunits separately is 93, only 81 of these can be observed in mixtures of the two sets of proteins. That is, 12 of the large-subunit proteins are electrophoretically indistinguishable from proteins of the small subunit. It is therefore possible that the two members of all or some of these overlapping protein pairs are in fact identical , and that

PAGE 141

130 each pair should be counted only once in the enumeration of the bovine mitoribosomal proteins. The evidence in favor of this possibility is only suggestive, and further characterization of the individual proteins involved will be required to resolve this question. The proteins of bovine mitoribosomes are found to have molecular weights averaging 21,200 in the large subunit and 22,600 in the small subunit, values similar to those of mammalian cytoribosomal proteins. Much higher molecular weights were reported for the mitoribosomal proteins of Xenopus , with averages of 27,000 and Vi,300 for the large and small subunits respectively (Leister and Dawid , 1974). Molecular weights were not determined in the published studies of rat mitoribosomal proteins, but the present experiments indicate that these are about the same size as bovine mitoribosomal proteins. The sum of the molecular weights of the bovine mitoribosomal proteins is somewhat greater than the total quantity of protein calculated to be present in these ribosomes on the basis of their buoyant density, particularly in the case of the small subunit. A similar result was obtained with Xenopus mitoribosomes (Leister and Dawid, 1974). The simplest interpretation of this phenomenon is that a proportion of the ribosomes in a given preparation are lacking some of the proteins, so that on the average there is less than one copy of each of these proteins per ribosome. This interpretation is consistent with results that have been obtained for E. coli ribosomes.

PAGE 142

131 The finding that bovine mitoribosouu J proteins are relatively acidic, compared to bacterial or mammalian cytoplasmic ribosomal proteins, is in qualitative agreement with the reported results of van den Bogert and de Vries (1976) and Czempiel et_ al_. (1976) on rat mitoribosomal proteins. The proteins of Xenopus mitochondrial ribosomes also appear to be more acidic than those of the cytoplasmic ribosomes, at least in comparisons between the large subunits (Leister and Dawid, 1974). Quantitatively, there are some differences between the results of the present experiments and those which have been published for rat mitoribosomal proteins. The cause of thes discrepancies is unclear. It does not seem to be related to differences between t he rat and bovine proteins: the direct comparisons of these two sets of proteins by electrophoresis show that they have similar charge properties at pH 4.3. A large number of electroph retic differences are found between bovine and rat mitoribosomal proteins. This observation is consistent with comparisons which have been made between species related both more closely and more distantly than rat and cow. Xenopus laevis and Xenopus mulleri differed from each other in at least seven of the proteins of the large subunit alone (Leister and Dawid, 1975). Rat mitoribosomal proteins showed no similarity with those of the fungus Neurospora (van den Bogert and de Vries, 1976). No obvious similarities can be seen when the patterns of bovine or rat mitoribosomes are compared with those published for Xenopus laevis (Leister and Dawid, 1974), obtained in an

PAGE 143

132 essentially identical electrophoretic system. Indeed even the molecular weights of the frog proteins are quite different from those of the mammalian proteins. The significance of these results lies in their contribution to studies on the structure of ribosomes and on the evolution of mitochondrial genes. First, it has been suggested that the low buoyant density oi animaJ mitochondrial ribosomes might be due to the presence of adherent membrane fragments (Borst and Grivell, 1971). The present results do iot support the hypothesis that a third component of lower density than protein (i.e., lipid) makes any significant contribution to the low buoyant density of these ribosomes: the quantity of protein alone required to account for the observed density dot:; not exceed the quantity which actually appears to be present by electrophoretic analysis. The remarkably large number (and total quantity) of proteins in these ribosomes raises some interesting questions about the similarities and differences that must exist between these particles and other structural kinds of ribosomes with respect to their biosynthesis, assembly, and detailed functional activities. For example, it seems possible that these ribosomes, in which the ratio of RNA to protein is only 1:2, are held together predominantly by different kinds of intermolecular bonding interactions than are found in E. coli ribosomes, which have an RNA: protein ratio of about 1:0.6. Clearly the structure of the mitoribosome must involve more prot cin-protein interactions and

PAGE 144

133 fewer protein-RNA interactions than that of the E. coli ribosome. Proteins which bind directly to polynucleotides, such as histones and many ribosomal proteins, frequently have exceedingly high isoelectric points; presumably a larg< part of the energy of binding in these cases is due to electrostatic attraction (and lack of electrostatic repulsion) between charged residues of the proteins and the phosphate backbone of the nucleic acid. Ioni< bonding between two proteins, on the other hand, requires the interaction of botii acidic and basic amino acids in the proteins. Thus the high protein content of animal mitochondrial ribosomes may be related to the lower basicity which is observed in their proteins. It may be predicted that other very protein-rich ribosomes, like the mitochondrial ribosomes of Tetrahymen a (Chi and Suyama, 1970) and Candida utilis (Vignais et al., 1972) , will likewise contain relatively acidic ribosomal proteins. Since hydrophobic interactions are probably more important in stabilizing protein-protein associations than protein-RNA binding, it seems reasonable that this kind of interaction plays a larger role in the fundamental structural organization of very protein-rich ribosomes. One characteristic of hydrophobic interactions is that they are resistant to disruption by high salt concentrations, whereas electrostatic interactions are highly susceptible. Thus it is of interest that most of the proteins of bovine mitochondrial ribosomes are found to be much more resistant to salt-stripping than those of the cytoplasmic ribosomes (O'Brien et al., 1976). These mitoribosomes thus share some properties

PAGE 145

I S4 (though to a lesser extent) with the ribosomes of the extreme halophile Halobacterium cutirubrum : the latter are structurally and functionally stable at the normal intracellular salt concentration of 4 molar, and their proteins have isoelectric points much lower than even those of animal mitoribosomes (Bayley, 1966). The large number of electrophoretic differences observed between the mitoribosomal proteins of two mammalian species complements the previous reports of even larger differences in the structural properties ui mitoribosomes from more distantly related organisms. Taken together, these observations clearly indicate a relatively high rate of evolutionary divergence among these ribosomes, and invite a consideration of the possible cellular, genetic, or molecular bases of this high evolutionary rate. Perhaps the simplest explanation for this phenomenon is the possibility that mitochondrial ribosomes are not subject to very strong selective pressures in nature. Relative to the cytoplasmic ribosomes, mitoribosomes make only a very few different proteins. Although these proteins are major (and essential) components of the mitochondrion, their total quantity is also much less than the total quantity of cellular proteins synthesized on cytoplasmic ribosomes. Only a small number of ribosomes in each mitochondrion are required to keep up with the work load of protein synthesis for mitochondrial growth aad turnover. Thus it would seem that a diminished efficiency due to the accumulation of slightly disadvantageous mutations might be more

PAGE 146

135 tolerable in the mitochondrial ribosomes; the lower efficiency could be compensated by increasing the number of riboson:-,*s, at only a modest expense in terms of tbe metabolic energy required to synthesize the extra ribosomes. In this regard it is necessary to consider also the ribosomes of chloroplasts, since the role of tbese ribosomes in cellular protein synthesis is similar to that of mitochondrial ribosomes. It is noteworthy that 'lie structural properties of these ribosomes indicate little evolutionary divergence among distantly related chloroplastcontaining protists and plants, or even between these organisms and prokaryotes. This observation argues against the significance of a low level of selective pressure as a contributing factor in the high evolutionary rate of mitochondrial ribosomes. However the argument is weakened by the fact that chloroplast ribosomes are responsible for a somewhat larger proportion of the total protein synthesis in the cell than mitochondrial ribosomes are: one subunit of the very abundant protein ribulose-diphosphate carboxylase is made on chloroplast ribosomes. Moreover, the number of chloroplast ribosomes per cell is greater than that of mitoribosomes. It is also possible that genetic mechanisms exist to effect the more rapid establishment of selectively neutral or advantageous mutations in mitochondrial ribo.sumes. No . ! i ferences can be visualized in the genetic processes that operate directly on mitoribosomal versus cytoribosomal proteins, since both groups of proteins appear to be coded in

PAGE 147

136 nuclear genes. But a rapid evolutionary rate in mitochondrial rRNA might force or encourage a complementary high rate in the proteins which must interact with this rRNA to form functional ribosomes. The mitochondrial rRNA is copied from mitochondrial genes, and several fundamental differences in the genetics of nuclear and mitochondrial DNA are known. Furthermore there is considerable experimental evidence demonstrating rapid evolution.' divergence not only of mitochondrial rRNA genes, but of the rest of the mitochondrial DNA as well. Most ot this evidence is based on DNA-DNA or DNA-RNA hybridization studies, usually including thermal denaturation analyses to evaluate the level of correspondence between the heterologous molecules hybridized. Such experiments measure differences in the primary sequence of the nucleic acids, and should be roughly comparable to electrophoretic comparisons of proteins with respect to their sensitivity in detecting interspecies variations. Unfortunately a direct comparison of the degree of variation found in mitochondrial DNA with that of nuclear DNA has only been reported in one instance. Groot et al. (1975) compared the nuclear and mitochondrial DNAs of the yeasts Saccharomyces carlsbergensis , Kluyveromyces lactis , and Candida utilis , finding that both mitochondrial rRNA and total mitochondrial DNA showed greater differences among these species than did the cytoplasmic rRNA and nuclear DNA. A number of indirect comparisons can also be made. Sinclair and Brown (1971) found detectable homologies between the cytoplasmic rRNA

PAGE 148

137 of Xenopus laevis and chat of many distantly related organisms, including yeast; but no correspondence in base sequence was detectable between the mitochondrial DNAs of X. laevis and yeast (Dawid and Wolstenholme, 1968). In comparisons of rat, mouse, guinea pig, monkey and chicken, Jakovcic e_t al. (1975a and b) found a lower degree of homology among mitochondria] tRNA genes than had been reported for nuclear genes coding for rRNA or proteins; most of the rest of the mitochondrial DNA showed even less homology. No significant differences were detected between the nuclear rRNA genes of X. laevis and X. mull eri, although the spacer DNA interspersed among these rRNA genes showed extensive non-homology (Brown et al. , 1972). Mitochondrial rRN.'. genes, on the other hand, were distinguishable in hybridization comparisons between these two species, and most of the remaining mitochondrial DMA showed a significantly higher degree of evolutionary divergence (Dawid, 1972). indeed, the high level of non-homology found between these two closely-related species in the bulk of the mitochondrial DNA sequences led Dawid (1972) to propose that these sequences represent nothing but spacer DNA, analogous to the spacer regions found in nuclear ribosomal DNA. But the weight of several other kinds of evidence contradicts this proposal, indicating that most of this mitochondrial DNA represents the structural genes for enzyme proteins (Schatz and Mason, 1974). Thus it must be concluded that all kinds of mitochondrial genes can show considerable divergence between closely related species. All these results are in contrast to

PAGE 149

138 those obtained in chloroplast studies, in which hybridization experiments revealed little difference in the base sequences of chloroplast rRNA from various monocotyledonous and dicotyledonous plants (Thomas and Tewari, 1974). What kinds of genetic mechanisms are known that might account for the high evolutionary rate of mitochondrial DNA? Among the known differences in the genetic processes acting on the mitochondria and nuclei of eukaryoiic cells is the phenomenon of vegetative segregation of mitochondrial genes. Yeast zygotes "heteroplasmic" for a mitochondrial gene (that is, containing both mutant and wild-type mitochondrial DNA molecules) give rise to some homoplasmic progeny within one or two mitotic cycles, and all descendants are homoplasmic within 10-20 cell divisions (Birky, 1976). Groot et_ al . (1975) have pointed out that this process would lead to the rapid phenotypic expression of mitochondrial mutations, thereby promoting the establishment of these mutations if they were positively selective. The latter authors have also postulated a mechanism for a relatively rapid establishment of selectively neutral mutations in mitochondrial genes. The presence of many copies of a gene per cell increases the probability of mutation in the gene. Multiple copies of genes for both mitochondria] and cytopJ.ismic i i
PAGE 150

139 would be expected to show a higher mutation rate. Some objections may be raised to this line of reasoning, however. First, the spacer regions present in the nuclear rDNA are considered to be subject to the same correction process as the rRNA genes, because both are part of the same repeating DNA sequence and both are essentially identical in all of the copies present in a given cell; but the spacer regions show very rapid evolutionary divergence, as noted above (Brown et al . , 1972). Second, all of the mitochondrial DNA molecules in a given organism are found to be identical to e:,.h other (Dawzd, 1972; PotLer et al., 1975), suggesting that some kind of correction mechanism may operate on the mitochondrial genes too (Birky, 1976). Another unique feature of mitochondrial genetics is the high rate of recombination observed between mitochondrial DNA molecules. In yeast the rate is high enough that no linkage can be observed among four gene loci, all of which are physically located on the same 5 x 10 d circular DNA molecule (Dujon et al., 1974). These results were interpreted to mean that the mitochondrial genomes undergo many repeated rounds of pairing and recombination. These events do not even appear to be limited in time to a short interval in the life cycle of the organism, such as meiosis. Such high rates of recombination should promote the spread of mutations through the population, and permit the combination of separate mutations into a single genome. Mitochondrial recombination can also occur in mammalian cells: Coon et_ al. (1973) have presented biochemical evidence for the forma-

PAGE 151

140 tion of recombinant mitochondrial DNA molecules in somatic cell hybrids of mouse and human origin. But the importance of this phenomenon in the exchange of mitochondrial genetic information between individuals in nature is questionable for mammals and most other organisms. In most species (unlike the case in yeast), mitochondrial genes are inherited uniparentally, affording little opportunity for recombination between the DNAs of the two parents. However some recombination can be detected be! ween the parental chloroplast genomes in Chlamydomonas crosses, despite the fact that chloroplast genes in this organism are also uniparentally inherited (Sager, 1972). The high recombination rate of mitochondrial DNA may be more significant as a possible inducer of mutations. The model of Dujon et al. (1974) includes a suggestion that mitochondrial recombination could account for the properties of the petite mutants which arise spontaneously at a rather high frequency in yeast. Faulty alignment of the DNA molecules during recombination might also lead to less dramatic mutations on occasion. In addition, mitochondrial DNA may be more mutable than nuclear DNA because of differences in repair processes. Yeast mitochondria are reported to lack an excision-repair system for thymidine dimers induced by ultraviolet radiation (Prakash, 1975). Mammalian mitochondria lack both this system and the photoreactivation system (Clayton et al. , 1974) . There is certainly no shortage of possible mechanisms to account for the rapid evolutionary divergence of mitochondrial DNA and mito-

PAGE 152

141 chondrial ribosomes. The question is which of these possibilities actually represent significant factors contributing to this divergence. Answers to this question can only come from further studies on the cellular and molecular processes acting on cytoplasmic genes. Particularly instructive would be a comparison of such processes between mitochondria and chloroplasts, since the latter show a much lower rate of evolutionary divergence. whatever the causes, the phylogenetic diversity of mitochondrial ribosomes should prove to be a useful tool for the detailed analysis of the biosynthesis, assembly, and mechanism of action of ribosomes in general. Important information about the relationships between structure and function in biological systems can frequently be derived from comparisons between naturally occurring systems which have the same function but differ structurally. Mitochondrial ribosomes offer a great variety of structural differences with considerable similarity in the details of their functional activity, and thus should be most valuable for such comparative studies.

PAGE 153

BIBLIOGRAPHY Aaij, C. N., N. Nanninga and P. Borst, 1972. The structure of ribosomelike particles from rat liver mitochondria. Biochim. Biophys. Acta 277:140-148. Agsteribbe, E. , R. Datema and A. Kroon, 1974. Mitochondrial polysomes from Neu rospora crassa . In The Biogenesis of Mitochondria , edited by \. Kroon and C. Saccone, Academic Press, New York, pp. 305-314. Avadhani, 'A. G. and D. E. Buetow, 1972. Isolation of active polyribosomes from the cytoplasm, mitochondria, and chloroplasts of Euglen a gracilis . Biochem. J. 128:353-365. Avadhani, N. G. and D. E. Buetow, 1974. Mitochondrial and cytoplasmic ribosomes: Distinguishing characteristics and a requirement for the homologus ribosomal salt-extractable fraction for protein synthesis. Biochem. J. 140:73-78. Bartoov, B., R. S. Mitra and K. B. Freeman, 1970. Ribosomal-type ribonucleic acid from rodent mitochondria. Biochem. J. 120:455-466. Bayley, S. T. , 1966. Composition of ribosomes of an extremely halophilic bacterium. J. Mol. Biol. 15^420-427. Bickle, T. A. and R. R. Traut, 1971. Differences in size and number of 80S and 70S ribosomal proteins by dodecyl sulfate gel electrophoresis. J. Biol. Chem. 246.: 6828-6834 . Birky, C. W. , 1976. The inheritance of genes in mitochondria and chloroplasts. Bioscience 26:26-33. Bodley, J. W. , F. J. Zieve, J. Lin and T. S. Zievel, 1970. Studies on translocation. III. Conditions necessary for the formation and detection of a stable ribosome-G factor-guanosine diphosphate complex in the presence of fusidic acid. J. Biol. Chem. 245 :5656-5661. Bonner, J. D. and J. E. Varner, 1965. Plant Biochemistry , Academic Press, New York. 14 2

PAGE 154

143 Bonner, W. M. and R. A. Laskey, 1974. A film detection method for tritium-labelled proteins and nucleic acids in polyacrylamide gels. Euro. J. Biochem. 46:83-88. Borst, P. and L. Grivell , L971. Mitochondrial ribosomes. FEBS Letters 13:73-88. Bourque, D. P., J. E. Boynton and N. W. Cillham, 1971. Studies on the structure and cellular location of various ribosomes and ribosomal RNA species in the green alga Chlamydomonas reinhardi . J. Cell Sci. 8:153-183. Brouwer, J. ind R. J. Planta, 1975. The origin of high molecular weight protein:-: in ribosomal preparations of Bacillus licheniformis . FEBS LeLters ^ : 73-75. Brown, D. D., P. C. Wensink and E. Jordan, 1972. A comparison of the ribosoi.utl DNA's of Xenopus laevis and Xenopus mulleri : the evolution of tandem genes. J. Mol. Biol. 63:57-73. Rrunk, C. F. and V. Leick, 1969. Rapid equilibrium isopycnic CsCl gradients. Biochim. Biophys. Acta 179:136-144. Cammarano, P., A. Felsani, A. Romeo and F. M. Alberghina, 1973. Particle weights of active ribosomal subunits from Neuro s pora crassa . Biochim. Biophys. Acta 308 :404-411. Chi, J. C. H. and Y. Suyama, 1970. Comparative studies on mitochondrial and cytoplasmic ribosumes of Tetrahymena pyriformis . J. Mol. Biol. 53:531-556. Chrambach, A., P. Doerr, G. R. Finlayson, L. E. M. Miles, R. Sherins and D. Roland, 1973. Instability of pH gradients formed by isoelectric focusing in polyacrylamide gel. Ann. N.Y. Acad. Sci. 209 :44-64. Clayton, D. A., J. N. Doda and E. C. Friedberg, 1974. The absence of a pyrimidine dimer repair mechanism in mammalian mitochondria. Proc. Natl. Acad. Sci. USA 71:2777-2781. Coon, H. C, I. Horak and I. B. Dawid, 1973. Propagation of both parental mitochondrial DNAs in rat-human and mouse-human hybrid cells. J. Mol. Biol. 81_:285-298. Cooper, C. and C. Avers, 1974. Evidence of involvement of mitochondrial polysomes and messenger RNA in synthesis of organelle proteins. In The Biogenesis of Mitochondria , edited by A. Kroon and C. Saccone, Academic Press, New York, pp. 289-303.

PAGE 155

144 Coote, J. L., T. H. Rabbitts and T. S. Work, 1971. The mitochondrial ribosomes of baby hamster kidney cells. Biochem. J. 123:279-281. Creusot, F. , J. Delaunay and G. Schapira, 1975. Eukaryotic ribosomal proteins. Molecular weights of proteins from rabbit liver subunits. Biochimie 5_7_: 167-17 3 . Czempiel, W. , J. Klose and R. Bass, 1976. Mammalian mitochondrial ribosomes: Characterization of ribosomal proteins by two-dimensional gel electrophoresis. FEBS Letters 62:259-262. Darnell, J. E., 1968. Ribonucleic acid from animal cells. Bacteriol. Rev. 32: 162-290. Dawid, I. ri. , 1972. Evolution of mitochondrial DNA sequences in Xenopus. Dev. Biol. 29:139-151. Dawid, I. B., D. D. Brown and R. H. Reeder , 1970. Composition and structure of chromosomal and amplified ribosomal DNA's of Xenopus laevis . J. Mol. Biol. 51:341-360. Dawid, I. B. and J. W. Chase, 1972. Mitochondrial RNA in Xenopus lae vis. II. Molecular weights and other physical properties of mitochondri;.l ribosomal and 4S RNA. J. Mol. Biol. 63:217-231. Dawid, I. B. and D. R. Wolstenholme, 1968. Renaturation and hybridization studies of mitochondrial DNA. Biophys. J. 8:65-81. Delaunay, J., F. Creusot and G. Schapira, 1973. Evolution of ribosomal proteins. Euro. J. Biochem. 39:305-312. Delaunay, J., F. Creusot and G. Schapira, 1974. Characterization of the proteins extracted with KC1 from eukaryotic ribosomes. Biochimie _56:189-191. Delaunay, F. and G. Schapira, 1974. Phylogenic distance between prokaryotes and eukaryotes as evaluated by ribosomal proteins. FEBS Letters 40:97-100. Denslow, N. D. and T. W. O'Brien, 1974. Susceptibility of 55S mitochondrial ribosomes to antibiotics inhibitory to prokaryotic ribosomes, lincomycin, chloramphenicol, PA114A. Biochem. Biophys. Res. Comm. 57:9-16. de Vries, H. , E. Agsteribbcand A. M. Kroon, 1971. The "Fragment Reaction": A tool for the discrimination between cytoplasmic and mitochondrial ribosomes. Biochim. Biophys. Acta 246:111-122.

PAGE 156

145 de Vries, H. and A. Kroon, 1974. Physiuchemical and functional characterization of the 55S ribosomes from rat liver mitochondria. In The Biogene sis of M itochondria , edited by A. Kroon and C. Saccone, Academic Press, New York, pp. 357-365. de Vries, H. and R. van der Koogh-Schuuring, 1973. Physiochemical characteristics of isolated 55S mitochondrial ribosomes from rat liver. Biochem. Biophys. Res. Co mm. 54^:308-314. Dubin, D. T. , 1974. Methylated nucleotide content of mitochondrial ribosomal RNA from hamster cells. J. Mol. Biol. 84:257-273. Dujon, B., P. P. Slonimski and L. Weill, 1974. Mitochondrial genetics IX: A model for recombination and segregation of mitochondrial genomes in Saccharomyc es cerevisiae . Genetics 78:415-437. Dzionara, M. , E. Kaltschmidt and H. G. Wittmann, 1970. Ribosomal proteins, XIII. Molecular weights of isolated ribosomal proteins of Escheri chia coli . Proc. Natl. Acad. Sci. USA 67:1909-1913. Edelman, M. , I. M. Verma and U. Z. Littauer, 1970. Mitochondrial ribosomal RNA from Aspergillus nidulan s: Characterization of a novel molecular species. J. Mol. Biol. 49:67-83. Fazekas de St. Groth, S., R. G. Webster and A. Datyner, 1963. Two new staining procedures for quantitative estimation of proteins on eleetrophoretic strips. Biochem. Biophys. Acta 71:377-391. Geisser, M. , G. W. Tischendorf, G. Sto'ffler and H. G. Wittmann, 1973. Immunological and electrophoret ic comparison of ribosomal proteins from eight species belonging to Enterobacteriaceae. Mol. Gen. Genet. 127:111-128. Gerding, J. J. T., A. Koppers, P. Hagel and H. Bloemendal, 1971. Cyanate formation in solutions of urea. II. Effect on the eye lens protein a-crystallin. Biochim. Biophys. Acta 243 :374-379. Gesteland, R. F. and T. Stahelin, 1967. Electrophoretic analysis of proteins from normal and cesium chloride-treated Escherichia coli ribosomes. J. Mol. Biol. 24:149-155. Gressner, A. M. and I. G. Wool, 1974. The phosphorylation of liver ribosomal proteins In vi_i.ro. J. Biol. Chem. 249 :6917-6925. Grivell, L. A., P. Netter, P. Borst and P. P. Slonimski, 1973. Mitochondrial antibiotic resistance in yeast ribosomal mutant-; resistant to chloramphenicol, erythromycin, and spiramycin. Biochim. Biophys. Acta 312:358-367.

PAGE 157

146 Grivell, L. A., L. Reijnders, and P. Borst, 1971. Isolation of yeast mitochondrial ribosomes highly active in protein synthesis. Biochim. Biophys. Acta 247:91-103. Groot, G. S. P., R. A. Flavell and J. P. M. Sanders, 1975. Sequence homology of nuclear and mitochondrial DNAs of different yeasts. Biochi,;. Biophys. Acta 378:186-194. Hamel, E., M. Koka and T. Nakomoto, 1972. Requirement of an Escheri chia coli 50S ribosomal protein component for effective interaction of the ribosome with T and (.' factors and with guanosine triphosphate. J. Biol. Chem. 247 :805-814. Hamilton, M. G., 1971. Isodensity equilibrium centr if ligation of ribosu.:ia.l particles; the calculation of the proiein content of ribosomes and other ribonucleoproteins from buoyant density measurements. Methods. Knzymol. 20:Part C 512-521. Hamilton, M. G. and T. W. O'Brien, 1974. Ultracentrifugal characterization of the mitochondrial ribosome and subribosomal particles of bovine ! Lver: molecular size and composition. Biochemistry 13:5400r .403. on Hamilton, M. G. and M. L. Petermann, 1959. Ultracentrif ugal studi. ribonucleoprotein from rat liver microsomes. J. Biol. Chem. 234: 1441-1448. ' Hardy, S. J. S., 1975. The stoichiometry of the ribosomal proteins of Escherichia coli . Mol. Gen. Genet. 140:253-274. Hardy, S. J. S. and C. G. Kurland, 1966. The relationship between Poly A polymerase and the r ibosomes. Biochemistry 5:3676-3684. Hartley, M. R. and R. J. Ellis, 1973. Ribonucleic acid synthesis in chloroplasts. Biochem. J. 134:249-262. Hill, W. E., G. P. Rossetti and K. E. van Holde, 1969. Physical studies of ribosomes from Escherichia coli . J. Mol. Biol. 44:263-277. Hosokawa, K. , R. K. Fujimura and M. Nomura, 1966. Reconstitut ion of functionally active ribosomes from inactive subparticles and proteins. Proc. Natl. Acad. Sci. USA 55:198-204. Howard, G. A., J. A. Traugh, E. A. Croser and R. R. Traut, 1975. Ribosomal proteins from rabbit reticulocytes: Number and molecular weights of proteins from ribosomal subunits. J. Mol. Biol. 93:391-404. Huynh-van-Tan, M. Gavrilovic and G. Schapira, 1974. Eukaryotic ribosomal proteins: The number of proteins in the subunits and their isoelectric points. FEBS letters 45:299-303.

PAGE 158

147 Jakovcic, S., J. Casey and M. Rabinowitz, 1975a. Sequence homology of the mitochondrial leucyl-tRNA cistron in different organisms. Biochemistry 14:20 i 7-2042. Jakovcic, S., J. Casey and M. Rabinowitz, 1975b. Sequence homology between mitochondrial DNA of different eukaryotes. Biochemistry 14:243-250. Johnson, J. D. and J. Horowitz, 1971. Characterization of ribosomes and RNAs from Mycoplasma hominis . Biochim. Biophys. Acta 247:262-279. Kaltschmidt, E. , 1971. Ribosomal Proteins. XIV. Isoelectric points of ribosomal proteins of E. coli as determined by two-dimensional polyacrylamide gel electrophoresis. Anal. Biochem. 43:25-31. Kaltschmidt, E. and H. G. Wittman, 1970. Ribosomal proteins. XII. Number 01 proteins in small and large ribosomal subunits of E_. coli as determined by two-dimensional gel electrophoresis. Proc. Natl. Acad. Sci. USA 67:1276-1280. Kirby, K. S., 1965. Isolation and characterization of ribosomal ribonucleic acid. Biochem. J. 96:266-269. Kleinow, W. , 1974. RNA from mitochondrial ribosomes of Locusta migratoria . In The Biogenesis of Mitochondri a, edited by A. M. Kroon and C. Saccone, Academic Press, New York, pp. 337-346. Kleinow, W. , W. Neupert and T. Bucher, 1971. Small sized ribosomes from mitochondria of Locusta migratoria . FEBS Letters 12:129-133. Kleinow, W. , W. Neupert and F. Miller, 1974. Election microscope study of mitochondrial 60S and cytoplasmic SOS ribosomes from Locusta migratoria . J. Cell Biol. 62:860-875. Klose, J., 1975. Protein mapping by combined isoelectric focusing and electrophoresis of mouse tissues. Humangenetik 26:231-243. Krawiec, S. and J. M. Eisenstadt, 1970. Ribonucleic acids from the mitochondria of bleached Euglena gracilis . Biochim. Biophys. Acta 217:132-141. Kiintzel, II. and H. Noll, 1967. Mitochondrial and cytoplasmic polysomes from Neurospora crassa . Nature (London) 215 :1340-1345. Kurland, C, 1960. Molecular characterization of ribonucleic acid from Escherichi a coli ribosomes. I. Isolation and molecular weights. J. Mol. Biol. _2:83-91. Kurland, C. G., 1966. The requirements for specific sRNA binding by ribosomes. J. Mol. Biol. 18:90-108.

PAGE 159

148 Kuter, D. J. and A. Rodgers, 1974. Ribosomal protein differences between animal cells. Exptl. Cell Res. 87:186-194. Laskey, R. and A. Mills, 1975. Quantitative film detection of H and ^C in polyacrylamide gels by f luorography. Eur. J. Biochem. 56:335-341. Leaver, C. J. and M. A. Harmey, 1973. Ribosomal RNAs of higher plant mitochondria. In Ribosomes and RNA Metabolism , edited by J. Zelinka and J. Balan, Publishing House of the Slovak Academy of Sciences, pp. 407-417. Leister, D. E. and I. B. Dawid, 1974. Physical properties and protein constituents of cytoplasmic and mitochondrial ribosomes of Xenopus laevis . J. Biol. Chem. 249:5108-5118. Leister, D. E. and I. B. Dawid, 1975. Mitochondrial ribosomal proteins in "enop as laevi s/X. mulleri interspecific hybrids. J. Mol. Biol. 96:119-123. Lin, A. and I. G. Wool, 1974. The molecular weights of rat liver ribosomal proteins determined by "three-dimensional" polyacrylamide gel electrophoresis. Mol. Gen. Genet. 134:1-6. Loening, U. E. , 1968. Molecular weights of ribosomal RNA in relation to evolution. J. Mol. Biol. 3_8: 355-365. Lolling, U. E., K. W. Jones and M. L. Birnstiel, 1969. Properties of the ribosomal RNA precursor in Xenopus laevis ; comparison to the precursor in mammals and in plants. J. Mol. Biol. 45:353-366. Loewenstein, J., H. R. Scholte and E. M. Wit-Peeters, 1970. A rapid and simple procedure to deplete rat-liver mitochondria of lysosomal activity. Bioehim. Biophys. Acta 223 :432-436. Lowry, 0. !!., N. J. Rosebrough, A. L. Farr and R. J. Randall, 1951. Protein measurement with folin phenol reagent. J. Biol. Chem. 193:265-275. Lyttleton, J. W. , 1962. Isolation of ribosomes from spinach chloroplasts. Exp. Cell Res. 26:312-317. McConkey, E. H. , 1974. Composition of mammalian ribosomal subunits: A re-evaluation. Proc. Natl. Acad. Sci. USA 71:1379-1383.

PAGE 160

149 Maizel, J. W. , 1971. Polyacrylamide gel electrophoresis of viral proteins. In Method s in Virology, Vol 5. Chap. 5., edited by K. Maramorosch and li. Kaprowski, Academic Press, New York. Malkin, L. I., 1971. Minor species of ribonucleic acid associated with rat liver mitochondria. Biochemistry 10:4752-4756. Mans, R. J. and G. D. Novelli, 1960. A convenient, rapid and sensitive method for measuring the incorporation of radioactive amino acids into protein. Biochem. Biophys. Res. Comm. 3^:540-543. Martini, 0. H. W. and H. J. Gould, 1971. Enumeration of rabbit reticulocyte ribosomal proteins. J. Mol. Biol. 62:403-405. Martini, 0. H. W. and H. J. Gould, 1975. Molecular weight distribution of ribosomal proteins from several vertebrate species. Mol. Gen. Genet. 142:317-331. Morimoto, H. and H. 0. Halvorson, 1971. Characterization u£ mitochondrial ribosomes from yeast. Proc. Natl. Acad. Sci. USA 68:324-328. Morimoto, H., A. H. Scragg, J. Nekhoroceff, V. Villa and H. 0. Halvorson, 1971. Comparison of the protein synthesizing systems from mitochondria and cytoplasm of yt ast. In Autonomy and Biogenesis of Mitochondria and Chloroplasts , edited by N. K. Boardman, A. W. Linnane, and R. M. Smillie, American Elsevier, New York, pp. 282-292. Morrison, C. A., R. A. Garrett, H. Zeichhardt and G. Stoffler, 1973. Proteins occurring at, or near, the subunit interface of E. coli ribosomes. Molec. Gen. Genet. 127:359-368. Neupert, W. , W. Sebald, A. J. Schwab, A. Pfaller and T. Bucher, 1969. Puromycin sensitivity of ribosomal label after incorporation of Clabeled amino acids into isolated mitochondria from Neurospora crassa . Eur. J. Biochem. 10; 58 5-588. O'Brien, T. W. , 1971. The general occurrence of 55S ribosomes in mammalian liver mitochonh La. J. Biol. Chem. 246 : 3409-3417 . O'Brien, T. W. , 1972. Occurrence of 55S miniribosomes in mitochondria of the shark. J. Cell Biol. 5J3_:191a. O'Brien, T. W. , N. D. Denslow and G. R. Martin, 1974. The structure, composition, and function of 55S mitochondrial ribosomes. In The Biogenesis of Mitochondria , edited by A. Kroon and C. Saccone, Academic Press, New York, pp. 347-356.

PAGE 161

150 O'Brien, T. W. and G. F. Kalf, 1967. Ribosomes from rat liver mitochondria. II. Partial characterization. J. Biol. Chem. 242:2180-2185. O'Brien, T. W. and D. E. Matthews, 1976. Mitochondrial ribosomes. In Handbook of Genetics : Vol . 5Molecular G enetics , edited by R. C. King, Plenum Press. New York, pp. 535-580. O'Brien, T. W. , D. E. Matthews and N. D. Denslow, 1976. Protein composition of the bovine mitochondrial ribosome. In The Genetics and Bi ogenes is of Chloroplasts and Mitochondria , edited by T. Bucher, W. Neupert, W. Sebald , and S. Werner. North Holland, Amsterdam, in press. O'Farrell, P. H. , 1975. High resolution two-dimensional electrophoresis of proteins. J. Biol. Chem. 250 :4007-4021. Olsnes, ., 1971. The isolation of polysomes from rat liver: contamination by membrane proteins with high affinity to ribosomes. Biochim. Biophys. Acta 232:705-716. Perlman, S. and S. Penman, 1970. Protein-synthesizing structures associated with mitochondria. .. Lure (London) 227 : 1 33-137 . Pettersson, T., S. J. S. Hardy and A. Liljas, 1976. The ribosomal protein L8 is a complex of L7/12 and L10. FEBS Letters 64:135-138. Potter, S. S., J. E. Newbold , C. A. Hutchinson III, and M. H. Edgell, 1975. Specific cleavage analysis of mammalian mitochondrial DNA. Proc. Nail. Acad. Sci. USA 7_2: 4496-4500. Prakash, L., 1975. Repair of pyrimidine dimers in nuclear and mitochondrial DNA of yeast irradiated with low doses of ultraviolet light. J. Mol. Biol. 98:781-795. Pring, D. R. , 1974. Maize mitochondria: Purification and characterization of ribosomes and ribosomal ribonucleic acid. Plant Physiol. 53:677-683. Pring, D. R. and D. W. Thornbury, 1975. Molecular weights of maize mitochondrial and cytoplasmic ribosomal RNAs under denaturing conditions. Biochim. Biophys. Acta 383 :140-146. Rabbitts, T. H. and T. S. Work, 1971. The mitochondrial ribosome and ribosomal RNA of the chick. FEBS Letters 14:214-218.

PAGE 162

151 Rawson, J. R. and E. Stutz, 1969. Isolation and characterization of Euglena gracili s cytoplasmic and chloroplast ribosomes and their ribosomal RNA components. Biochim. Biophys. Acta 190:368-380. Reijnders, L., P. Sloof, J. Sival and P. Borst, 1973. Gel electrophoresis of RNA under denaturing conditions. Biochim. Biophys. Acta 324:320-333. Reisfeld, R. A., U. J. Lewis and D. E. Williams, 1962. Disk electrophoresis of basic proteins and peptides on polyacrylamide gels. Nature 19_5 :281-282 . Robberson . '). , Y. Aloni, G. Attardi and N. Davidson, 1971. Expression of the r ; .ochondrial genome in HeLa cells. VI. Size determination of mitochondrial ribosomal RNA by electron microscopy. J. Mol. Biol. 60:473-4^4. Sacchi, A., F. Cerbone, P. Cammarano and U. Ferrini, 1973. Physiochemical characterization of ribosome-like (55S) particles from rat liver mitochondria. Biochim. Biophys. Acta 308 :390-403. Sager, R. , 1972. Cytoplasmic Genes and Organelles. Academic Press, New York. Sager, R. and M. G. Hamilton, 1967. Cytoplasmic and chloroplast ribosomes of Chlamydomonas : Ultraceni rifugal characterization. Science 157:709-711. Schatz, G. and T. L. Mason, 1974. The biosynthesis of mitochondrial proteins. Ann. Rev. Biochem. 43:51-87 . Schmitt, H. , 1970. Characterization of a 72S mitochondrial ribosome from Saccharomyces cere v isiae. Eur. J. Biochem. 17 : 278-283. Schmitt, H., H. Grossfeld, J. S. Beckmann and U. Z. Littauer, 1974. Biogenesis of mitochondria from Artemia salina cysts and the transcription in_ vivo of the DNA. 1 n The Biogenes is of_ Mitochondr ia, edited by A. M. Kroon and C. Saceone, Academic Press, New York, pp. 135-146. Schnaitman, C. and J. W. Greenawal t , 1968. Enzymatic properties of the inner and outer membranes of rat liver mitochondria. J. Cell Biol. 38:158-175. Sherton, C. C. and I. G. Wool, 1974. A comparison of the proteins of rat skeleLal muscle and Liver ribosomes by two-dimensional polyacrylamide gel electrophoresis. J. Biol. Chem. 249:2258-2267.

PAGE 163

152 Sinclair, J. H. and D. D. Brown, 1971. Retention of common nucleotide sequences in the ribosomal deoxyribonucleic and of eukaryotes and some of their physical characteristics. Biochemistry 10:2761-2769. Spencer, M. and F. Poole, 1965. On the origin of crystali zable RNA frcm yeast. J. Mol. Biol. Ll:315-326. Spitnik-Elson, P. and A. Atsmon, 1969. Detachment of ribosomal proteins b\ salt. I. Effect of conditions on the amount of protein detacheJ. J. Mol. Biol. 45:113-124. Staehelin, T., D. Maglott and R. E. Monro, 1969. On the catalytic center of peptidyl transfer: A part of the 50S ribosome structure. Cold Spring Harbor Symposium 34 :39-48. Stutz, E. and A. Boschetti, 1976. Chloroplast ribosomes. In Handbook of Gene tics: Vol . 5Molec u lar G enetics , edited by R. C. King, Plenum Press, New York, pp. 425-449. Subramanian, A. R. , 1974. Sensitive separation procedure for Escherichia coli ribosomal proteins and the resolution of high-molecular-weight components. Eur. J. Biochem. 45: 541-546. Swanson, R. F. and I. B. Dawld, 1970. The mitochondrial ribosome of Xenopus laevis . Proc. Natl. Acad. Sci. USA 66: 117-124. Tappel, A. L., 1973. Lipid peroxidation damage to cell components. Fed. Proc. 12:1870-1874. Taylor, M. M. and R. Storck, 1964. Uniqueness of b.icterial ribosomes. Proc. Nail. Acad. Sci. USA 52_: 958-965. Terhorst, C. , W. Mb'ller, R. Laurs. n and B. Wittmann-Liehold, 1972. Amino acid sequence of a 50S ribosomal protein involved in both EFG and EFT dependent GTP-hydrolysis . FEBS Letters _28: 325-328. Thomas, J. R. and K. K. Tewari, 1974. Conservai ion of 70S ribosomal RNA genes in the chloroplast DNAs of higher plants. Proc. Natl. Acad. Sci. USA 7_1: 314 7-3151. Tischcndorf, G. W. , G. Stoffler and H. G. Wi . imann, 1974. The comparative characterization of the proteins of Escherichia coli 30S and 50S ribosomal subunits. The identity of S20 and L26. Eur. J. Biochem., in pre.-;s. Tissieres, A., J. Watson, D. Schlessinger , and B. Hollingworth, 1959. Ribonucleoprotein particles from Escherichia coli . J. Mol. Biol. 1:221-234.

PAGE 164

153 Ts'o, P. 0. P., J. Bonner and J. Vinograd, 1958. Structure and properties of microsomal nucleoprotein particles from pea seedlings. Biochim. Biophys. Acta 30:570-582. van den Bogert, C. and H. de Vries, 1976. The mitochondrial ribosomes of Neurospora crassa. II. Comparison of the proteins from Neuros pora mitochondrial ribosomes with ribosomal proteins from Neurospora cytoplasm, rat-liver mitochondria and from bacteria. Biochim. Biophys. Acta, in press. Vasconcelos, A. C. L. and L. Bogorad, 1971. Proteins of cytoplasmic, chlorop! ,>L and mitochondrial ribosomes of some plants. Biochim. Biophys. Acta 228:492-502. Verma, I. M., M. Edelm.m, M. Herzhog and U. Littauer, 1970. Size determi;: ition of mitochondrial ribosomal RNA from Aspergillus nidulans by election microscopy. J. Mol. Biol. 52:137-140. Vesco, C. and S. Penman, 1969. The cytoplasmic RNA of HeLa cells: New disci etc species associated with mitochondria. Proc. Natl. Acad. Sci. USA 62:218-225. Vignais, P. V., B. J. Stevens, J. Huet and J. Andre, 1972. Mitoribosomes from Candida utilis . Morphological, physical and chemical characterization of the monomer form and its subunits. J. Cell Biol. 54:468-492. Voynow, P. and C. G. Kurland, 1971. Stoichiometry of the 30S ribosomal proteins of Escherichia coli . Biochemistry 10:517-524. Weber, H. J., 1972. Stoichiometric measurements of 30S and 50S ribosomal proteins. Mol. Gen. Genet. 119:233-248. Wengler, G., G. Wengler and K. Scherrer, 1972. Ribonucleoprotein particles in HeLa cells: The contamination of the postmitochondrialcytoplasmic fraction from HeLa cells with ribonucleoprotein particles of mitochondrial origin. Eur. J. Biochem. 24:477-484 . Wilson, R. H., J. B. Hanson, and H. H. Mollenhauer, 1968. Ribosome particles in corn mitochondria. Plant Physiol. 43:1874-1877. Wittmann, H. G. , 1974. Purification and identification of Escherichia coli ribosomal proteins. In Ribosomes , edited by M. Nomura, A. Tissieres, and P. Lengyel, Cold Spring Harbor Laboratory, Cold Spring Harbor, pp. 93-114.

PAGE 165

154 Wittmann, H. G., G. Stoffler, E. Kaltschmidt, V. Rudloff, H. Janda, M. Dzionara, D. Bonner, K. Nierhaus, M. Cech, T. Hindennach and B. Wittmann, 1970. Protein chemical and serological studies on ribosomes of bacteria, yeast and plants. FEBS Symp. 21:33-46. Wittmann, H. G., G. Stoffler, W. Piepersberg, P. Buckel, D. Ruffler and A. Bock, 1974. Altered S5 and S20 ribosomal proteins in revertants of an alanyl-tRNA-synthetase mutant of Eschericia coli . Mol. Gen. Genet. 134:225-236. Wool, I. G. and G. Stoffler, 1974. Structure and function of eukaryotic ribosomes. In Ribosomes, edited by M. Nomura, A. Tissieres, and P. Lengyel, Cold Spring Harbor Laboratory, Cold Spring Harbor, pp. 417-460. Yurina, N. P. and M. S. Odintsova, 1974. Buoyant density of chloroplast ribosrmes in CsCl. Plant. Sci. Lett. 3_ :2 29-234. Zitomer, R. S. and J. G. Flaks, 1972. Magnesium dependence and equilibrium of the Escherichia coli ribosomal subunit association. J. Mol. Biol. 71:263-279.

PAGE 166

BIOGRAPHICAL SKETCH David Earl Matthews was born May 10, 1948, in Birmingham, Alabama. He was raised in Alabama, Georgia, and Florida, and graduated from Lakeland High School, Lakeland, Florida in 1966. His undergraduate education was at Davidson College and Florida Southern College, with support from a National Merit Scholarship. In 1972 he received a B.S. degree in Chemistry, summa cum laude, from Florida Southern College. Shortly thereafter he was married to Patty S. Salisbury of St. Petersburg, Florida, His graduate studies at the University of Florida have been supported by a National Science Foundation Graduate Fellowship and a Graduate Council Fellowship. 155

PAGE 167

I certify that I have read this study and that, in my opinion, it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^kf>^ Thomas W. O'Brien, Chairman Associate Professor of Biochemistry and Molecular Biology I certify that I have read this study and that, in my opinion, it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Md/ a^ Robert J>. Cohen Associate Professor of Biochemistry and Molecular Biology I certify that I have read this study and that, in my opinion, it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. /6 ( L. O'Neal Ingram Assistant Professor of Microbiology

PAGE 168

I certify that I have read this study and that, in my opinion, it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Rusjty J. Mans Projfesqor/ of Biochemistry and' Molecular Biology I certify that I have read this study and that, in my opinion, it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. R. Michael Roberts Professor of Biochemistry and Molecular Biology This dissertation was submitted to the Graduate Faculty of the Department of Biochemistry in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December l97o Dean, Graduate School

PAGE 169

UNIVERSITY OF FLORIDA 3 1262 08553 3049


PROTEINS OF THE MAMMALIAN MITOCHONDRIAL RTBOSOME
by
DAVID EARL MATTHEWS
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1976

ACKNOWLEDGEMENTS
The author gratefully acknowledges the financial support
awarded him in the form of fellowships from the National Science
Foundation and the Craduate Council.
Thanks are also due to Mark Critoph, Warren Clark, and Mark
Moffitt for their technical assistance in these experiments; and to
Nancy Denslow, Robert Hessler, and Mary Conde for helpful discussions.
The senior research project of Dean Kane contributed much to the
studies on the isoelectric points of mitochondrial ribosomal proteins.
The encouragement and assistance of Patty Matthews in the preparation
of this dissertation were invaluable.
The author also wishes to express his gratitude for the helpful¬
ness of the faculty of the Department of Biochemistry, and of his
supervisory committee in particular; their good will, advice, and
scientific insights have made the past four years of graduate study
a most rewarding experience. Most of all, the profound influence of
Dr. Thomas W. O'Brien In shaping this research, as well as the author's
scientific training, goals, and attitudes, Is humbly and gratefully
acknowledged.

TABLE OF CONTENTS
Acknowledgements il
List of Tables v
List of Figures vi
Abbreviations Used viii
Abstract ix
Introduction 1
I.Structural properties of non-mitochondrial
ribosomes 3
II. Structural properties of mitochondrial ribosomes 7
A. Mitochondrial ribosomes of protists, fungi
and plants 7
B. Mitochondrial ribosomes of animals 12
III. Phylogenetic relationships in ribosome structure 17
IV. Summary 21
Materials and Methods 23
I. Materials 23
II. Preparation of ribosomes 25
A. Bovine mitochondrial ribosomes 25
1. Method A 26
2. Method B 27
3. Method C 27
B. Rat mitochondrial ribosomes 28
C. Bovine cytoplasmic ribosomes 28
D. EL coli ribosomes 29
III. Extraction of ribosomal proteins 29
IV. Radioactive labelling of ribosomal proteins 30
A. Radioidination 30
B. Reductive methylation 31
V. Two-dimensional electrophoresis 31
VI. Isoelectric focusing 34
VII. Assays of ribosome function 35
VIII. Buoyant density determinations 36
IX. Quantitative measurements 37
Results 38
I. Preparation and characterization of
mitochondrial ribosomes 39
II. Electrophoretic analysis of mitochondrial
ribosomal proteins 47
iii

A. Criteria for the identification of
ribosomal proteins 52
B. Identification of ribosomal proteins 57
1. Large subunit 57
a. Reproducibility 61
b. Resistance to salt-washing 66
Low-salt treatment 68
High-salt treatment 72
Extreme-salt treatment 73
c. Summary of the large-subunit proteins 78
2. Small subunit 79
3. Comparison of large and snail subunits 91
C. Experimental evaluation of possible artifacts 96
III.Comparison with proteins of non-mitochondrial
ribosomes 100
IV.Molecular weights of mitochondrial ribosomal
proteins 110
V.Comparison with proteins of rat mitochondrial
ribosomes 119
Discussion 128
Bibliography 142
Biographical Sketch 155
iv

LIST OF TABLES
I. Properties of cytoplasmic ribosomes of eukaryotes 4
II. Properties of Morieran and chloroplast ribosomes 6
III. Properties of mitochondrial ribosomes 8
IV.Compositions of buffers used in preparation of
ribosomes 24
V.Functional activity of mitochondrial ribosomes
prepared with or without DEAE-cellulose 48
VI.Appearance of large-subunit proteins in separate
experiments 58
VII.Appearance of small-subunit proteins in separate
experiments 82
VIII.Molecular weights of mitochondrial large subunit
proteins Ill
IX.Molecular weights of mitochondrial small-subunit
proteins 112
X.Calculation of the protein content of bovine
mitochondrial ribosomes 115
v

LIST OF FIGURES
1. Phylogenetic relationships in ribosome structure 20
2. Preparation of mitochondrial ribosomes by Method A 41
3. Preparation of mitochondrial ribosomes by Method B 43
4. Preparation of mitochondrial ribosomes by Method C 45
5. EleofrophoreLic pattern of large-subunit proteins
fr.itii bovine mitochondrial ribosomes 50
6. Electrophoretic pattern of small-subunit proteins
from bovine mitochondrial ribosomes 51
7. Effects of treatment with buffers of various ionic
composition on the peptidyl transferase activity
and buoyant density of bovine mitochondrial large-
subunits 55
8. Schematic diagram of bovine mitochondrial large-
subunit proteins 60
9. Electrophoretic pattern of proteins from large
subunits prepared in Buffer C (Experiment 2) 63
10. Electrophoretic pattern of proteins from large
subunits prepared in Buffer C (Experiment 4) 64
11. Electrophoretic pattern of proteins from large
subunits prepared in Buffer D (Experiment 6) 65
12. Effect of high-salt treatment on the protein content
of E. coli ribosomes 67
13. Electrophoretic patterns of proteins from low-salt
treated large and small subunits 69
14. Electrophoretic pattern of proteins from large
subunits prepared in Buffer C (Experiment 3) 70
vi

15. Electrophoretic pattern of proteins from large
subunits prepared in Buffer F 75
16. Electrophoretic pattern of proteins from large
subunits prepared in Buffer G 76
17. Electrophoretic pattern of proteins from small
subunits prepared in Buffer C (Experiment 2) 80
18. Electrophoretic pattern of proteins from small
subunits prepared in Buffer C (Experiment 3) 81
19. Schemat ic diagram of bovine mitochondrial small-
subunit proteins 84
20. Electrophoretic pattern of proteins from small
subunits prepared in Buffer E (Experiment 9) 87
21. Electrophoretic pattern of proteins from small
subunits prepared in Buffer G 89
22. Relative electrophoretic positions of bovine
mitochondrial large-subunit and small-subunit
proteins 92
23. Relative eLectrophoretic positions of bovine
mitoribosomal and cytoriboscmal large-subunit
proteins 102
24. Relative electrophoretic positions of mitoribosomal
and cytoribosomal small-subunit proteins 103
25. Isoelectric focusing of mitoribosomal proteins 107
26. The pH gradient formed during isoelectric focusing
of mitoribosomal proteins 109
27. Electrophoretic pattern of proteins from rat mito¬
chondrial large subunits 120
28. Electrophoretic pattern of proteins from rat mito¬
chondrial small subunits 121
29. Schematic diagram of rat mitochondrial large-
subunit proteins 123
30. Schematic diagram of rat mitochondrial small-
subunit proteins 124
vii

ABBREVIATIONS USED
ATP
bisacrylamide
C
cytoribosome
d
DNA
EDTA
g
G
GTP
1 eu
mitoribosome
Mg Ac 2
ML
MS
mRNA
phe
poly U
POPOP
PPO
RNA
rRNA
S
SDS
TEA
TEMED
Tris
tRNA
B-ME
adenosine-5T-triphosphate
N, N1-methylenebisacrylamide
cytosine
cytoplasmic ribosome
daltons
deoxyribonucleic acid
(ethylenedinitrilo)tetraacetic acid
gravity
guanine
guanosine-5'-triphosphate
leucine
mitochondrial ribosome
magnesium acetate
mitoribosomal large-subunit protein
mitoribosomal small-subunit protein
messenger RNA
phenylalanine
polyuridylic acid
1, 4-bis[2-(5-phenyloxazolyl)]benzene
2, 5-diphenyloxazole
ribonucleic acid
ribosomal RNA
Svedberg unit
sodium dodecyl sulfate
triethanolamine
N, N, NT, N1-tetramethyl-ethylenediamine
tris(hydroxymetbyl)aminomethane
transfer RNA
B-mercaptoethanol
vii X

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida
in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
PROTEINS OF THE MAMMA MAN
MITOCHONDRIAL RIBOSOME,
By
David Earl Matthews
December, 1976
Chairman: Thomas W. O'Brien
Major Department: Biochemistry
For some years it was widely believed that all ribosomes were
either of the 70S type found in bacteria or of the 80S type occurring
in the cytoplasm of eukaryotic cells. With the discovery and subsequent
characterization of ribosomes in the mitochondria of eukaryotes, this
simple generalization had to be rejected. The mitochondrial ribosomes
of various organisms display a great diversity of structural properties,
even though they all show considerable functional homology with the 70S
prokaryotic ribosome. Animal species possess mitochondrial ribosomes
with particularly interesting physical-chemical properties, by virtue
of their high protein content relative to the quantity of RNA they
contain.
The present research is an investigation into the nature of the
individual proteins which make up the large complement of total protein
ix

found in mammalian mitochondrial ribosomes. Two-dimensional electro¬
phoresis in polyacrylamide gels was used to separate, identify, and
characterize the proteins of mitochondrial ribosomes from bovine liver.
It was found that these ribosomes contain a relatively large number of
proteins: 52 are present in the large subunit and 41 in the small sub¬
unit. Several kinds of criteria and experimental evidence argue that
these proteins are true ribosomal components rather than contaminants
in the ribosome preparations. However, only 81 electrophoretically
distinguishable proteins are present in a mixture of large and small
subunits; the possibility that some of the large-subunit proteins are
identical to proteins found in the small subunit is discussed.
The molecular weights of these proteins are similar to those of
bovine cytoplasmic ribosomal proteins, and considerably larger than
those of Escherichia coli ribosomal proteins. Comparisons of electro¬
phoretic properties show that the mitochondrial proteins are more acidic
than those of mammalian cytoplasmic or most bacterial ribosomes. The
sum of the molecular weights of the 93 mitochondrial ribosomal proteins
is somewhat greater than predicted by physical-chemical measurements of
the total mass of protein in this ribosome; to account for this differ¬
ence it is suggested that a proportion of the ribosomes as they are iso¬
lated may be lacking some of the ribosomal proteins.
Other experiments demonstrated that exposure of the ribosomes to
certain ionic conditions could remove a small number of the proteins,
with a corresponding loss of the peptidyl transferase activity of the
x

ribosomes. One or more of the proteins removed may therefore be
involved in this particular ribosomal function. Finally, comparisons
between the proteins of bovine and rat mitochondrial ribosomes revealed
a large number of differences in their electrophoretic properties. The
level of divergence in these properties is greater than that reported
in similar comparisons of cytoplasmic ribosomal proteins. This obser¬
vation parallels the great phylogenetic diversity observed in the over¬
all structural properties of mitochondrial ribosomes from more distantly
related organisms, and leads to the conclusion that these ribosomes have
diverged more widely during evolutionary history than have their extra-
o
mitochondrial counterparts located only a few Angstroms away.
The large number and low isoelectric points of the proteins of
mammalian mitochondrial ribosomes are discussed with reference to their
implications for the structural organization of these particles. Some
mechanisms which may account for the rapid evolutionary divergence of
mitochondrial ribosomes are proposed.
xi

INTRODUCTION
Mitochondria are multifunctional organelles found in all eukary¬
otic cells, their primary function being the aerobic production of ATP.
In addition to their important role in cellular energy metabolism,
mitochondria make an essential contribution to their own biogenesis.
Several of the multi-subunit enzymes of oxidative phosphorylation con¬
tain one or more subunits synthesized within the mitochondrion (Schatz
and Mason, 1974). For this purpose mitochondria possess a large com¬
plement of biosynthetic enzymes and other macromolecules, distinct from
their analogs in the nucleus and extramitochondrial cytoplasm. These
components of the mitochondrial biogenetic system include DMA, DNA and
RNA polymerases, messenger RNA, ribosomes, translation factors, transfer
RNA species and aminoacyl-tRNA synthetases. Mitochondrial DNA codes
for the ribosomal RNA and at least some of the tRNA and mRNA species
found in mitochondria, and these macromolecules are indispensable in the
biogenesis of functionally active mitochondria. On the other hand most
of the (equally essential) protein components of the mitochondrial bio¬
genetic system ribosomal proteins, factors, and enzymes appear
to be coded by nuclear DNA and synthesized on cytoplasmic ribosomes.
Thus the mitochondrial and nuclear-cytoplasmic macromolecule-synthesizing
systems must cooperate as intimately in the production of the mitochon-
1

2
drial biogenetic system itself as they do in the synthesis of the
enzymes of oxidative phosphorylation.
The ribosomes in a given organism’s mitochondria are generally
distinguishable from its cytoplasmic ribosomes on the basis of sev¬
eral functional or physical-chemical criteria. Indeed, especially in
their functional properties, mitochondrial ribosomes have been found
to be more like bacterial ribosomes than cytoplasmic ones. Some
structural similarities between mitochondrial and Moneran ribosomes
were noted early (Kiintzel and Noll, 1967), and complemented reports
of other biochemical homologies between organelles and prokaryotes
that had already aroused considerable interest in the question of the
evolutionary origin of mitochondria and chloroplasts. More recent
comparisons of the structural parameters of bacterial and mitochondrial
ribosomes have shown some similarities hut also a surprising number of
differences, both between the two groups and among mitochondrial ribo¬
somes from different species. Tims, it is true that the mitochondrial
ribosomes of most organisms studied to date sediment more slowly than
the corresponding cytoplasmic ribosomes, and that some of them have
sedimentation coefficients close to that of prokaryotic ribosomes (70S).
However, mitoribosomes from various species range in sedimentation
rates from 55S to 80S, a much wider variation than is found among
bacterial, eukaryotic-cytoplasmic, or chloroplast ribosomes obtained
from different organisms. Indeed, when all ribosomal attributes are
considered, it seems that only in mitochondria do so many different
kinds of ribosomes occur.

3
I. STRUCTURAL PROPERTIES OF NON-MITOCHONDRIAL RIBOSOMES
Before proceeding to the description of the physical and chemical
characteristics of mitochondrial ribosomes, it will be useful to sum¬
marize the characteristics of the other kinds of ribosomes that exist.
This task is simplified by the fact that the characteristics of non-
mitochondrial ribosomes are not as divergent as might be thought. They
all fit reasonably well into two large categories within which the mem¬
bers seem to share more similarities than differences. Table 1 shows
the relative homogeneity of the properties of cytoplasmic ribosomes,
whatever eukaryotic organism they are obtained from protists, fungi,
plants or animals. Despite small differences among the cytoplasmic
ribosomes from these four taxonomic kingdoms, all of them appear to be
members of a single structural class typified by a sedimentation co¬
efficient of 80S, a buoyant density of 1.57 g/cc, and rRNA molecules
of 0.7 and 1.3 million daltons containing 50 percent G + C,
The second category of non-mitochondrial ribosomes is also rela¬
tively uniform in physical and chemical properties. Described in
Table II are several prokaryotic ribosomes from E. coli, a mycoplasm,
and a blue-green alga and those of various chloroplasts. The pro¬
karyotic particles can all be adequately described by the values 70S,
1.64 g/cc, (0.56 + 1.10) x 10^ daltons, and 50% for the structural
parameters tabulated. Chloroplast ribosomes are similar in all respects
except that some of them appear to be significantly lower in buoyant
density. Perhaps it should be mentioned that this homogeneity of

Table 1. Properties of cytoplasmic ribosomes of eukaryotes.
Organism
Sedimentation
Coefficient
Buoyant Density
in CsCl, g/cc
Molecular Weight
of rRNA (x 106)
Base Compos:
of rRNA, % <
PROTISTS
Euglena
87S1
L1.53]2
(0.85 + 1.20)1
54%3
Tetrahymena
80S4
1.564
0.52 + 1.185
47%4
Chlamydomonas
83S6
[1.52]7
(0.70 + 1.28)6
FUNGI
Neurospora
77S8
1.589
0.67 + 1.2810
51% 8
Saccharomyces
80S11
1.55 12
0.72 + 1.215
53%' 3
PLANTS
Pea
80S14
1.589
0.7 + 1.31 5
52%16
ANIMALS
Toad
87S17
1.5918
0.7 + 1.51 9
59%19
Rat
83S20
1.5521
0.66 + 1.485
64%22
[ ] Calculated from the data published in the cited reference, by the
1 89
formula Buoyant Density = 0Tó¿40 (% Protein') (Hamilton> 1971)'
( ) The data were published as ultracentrifugal or electrophoretic S
values relative to Escherichia coli rRNA standards taken as 16S and
23S. To simplify the tabulation, these S values have been converted
to molecular weights according to the formula M=KSa; the constants K
and a were evaluated from the values 0.56 x 10s and 1.10 x 10(’ for
Lhe molecular weights of 16S and 23S RNA.
'Avadhani and Buetow, 1972. 12Grivell et al., 1971.
2Avadhani and Buetow, 1974. 13Morimoto and Halvorson, 1971.

3Krawiec and Eisenstadt, 1970.
**Chi and Suyama, 1970.
5Reijnders e_t jal. , 1973.
6Bourque et al., 1971.
7Sager and Hamilton, 1967.
®Kiintzel and Noll, 1967.
9Cammarano et aJL. , 1973.
10Neupert e£ al.. , 1969.
13Cooper and Avers, 1974.
14Ts'o et al., 1958.
1 5Loening at al_. , 1969.
16Bonner and Varner,
1965.
17Swanson and
Dawid,
1970.
18leister and
Dawid,
1974.
19Dawid et al.
, 1970.
20Hamilton and
Peterraatm,
21de Vries and
Kroon,
, 1974
22Kirby, 1965.
-n

Table II. Properties of Moneran and chloroplast ribosomes.
Sedimentation Buoyant Density Molecular Weight Base Composition
Source Coefficient in CsCl, g/cc of rRNA (x 10 8) of rRNA, Z G + C
BACTERIA
Escherichia coli
69S1
1.642
0.56
+
1.103
52%m
Mycoplasma hominis
71S 5
[1.63]5
(0.56
+
1.01)5
46%5
BLUE-GREEN ALGAE
Anabaena
72S6
1.637
0.55
+
1.07®
CHLOROPLASTS
Euglena
69S9
[1.67]10
(0.63
+
1.01)9
52%10
Chlorella
67S7
1.57 7
0.56
+
1.108
Pea
70S7
1.587
0.56
+
1.108
Spinach
66S* 1
0.56
+
1.0512
54%u
Parenthesized and bracketed entries as in Table i.
^issiéres et al. , 1959.
2Sacchi £t al., 1973.
3Kurland, 1960.
^Horimoto and llalvorson, 1971.
^Johnson and Horowitz, 1971.
^Taylor and Storck, 1964.
'Yurina and Odintsova, 1974.
®Loening, 1968.
9Avadhani and Buetow, 1972.
10Rawson and Stutz, 1969.
1:LLyttleton, 1962.
l2Hartley and Ellis, 1973.

7
gross structural properties almost certainly conceals a great deal of
diversity in fine structure* Differences in such properties as ribo-
somal protein electrophoretic mobilities and immunologic identities are
the rule even for prokaryotic ribosomes related to each other more
closely than are the entries of Table II (Geisser et_ aj_., 1973;
Wittmann ^t al_., 1970).
II. STRUCTURAL PROPERTIES OF MITOCHONDRIAL RIBOSONES
In contrast to the rather simple classification scheme possible for
ribosomes from all other sources, mitochondrial ribosomes do not seem to
fall into one or even a few structural categories. Reference to Ta¬
ble III shows the degree of diversity found in mitochondrial ribosomes
from different species. Many of the individual ribosome species in this
table are as distinct in physical and chemical properties from each
other as the 80S (Table I) and 70S (Table II) classes are. Furthermore,
among the protists and fungi no two genera have yet been shown to con¬
tain similar mitoribosomes, so it seems likely that many more structur¬
ally different ribosomes will be found as other species are investi¬
gated. On the other hand, considerable homology is seen among the
mitoribosomes of several species of higher animals, from locust to man.
A. Mitochondrial Ribosomes of Protists, Fungi, and Plants
The best-characterized protist mitochondrial ribosomes are those
of Euglena gracilis and Tetrahymena pyriformis. The Euglena mitoribo-
some is unusual in that it is the only mitochondrial ribosome yet
described that shows a large degree of structural homology with the

Table III. Properties of mitochondrial ribosomes.
Source
Sedimentation
Coefficient
Buoyant Density
in CsC.1, g/cc
Molecular Weight
of : Tí A (x 106)
Baise Composition
of rRNA, % G + C
PROT I STS
Euglena
71S1
ri.6i]2
(0.56 + 0.93)'
27%3
Tetrahymena
80S14
1.464
0.47 + 0.905
29%l'
fungi
Neurospora
73S6, 80S7
1.52s
0.72 + 1.289
38
Saccharoravces
72S10, 75S11, 80S12
1.641^
0.70 + 1.305
30113
Candida utilis
72S15
1.4815
0.71 + 1.2115
34115
Aspergillus
0.66 + 1.2716
32% -7
PLANTS
Maize
77S18
1.5619
0.76 + 1.2520
Eean
78S21
0.78 + 1.15 21
ANIMALS
Locust
60S22
0.28 + 0.5273
Slirimp
0.35 + 0.5024
43%24
Toad
60S25
1.4526
0.32 + 0.5823
417;2 7
Rat
55S28
1.4329
0.30 + 0.5030
47%31
He La CelLs
55S32
1.4032
0.35 + 0.5433
43%3L
Parenthesized and bracketed entries as in Table 1.
1Avadhani and Buetow, 1972. 18Pring, 1974.
2Avadhani and Buetow, 1974. 19Pring and O'Brien, 1974,
3Krawiec and Eisenstcidt, 1970. private communication.

14 Chi and Suyama, 1970.
5Reijnders et al., 1973.
6Kuntzel and Noll, 1967.
7Agsteribbe e_t^ aK , 1974.
8Agsteribbe and Kroon, 1974,
private communication
^Kc'upert et^ al. , 1969.
Schmitt, 1970.
^Cooper and Avers, 1974,
12Morimoto et_ _al_. , 1971.
1 3Grivell e_t_ al_. , 1971.
*l,Morimoto and Halvorson, 1971.
15Vignais e_t_ ctL. , 1972.
16Verma £t_ _al. , 1970.
17Edelman et al.» 1970.
20Pring and Thornbury, 1975.
21Leaver and Harmey, 1973.
22Kleinow et_ al., 1971.
23Kleinow, 1974.
211 Schmitt et_ al_. , 1974.
25Swanson and David, 1970.
28Leister and Dawid, 1974.
27Dawid and Chase, 1972.
280’Brien and Kalf, 1967.
23de Vries and Kroon, 1974.
38Sacchi e_t^ al_. , 1973.
31Bartoov et_ al. , 1970.
32Perlman and Penman, 1970.
33Robberson et_ al, , 1971.
314Vesco and Penman, 1969.

10
ribosomes of prokaryotes. The fact that only one of the many mitori-
bosomes shown in Table III is similar to bacterial ribosomes in its
physical-chemical properties is quite remarkable, in view of the
extensive functional homologies between mitochondrial and bacterial
ribosomes that have been found in every case tested. And even in this
instance the structural similarity is by no means complete. Although
the mitori hnsome of Euglena is similar to the E). coli ribosome in its
sedimentation coefficient, the molecular weights of its rRNA, and its
high buoyant density, the base composition of its rRNA (27% C, + C) is
very different.
Tetrahymena mitoribosomes are quite dissimilar from those of
Euglena, and indeed from any other ribosome yet studied. Although the
Tetrahymena mitoribosome shares with that of Euglena an exceptionally
low rRNA GC content, it more nearly resembles cytoplasmic ribosomes
with respect to its sedimentation coefficient. And its buoyant density
and rRNA molecular weight values are smaller than those of any other
ribosome except the mitoribosomes of animals.
The mitochondrial ribosomes of various fungi also show a rather
large diversity in sedimentation coefficients and buoyant densities.
Unfortunately, a nearly equal diversity in these parameters has been
reported by different research groups studying the same fungal species.
Thus the reported sedimentation coefficients range from 72S to 80S for
Saccharomyces mitoribosomes, and from 73S to 80S for those of Neuro-
spora. Although the buoyant density of a 74S form of the Saccharomyces

n
mitoriboscme was found to be 1.64 g/cc (Grivell > 1971), the
composition of the 80S version (Morimoto and Halvorson, 1971) corre¬
sponds to a buoyant density of 1.56 g/cc. There is reasonably good
evidence that the 80S ribosomes described in these species really are
mitochondrial ribosomes rather than cytoplasmic contaminants (O’brien
and Matthews, 1976). It therefore seems most likely that the contra¬
dictory reports are due to artifactual alterations of the structure
of these ribosomes by son. of the preparative procedures used. How¬
ever it is not yet clear which of the results are artifactual, so it
is difficult to compare the sedimentation coefficients and buoyant
densities of fungal mitoribosomes with those of other kinds of ribo¬
somes at present.
There is much less controversy about the properties of the rRNA
of fungal mitoribosomes. Indeed, the molecular weights and GC contents
of these molecules are quite similar in all the fungal species tested.
The CC contents are rather low, nearly as low as those found in pro-
tist mitoribosomal RNA. The molecular weights, on the other hand, are
quite high: these molecules are of about the same size as the rRNAs
found in cytoplasmic ribosomes.
The properties of mitochondrial ribosomes from higher plants are
not firmly established. The best-documented results are those shown
in Table III. It may be seen that the physical-chemical character¬
istics of these ribosomes are quite similar to those of cytoplasmic
ribosomes in general. In fact, they differ only slightly or not at

12
all from those of the cytoplasmic ribosomes found in these same
organisms (Leaver and Harmey, 1973; Pring and Thornbury, 1975; Pring
and O’Brien, 19/-*» private communication). Such results raise the
possibility that these ribosomes are actually of cytoplasmic rather
than mitochondrial origin. This interpretation is supported by the
observation that these ribosomes are inhibited by anisomycin and not
by chloramphenicol, like cytoplasmic ribosomes and unlike any other
known mitochondrial ribosomes (Pring, Denslow, and O’Brien, 1975,
private communication).
There have been isolated reports of plant mitoribosomes with
properties different from those described above. A 70S particle was
obtained from mung bean mitochondria (Vasconcelos and Bogorad, 1971),
and a 66S mitoribosome from maize (Wilson et_ al., 1968). Mitochondrial
RNA from Virginia creeper was found to be relatively small, 0.42 x 10^
and 0.84 x 10^ in molecular weight. However, the functional properties
of these ribosomes have not been described, and the uncertainty about
the nature of mitochondrial ribosomes in higher plants remains unresolved.
B. Mitochondrial Ribosomes of Animals
In contrast to the fascinating and perplexing diversity of mito¬
chondrial ribosomes from protist, fungal and plant species, Metazoan
mitoribosomes appear strikingly uniform in structure. Although most
of the animal species investigated have been mammalian, the mitoribo¬
somes of the toad Xenopus have also been thoroughly characterized, and
enough data have been presented for several invertebrates to justify

li
a tentative conclusion that the mitoribosomes of all multi-cellular
animals may be quite similar. Furthermore, this relatively liomo-
gSBeous group of ribosomes is distinctly different in structural
properties from any other ribosomes yet described.
The first difference to be noticed was the low sedientation
coefficient of animal mitoribosomes. Values within the range 54S to
61S have been obtained for these particles from rat (O'Brien and Kalf,
1967), IleLa cells (Perlman and Penman, 1970), rabbit, pig, cow (O'Brien,
1971), hamster (Coote at al_., 1971), chicken (Rabbitts and Work, 1971),
toad (Swanson and Dawid, 1970), shark (O'Brien, 1972), and locust
(Kleinow e_t al^. , 1971). The early inference drawn from these low
sedimentation coefficients was that animal mitoribosomes were smaller
(lower in molecular weight) than other ribosomes (Borst and Crivell,
1971), and this inference was reinforced by the later observations of
unusually small rRNA in these particles. An alternate possibility,
that the 55S particle is actually a subunit of the functional mono¬
ribosome, has been excluded by the dissociation of the 55S structure
into two subunits and the demonstration that either the 55S particle
or a mixture of both subunits is competent for poly U-dependent
phenylalanine incorporation (Leister and Dawid, 1974; O'Brien et al.,
1974).
But there is a third interpretation of the low sedimentation
coefficient of animal mitoribosomes, because in point of fact these
55S particles are not exceptionally small, either in molecular weight

14
or in physical dimensions. The particle weight of the bovine mito-
ribosome as determined by high-speed equilibrium centrifugation is
2.8 million daltons (Hamilton and O'Brien, 1974), slightly greater
than the value reported for the ribosome of E. coll (Hill et al.,
1969). De Vries and Kroon (1974) have presented evidence that rat
mitoribosomes are even larger in volume than the E. coli particles,
though smaller than rat cytoribosomes. These investigators electro-
phoresed ribosomes into gels composed of a gradient of polyacrylamide
concentration until the particles could make no further progress
through the decreasing pore size of the gel matrix. The mitoribo¬
somes penetrated farther into the gels than cytoribosomes but not as
far as bacterial ribosomes. The physical dimensions of ribosomes can
also be determined by electron-microscopic measurements. The results
confirm the conclusion that mitoribosomes are smaller than cytoribo¬
somes in rat (O'Brien and Kaif, 1967; Aaij et al., 1972) and locust
(Klelnow et al., 1974), but d irect comparisons of mitochondrial and
bacterial ribosomes (under the same conditions of fixation and stain¬
ing) have not been performed.
How can two particles of the same molecular weight sediment
respectively at 55S and 70S? The significant difference appears to be
the much lower buoyant density of the animal mitochondrial ribosome.
Buoyant density values from 1.40 to 1.46 g/cc have been found for
mitoribosomes of IleLa cells (Perlman and Penman, 1970; Wengler et al.,
1972), rat (Sacchi et al., 1973; de Vries and Kroon, 1974), cow

(Hamilton and O’Brien, 1974), and toad (Leister and Dawid, 1974), The
buoyant density of a ribosome can be used to calculate the relative
proportions of RNA and protein in the particle (Hamilton, 1971); from
the relationship given in the notes to Table I an RNA content of about
30 percent can be calculated for animal mitoribosomes, as contrasted
with 63 percent for bacterial ribosomes (Tissiéres et_ nj_. , 1959). From
the RNA content and the sum of the molecular weights of the rRNA mole¬
cules, the particle weight of the ribosome may be calculated. By this
means molecular weight estimates equal to or greater than those for
bacterial ribosomes were obtained for mitoribosomes from rat (Sacchi
et. _al., 1973; de Vries and Kroon, 1974), cow (O’Brien e^t a^L., 1974), and
toad (Leister and Dawid, 1974) even before the molecular weight was
determined directly by sedimentation equilibrium (Hamilton and O'Brien,
1974).
Doubts have been raised about the validity of buoyant density
values as measures of the protein content of ribosomes (McConkey, 1974).
Others have suggested that the low buoyant densities of animal and
Tetrahymena mitoribosomes might be due to membrane fragments adhering
to these particles (specifically, due to membrane lipids, which are
assumed to be absent in the calculation of protein content from buoyant
density) (Borst and Grivell, 1971). Several lines of evidence indicate
that these factors do not represent significant objections to the
description of animal mitoribosomes presented above. Determinations
of the protein content by either chemical analysis or ultra-violet

16
absorption spectra of the ribosomes are in agreement with estimates
from buoyant density, yielding values of 70-80 percent protein in rat
(O'Brien and Kalf, 1967) and toad (Leister and Dawid, 1974). No phos¬
pholipids were detectable in rat mitoribosomes (de Vries and Kroon,
1974). The molecular weight estimate for bovine mitoribosomes based
on their buoyant density and the size of their rRNA agrees quite closely
with the molecular weight determined by sedimentation equilibrium
(Hamilton and O'Brien, 1974). Finally, similar values for the total pro¬
tein content of Xenopus mitoribosomes have been obtained directly by
summing the molecular weights of the individual ribosomal proteins
(Leister and Dawid, 1974).
Besides their unusual sedimentation behavior and high protein
content, another unusual characteristic of animal mitoribosomes is the
small size of their rRNA molecules. Values of 0.35 and 0.54 million
daltons have been obtained for Hel.a mitoribosomal RNA by electronmicro-
scoplc length measurements (Robberson et al., 1971), and these numbers
are in good agreement with those found by other methods for rat (Sacchi
et ai^., 1973), toad (Dawid and Chase, 1972; Leister and Dawid, 1974),
shrimp (Schmitt et al., 1974), and locust (Kleinow, 1974). Thus animal
mitoribosomes contain scarcely more than half as much RNA as any non-
mltochondrial ribosome known. To a first approximation, these particles
may be pictured as 1£. coli ribosomes modified by converting half of the
RNA into an equal mass of protein.

17
The base composition of the rRNA of animal mitoribosomes is
distinctly higher in guanine and cytosine than that of protist or fungal
mitoribosomes, though still lower than that of animal cytoplasmic ribo¬
somes. G + C contents of 40-47 percent have been found for mitochon¬
drial rRNA of HeLa (Vesco and Penman, 1969), rat (Bartoov et cLL., 1970),
toad (Dawid and Chase, 1972), and shrimp (Schmitt et al., 1974). The
degree of methylation is also lower for animal mitoribosomal RNAs than
for either cytoplasmic or prokaryotic rRNAs (Dubin, 1974).
Although mitochondrial ribosomes from all animal species examined
thus far appear quite similar in their physical and chemical properties,
it may be anticipated that they will differ in their detailed structure.
In fact, detectable non-homology has already been found between the base
sequences of mitoribosomal RNA from two species of toad, Xenopus laevis
and X_. mulleri (Dawid, 1972). Electrophoretic differences have also been
demonstrated in several of the mitoribosomal proteins of these two spe¬
cies (Leister and Dawid, 1975).
111. PHYLOGENETIC RELATIONSHIPS IN RIBOSOME STRUCTURE
The comparisons presented above permit some conclusions about the
variation in structural properties among mitochondrial ribosomes of
different organisms, and the differences between the ribosomes of mito¬
chondria and those of prokaryotes, eukaryotic cytoplasm, and chloroplasts.
The most obvious generalization is that mitochondrial ribosomes show
more diversity in all of their structural properties than do any other
kinds of ribosomes. Cytoplasmic ribosomes of all four eukaryotic kingdoms

18
are relatively similar with respect to sedimentation coefficient,
buoyant density, rRNA si lie, and guanine plus cytosine content. Pro¬
karyotic ribosomes display even more uniformity in these characteris¬
tics, while chloroplast ribosomes, whether obtained from protists or
from higher plants, show remarkable homologies not only among them¬
selves but to a large extent between themselves and prokaryotic ribo¬
somes. Mitochondrial ribosomes, on the other hand, can vary in sedi¬
mentation coefficient even within a kingdom: such differences do exist
between the protists Euglena and Tetrahvmena, and almost certainly
exist among the fungi. Similarly, differences in the buoyant densi¬
ties of mitoribosomes are found within both the protist and the fungal
groups. Mitoribosomal RNA molecular weights and G + C contents, in con¬
trast, appear to vary significantly between kingdoms but not within them.
The correlation between these last two structural properties of
mitochondrial ribosomes and the taxonomic kingdoms in which they are
found is illustrated in Figure 1. Also plotted in Figure 1 are the data
for chloroplast and prokaryotic ribosomes, which cluster together as
expected, and for four kingdoms of cytoplasmic ribosomes. The latter
group also forms a cluster, with the exception of the cytoribosome from
Tetrahymena. Both of these groups are dissimilar from any of the three
mitoribosome kingdoms plotted, and these are in turn distinct from each
other. Thus the two parameters, rRNA size and G + C content, discrim¬
inate ribosomes along the lines of their phylogeny and intracellular
location.

Figure 1. Phylogenetic relat
10 Mitochondrial Ribosomes
Fungi
S Saccha romyces
A Aspergillus
C Candida utills
N Neurospora
Protists
Eu Euglena
T Tetrahymena
Animals
X Xe.no pus
H HeLa
R Rat
£3 Prokaryotic Ribosomes
E Escherichia
coli
M Mycoplasma
hominis
Achloroplast Ribosomes
Eu Euglena
Sp Spinach
o CytupLasmic Ribosomes
Fungi
S Saccharomyces
A Aspergillus
C Candida utilis
N Neurospora
Protists
Eu Euglena
T Tetrahymena
Animals
X Xenopus
H HeLa
R Rat
Plants
P Pea
ionships in ribosome structure.
Reijnders ejt al. , 1973; Morimoto and
HaIvorson, 1971.
Verma et_ al. , 1970; Edelman et_ al. , 1970.
Vignais <2t al. , 1 972.
Neupert e_t aM. , 1969; KUntzel and Noll,
1967.
Krawiec and Eisenstadt, 1970; calculated
from Avadhani and Buetow, 1972.
Reijnders t^t al. , 1973; Chi and Suyama,
1970.
Leister and Dawid, 1974; Dawid and Chase,
1972.
Robberson e_t al_. , 1971; Vesco and Penman,
] 969.
Sacchi et al., 1973; Bartoov et al., 1970.
Kurland, 1960; Morimoto and Halvorson,
1971 .
Johnson and Horowitz, 1971.
Rawson and Siutz, 1969; calculated from
Avadhani and Buetow, 1972.
Hartley and Ellis, 1973; Lyttleton, 1962.
Reijnders e_t al., 1973; Morimoto and
Halvorson, 1971.
Verma e_t al. , 1970; Edelman et^ aj_. , 1970.
Vignais eit^ aj . , 1972.
Neupert et_ a 1 . , 1969; KUntzel and Noll,
1967.
Krawiec and Eisenstadt, 1970; calculated
from Avadhani and Buetow, 1972.
Reijnders et al., 1973; Chi and Suyama,
1970.
Loening e_t al_. , 1969; Dawid et al. , 1970.
Darnell, 1968; calculated from Vesco and
Penman, 1969.
Reijnders et a_L. , 1973; Kirby, 1965.
Loening e_t aj_. , 1969; Bonner and Varner,
1965.

RNA Content (daltons
20
RNA G + C Content (moles percent)

21
Sedimentation coefficients and buoyant densities, on the other
hand, do not group mitoribosomes from the same kingdom together; nor
do they differentiate them from the various non-mitochondrial ribo¬
somes. This observation suggests that these two structural proper¬
ties have been less conserved than rRNA size and G + C content in the
course of ribosome evolution. It is noteworthy that buoyant density
is the only one of these four structural characteristics which
discriminates chloroplast from prokaryotic ribosomes (Stutz and
Boschetti, 1976). However, some of the intrakingdom variability
reported for sedimentation coefficients and buoyant densities may be
due to the possibly greater sensitivity of these parameters to differ¬
ing conditions used in the preparation of the mitoribosomes.
IV. SUMMARY
Perhaps the most remarkable structural property of mammalian
mitochondrial ribosomes is the very high ratio of protein to RNA found
in these particles. On the basis of various kinds of physical-chemical
measurements it appears that this ratio is approximately 2:1, or 1.8
x 10^ dal tons of protein to 0.9 >: 10^ dal tons of RNA. These ribosomes
thus contain about twice as much protein as E. coli ribosomes, and
about half as much RNA.
Such an unusual composition poses questions about the molecular
architecture of these particles, and about the functional roles played
by this large quantity of protein in the process of protein synthesis.
Is the difference in protein content between the ribosomes of E. coli

22
and those of mammalian mitochondria due to a difference in the number
of the ribosomal proteins or to a difference in their size? Do the
different kinds of molecular interactions which must be involved in
maintaining the structural integrity of such a protein-rich ribosome
correlate with any identifiable differences in the properties of the
proteins? On the other hand, might the high protein content found in
this ribosome represent nothing more significant than the presence of
large quantities of contaminating non-ribosomal proteins? A portion
of the present research is directed toward these questions.
Also of interest is the phylogenetic diversity in structural
properties which is found in comparisons of mitochondrial ribosomes
from distantly related organisms. Only small differences in the over¬
all physical and chemical properties of mitoribosomes are seen when
species within the animal kingdom are compared. But if a high degree
of evolutionary divergence is indeed the rule for mitochondrial ribo¬
somes in general, it should be possible to detect significant differ¬
ences between the mitoribosomes of more closely related species at a
sufficiently detailed level of analysis. The rule further predicts
that such differences will be greater between mitochondrial ribosomes
than between the cytopLasmic ribosomes of the same species. The
experiments presented below include a test of these predictions.

MATERIALS AND METHODS
I. MATERIALS
Sucrose (density gradient grade, ribonucLease free) was obtained
from Schwarz/Mann. Urea (reagent-grade) was purchased from J. T. Baker
or frcm Mallinckrodt. Stock solutions of 10 M urea were prepared,
filtered, and stored at room temperature for no more than 24 hr before
use; if the conductivity of the stock solution was greater than 50 prnho,
it was deionized by stirring witli AC501-X8 resin (Bio-Rad). Acrylamide
and N, N'-methylene-bisacrylamide (Eastman) were recrystallized from
chloroform and acetone, respectively; stock solutions were deionized
with Rexyn-300 (Fisher) and stored at room temperature in the dark.
Ampholine was obtained from LKB. Bovine serum albumin and human
y-globulin were obtained from Nutritional Biochemicals Corporation, and
egg albumin, equine myoglobin and egg-white lysozyme were from Sigma
Chemical Company. Purumycin dihydrochloride was from Nutritional Bio-
chemlcals Corporation. Na^^T (carrier-free, in 0.1 N NaOH) and [^C]
formaldehyde (44 mCi/mmol) were purchased from New England Nuclear;
and [4,5-^H] L-leucine (55 Ci/mniol) , [8-^H] (¡TP (12 Ci/mmol, tetrasodium
salt), and [JH] I.-phenylalanine (7 Ci/mmol) from Schwarz/Mann.
23

Table IV. Compositions of buffers used in the preparation of ribosomes.
Buffer Composition
A
0.1
M
KC1,
20 mM MgCl2
, 5 mM g-ME, 20 mM TEA, p:
7.5
B
0.5
M
KC1,
A0 mM MgCl2
, 5 mM g-ME, 10 mM Tris, pH 7.5
C
0.5
M
KC1,
10 mM MgCl2
, 5 mM p-ME, 10 mM Tris, pH 7.5
D
0.3
M
KCl,
5 mM MgCl2,
5 mM g-ME, 20 mM TEA, pH
7.5
E
0.5
M
KC1,
5 mM MgCl2>
5 mM g-ME, 20 mM TEA, pH
7.5
F
1.0
M
KCl,
5 mM MgCl?,
5 mM g-ME, 20 mM TEA, pH
7.5
G
1.0
M
KCl,
1 raM MgCl2,
5 m>l g-ME, 20 mM TEA, pH
7.5
II
0.1
M
KCl,
10 mM MgCl2
, 0.1 mM EDTA, 5 mM g-ME,
10 mM
Tris,
Pn 7.5
I
0.2
M
KCl,
40 mM MgCl2
, 1.6% Triton X-100, 5 mM
g-ME,
10 mM
Iris, pH 7.5
J
0.5
M
KCl,
40 mM MgCl9
, 1.6% Triton X-100, 5 mM
S-ME,
10 mM
Tris, pH 7.5
K
0.34 M sucrose, 5 raM Tris, pH 7.4
L
O,25 M sucrose, 1 mM EDTA, 5 mM Tris, pH 7.5
M
0.1
M
KCl,
5 mM MgCl2>
5 mM g-ME, 20 mM TEA, pH
7.5
N
60 mM
KCl,
14 mM KgAc 2
, 6 mM g-ME, 10 mM Tris, pH 7.8
1.0 M NH^Cl, 10 mM HgCl2> 5 mil S-ME, 10 mil Tris, pH 7.
0
5
Js

25
II. PREPARATION OF RIBOSOMES
A. Bovine Mitochondrial Ribosomes
In the course of this research the procedure for the preparation
of mitochondrial ribosomes from bovine liver was modified in several
ways to improve the yield and purity of the ribosomes obtained. The
major changes are discussed under Methods A, B, and C below. All of
the preparative methods shared the following common features. Livers
of freshLy killed animals were obtained from a slaughterhouse and
transported to the laboratory on ice. All subsequent procedures were
performed in the cold unless otherwise noted. Four to 8 kg (fresh
weight) of liver were passed through a meat grinder, diluted with 4
volumes of Buffer K or Buffer L, strained through a coarse-mesh cloth
screen, and homogenized. In the earlier preparations homogenization was
performed with a Pottet-Elvehjem homogenizer modified so that the ground
tissue could be pumped through it continuously. Increased cell breakage
and increased final yields of mitochondria were obtained with the use of
a high-frequency dispersion device (a Tekmar Company Super Dispax,
Model SD-45K).
Unbroken cells and nuclei were removed by pumping the homogenate
at 880 ml/min through a Vernitron CFR-2 continuous-flow rotor rotating
at 11,000 rp:;i in a Vernitron l.CA-2 centrifuge. Mitochondria were
harvested from the supernatant by continuous-flow centrifugation in a
Beckman JCF-Z rotor at 18,000 rpin and a flow rate of 440 ml/min. Alter
the mitochondria had been washed as described below, they were

26
resuspended to a concentration of 20 mg protein/ml in buffer and lysed
by the addition of non-ionic detergent. The lysate was clarified by
centrifugation in a Beckman Type 35 rotor at 28,000 rpm for 10 rain.
The supernatant (after treatment with DEAE-cellulose, in the case of
Method C below) was centrifuged in a Beckman Type 35 rotor at 35,000 rptn
for 12 hr. The ribosome pellets obtained at this step are referred to
as "crude ribosomes" below.
1. Method A
The buffer used for the homogenization of the liver and the prepa¬
ration of mitochondria was Buffer K. The mitochondria were washed three
times by resuspending in Buffer K and centrifuging in a Beckman JA-10
or Sorvall GS-3 rotor at 8,000 rpm for 10 min. Washed mitochondria
were suspended in Buffer A containing 34% sucrose (ribonuclease-free),
50 pg/ml heparin and 200 pg/ml oligonucleotides (prepared by partial
base hydrolysis of yeast tRNA according to Spencer and Poole (1965)).
Triton X-100 and sodium deoxycholate were added to final concentrations
of 2% and 0.5% respectively, and crude ribosomes were prepared iron the
mitochondrial lysate as described above.
The ribosome pellets were resuspended in Buffer A containing
50 pg/ml heparin and 550 pg/ml puromycin, and incubated at 37°C for
5 min. The ribosomes were then purified by sedimentation into a linear
10-30% sucrose density gradient made up in Buffer A, in a Beckman SW27
rotor. After centrifugation the gradient was pumped through the flow
cell of a Gilford Model 2400 spectrophotometer, and its absorbance at
260 nm was recorded. One-ml fractions were collected.

27
2. Method B
Buffer L was used for the preparation and washing of the mito¬
chondria, After two washes the mitochondria were resuspended to a
concentration of 10 mg protein/ml in Buffer L containing 50 pg/ml
digitonin. The suspension was stirred for 15 min, and the mitochondria
pelleted at 8,000 rpm for 10 min. The mitochondria were washed once
more with Buffer L, resuspended to 20 mg protein/ml in Buffer H, and
lysed by the addition of Triton X-100 to a concentration of 1.6%. After
clarification as described above, the lysate was layered onto 20 ml of
Buffer H containing 34% sucrose and 1,6% Triton X-100, and centrifuged
to prepare crude ribosomes. Purification of the ribosomes by sucrose-
density-gradient centrifugation was as described above, except that
Buffer H was used instead of Buffer A.
3. Method C
To the clarified mitochondrial lysate prepared by Method B was
added a moist cake of DEAE-cellulose equilibrated in Buffer H containing
1.6% Triton X-100. The quantity of DEAE-cellulose used was about 1 g
(dry weight) per 50 ml of lysate. The slurry was stirred for 30 min,
placed in a large Buchner funnel, and filtered just until all excess
liquid was removed. Ten ml of Buffer H/Triton per gram DEAE-cellulose
was added and filtered off as before. The filtrates were discarded
and the cake of DEAE-cellulose was stirred for 30 min in Buffer I
(10 mi/g) to elute the ribosomes. The slurry was filtered and rinsed
with 2.5 ml/g of Buffer I. The ribosomes in the filtrate were then

28
centrifuged through a layer of Buffer H/l.6% Triton/34% sucrose and
purified as in Method B. By this procedure about 10 mg of ribosomes
could be obtained from 6 kg of liver.
B. Rat Mitochondrial Ribosomes
The livers of 20 to 50 young (80-100 g) female Sprague-Dawley
rats were homogenized in 4 volumes of Buffer L with a Potter-Elvehjem
homogenizer. Rapidly sedimenting material was removed by centrifuga¬
tion at 3,000 rpm for 10 min in a Beckman JA-10 rotor, and mitochon¬
dria were obtained from the supernatant by centrifugation at 8,000 rpm
for 10 min in the same rotor. The mitochondria were washed as de¬
scribed for bovine mitochondria above (Method B) except that the
concentration of digitonin used was 17 pg/ml. The washed mitochondria
were resuspended to a concentration of 5 mg protein/ml in Buffer H
and lysed by the addition of Triton X-100 to 1%. Ribosomes were then
prepared and purified as in Method B above. The yield was about 50 pg
of ribosomes per tat liver.
C. Bovine Cytoplasmic Ribosomes
Cytoplasmic ribosomes were prepared from a microsomal fraction of
bovine liver. The liver was homogenized in 4 volumes of 25 mM KC1,
5 mM MgCl,, 0.34 M sucrose, 5 mM B-mercnptoethanol, 10 mM Tris, pH 7.5.
Mitochondria and larger particles were removed at 8,000 rpm for 10 min
in a Beckman JA-10 rotor, and microsomes were pelleted from the super¬
natant at 9,000 rpm for 45 min in the same rotor. The pellet was
suspended in luffer M and Triton X-100 was added to 2%. Ribosomes

29
were then prepared as described above except that Buffer M was used
instead of Buffer A or H.
D. E. coli Ribosomes
Escherichia coli K-12, strain 1200F~ end A 1100 rns A Su~ or
strain Hfr DIO RNase-, were grown in nutrient broth at 37°C. The
cells were harvested by centrifugaLion at 6,000 rpm for 10 min in a
Beckman .TA-10 rotor, suspended in a small volume of cold Buffer N, and
ruptured by sonication. After centrifugation in a Beckman Type 65
rotor at 15,000 rpm for 10 min, the supernatant was centrifuged again
at 60,000 rpm for 2 hr in the same rotor to pellet the ribosomes.
These weiu- then washed once by resuspending and recentrifuging in
either Buffer N (low-salt ribosomes) or Buffer 0 (salt-washed ribo¬
somes) .
III. EXTRACTION OE RIBOSOMAL PROTEIKS
Proteins were prepared for electrophoresis by a modification of
the method described by Leister and Dawid (1974). For each gel sample,
a pellet of ribosomes containing 200-300 pg of protein was suspended
in 40 pi of 0.1 M KCl, 10 mM MgClj. 40 pi of 10 M. urea, 4 M LiCl,
HC1, pH 3.5 was added and the mixture was stirred at 5°C for 12 hr.
The RNA-containing precipitate was removed by centrifugal ion in a
Beckman Type 65 rotor at 50,000 rpm for 1 hr, and re-extracted by stir¬
ring with 80 pi of 6 M urea, 3 M LiCl, HC1, pH 3.5 for 2 hr. The
supernatants from the two extractions were combined, and disulfide
bonds were reduced by the addition of 40 pi of 7.5 M urea, 0.25 M

30
dithiothreitol, 0.5 M F.DTA, 1 M Tris, pH 8.8 and incuhation aC 37°C
for 1 hr. The proteins were then dialyzed against Sample Buffer
(8 M urea, 60 nil potassium acetate, 0.01% aminoethanethiol, pH 6.7).
IV. RADIOACTIVE LABELLING OF RIB0S0MAL PROTEINS
A. Radioiodinatlon
125
Chemical labelling of ribosomal proteins with Na —I and Chlor¬
amine T was performed according to Leister and Dawid (1974). A 15 pi
aliquot (20 to 30 pg protein) of the ribosomal protein extract described
above was set aside before the disulfide reduction step. The pH was
adjusted with an equal volume of 6 M urea, 3 M Li.Cl, 100 mil Tris,
125
pH 7.5. 100 pCi of Na I (100 Ci/mmol) was added, followed immediately
by 30 pg of Chloramine T. The reaction (total volume = 55 pi) proceeded
at room temperature for 10 minutes, and was stopped with 1 pi of 1 M
0-mercaptoethanol. Disulfide bonds were then reduced as described above,
and the radioactive protein was separated from unreacted I by chroma¬
tography on Sephadex G-25 equilibrated in Sample Buffer.
Control experiments were performed to determine whether the label¬
ling reaction affected the electrophoretic properties of the proteins.
Radioiodinated ribosomal proteins were added back, to a large non-radio-
active sample of the same proteins, and the mixture was subjected to
electrophoresis, staining and autoradiography as described below. No
alteration in the electrophoretic mobilities of the proteins could be
detected. However some of the stained proteins were not detectably
labelled, and vice versa. Radioactive spots which did not correspond

31
to proteins regularly seen by staining were ignored. The electro¬
phoretic positions of the proteins that could not be radiolabelled
were determined by interpolation between the nearest spots that were
labelled.
B. Reductive Methylation
A pellet of ribosomes containing 5-10 pg of protein was suspended
in 2 pi of 4 M guanidine hydrochloride, 100 mM sodium borate, 10 mil
MgCl^, 20 mM KC1, 6 mM (3-ME, pH 8.5. Thirty nmol of [^Cj formalde¬
hyde (44 Ci/mol) in 1 pi of water were added, and the mixture was
incubated on ice for 30 sec. One pi of 30 nmol/pl sodium borohydride
was added, and this addition was repeated after one minute. The
reaction mixture was left standing in the cold for 5 hr. A resuspended
pellet of ribosomes containing 200-300 pg of protein in 40 pi of 0.1 M
KC1, 10 mM MgCl2 was then added, and the proteins were extracted for
electrophoresis as described above.
V. TWO-DIMENSIONAL ELECTROPHORESIS
The procedure was modified from that of Leister and Dawid (1974).
The first dimension was essentially the discontinuous-buffer electro¬
phoretic system of Reisfeld et al. (1962), providing electrophoretic
stacking of the protein sample. The composition of the first-dimension
separation gel (1.5 mm in diameter, 11 cm long) was 7.5% acrylamide,
0.023% bisacrylamide, 0.063% TEMED, 8 M urea, 60 mM potassium acetate,
pH 4.3. It was polymerized by adding ammonium persulfate to a concen¬
tration of 0.1%, and pre-electrophoresed before use for 12 hr at 0.2 ma

per gel, in a tank buffer of 0.01% aminoethanethiol, 8 M urea, 60 mM
potassium acetate, pH 4.3. The stacking gel (2 cm long) was 3.86%
acrylamide, 0.14% bisacrylamide, 0.05% TEMED, 8 M urea, 60 mM
potassium acetate, pH 6.7, polymerized by the addition of ammonium
persulfate to 0.02% and riboflavin to 0.001%. The tank buffer was
0.01% aminoethanethiol, 35 mil 3-alanine acetate, pH 5.0.
The dialyzed protein samples (about 200 pi) were loaded onto the
gels and lectrophoresed at 0.1 ma per gel. When the tracking dye
(0.001% Pyronin Y in 200 pi of Sample Buffer) in a parallel gel tube
had reached the top of the separation gel, the current was increased
to 0.2 ma per gel. Electrophoresis was continued until the tracking
dye reached the bottom of the gel. The gels were extruded from the
glass tubes with a 6-inch, 22 gauge needle through which water was
flowing under pressure.
The second-dimension gel slabs were 13 cm high, 20 cm wide, and
1.5 mm thick. Each gel was poured between a pair of glass plates
separated by two 1.5 nan-thick plexiglass strips, sealed at the bottom
and sides with a piece of silicone-rubber tubing, and held together
with spring clamps. After the gel was polymerized> the silicone tubing
was removed. The gel composition was 9.65% acrylamide, 0.35% bisacryl¬
amide, 0.1% TEMED, 5 M urea, 0.5% SDS, 0.1 M sodium phosphate, pH 7.2,
polymerized with ammonium persulfate (0.025%).
The extruded first-dimension gels were laid on top of the second-
dimension gel slabs and overlaid with 0.05% mercaptoacetic acid, 5 M

33
urea, 1% SDS. At each end of the first-dimension gel was placed a
small piece of agarose (1,5 mm in diameter, 5 mm long) containing
several marker proteins. The composition of the marker mixture was
0.5 mg/ml bovine serum albumin, 1 mg/ml human ^-globulin, 0.4 mg/ml
ovalbumin, 0.33 mg/ml equine myoglobin, 0.4 mg/ml egg-white lyso¬
zyme, 0.33% SDS, 0.33% B-ME, 0.016% mercaptoacetic acid, 2% agarose.
Electrophoresis was at 30 ma/gel with a tank buffer of 0.024% mer¬
captoacetic acid, 0.5% SDS, 0.1 M sodium phosphate, pH 7.2, until
the dye marker (Bromophenol Blue) reached the bottom of the gel. The
total time for electrophoresis in both dimensions was about 24 hr.
After electrophoresis the gels were soaked in the following
solutions with continuous agitation: 25% isopropanol, 10% acetic
acid (18 hr); 0.25% Coomassie Brilliant Blue R, 50% ethanol, 7.5%
acetic acid (6 hr); and 5% ethanol, 10% acetic acid (several changes,
until the gels were adequately destained). For convenient storage
and for autoradiography, the gels were soaked in 3% glycerol for 30
min and then dried onto Whatman ir3MM filter under vacuum on a steam
bath, by the method of Maizel (1971).
For autoradiography the dri d gel was placed in contact with
Kodak RP14 medical X-ray film in an X-ray film cassette. Two small
holes had previously been drilled through the cassette. After the gel
and film were loaded and the cassette was closed, a needle was inserted
through these holes to mark the alignment of the gel with the film.

34
Gels containing reductively methylated proteins were prepared for
autoradiography as described by Bonner and Laskey (1974). They were
shaken in two changes of dimethyl sulfoxide, for 30 min each time,
then in 4 volumes of a 20% solution of 2,5-diphenyloxazole in dimethyl-
sulfoxide for 3 hr, and m water for 1 hr. The gels were then dried
onto filter paper as described above. The film used for autoradio¬
graphy was Kodak RP/R-54 medical X-ray film, pre-exposed with a flash
from a photographic strobe light according to Laskey and Mills (1975).
Exposure at -70°C for 5 weeks was sufficient for gels containing 10,000
cpm of reductively methylated protein.
VI. ISOELECTRIC FOCUSING
A previously published procedure (Czempiel et^ aT., 1976; Klose,
1975) was followed closely. The gel composition was 4.8% acrylamide,
0.2% bisacrylamide, 0.065% TEMED, 1.0% Ampholine (pH 3.5 - 10), 5%
sucrose, 8 M urea, polymerized by the addition of ammonium persulfate
to 0.019%. The gels were 0.4 cm in diameter and 7 cm long. The upper
tank buffer was 5% phosphoric acid and the lower tank buffer was 5%
ethylenediamine, with the anode in the upper tank. The protein sample
for each gel was dissolved in 20X of 8 M urea and then mixed with 30X
of Sephadex G-200 (superfine) swollen in 20% sucrose, 8 M urea, 10%
B-ME, 1% Ampholine (pH 3.5 - 10).
The details of the procedure listed above differed from those of
the published procedure only in three minor respects. In the latter
procedure the ammonium persulfate concentration was slightly greater

35
(0*023%), the gel was longer (8 cm), and the upper tank buffer contained
urea at a concentration of 3 M. Besides these insignificant differences,
there was one further modification which was significant. Czempiel
al. (1976) used an Ortec 4100 Pulsed Constant Power Supply, which has
power output characteristics different from those of the constant-voltage
power supply used in the present experiments. The published procedure
involved a program of increasing voltage (50V for 1 hr, 100V for 1 hr,
150V for 1 hr, 200V for 2 hr, 300V for 2 min, and 400V for 2 min). In
the present experiments several parallel samples were electrofocused
according to this schedule but for different lengths of time. The first
sample was run for a total of 3 hr (the first 3 hr of the schedule),
the second sample for the whole 5 hr of the schedule, and the third
sample for the whole 5 hr plus an additional 2 hr at 200V.
After electrofocusing, the gels were extruded and subjected to a
second dimension of electrophoresis in SDS as described in Section V.
VII. ASSAYS OF RIBOSOME FUNCTIONS
Peptidyl transferase activity was assayed by the modified fragment
reaction (Denslow and O'Brien, 1974; de Vries ert al_. , 1971). The
reaction mixture contained 0.1 mg ribosomes, 83 nM (10,000 cpm) N-acetyl-
[^H] leucyl-tRNA, 0.66 mM puromycin, 267 mM KOI, 13.3 mM MgAc^, 33%
ethanol, 33 mM Tris-HCl, pH 7.5 in a total volume of 0.15 ml. After
incubation at 25°C for 10 min, KOI! was added to 0.6 M and the mixture
3
warmed to 40°C for 3 min. The N-acetyl-[ H]leucyl-puromycin synthesized
in the reaction was extracted into 1.5 ml of ethyl acetate, and

36
radioactivity was determined by liquid scintillation counting in Triton
X-100/toluene (1:1) containing 0.5% PPO and 0.05% POPOP.
3
[ H] GTP binding was measured by the Millipore filter assay of
Bodley al. (1970). Fifty )il of reaction mixture containing 15 pg of
ribosomes, 42 pmol (0.5 pCi) of [8-^H] GTP, 10 mM NH^Cl, 20 mM MgAc?,
5 mM 3-ME, 10 mM Tris-HCl, pH 7.4 was incubated at 0°C for 5 min and
then filtered thru a Millipore filter. The filter was washed with 10 mM
NH^Cl, 10 m.T MgAc2, 10 mM Tris-HCl, pH 7.4, and the ribosome-bound
radioactff*y determined by liquid scintillation counting.
Poly L-dependent polyphenylalan ine synthesis was assayed according
to Hosokawa et al, (1966). The composition of the 0.25 ml reaction
mixture was 0.64 mg/ml poly U, 5.4 pM [ H]phenylalanine (1.82 Ci/mmol),
25 mM tyrosine, 50 mM of each of the other 18 amino acids, 0.5 mg/ml
tRNA, 32 pM GTP, 1 mM ATP, 5 mM phosphoenolpyruvate, 0.1 mg/ml pyruvate
kinase, 1 mg/ml ji. coli factors, 0.4 mg/ml ribosomes, 50 mM KC1, 20 mM
MgAc , 6 mM g-ME, 1 mM dithiothreitol, 10 mM Tris-HCl, pH 7.8. Aliquots
of 50 pi were withdrawn at 5 min intervals, and radioactivity insoluble
in hot trichloroacetic acid was determined by the method of Mans and
Novelli (1960).
VIII. BUOYANT DENSITY DETERMINATIONS
Ribosomes were dialyzed in 50 mM KC1, 5 mM MgC^, 20 mM TEA,
pH 7.5, and then fixed by the addition of formaldehyde to a concentra¬
tion of 5%. They were analyzed by equilibrium centrifugation in gradients
of CsCl containing 50 mil KC1, 5 mM MgCl, 20 mM TEA, pH 7.5, 0.3%

37
formaldehyde in a Beckman SW39 rotor, as described by Brunk and Leick
(1969). The gradients were pumped through the flow cell of an Isco
UA-2 absorbance monitor, and 0.25 ml fractions were collected. The
refractive index of each fraction was measured with a Bausch and Lomb
refractometer, and used to calculate the concentration of CsCl and the
buoyant density in the fractions.
IX. QUANTITATIVE MEASUREMENTS
Quantities of mitochondria were estimated by absorbance at 550 nm.
Samples of mitochondria of known 550 nm absorbance were analyzed by
the method of Lowry e_t al. (1951), to establish a calibration curve
relating the absorbance at this wavelengch to the concentration of
mitochondrial protein in mg/ml.
Ribosome quantities were determined by absorbance at 260 nm, using
extinction coefficients (; a' ) of 110 for mitochondrial ribosomes,
1 cm
135 for cytoplasmic ribosomes, and 160 for !!. coli ribosomes.

RESULTS
The purpose of this research was to characterize the proteins of
mammalian mitochondrial ribosomes, especially in comparison with the
proteins of other kinds of ribosomes. The major characteristics to be
investigated were the number of ribosomal proteins, their molecular
weights and their electrophoretic properties. For this purpose It was
necessary to establish and apply a number of criteria to distinguish
the ribosomal proteins from the contaminating non-ribosomal proteins
which might be present in the mitochondrial ribosome preparation.
The results will be presented in five sections. First, the means
of preparing mitochondrial ribosomes in adequate quantity and purity
for these experiments will be described. The two-dimensional electro¬
phoretic patterns of bovine mitochondrial ribosomal proteins will be
presented and analyzed in terms of criteria for the identification of
the proteins which are most likely to be true components of the ribo¬
some in vivo; the effect of the ribosome preparation procedure (partic¬
ularly the effect of high salt treatment) on the electrophoretic pattern
of ribosomal proteins will be evaluated, with a view to the possibility
of various kinds of artifacts. The electrophoretic properties of the
proteins will be compared with those of mammalian cytoplasmic and
bacterial ribosomal proteins. The number and sizes of the mitochondrial
38

39
ribosomal proteins will be used to estimate the total mass of protein
in the ribosome. Finally, the proteins of bovine mitochondrial ribo¬
somes will be compared with those of other mammalian species to assess
the level of evolutionary divergence in these proteins.
I. PREPARATION AND CHARACTERIZATION OF MITOCHONDRIAL RIBOSOMES
At the time this research was begun, the method used for the prep¬
aration of mitochondrial ribosomes (Method A in MATERIALS AND METHODS)
was not cambie of producing adequate quantities of material for the
electrophoretic studies to be described below. Three separate prepa¬
rations were required to produce enough ribosomes for a single electro¬
phoretic analysis, each preparation being a week-long procedure.
Moreover, the crude mitochondrial ribosome preparation was heavily
contaminated with cytoplasmic ribosomes, as evidenced by the large
amount of material sedimenting at 80S when the ribosomes were subjected
to sucrose density gradient centrifugation (Figure 2A). The presence
of such quantities of 80S ribosomes suggested that even the 55S mito¬
chondrial ribosome region of the gradient contained significant amounts
of cytoplasmic ribosomal subunits, as well as possibly other unknown
structures with similar sedimentation coefficients. For this reason
the intact 55S ribosomes were not suitable as samples for electro¬
phoretic analysis.
Rather, a further purification step was performed by pooling the
material in the 55S peak, treating it with a higher concentration of
KC1 and a lower concentration of MgCl^ to dissociate the mitochondrial

Figure 2. Preparación of mitochondrial ribosomes by Method A.
(A) Crude mitochondrial ribosomes were prepared from bovine
liver according to Method A in MATERIALS AND METHODS. The ribosomes
were suspended in 2 ml. of Buffer A and layered onto a linear 10-30%
sucrose density gradient made up in Buffer A. After centrlfugatIon
in a Beckman SW27 rotor at 27,000 rpm for 5 hours, the gradient was
pumped through the flow cell of a spectrophotometer and its absor¬
bance at 260 nm was recorded. Gne-ml fractions were collected. The
direction of sedimentation is from left to right.
(B) Fractions corresponding to the 55S absorbance peak of the
sucrose density gradient shown in Figure 2A were pooled and centri¬
fuged in a Beckman Type 65 rotor at 65,000 rpm for 3 hours. The
pellet of mitochondrial ribosomes was suspended in 2 ml of Buffer E
and layered onto a sucrose density gradient made up in the same buf¬
fer. Centrifugation was at 20,000 rpm for 13.5 hours.

Absorbance at 260 nm Absorbance at 260nm

42
ribosomes to their subunits, and re-ciatrifuging on a second sucrose
density gradient. The only contaminants in the preparation of subunits
obtained in this manner, besides those that might be bound directly
to the subunits, would be those which sedimented near 55S under the
ionic conditions of the first centrifugation and in the 25S to 45S
region under the second conditions. The kinds of possible contaminants
which might behave in this manner seem intuitively to be few, and in
particular do not include cytoplasmic ribosomal subunits. In fact,
the discrete peaks of mitochondrial small and large rllosomal subunits
were found to be by far the major species visible on the absorbance
profile of the second centrifugation (Figure 2B) , suggesting that the
preparation was indeed reasonably free of co-sedimenting material.
Later improvements in the procedure increased both the yield and
the purity of the mitochondrial ribosomes. The yield was approximately
tripled by the use of a more efficient homogenizer for the disruption
of the cells. Treatment of Lhe mitochondrial preparation with digitonin
(Method B) dramatically reduced the quantity of cytoplasmic ribosomes
in the crude mitochondrial ribosome preparation (Figure 3). This
detergent has been used previously to remove cytoplasmic ribosomal RNA
(Malkin, 1971) and cytoplasmic ribosomes (de Vries and van der Koogh-
Schuuring, 1973) from rat liver mitochondrial preparations. In addi¬
tion, treatment under these conditions has been reported to solubilize
latent lysosomal enzymes (Schnaitman and Greenawalt, 1968; Lowenstein
£t a_l., 1970), thereby diminishing the possibility of degradation of
the mitochondrial ribosomes during the preparation.

Absorbance at 260 nm
43
Figure 3. Preparation of mitochondrial ribosomes by Method B.
Crude bovine mitochondrial ribosomes obtained by Method B
(MATERIALS AND METHODS) were suspended in Buffer H and analyzed
by sucrose density gradient centrifugation in this buffer as des¬
cribed in Figure 2A, except that centrifugation was for 4 hours
at 27,000 rpm. (No differences have been observed between Buffer H
and Buffer A with respect to their effects on the mitochondrial
ribosomes.)

44
At this point the maximum yield of mitochondrial ribosomes from
a single preparation was limited by the volume of the largest prepar¬
ative ultracentrifuge rotors available. After preparation and lysis
of the mitochondria, the best means available for concentrating the
ribosomes for purification on sucrose density gradients was to
centrifuge them to a pellet at about 100,000xg. The volume of an
average yield of mitochondria (about 35g protein), suspended in buffer
to a concentration low enough for efficient detergent lysis, is far
greater than the capacity of three Beckman Type 35 rotors. Therefore,
a method was developed to preconcentrate the ribosomes before harvest¬
ing them by centrifugation (Method C).
The mitochondrial lysate was stirred with DEAE-cellulose as de¬
scribed in MATERIALS AND METHODS to adsorb out the mitochondrial ribo¬
somes, and then the ribosomes were eluted by stirring with a smaller
volume of buffer at a higher ionic strength. This procedure is very
rapid and effects a fourfold concentration of the ribosomes, thereby
quadrupling the yield of each preparation. Furthermore, any remaining
traces of cytoplasmic ribosomes, as well as a considerable proportion
of the other contamimmts of the mitochondrial ribosome preparation,
are removed by this means.
An assessment of the degree of purification achieved by this method
is shown in Figure 4. A sample of purified, digitonin-treated mito¬
chondria was split into two equal aliquots. Ribosomes were prepared
from the first aliquot without DEAE-cellulose treatment and analyzed by

45
Figure 4. Preparation of mitochondrial ribosomes by Method C.
All samples were analyzed by sucrose density gradient centrifu¬
gation in Buffer H.
(A) Ribosomes from 1.4 g (protein) of bovine liver mitochondria
prepared by Method B, The 55S absorbance peak contains 5,7 A0/.a units
c 'u 4oU
or ribosomes.
(B) Cytoplasmic ribosomes from 5 ml of a preparation of bovine
liver microsomes. The 80S peak contains 5.6 A2^q units of ribosomes.
(C) 1.4 g of mitochondria and 5 ml of microsomes in Buffer H
were mixed together, lysed with Triton X-100, and stirred with 2 g
(dry weight) of DEAE-cellulose equilibrated in Buffer H. The slurry
was filtered and the filtrate discarded. Mitochondrial ribosomes
were then eluted by stirring the DEAE-cellulose with Buffer I, and
were prepared for sucrose density gradient analysis according to
Method B. The 55S peak contains 6.1 A^q units of ribosomes.
(D) The once-eluted DEAE-cellulose from (C) was stirred with
Buffer J to elute the cytoplasmic ribosomes. The 80S peak contains
4.6 A26q units, or 82% of the input cytoplasmic ribosomes.

46
sucrose density gradient centrifugation, giving the absorbance profile
shown in Figure 4A. A preparation of bovine cytoplasmic ribosomes
(having the sucrose density gradient profile shown in Figure 4B) was
added to the second aliquot of mitochondria, to provide a more strin¬
gent test of the method. The mitochondria were then lysed and the
ribosomes prepared by the DEAE-cellulose procedure described above. As
seen in Figure 4C, the product consists primarily of mitochondrial
ribosomes and subunits, much of the slowly-sedimenting material and all
detectable 80S cytoplasmic ribosomes having been removed. Probably
because of the larger proportion of rRNA in cytoplasmic ribosomes and
their consequent higher density of negative charges, these particles
remain bound to the DEAE-cellulose under the conditions used for elution
of the mitochondrial ribosomes, as shown by re-elution of the DEAE-
cellulose with a buffer of higher ionic strength (Figure 4D).
As may be seen in Figure 4, the yield of mitochondrial ribosomes
per gram of mitochondria obtained with the DEAE-cellulose procedure is
as great as or slightly greater than that obtained without this step.
Since this method permits the processing of a four-fold greater quantity
of mitochondria, and gives a cleaner preparation of ribosomes as well,
it is clearly the method of choice unless it adversely affects the
structural or functional integrity of the ribosomes. Of particular con¬
cern is the possibility that some of the more acidic ribosomal proteins
might bind more strongly to the DEAE-cellulose than to the ribosome
itself and thus might be stripped off when the ribosomes are eluted.

47
The most acidic proteins of E). col 1 ribosomes, L7 and L12, have been
found to be relatively loosely bound to the ribosomes (Hamel et al.,
1972).
To answer this question, the ribosomes were characterized by a
number of functional and physical tests. The peptidyl transferase and
GTP-binding activities of the ribosome were essentially unaltered by
treatment with DEAE-cellulose (Table V). DEAE-cellulose-prepared ribo¬
somes were also found to be active in the translation of poly U (400
pinoles phe incorpornted/mg rRNA/15 min). The sediment ition coefficients
and buoyant densities of mitochondrial ribosomes and the subunits de¬
rived from them were essentially the same as those obtained from previous
preparations. The effect of this procedure on the two-dimensional
electrophoretic pattern of the ribosomal proteins will be evaluated in
Section IIC below.
11. ELI-CTROPHORETIC ANALYSIS OF MITOCHONDRIAL RIBOSOMAL PROTEINS
The very high proportion of protein which physical-chemical measure¬
ments have shown to be present in bovine mitochondrial ribosomes (see
INTRODUCTION) predicts that a rather large number of individual proteins
are to be found in these ribosomes. If the 1.8 x 10^ d. of total protein
per particle were made up of proteins similar in size to E. coli ribo¬
somal proteins (average molecular weight about 17,000), for example,
each ribosome would contain more than a hundred protein molecules.
For this reason it seemed unlikely that electrophoresis in one
dimension would separate all the proteins sufficiently for the purposes

48
Tab1
V. Functional activity of mitochondrial ribosomes
prepared with or without DEAE-cellulose.
Preparation
^H-Leu-Puromycin
Synthesis %-GTP Binding
(cpm/nmolu ribs/10') (moles CTP/mole ribs)
No DEAE
DEAE
31,600 1.12
36,200 1.09

49
of this study. Several two-dimensional electrophoretic systems have
been devised to provide adequate resolution for the analysis of ribo-
somal proteins. The system used for most of the experiments to be
described below was that of Leister and Dawid (1974), employing a
separation partially on the basis of charge at pH 4.3 in the first
dimension and on the basis of size in the second (SDS) dimension. For
the purposes of this research, this electrophoretic system has several
advantages over the one that if. most widely used for ribosomal protein
studies at present (Kaltschmidt and Wittmann, 1970). It permits a
direct determination of the molecular weights of the proteins. It is
more sensitive, requiring only about 5 pg of each protein. And it has
been used for the analysis of Xenopus mitochondrial ribosomal proteins
(Leister and Dawid, 1974), making it possible to compare these published
results with those obtained in the present experiments on mammalian
mitochondrial ribosomes.
The results of one of the early electrophoretic analyses (shown
in Figures 5 and 6) amply confirmed the expectation that a large number
of proteins would be found in mitochondrial ribosomes. 40 to 50 reason¬
ably intense and distinct spots may be discerned in the photograph of
the electrophoretic pattern of proteins from each subunit. Somewhat
larger numbers could be seen on the original gels, since some of the
resolution of closely adjoining spots has been lost in the photographs.
It is evident in these figures that the description "reasonably
intense" involves a rather arbitrary judgment. In fact, particularly

Figure 5. Electrophoretic pattern of large-subunit proteins from bovine
mitochondrial ribosomes.
Mitochondrial ribosomes were dissociated to subunits by sucrose
density gradient centrifugation in Buffer E. Gradient fractions corre¬
sponding to the large (39S) subunits were pooled and centrifuged in a
Beckman Type 65 rotor at 65,000 rpm for 5.5 hours. The proteins were
extracted from the ribosome pellet and analyzed by two-dimensional
polyacrylamide gel electrophoresis. The first dimension (left to right)
was run in urea at pH 4.3, and the second dimension (top to bottom) in
SDS.
Just before the second dimension was begun, pieces of agarose gel
containing a mixture of several proteins (bovine serum albumin, human
y-globulin heavy and light chains, ovalbumin, myoglobin, and lysozyme)
were placed on top of the gel slab, immediately adjacent to each end
of the first-dimension gel. These proteins may be seen at the left side
of the photograph.

51
\
%
*
I
m
V m
*â–  *
Figure 6. Electrophoretic pattern of small-subunit proteins from bovine
mitochondrial ribosomes.
Small (28S) subunits were prepared by sucrose density gradient
centrifugation in Buffer E, and their proteins were subjected to two-
dimensional electrophoresis as described in Figure 5.

52
in the small-subunit pattern, there is a large range of spot inten¬
sities with a more or less continuous variation from the most intense
tc the very faintest. This result was not what had been anticipated.
Since nearly all ribosomal proteins probably are present in one copy
per ribosome, at least in vivo (Hardy, 1975), it had been expected
that the gels would show a discrete group of ribosomal protein spots
of roughly similar staining intensities. In addition there might be
some other spots of varying intensity, corresponding to contaminating
proteins. The results obtained indicated that there were considerable
differences in the relative quantities of the ribosomal proteins pres¬
ent in the sample, making it impossible to distinguish them from con¬
taminating proteins on the basis of their staining intensity. Some
possible explanations for this phenomenon will be considered below
(Section IV).
A. Criteria for the Identification of Ribosomal Proteins
For this reason it seemed necessary to seek, other criteria for the
purpose of identifying which of the spots appearing on the gels actually
represented ribosomal proteins. Several possible criteria were con¬
sidered, including reproducibility of occurrence, subunit-specific local¬
ization, and resistance to removal from the ribosomes by high salt
treatment. It seemed reasonable to expect that the ribosomal proteins
would be fcund reproducibly in separate preparations of the ribosomes,
whereas the non-ribosomal contaminants might be more variable in occur¬
rence .
In fact, it was found that many of the proteins s**en in individual

53
gels such as those of Figures 5 and 6, particularly some of the
fainter ones, could be disregarded on the basis that they were not
seen in the majority of the electrophoretic analyses. This rule was
thus found to be a useful one, and the application of it will be
described below.
The idea that some contamination might arise from the non-specific
binding of some proteins to the ribosomes suggested that contaminants
of this type might be found nearly equally in both ribosomal subunits.
However, the existence of some plausible mechanisms by which ribosomal
proteins might also show an apparent lack of subunit-specific local¬
ization (Section IIB3) made the application of this criterion difficult,
and it was used only in conjunction with other kinds of evidence.
Yet another way to discriminate the ribosomal and non-ribosomal
proteins, and the most unambiguous way, would be to purify all the pro¬
teins and then reconstitute the ribosome, withholding each protein in
turn, to establish which proteins were essential for the generation of
a structurally and functionally normal ribosome. Unfortunately, this
approach is not practical in the case of ribosomes such as those of
mammalian mitochondria, which can be obtained only in relatively small
quantities. However the observations (Olsnes, 1971; Kurland, 1966;
Hardy and Kurland, 1966) that contaminating proteins are in general
more loosely bound to ribosomes than are the ribosomal proteins them¬
selves, suggested that an analogous approach might be feasible.
Specifically, it has been found that treatment of ribosomes with

54
moderately high concentrations (about 0.5 >0 of monovalent salts dis¬
lodges adventitiously bound proteins from the particles, while affecting
the binding of the ribosomal proteins to a lesser extent. Of course,
higher salt concentrations do remove ribosomal proteins (Delaunay et al.,
1974; Gesteland and Staehelin, 1967) and the optimum concentration range,
that which removes the largest quantity of contaminants and the smallest
amount of ribosomal protein, may differ from one type of ribosome to
another.
The effect of increasing salt concentration on the content of
individual proteins in the bovine mitochondrial ribosome was assessed by
centrifuging the ribosomes through various high-salt buffers and then
analyzing the proteins remaining in the treated particles by two-dimen¬
sional electrophoresis. To establish the point at which these treatments
began to remove ribosomal proteins, the peptidyl transferase activity of
the ribosomes prepared under each condition was assayed. The electro¬
phoretic protein patterns and specific activities obtained v;ere then
correlated, and any proteins which were found to be removed from the
ribosome by a given treatment, without diminishing its functional integ¬
rity, were tentatively regarded as non-ribosomal contaminants. Since
the peptidyl transferase activity is a property of the large subunit
and is not dependent on the presence of the small subunit, this criter¬
ion was directly applicable only to the large-subunit proteins.
Figure 7 shows the effect of treatment with buffers containing
increasing ratios of KCi concentration to MgC^ concentration on the

Buoyant density Activity
(g/cc) (% of control)
55
KCI/MgClj
Figure 7. Effects of treatment with buffers of various ionic compo¬
sitions on the peptidyl transferase activity and buoyant density of
bovine mitochondrial large subunits.
Large subunits were prepared by sucrose density gradient centri¬
fugation in various buffers. Peptidyl transferase activity was
assayed as described in Materials and Methods, and expressed as a
percentage of the activity found in Buffer A treated subunits. The
buoyant densities of the formaldehyde-fixed subunits in CsCl were
determined by isopycnic centrifugation.
The buffer conditions used are plotted along the horizontal axis
according to their ratio of KC1 to MgCl? concentrations. The actual
concentrations of these two salts were:
Buffer
KC1(M)
MgCl„(mM)
A
0.1
20
C
0.5
10
D
0.3
5
E
0.5
5
F
1
5
G
1
1
complete compositions
of these buffers
are given in Table

56
activity and buoyant density of the large subunit. In this series,
buffeis containing increasing concentrations of KC1 and decreasing
concentrations of MgCl^ were used to wash the ribosomes with increas¬
ing stringency. For other ribosomes monovalent and divalent cations
have been found to produce opposing and competitive effects on the
structure of the particle, with respect to the association of the two
subunits with each other (Hamilton and Petermann, 1959; Zitomer and
Flaks, 1L< ‘J.) and the binding of proteins to the ribosome (Spitnik-
Elson and Atsmon, 1969; Staehelin e_t aJ., 1969). Such effects on
subunit association have also been found in mammalian mitochondrial
ribosomes (O’Brien, 1971), and Figure 7 indicates that either increasing
KC1 or decreasing MgClr, can diminish both the functional activity and
the protein content (as measured by the increase in buoyant density)
of these ribosomes.
The upper curve of Figure 7 shows that the large subunit retained
most of its peptidyl transferase activity after treatment with either
Buffer C or Buffer D. The activity was significantly reduced by treat¬
ment with Buffer F, and higher KCl/MgC^ ratios yielded particles with
negligible activity. Therefore in the analysis of the electrophoretic
results presented below, proteins which were found to be absent from
large subunits treated with either Buffer C or Buffer D are considered
to be dispensable for this particular large-subunit function. These
proteins are regarded as possible non-ribosomal contaminants, and are
discussed separately.

57
B. Identification of the Ribosomal Proteins
1. Large Subunit
In order to apply the criteria of reproducibility and resistance to
high-salt treatment, several electrophoretic analyses were performed on
large-subunit proteins taken from separate preparations of ribosomes,
after treatment with various high-sa]t buffers. The gels were scored for
the presence or absence of individual proteins, and the results are given
in Table VI. The electrophoretic positions of the proteins designated
in the table are shown in Figure 8. The 52 proteins which were seen
reproducibly and which were present both in large subunits treated with
Buffer C and in those treated with Buffer D are shown as numbered spots.
Lettered spots represent some of the proteins which did not satisfy
these criteria.
For the purposes of Table VI tiie relatively subjective question of
the intensities of the protein spots was not considered, and a
indicates only that the protein was clearly visible in the gel, whether
it was very intense or quite faint. Proteins which were near the limit
of visible detection, however, are marked with a n-n; some of these are
not visible in the photographs of the gels. In a few cases ("nr") the
electrophoretic resolution in a given experiment was inadequate to
determine whether a protein was present or not. The three proteins
which move fastest in the first dimension of electrophoresis (ML45,
ML51, and ML52) could have been run off the end of the first-dimension
gel in some experiments, and these cases are marked nro".

Table VI. Appearance of large-subunit proteins in separate experiments.
Large subunits were prepared by sucrose density gradient centrifugation in the buffers
specified. The proteins were then analyzed by 2-dimensional electrophoresis and the gels were
scored for the presence of the proteins shown in Figure 8. +, present; trace; 0, absent;
nr, not resolved; ro, run off the first dimension gel.
Buffer
B
Buffer C
Buffer D
Buffer
E
Expt.
1
Expt.
2
Expt. 3
Expt. 4
Expt.
5 Expt. 6
txpt. 7
Expt.
8
Expt. 9
Protein
(Fig.
9)
(Fig.14)
(Fig.10)
(Fig.11)
(Fig.
5)
i
+
+
+
+
+
_
+
+
+
2
+
+
+
+
+
-
+
+
+
3
-
-
-
+
+
-
+
-
+
4
+
+
+
+
+
-
nr
+
+
5
+
+
+
+
+
+
+
+
+
6
+
+
+
+
+
+
+
+
+
7
+
+
+
+
+
0
nr
+
+
8
+
+
+
+
+
0
+
+
+
9
+
+
+
+
+
+
+
+
+
10
+
+
+
+
+
+
+
+
+
11
+
+
+
+
+
+
+
+
+
12
+
+
+
+
+
0
+
+
+
13
+
+
+
+
+
+
+
+
+
14
+
+
+
+
+
+
+
+
+
15
+
+
+
+
+
-
nr
+
+
16
+
+
+
+
+
-
+
+
+
17
+
+
+
+
+
+
+
+
+
18
+
+
+
+
+
+
+
+
+
19
+
+
+
+
0
+
+
+
+
20
+
+
+
+
+
+
+
+
+
21
+
+
nr
0
+
+
+
+
+
22
+
+
+
+
+
+
+
+
+
23
+
+
+
+
+
+
+
+
+
24
+
+
+
+
+
+
+
+
+
25
4-
+
+
+
+
+
+
+
+
26
+
+
+
+
+
+
+
nr
27
+
+
+
+
+
+
+
+
+

59
i i + + + + + i + + i + + + + + + + + + + + + + +
o+o
+ + + + + + + I + + + + + 1+ + +0 + 01+ + I + o o + IO + 1 +
1 GC++l+ + + + + + + + + + + 0 + +0++}-i}-i ++0++G+ +
+ + + + + + +00 + + + + + + + +0 + + I + + o I
+ + O O O + O I
+ + + + + + + + + + + + + + + + + M + + 0 I +
OI+ + + + + I+ + I + + 1+ + +I+ + I+ + I
o
u
I O I O I O I c
+ + + + + + + I + + +• + +
I + G + I+++ + + 1 +
O O I O O G + O
I + + + + + + I+ + I + + i + + + + + + + + + + + lo»ooo + o
1 + + + + + 0C + 0 I + + + + +0 + +0 + + G G
I O I O + G + O
COO'OHC'l^sMTikDrsCO(JiOH(Nfrl^iO^NCO(T>Oi-IN
(i3 xi u 'd o) uh go

60
+ I-D -
>
Figure 8. Schematic diagram of bovine mitochondrial large-subunit
proteins.
The origin of electrophoresis is at the upper left corner of the
figure. The second-dimension electrophoretic positions of bovine serum
albumin (molecular weight 68,000), ovalbumin (44,000), human y-globulin
light chain (23,500), horse myoglobin (17,200), and egg-white lysozyme
(14,400) are indicated at the left of the figure.
Proteins which were reproducibly present in functionally active
large subunits are shown as numbered spots. Lettered spots designate
some of the other proteins which were seen in some of the electrophoretic
analyses.

61
a. Reproducibility
In order to discriminate ribosomal proteins from contaminants, an
arbitrary level of reproducibility of 50% was taken as the cut-off.
That is, a protein was considered to be non-ribosomaL unless it was
present in over half of the ribosome preparations analyzed. In fact,
however, very few of the proteins in large-suhunit preparations were
near this borderline. Most of the proteins which were found in reason¬
able quantity, and many of the faint ones, in any one gel proved to be
quite reproducible. A small number of proteins in the unreproducible
group were seen so rarely that it appears most unlikely that they
represent ribosomal proteins: the spot seen between ML16 and ML17 in
Figure 5, for example, was not found in any other experiment. Such
proteins are not listed in Table VI. The proteins designated by letters
in the table were seen more frequently, especially in ribosomes pre¬
pared in certain buffers. Because the presence or absence of these
proteins was correlated with the ionic conditions with which the ribo¬
somes bad been treated, they are considered below under "Resistance to
salt-washing".
Among the numbered proteins in Table VI, many appeared in every
ribosome preparation, and with reasonable staining intensity. Some
faint spots, like ML29 and ML41, were likewise very regular in occur¬
rence although sometimes present only in trace quantities. Occasion¬
ally a protein which was usually present in moderate quantity would be
unaccountably lacking in a particular experiment: ML-47, for example,

62
can be clearly seen in experiments 2 (Figure 9), 4 (Figure 10), and
6 (Figure 11) and is clearly absent from experiment 8 (Figure 5).
Similarly, ML1 and ML2 are always found to have about the same stain¬
ing intensity seen in Figures 5, 9, and 10, except that in experiment
6 (Figure 11) they were both reduced to barely detectable traces.
Although such results are quite surprising, the overall reproducibil¬
ity of those proteins was considered adequate to satisfy the criterion.
Predictably, the greatest variability was found among the faintest
spots, including the only two real borderline cases. ML48 was seen
with low but clearly visible intensity in three experiments (as in
Figure 9), as only a trace on three other occasions, and three times
not at all. ML45 was found even Jess often, again was sometimes only a
trace when it was present, but was sometimes quite intense (Figure 9).
Overall, relatively little variability was observed in the
pattern of large-subunit proteins from one experiment to another. Most
of the proteins seen on any one gel (including the faintly-staining
ones) were regularly found in separate preparations of ribosomes, and
thus probably represent ribosc-mal proteins, according to this criter¬
ion. Some possible explanations for those variations that were obtained
in these experiments will be discussed below In the analysis of the
small-subunit proteins, for which variability is a more serious prob-
1 em.

i
6 i
Figure 9. Electrophoretic pattern of proteins from large subunits
prepared in Buffer C (Experiment 2).

Figure 10. electrophoretic pattern of proteins from large subunits
prepared in Buffer C (Experiment 4).

Figure 11. Electrophoretic pattern of proteins from large subunits
prepared in Buffer D (Experiment 6).

66
b. Resistance to salt-washing
The salt-washing procedure used in the preparation of mitochondrial
ribosomes was designed to remove two types of contaminants, those which
were bound directly to the ribosomes and those which were not bound,
but nonetheless accompanied the ribosomes through the early stages of
the preparation. As mentioned in Section IIA, treatment of ribosomes
with high concentrations of monovalent salts should preferentially
release non-ribosomal material bound to them. Figure 12 shows the
effect of washing E. coli ribosomes with 1 M NH^Cl. It is evident that
the major difference between (he two protein patterns is the presence
of many high-molecular-weight proteins in the unwashed ribosomes. This
phenomenon is well-documented in the literature (Subramanian, 1974;
Hardy, 1975; Brouwer and Planta, 1975), and the proteins removed by
salt-washing are universally regarded as non-ribosomal for a number of
reasons, including the essentially undiminished functional activity of
the salt-washed ribosomes.
The fact that most high-salt buffers also dissociate mitochondrial
ribosomes to their subunits permits the removal of non-ribosome-bound,
co-sedimenting contaminants by the use of two cycles of sucrose-density-
gradient centrifugation under different ionic conditions. As discussed
above (Section I) the isolation of the ribosomes first at low ionic
strength (0.1 M KC1) as intact 55S particles, and then a higher salt
concentrations as subunits should separate them from most non-ribosome-
bound contaminants.

Figure 12. Effect of high-salt treatment on the protein content of E. coli ribosomes.
E. coli ribosomes were washed with Buffer N (60 mM KC1) or Buffer 0 (1 M NH^Cl) as described
in MATERIALS AND METHODS. The electrophoretic pattern of the proteins of Buffer N ribosomes is
shown in the photograph on the left, and of Buffer 0 ribosomes on the right. (The complete
compositions of these buffers are given in Table IV.)

68
Low-salt treatment. The affectiveness of this procedure can be
evaluated by comparing subunits prepared in this way ("derived subunits")
with the subunits which are present in small amounts in the first, low-
salt sucrose density gradient ("native subunits”; see Figure 3).
Figure 13 shows the electrophoretic patterns obtained from native large
and small subunits, and Figure 14 shows the pattern of the derived large
subunit from the same preparation of ribosomes. The most obvious
difference is the presence of two high-molecular-weight proteins in
large quantity in the native subunits. Several other proteins of sim¬
ilar and higher molecular weights are also present in native but not in
derived subunits; many of these, like the two major proteins, are found
equally in both the large and small subunits. Another native-subunit-
specific protein, seen to the left of ML30, is also feund in somewhat
smaller quantity in the small subunit. Additionally, numerous proteins
of various molecular weights and staining intensities appear in the
native large subunit but not in either of the other two patterns.
It seems reasonable to conclude that most if not all of the proteins
found in native but not derived subunits represent non-ribosomal contam¬
inants. The similarity of these results to those obtained with F,. coli
ribosomes (Figure 12) is clear, particularly with respect to the high-
molecular-weight proteins. The lack of subunit-specificity of many of
these proteins is also incriminating. Most importantly, they are
essentially absent from large subunits derived in Buffer C or Buffer D,
and such subunits are functionally active. By visual comparison of the

u
5
5
'
*
*
i
figure 13. Electrophoretic patterns of proteins from low-salt treated large and small subunits.
The proteins of mitochondrial large (left-hand photograph) and small (right-hand photograph)
subunits, prepared by sucrose density gradient centrifugation in Buffer H, were analyzed by two-
dimensional electrophoresis.
O'
vC

70
Figure 14. Electrophoretic pattern of proteins from large subunits
prepared in Buffer C (Experiment 3).
The irregular white line in the upper left is a crack in the gel.

71
relative staining intensities of tliese proteins in native and derived
subunits most of them appear to be reduced in quantity by 90% or more
in derived subunits. Yet the activity of Buffer C derived subunits
is 75% of that found in native subunits (shown as Buffer A subunits
in Figure 7). Thus none of tliese proteins can be essential for the
peptidyl transferase activity of the ribosome.
The kinds of contaminants which have thus been identified are
various, although some loose generalizations can be made about them.
The contaminants may be present in a very wide range of staining in¬
tensities, have molecular weights ranging from values similar to those
of the ribosomal proteins to considerably larger values, and may be
present subunit-specifically or not. However, a major part of the
contamination seems to be represented by a group of relatively intense,
high-molecular-weight, non-subunit-specific proteins. It is of some
interest to ask whether this contamination is due primarily to adherent
proteins or to protein-containing structures that are not bound to the
ribosomes but are large enough to sediment in ttie same region of the
sucrose density gradient. The buoyant density of the native subunit
is significantly lower than that of the Buffer C or Buffer D derived
subunits (Figure 7), suggesting that at least some of the native-
subunit-specific protein must be bound directly to the particle under
these low-salt conditions. Further, in two experiments the subunits
were derived directly from the crude ribosome pellet, rather than from
the 55S fractions of a sucrose density gradient. Such subunits showed

72
only small quantities of these contaminants (Figure 10). Most of
these proteins thus seem to he bound directly to the ribosomes under
low-salt but not high-salt conditions, rather than components of
separate structures which also sediment in the 25S to 45S region of
the gradient.
High-salt treatment. In addition to these proteins which were
absent from all derived large subunit preparations, several proteins
were found to be removed when the ribosomes were dissociated under
some ionic conditions but not others. These proteins are the ones
designated by letters in Table VI. The buffer conditions of interest
are Buffer C and Buffer D: since large subunits prepared in either
of these conditions are functionally active (Figure 7), the absence
of a protein from either kind of preparation indicates that that pro¬
tein is not required for the function. Thus MLc and MLg are not
essential for peptidyl transferase activity because Buffer D subunits
lack them but possess the activity. MLb, d, e, f, and h are specif¬
ically absent from Buffer C subunits. MLa is found in Buffer C
preparations, but only in much smaller quantity than in Buffer D sub¬
units; since this large quantitative difference does not correlate with
the similarity between the two kinds of subunits in functional activity,
this protein too may be regarded as nonessential.
The fact that these proteins are dispensable for the peptidyl
transferase activity of the ribosome does not necessarily imply that
they are contaminating non-ribosomal proteins. It only indicates that

73
they are not components oí the active site for this function, and are
not structurally required to maintain the proper conformation of the
active site. Peptidyl transferase is only one of many reactions and
interactions involved in protein synthesis, including translocation,
initiation, termination, and binding to factors and mRNA. And not all
of the ribosomal proteins are required for each function. For example,
the El. coll ribosomal protein L7/12, which is needed for the inter¬
action of the elongation factors EF-T and EF-G with the ribosome, can
be removed without significantly affecting the peptidyl transferase
activity (Hamel et al., 1972). Therefore, in the absence of data
demonstrating that both Buffer C and Buffer D ribosomes are active in
a more demanding assay (such as the translation of poly U or a natural
mRNA), it remains quite possible that all of the lettered spots in
Figure 8 do represent ribosomal proteins.
Extreme-salt treatment. As an attempt to identify some proteins
which were required for peptidyl transferase activity (and which there¬
fore must clearly be ribosomal proteins), the electrophoretic patterns
of large subunits prepared with Buffers E, F, and G were evaluated.
Since Buffer F and Buffer G subunits are essentially inactive, it seemed
likely that they would be missing one or more proteins, and that among
these proteins at least one was required for the activity. One or sev¬
eral of these proteins might also be found in diminished quantities in
Buffer E subunits, which have diminished activity.

74
For Chis purpose, two preparations of Buffer F subunits and two
of Buffer G subunits were subjected to two-dimensional electropho¬
resis. One of the Buffer F analyses is shown in Figure 15,and one of
the Buffer G gels in Figure 16. Some differences were observed be¬
tween the Buffer F preparations and between the Buffer G preparations,
but there were several proteins which were absent or significantly
diminished in staining intensity in both of the analyses in each pair.
Buffer F did not completely remove any of the numbered large-subunit
proteins, but it did reduce five of them to trace-level staining inten¬
sity. ML1, 15, 16, and 18 were diminished to a tenth or less, and ML2
to less than half, of the quantity normally present in Buffer C or
Buffer D subunits. In subunits treated with Buffer G, small quantities
of ML16 and ML18 were still present, but ML1 and ML15 were not detectable.
In ¿iddicion ML23, 29, 30, 33, 43, and 48 were absent and the quantity
of ML47 was considerably reduced.
Interestingly, the amount of ML2 found in Buffer G subunits did not
appear to be significantly less than that in subunits treated with
Buffer C or Buffer D, even though the less stringent treatment with
1 M KC1 at 5 mH MgClj (Buffer F) did remove some of this protein. A
similar phenomenon was observed among the large-subunit proteins which
are designated by letters. >ILg, which is absent from sub . , its treated
with the moderate-salt Buffer D and is present only in trace quantities
after treatment with Buffer F, is not removed by Buffer G. Such results
are at odds witli the expectation that the set of proteins removed by a

Figure 15. Electrophoretic pattern of proteins from large subunits
prepared in Buffer F.

76
Figure 16. Electrophoretic pattern of proteins from large subunits
prepared in Buffer G.

77
given ionic medium should be a subset of the proteins removed by more
stringent conditions. This expectation derives from the continuous
decrease in the total protein content of the subunit which is ob¬
served as the KCl/MgCl^ ratio in the medium is increased (Figure 7).
Apparently the generalization that increasing KCl/MgCl^ ratios remove
increasing amounts of protein applies to the total protein content (as
measured by the buoyant density) and to some of the individual pro¬
teins, but ML2 and MLg are exceptions. That a few such exceptions to
this rule might exist, due to specific effects of particular buffer
compositions on the binding of some of the proteins to the ribosome,
is not too surprising.
The protein complement found in Buffer E treated large subunits,
however, represents a more serious anomaly. None of the numbered pro¬
teins were found to be removed or significantly diminished in quantity
by washing with this buffer, offering no explanation for the reduction
in functional activity produced by this treatment. Furthermore, even
the lettered proteins (which are all removed by either Buffer C or
Buffer D) were each found at least once in preparations of Buffer E
subunits (Table VI). Thus the buoyant density data (Figure 7) indicate
that these particles must be lacking some proteins present in Buffer C
or Buffer D subunits, but no missing proteins can be identified by
electrophoretic analysis. The reasons for this discrepancy are not
clear, although some considerations described below (Section IV) may be
relevant to this question.

78
As mentioned above, several proteins are specifically missing or
reduced in quantity in large subunits that have been treated with ionic
conditions sufficiently stringent to inactivate them* In particular,
the loss of ML1, 15, 16 and 18 correlates with the loss of functional
activity. In order to prove that the inactivity of the stripped sub¬
unit is dm- to the loss of these proteins (and not, for example, due
to a conformational alteration of the subunit induced by these salt
conditions) it is necessary to demonstrate that the activity can be
reconstituted by adding the proteins back. Such a partial reconstitu¬
tion would also be useful as an assay to determine which of the missing
proteins is (are) required for the activity. However, attempts to
reconstitute the subunit were unsuccessful,
c. Summary of the large-subunit proteins
Electrophoretic analysis of the large subunit of the bovine mito¬
chondrial ribosome revealed 52 protein components which were repro-
ducibly present in functionally active large subunits and which there¬
fore probably represent ribosomal proteins. Eight additional proteins
were present in large subunits prepared only under certain ionic
conditions; these may also be ribosomal proteins, but they are not
required for the peptidyl transferase activity of the large subunit.
Treatment with very-high-salt buffers removes a small number of the 52
proteins and destroys the peptidyl transferase activity, suggesting
that one or more of the missing proteins is required for this function.

79
2. Small Subunit
As shown In Figure 13 above, native small subunits prepared in low-
salt buffers contain a set of mainly high-molecular-weight proteins
which are also found in the native large subunit. Although lesser
quantities of these proteins are sometimes found in derived small sub¬
units treated with moderate salt concentrations (Figures 6 and 17),
they are usually much diminished relative to the quantities present In
the native subunit. Figure IS, obtained from the derived small subunit
from the same ribosome preparation whose native subunits are shown in
Figure 13, is completely lacking in these proteins. For some of the
same reasons given in the discussion of the large subunit, it seems
unlikely that these proteins represent functional components of the
small subunit either.
Table VII shows the results of several electrophoretic analyses of
high-salt-washed small subunits. The proteins which were found regu¬
larly in separate ribosome preparations and thus satisfy the criterion
of reproducibility are designated by numbers. Some of the proteins
which were seen less frequently are also tabulated and are designated
by letters. (To avoid confusion with the numerals 1 and 0, the letters
1 and o have been skipped.) The electrophoretic positions of these pro¬
teins are given in Figure 19.
The overall level of reproducibility of the small-subunit proteins
was less than that of the large-subunit proteins. To some extent this
variability is due to the absence of many of the proteins from a partic-

Figure 17. Electrophoretic pattern of proteins from small subunits
prepared in Buffer C (Experiment 2).

81
Figure 18. Electrophoretic pattern of proteins from small subunits
prepared in Buffer C (Experiment 3).

Table VII. Appearance of small-subunit proteins in separate experiments.
Small subunits were prepared by sucrose density gradient centrifugation in the buffers
specified, and the proteins were analyzed by 2-dimensional electrophoresis. The electro¬
phoretic positions of the proteins listed are shown in Figure 19. +, present; -, trace;
nr, not resolved.
Buffer B
Buffer C
Buff
er D
Buffer E
Expt. 1
Expt. 2
Expt. 3
Expt. 4
Expt. 5
Exp t.
6
Exp t. 7
Expt. 8
Expt. 9
Protein
(Fig.17)
(Fig.18)
(Fig.6)
(Fig.20)
1
-
4
_
+
0
0
4
4
4
2
nr
nr
+
nr
4
-
nr
4
4-
3
nr
nr
+
nr
4
-
nr
4
4
4
4
4
4
+
4
4
4
4
4
5
+
+
+
+
4
4
4
4
4
6
4
+
4
+
4
4
4
4
4
7
0
4
4
+
4
4
4
4
4
8
4
+
0
+
4
-
nr
4
4
9
0
+
0
-
4
4
4
4
4
10
+
4
4
4-
4
4
4
4
4
11
4
-
+
+
4
4
4
4
-
12
4
4
0
+
4
4
4
4
4
13
+
+
+
+
0
0
0
-
4
14
0
4
0
4
4
4
4
4
4
15
+
4
nr
+
4
4
4
4
4
16
4
+
+
+
4
0
0
-
4
17
4
4
+
4
4
4
4
4
4
18
4
4
-
4
4
0
4
4
4
19
4
4
+
4
4
4
4
4
4
20
4
4
-
-
4
0
4
4
4
21
4
4
0
4
0
0
4
4
4
22
4
4
-
4
4
0
4
4
4
23
4
4
+
4
0
-
nr
4
-
24
+
+
+
4
4
4
4
4
4
25
-
4
+
4
4
0
4
4
4
26
4
4
+
4
4
4
4
4
4
27
4
nr
0
4
4
0
4
4
4

x> 3 3 ft' lj. h* cn Hi fD clo cro
Í'-MJLOUUUiWU'LiJWOjWM
HO^OOSvl'T'ln-C'O.'fOl—1 O '■£> 00
00 + 0 1 + +0 + 003 00 + I+ +0 + 0 + 0 + 3 3
*-l 3 t-i
000 + 000000 + 0+ + + I + + + + + + + + + +
000 + 0000 + 000++ + I + + + + ++ I + + 0
3
IOOOOOOOOOOOI I +I+++ + + +I++O
+ 0+ I+ 0+ 00 + + 000 + I I+ + 0 + OI++ +
0000000++00000 +0+ + + I V+ + + + 0
+ 003 00 + 00++000 ++0 + + 0+I+ + + +
3
+ IOOI 100000 + 00 +I+ + + + +I + I+ + + +
00010000 + 0001 I +I++ + + + +I+ + + + +
£8
nr + + 0 + 0 nr

84
+ l-D -
Figure 19. Schematic diagram of bovine mitochondrial small-subunit
proteins.
Proteins which were reproducibly observed in preparations of small
subunits are designated by numbers, and proteins seen less frequently
are marked with letters.

85
ular experiment. In Experiment 6 for example, the quantity of protein
in the sample was inadequate; as a result all of the spots obtained
were faint, so the absence of many of the spots which are normally
faint in adequately loaded gels is understandable. The absence of
6 of the numbered proteins from Experiment 3 (Figure 18), however, is
not so simply explained. On the other hand, there was a subset of the
experiments in which variability was not a problem at all: none of
the numbered proteins were lacking in Experiments 2, 8, or 9, and only
two were missing in Experiment 4.
The variability which was observed in the other experiments seemed
to be random with respect to which proteins were found to be missing.
That is, it was not due to a high level of variability in a few specific
proteins. Only two of the numbered proteins, MS13 and MS21, were absent
from as many as three of the nine experiments. In the case of these
two proteins (as well as MSI), much of the lack of reproducibility was
due to their absence from suhunits prepared in buffer D; thus there may
be a specific ionic effect on the binding of these proteins to the ribo¬
some, as was seen for MLc and MLg above. A complementary effect was
found for MSa and MSb. These two proteins were always present in sub¬
units prepared in Buffer C, but only once in any other preparation.
The general impression of unreproducibility in the small subunit
pattern derives partly from the relatively large number of spots which
appear in fewer than half of the analyses and thus are designated by
letters in Table VII. Some of these, like. MSc and MSn (Figure 6),

86
appeared in only one experiment and can be easily disregarded. Others
(MSa, b, f, k, and p) were present in four of the nine experiments and
thus represent borderline cases which are not clearly excluded from the
ribosomal protein category.
Another factor which contributes to this impression of unrepro-
ducihility is the variation in staining intensity among the different
spots on a given gel. In most experiments (Figure 17 is a conspicuous
except ion) the electrophoretic pattern was found to consist of a minor¬
ity of heavily-stained spots, usually including MS4, 5, 6, 10, 17, 19,
24, 28, 31, 37, and 38, and a majority of fainter spots. Furthermore
the relative intensity of a given spot, particularly among the fainter
group, varied considerably among different gels. (Compare MS13, 15,
and 16 in Figures 6 and 20, for example,) Some differences in the
stain-binding abilities of different proteins have been reported
(Fazekas de St. Groth e_t al. , 1963; Bickle and Traut, 1971), but the
different os are not of the magnitude required to explain the present
results if all of the sma11-subunit proteins were present in equimolar
quantities in these gels. And in any case this phenomenon could not
account for the variations observed for a particular protein in
separate experiments. It is therefore necessary to conclude that the
small-subunit proteins are present in differing molar quantities in
these preparations. It follows that the individual subunits in each
preparation are heterogeneous with respect to the set of proteins they
contain.

M7
Figure 20. Electrophoretic pattern of proteins from small subunits
prepared in Buffer E (Experiment 9).

88
This heterogeneity could be due to the non-specific binding of con¬
taminating proteins to some of the subunits. However, in view of the
small number of heavily-staining proteins found in these particles, it
seems unlikely that all of the fainter spots represent contaminants. A
more plausible hypothesis is that a proportion of the subunits are
lacking some of the ribosomal proteins, as a result of an i_n vitro
artifact. This hypothesis is supported by the similar results which
have been oh ained (Kaltschmidt and Wittmann, 1970) and explained
(Hardy, 1975) for El. coli ribosomal proteins. Some possible causes of
such an artifact will be considered below (Section IV); but whatever
the cause, it is probably responsible for the variability between
ribosome preparations as well as the heterogeneity within each prepa¬
ration. Similar considerations also apply (to a lesser extent) to the
proteins of the large subunit, which show a lower level of variability
and non-uniformity of staining intensities.
The effect of extreme-salt treatment (Buffer G) on the protein
content of the small subunit was also investigated. One of the gels
obtained is shown in Figure 21. The most notict.ible result was a
general reduction in the quantity of the low-molecular-weight proteins,
those below MS31. Several proteins (MS13, 16, and 35) were completely
absent from all three preparations of Buffer G subunits, and MS27, 28,
33, 36, 37, and 41 were diminished to negligible levels. Some proteins,
on the other hand, were increased relative to the quantities present
in high-salt-treated (Buffers C and D) subunits: this was particularly

K9
Figure 21. Electrophoretic pattern of proteins from small subunits
prepared in Buffer G.
The irregular spots seen above and to the right of MS4 are
artifacts due to precipitation of the stain on the surface of the gel.

true of MS8, 12, and 14. In addition, some of the lettered, unreprodu-
eible proteins were found reproducibly in Buffer G subunits. MSd, e, f,
and m were each found in all three preparations. It was also observed
that several of the high-molecular-weight proteins present in low-salt
but not high-salt subunits are present in considerable quantity in these
extreme-salt subunits. This rather confusing set of observations is fur
ther evidence for a conclusion reached earlier in the discussion of the
large subunit. The general effect of treating these ribosomes with
buffers of increasing KCl/MgC^ composition is to reduce both the total
amount of protein as measured by the buoyant density of the particles
and the number of proteins as detected by gel electrophoresis. But some
of the individual proteins may show the opposite response- to such treat¬
ments, being preferentially removed by buffers of lower KCl/MgCI^ ratios
As was the case with the large subunit, no proteins were found to
be removed or significantly diminished by treatment with Buffer E
(Table VII), even though the buoyant density of the Buffer E subunit
is somewhat higher than that of Buffer C or Buffer D subunits. Thus
buoyant density measurements appear to be a more sensitive means of
detecting the loss of protein as a function of KCl/MgClo than the
electrophoretic anal,ses are. This lack of sensitivity may be due Lo
the high background level of variability between electrophoretic
experiments, especially in the case of the small subunit. However,
it also seems likely that these treatments are acting non-selectively
to some extent, removing small quantities of many of the proteins

91
(and thus increasing the heterogeneity of the subunits with respect
to their protein content) rather than larger quantities of only a
few proteins. Any such non-seLective effects would not be detected
by comparing the relative staining intensities of the proteins Jn the
gel.
3. Comparison of Large and Small Subunits
In order to determine whether the proteins of the large subunit
were all distinct from those of the small subunit, Che two electro¬
phoretic patterns were compared with each other. To obtain a suffi¬
ciently accurate relative positioning of the patterni, samples of
large- and small-subunit protein» were mixed together and electro-
phoresed on the same gel: one sample was a trace quantity ol protein
which had been radfoiodinated as described in MATERIALS AND METHODS,
and the other sample was a larger (stainable) quantity of protein.
After electrophoresis, staining, and autoradiography, the patterns
of stained and radioactive proteins were compared. Control experi¬
ments in which a small aliquot of a protein sample was radioiodinated
and then mixed with a large aliquot of the same sample before electro¬
phoresis showed that the labelling reaction had no detectable effect
on the electrophoretic mobilities of the proteins.
As shown in Figure 22, it was found that many of the small-subunit
proteins overlapped at least partially with proteins of the large
subunit, and there were twelve pairs of proteins which were electru-
phoretically indistinguishable. These were MSI and ML2, MS4 and ML3

y 2
4- l-D -
•>
Figure 22. Relative electrophoretic positions oi bovine mitochondrial
large-subunit and small-subunit proteins.
Filled spots, large-subunit proteins; open spots, small-subunit
proteins; hatched spots, overlapping large-subunit and small-subunit
proteins.

MS7 and ML5, MS11 and ML12, MS15 and Mi l 7, MS23 and ML20, MS25 and
ML25, MS26 and ML27, MS30 and ML32, MS31 and ML31, MS32 and ML35, and
MS40 and ML46. Only the pro Leins which have been designated by numbers
in the discussion above are shown in Figure 22. In addition the
electrophoretic positions of MSk and ML38 overlap, as do those of
MS37 and MLh.
There are several possible explanations for the relatively large
number of olectrophoretically identical proteins in the two subunits.
First, it might be that two copies of some proteins are present in the
ribosome, one in each subunit, even though such a double role for a
single protein has not been demonstrated in any other ribosome to date.
Alternatively it may be that the two proteins in each pair are non¬
identical but are not resolved by the electrophoretic system; in view
of the generally crowded appearance of Figure 22 it would not be
surprising if some proteins occupied the same electrophoretic position
by coincidence. A third possibility is that some of these proteins
are contaminants similar to the non-subunit-specific proteins found
in native subunits (Figure 13), except that they have molecular weights
similar to those of the ribosomal proteins and remain bound to the
ribosomes even after washing with high-salt buffers.
Finally, it is possible that each of the protein pairs represents
a single ribosomal protein which is present at the level of one copy
per intact ribosome, but fails to bind exclusively to one subunit or
the other when the ribosome is dissociated. Such a partitioning of

94
individual proteins between the dissociated subunits has been observed
in other ribosomes. The E,. coli small-subunit protein S20, which co¬
electrophoreses with the large-subunit protein L26, has been shown
to be identical to it by immunochemical and genetic studies
(Tischendorf et al_. , 1974; Wittmann et_ al., 1974). Significant quan¬
tities of S5 have been found in the large subunit as well (Tischendorf
et a] ., 1974). Likewise, three proteins of rat cytoplasmic ribosomes
appear to be shared between the two subunits upon dissociation
(Sherton and Wool, 1974).
If this phenomenon is responsible for some of the overlapping
protein spots observed, several related effects might also be seen.
First, this hypothesis requires that the total quantity of protein
present in each pair represent no more than one copy per ribosome.
This requirement could be satisfied by almost all of the protein pairs
(MS31/ML31 is the exception), since in each case one or both members
of the pair are seen in quantities significantly smaller than the
average of the proteins present on the gel. It might also be expected
that even though these proteins do not bind exclusively to either of
the subunits, they do bind somewhat more strongly to the subunit with
which they are associated in the intact ribosome and thus are found
preferentially in this subunit upon electrophoretic analysis. In
fact, in six of the protein pairs, one protein of each pair does appear
in much greater quantity than the corresponding protein in the other
subunit. The overlapping proteins which are present in relatively

95
larger quantity on one subunit are ML2, 5, 17, 25, and 27, which are
found in greater amounts than the corresponding small-subunit proteins,
and MS4, which occurs in greater amounts than ML3.
Since the failure of such proteins to segregate exclusively to
one subunit or the other upon dissociation probably represents an
artifact of the method used to effect the dissociation, it seems likely
that differences in the ionic composition of the buffer used for this
purpose could alter the extent of the protein partitioning. In fact
it was reported (Sherton and Wool, 1974) that the suhunit localization
of one of the three partitioning proteins in rat cytoplasmic ribosomes
was determined by the ratio of potassium to magnesium ions in the
dissociating buffer. In the present results such a quantitative shift
of a protein from presence only In one subunit to presence only in the
other, due to differences in the ionic medium, was not observed. How¬
ever MSI was found to bo specifically absent from small subunits
prepared in Buffer D, and MS23 and 25 were significantly reduced in
quantity in this ionic condition as well.
Clearly, this phenomenon could contribute to the heterogeneity in
protein content observed in these subunits, although it could not
completely account for it. In any case more direct evidence to test
this hypothesis will not be easy to obtain. Whether the proteins in
each pair are in fact identical can perhaps be determined by chemical
or immunochemical analysis of the isolated proteins. If they are
identical, however, the question whether they are ribosomal proteins

9(1
or non-ribosoraal contaminants still rests primarily on the criteria
discussed above. Some additional evidence that they are rihosomal
proteins would be obtained if they could he shown to be located at
the interface between the two subunits in the intact ribosome; if
these proteins are responsible for the binding of the subunits to
each otli.-r, their partitioning behavior is more easily understood.
The _E. coi i S20/L26 pair has been found to be a subunit-interface pro¬
tein (Morrison et_ al_., 1973).
u. Experimental Evaluation of Possible Artifacts
A total of 93 proteins (the numbered spots in Figures 8 and 19)
were found reproducibly in high-salt-treated mitochondria! ribosome
subunits. Eighty-one of these proteins were distinguishable from each
other on the basis of their electrophoretic mobilities. Although
neither of these numbers is greater than had been expected on the basis
of measurements of tile total protein content of this ribosome, they
are considerably in excess of the numbers of proteins which have been
reported for any other kind of ribosome. For this reason it was neces¬
sary to consider the possibility of various artifacts which might
generate an unrealistically ! ;¡ge number of apparent ribosomal proteins.
One such possibility, the presence of adventitiously bound contaminants
in the ribosome preparation, has been discussed above.
Another possibility is that some single proteins may be represented
by more than one spot on a gel. This could occur if a protein had been
partially modified as the result of either a normal in vivo process or

97
an artifact of the isolation or electrophoretic procedure. Thus E. coli
L7 is actually identical to L12, except that it has been acetylated at
the N-terminus in vivo (Terhorst et_ chL. , 3 972); similarly, the rat liver
cytoplasmic ribosomal protein S6 sometimes appears as multiple electro¬
phoretic species due to in vivo phosphorylation (Cressner and Wool,
1974). Two electrophoretieally distinct forms, thought to represent
different states of oxidation, have been observed for each of the
coli proteins Sil, S12, and S17 (Wittmann, 1974). Carbamylation by
cyanate ions formed spontaneously in urea solutions can alter the
electrophoretic mobility of proteins (Gerding et al., 1971). In the
electrophoretic system used in the present experiments, these kinds of
chemical modifications could result in a small alteration of a pro¬
tein's mobility in the first electrophoretic dimension (in urea at
pH 4.5), but would not affect the migration in the second dimension
(in SDS). Some groups of spots which do show this electrophoretic
pattern include ML42, 43, and 44; MS30 and 31; and MS37 and 38.
Extra spots might also be produced by proteolytic degradation of
some of the proteins to discrete polypeptide products. Alternatively,
some of the proteins may be covalently crosslinked to give specific
aggregates. Disulfide crosslinks have been reported to form during
two-dimensional electrophoresis (Kaltschmidt and Wittmann, 1970).
Peroxidation of lipids in the crude ribosome preparation could also
lead to protein crosslinking (Tappel, 1973).

98
Some precautions against these possihle artifacts were taken in
the experiments presented above. BHT was added to all solutions used
in the ribosome preparation to prevent peroxidation-induced crosslinks.
Disulfide bonds x^ere reduced with dithiothreitol, and the proteins were
maintained in the reduced state by the inclusion of reducing agents
throughout the electrophoretic procedure. In addition, the electro¬
phoretic results themselves argue against the likelihood of some of
these artifacts. Carbamylation, for example, would be expected to
affect most or all of the proteins to a similar extent; thus the pres¬
ence of many proteins which do not show the peculiar electrophoretic
pattern represented by ML42, 43, and 44 argues that such patterns are
not produced by a non-specific chemical modification.
To obtain more evidence about the possibilities of proteolytic,
crosslinking, or other chemical alterations, as well as the non¬
specific binding of contaminants to the ribosomes, the effect of the
ribosome isolation procedure on the proteins of purified E. coli
ribosomes was assessed. To a preparation of bovine liver mitochon¬
dria was added a small quantity of salt-washed E. coll ribosomes,
equal to the quantity of mitochondrial ribosomes the mitochondria
were estimated to contain. The mitochondria were then lysed with
Triton X-100 and ribosomes pelleted at 100,000 x g as usual. After
sucrose density gradient centrifugation in Buffer H, the ribosomes in
the 70S (E!. coli ribosome) peak were dissociated to subunits by a
second centrifugation in a sucrose density gradient in Buffer E. The

99
proteins extracted from these subunits were then analyzed by two-
dimensional electrophoresis.
The results suggested that the types of artifacts mentioned above
are not a problem in these experiments. The electrophoretic pattern
of the proteins of the _E„ coli 50S subunit showed no significant
difference from that obtained from ribosomes subjected to the same re¬
isolation ,-rocedure in the absence of mitochondria, or from freshly
prepared ribosomes. The 30S subunits were apparently slightly con¬
taminated with mitochondrial large (39S) subunits, since traces of all
of the most intensely staining 39S proteins were found on the gel of
30S proteins; with this exception, the 30S subunit protein pattern was
likewise unaltered in the number and electrophoretic mobilities of the
proteins. Thus incubation with components of the mitochondrial lysate
under the conditions of our normal preparation procedure does not
cause any increase in the apparent number of ribosomal proteins of
exogenously added ribosomes. This result argues especially against
the possibilities that the number of proteins found in mitochondrial
ribosomes is overestimated due to proteolysis, cross!inking, or the
adsorption of non-ribosomal proteins.
Another possible artifact of the ribosome isolation procedure was
mentioned in Section I above: tlie binding of the ribosomes to DEAE-
cellulose, or the subsequent elution, during the preparation of ribo¬
somes by Method C might dislodge some of the ribosomal proteins. This
possibility seems unlikely in view of the undiminished functional

100
activity of ribosomes prepared in this way. However, the electro¬
phoretic results presented above include analyses of subunits prepared
both with and without the use of DEAE-cellulose„ so a comparison of
the protein patterns produced by these two methods is easily made. Tn
Tables VT and VII the samples for Experiments 2, 3, and 4 were DEAE-
cellulose prepared ribosomes. Comparing these three experiments with
tie other experiments tabulated shows no proteins which are clearly
removed by treatment with DEAE-cellulose, although some possibly sig¬
nificant effects can be seen. They most likely involve proteins MS9
and MS14, which are both absent from two of the DEAE-cellulose prepara¬
tions whereas they are otherwise quite reproducible. Somewhat weaker
correlations are observed for MS30, MS33, and ML21. Thus it remains
possible that a few proteins, particularly in the small subunit, are
specifically removed by treatment with DEAE-cellulose. It may also be
that MS9 and MS14 are unnecessary for the translation of poly U; however,
a more direct comparison between the activities of ribosome preparations
known to contain and to lack these proteins will be necessary to estab¬
lish this point.
IT!. COMPARISON WITH PROTEINS OF NON-MITOCHONDRIAI, RIBOSOMES
Yet another way in which the number of proteins in mitochondrial
ribosomes could be over-estimated by the results of these experiments
would be the possible contamination of the preparations by significant
quantities of cytoplasmic ribosomes. Although the more purified prep¬
arations described in Section T contain no detectable cytoplasmic

101
ribosomes, it seemed worthwhile Lo demonstrate directly the absence of
cytoplasmic ribosomal proteins from the electrophoretic patterns of
mitochondrial ribosomes. It was expected that the results would also
provide a comparative basis for the description of the properties of
mitochondrial ribosomal proteins.
The relative electrophoretic positions of the two sets of ribosomal
proteins were determined by co-electrophoresing radioiodinated and
stainable protein samples as described in Section IIB3. The results
for the large subunits of Lhe two ribosomes are shown in Figure 23, and
for the small subunits in Figure 24. With the exception of one protein
which comigrates with MS25, all of the mitoribosomal proteins are
electrophoretically distinct from those of the corresponding subunit of
the cytoplasmic ribosome. Indeed most of the cytoribosomal proteins are
found in a different region of the gel from that which contains most of
the mitoribosomal proteins. Comparing proteins of similar molecular
weights (similar positions in the second electrophoretic dimension),
it is evident that most of the cytoribosomal proteins migrate more
rapidly in the first dimension than do the mitoribosomal proteins. This
tendency, which is most evident in the comparison of the large subunits,
implies that the cytoplasmic ribosomal proteins bear a greater positive
charge at pH 4.3 than the mitoribosomal proteins. It also suggests that
the isoelectric points of many of the cytoribosomal proteins may be
higher than those of the mitoribosomal proteins.

102
+
l-D
â– >
Figure 23. Relative electrophoretic positions of bovine mitoribosomal
and cytoribosomal large-subunit proteins.
Filled spots, mitoribosomal proteins; open spots, cytoribosomal
proteins.

103
l-D
â– >
Figure 24. Relative electrophoretic positions of mitoribosomal and
cytoribosomal small-subunit proteins.
Filled spots, mitoribosomal proteins; open spots, cytoribosomal
proteins; hatched spot, overlapping mitoribosomal and cytoribosomal
proteins.

104
Comparing the mobilities of the two sets of proteins in the second
electrophoretic dimension shows that the distributions of molecular
weights of the mltoribosomal and cytoribosomal proteins are similar.
The large-subunit proteins of the mitochondrial ribosome are slightly
smaller on the average than those of the cytoplasmic ribosome, and the
reverse is true for the small-subunit prcteins.
The major difference in electrophoretic characteristics between
mitochondrial and bacterial ribosomal proteins lies in the much greater
second-dimension mobility of the latter. As seen in Figure 12, over
half of the ribosomal proteins of Eh coli have molecular weights less
than tbit of myoglobin (17,200), whereas relatively few mitoribosomal
proteins are this small. A more detailed description of the molecular
weight properties of the mitoribosomal proteins is given in Section IV.
The suggestion above that mitoribosomal proteins may have iso¬
electric points significantly lower than those of other kinds of ribo¬
somal proteins has some support in the literature. The relatively slow
first-dimension migration of mitoribosomal proteins can be seen in the
comparison of the large subunits of Xenopus laevis ribosomes (Leister
and Dawid, 1974, Figure 7). About half of the proteins of rat mito¬
chondrial ribosomes are anionic at pll 8.6 (van den Rogert and de Vries,
1976), whereas nearly all eukaryotic cytoplasmic and bacterial ribosomal
proteins are cations at this pH. Indeed, while most ribosomal proteins
have isoelectric points that are too high to permit analysis by iso¬
electric focusing, it has been reported that rat mitoribosomal proteins

105
can be analyzed by this technique (Czempiel et_ aJL. , 1976). The iso¬
electric points of most of the proteins were reported to lie below
a pi of 8.5, with none above a pi of 10.
Since the two-dimensional electrofocusing/electrophoresis system
described in the latter report should be capable of much greater
resolution than the system used in the experiments described above,
an attempt was made to reproduce these results. The methods given for
the electrofocusing procedure (Czempiel et^ al^., 1976; KLose, 1975)
were followed as closely as possible; the only significant exception
was the use of a constant-voltage power source instead of a pulsed-
power source. The electrofocusing was performed on parallel samples
for various periods of time in order to determine the length of time
required to focus the proteins. This question is of considerable
importance in isoelectric focusing experiments. Short periods of time
may be adequate for the carrier ampholytes to reach their equilibrium
positions and establish the pH gradient, but not for the proteins,
which are retarded by the gel due to their higher molecular weights, to
reach their equilibrium positions. Thus false estimates of the iso¬
electric points of the proteins would be obtained. On the other hand
after long times the pH gradient begins to break down, probably due to
the migration of the carrier ampholytes out of the gel (Chrambach et al.,
1973).
The results of this experiment are shown in Figure 25. After 3
hours of electrofocusing the pattern of protein spots looked similar to

Figure 25. isoelectric focusing of mitoribosomal proteins.
Proteins from bovine mitochondrial ribosomes prepared in Buffer A
were electrofocused as described in MATERIALS AND METHODS for 3 hr
(upper photograph), 5 hr (middle photograph), or 7 hr (lower photo¬
graph) . The electrofocusing gels were then extruded and placed at the
top of SDS-gel slabs for electrophoresis in the second dimension. The
isoelectric focusing dimension is horizontal in the photographs, and
the SDS-electrophoresis dimension is from top to bottom. The point of
application of the protein sample is at the upper left-hand corner of
each photograph.

107

108
that shown by Czempiel e_t al_. (1976). It may be seen that the spots
tend to be arranged diagonally, especially on the right-hand side of
the pattern, rather than randomly. This diagonal tendency, which is
seen to a greater extent in the standard two-dimensional system used
in Section II, is an indication that the proteins are being differen¬
tially retarded .1. cording to their molecular weights as they move
through th- first-dimension gel. Electrofocusing/SDS-electrophoresis
patterns obtained after the proteins have reached their equilibrium
positions in the first dimension show ¿in apparently random distribution
of the spots (O'Farrell, 1975).
The pattern observed after 5 hours of electrofocusing confirms
the conclusion that most of the proteins were not focused at the 3-hour
time point. Many of the proteins continued to move until they were
caught at the basic end of the gel. Furthermore several proteins of
relatively high molecular weight, seen in the upper right-hand corner
of this pattern, also moved to the basic end of the gel after 7 hours.
In the 7-hour pattern, those proteins which remained in the body of the
first-dimension gel showed the randomly-dispersed appearance expected
of focused proteins. It may be noted (Figure 26) that the pH gradient
was already formed at 3 hours and remained stable through the 7-hour
time point, although it had deteriorated by 10 hours.
The results thus seem to contradict those of Czempiel e_t al. (1976),
and to suggest that the very low isoelectric points reported by these
authors are in error due to the failure of the proteins to reach their

Figure 26. The pH gradient formed during isoelectric focusing of mitoribosomal proteins.
Measurements of pH were made on gels run in parallel with the first-dimension gels of the
experiment shown in Figure 25. After electrofocusing for the periods of time indicated, the gels
were sliced into 5 mm sections and each section placed in 1 ml of water. After 12 hours of soak¬
ing at room temperature, the pH of each solution was measured with a Fisher pH microelectrode.
L09

LIO
isoelectric positions in the gel. It also appears that considerably
fewer than half of the mitoribosomal proteins have isoelectric points
below 8.5, in contrast to the results of van den Bogert and de Vries
(1976). Although the latter results were obtained by electrophoresis
at pH 8.6 rather than by isoelectric focusing, it would be surprising
if these methodological differences were the source of the discrepancy.
Some preliminary electrophoretic experiments, however, appear to con¬
firm that about half of the large-subunit proteins are anions at
pH 8.6 (Robert Hessler, 1976, private communication). Thus the question
of the actual isoelectric points of the mitoribosomal proteins remains
somewhat uncertain. But even the electrofocusing experiments presented
above show that a relatively large proportion of these proteins are
isoelectric below pH 8.5. All of the evidence therefore jgrees that
these proteins are significantly more acidic than the proteins of E. coli
or mammalian cytoplasmic ribosomes, over half of which have isoelectric
points above 10 (Kaltschmidt, 1971; Huynh-van-Tan et aJL., 1974).
!V. MOLECULAR WEIGHTS OF MITOCHONDRIAL RIBOSOMAL PkuTEINS
The molecular weights of the mitoribosomal proteins were determined
from their mobilities in the second electrophoretic dimension and are
given in Table VIII (large-subunit proteins) and Table IX (small-subunit
proteins). The molecular weights of the large-subunit proteins range
from 11,100 to 45,500 (average: 21,200); the small subunit contains
proteins from 12,300 to 43,000, averaging 22,600. For comparison, the
average molecular weights of _E, coli large-subunit and small-subunit

1 I L
Table VIII. Molecular weights of mitochondrial large-subunit
Protein
MW x 10~3
Protein
i
45.5
27
2
42.5
28
3
38.5
29
4
32
30
5
31.5
31
0
31.5
32
7
31
33
8
29.2
34
9
28.1
35
10
27.6
36
11
26.3
37
12
26.0
38
13
26.0
39
14
25.7
40
15
25.0
41
16
23.6
42
17
23.5
43
18
22.2
44
19
21.6
45
20
20.6
46
21
20.4
47
22
19.9
48
23
19.5
49
24
19.2
50
25
19.0
51
26
18.8
52
proteins.
Mil x 10~3
18.3
18.2
18.2
18.0
17.8
17.6
17.4
17.1
16.8
17.0
16.5
16.2
15.8
15.9
15.3
15.2
15.2
15.4
14.7
14.0
13.6
13.4
13.5
13.3
11.2
11.1

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
112
Molecular weights
MW x 10~3
43
40
38
38
33.5
32
32
29
28.1
26.7
25.6
24.8
24.1
23.7
23.5
23.3
22.0
21.8
21.6
21.1
of mitochondrial small-subunit proteins.
Protein MW x 10
21
21.2
22
21.0
23
20.3
24
19.7
25
18.9
26
18.3
27
18.3
28
17.8
29
17.6
30
17.5
31
17.5
32
16.9
33
16.3
34
15.9
35
15.5
36
15.3
37
14.6
38
14.7
39
14.0
40
13.7
41
12.3

113
proteins are 16,300 and 19,000 (Uzionara et al., 1970). This rather
large difference in the molecular weight distributions of mitochondrial
and bacterial ribosomal proteins can be easily observed by visual com¬
parison of their electrophoretic patterns (Figures 5 and 6 vs. Figure
12). Reported average molecular weights of mammalian cytoplasmic ribo-
somal proteins range from values similar to those of bovine mitoribo-
somal proteins (Martini ana Could, 1971; Creusot et_ c^l. , 1975) to
significantly larger values (Howard et al., 1975; Lin and Wool, 1974).
The data in Tables VIII and IX,with an assumption, permit a calcu¬
lation of the total content of protein in the mitochondrial ribosome.
If one copy of each of these proteins is present per ribosome, the sum
of their molecular weights should equal the total protein content in
c
daltons. This sum is 2.03 x 10°. The sum of the molecular weights of
the E. coli ribosomal proteins, for comparison, is only 0.94 x 10^
(Dzionara et_ ¿lL., 1970). This two-fold difference is due partially
to the larger size of the mitoribosomal proteins, but mainly to their
greater number. In fact, mitochondrial ribosomes appear to contain at
least as much protein as 80S mammalian cytoplasmic ribosomes. As
mentioned above, there is some uncertainty about the molecular weights
of the cytoribosomal proteins; but because of the smaller number of
proteins in these ribosomes, even the larger values reported for their
molecular weigbLS lead to estimates for the total protein content of
1.73 x 10^ to 2.05 x 10^ daltons (Howard £t^ , 1975; Lin and Wool,
1974), about the same as that of bovine mitoribosomes.

114
This relatively large protein content had been expected on the
basis of physical-chemical measurements on the mitochondrial ribosomes.
Since these measurements provided quantitative predictions, it was of
interest to determine how closely the different kinds of estimates for
the total protein content agreed. The protein content can be calcu¬
lated from the buoyant density, p, of the ribosome in CsCl and the
molecular weight of the ribosomal RNA, by the equations
% RNA = 309,8 - 396. 6 (Hamilton and O’Brien, 1974)
P
M
and Protein content = x 100 - M.,.,,. The results of such i cal-
%RNA RNA
dilation, using a value of 1.43 g/ml for the buoyant density of either
subunit and 0.34 x 10^ and 0.35 x in6 for the molecular weights of the
large and small ribosomal RNA'k (Robberson et al., 1971), are given in
the first line of Table X.
A second, independent measurement of the protein content can be
obtained by determining the molecular weight of the ribosome by high¬
speed equilibrium centrifugation, and subtracting M . The second
RNA
line of Table X is based on the values 1.65 x 10^J and 1.10 x 10^ for
the molecular weights of the large and small subunits, respectively
(Hamilton and O’Brien, 1974).
The results obtained from the electrophoretic analysis artr given in
the last line of the table. The sum of the molecular weights of the 52
proteins in the large subunit is 1.10 x 10^, a value which agrees well
with the protein content as calculated from the buoyant density of this
subunit or from its particle weight. However, the total of the 41

115
Table X. Calculation of the protein content of bovine
mitochondrial ribosomes.
Data used in calculation Protein content x 10 ^ (daltons)
Large subunit Small subunit Total
Buoyant density of subunit, 1.10 0.71 1.81
molecular weight of rRNA
Molecular weight of subunit, 1.11 0.75 1.86
molecular weight of rRNA
Molecular weights of 1.10 0.93 2.03
individual proteins

116
small-subunit proteins, 0.93 x 10^ daltons, is somewhat greater than
the protein content determined by other methods.
One interpretation of this discrepancy is that it indicates an
overestimate of the number of proteins in the small subunit by about
20%. An argument in favor of this interpretation can he found in the
results presented in Section IIB3. Several of the proteins in the
large and small subunits were found to be electrophoretically indis¬
tinguishable from each other, possibly because each pair represents
a single protein which partitions between the two subunits when the
ribosome is dissociated. Five of these proteins are present in much
greater quantity in the large subunit than in the small, suggesting
that they are actually large-subunit proteins and thus should not be
included in the small-subunit list. If MSI, 7, 15, 25, and 26 were
disregarded for this reason, the total of the small-subunit protein
molecular weights would he 0.79 x 10^ daltons rather than 0.93 x 1011
daltons, agreeing more closely with the total protein content deter¬
mined by physical-chemical measurements (Table X).
A more convincing interpretation, however, is that the original
assumption (that each ribosome contains one copy of each of the pro¬
teins) is false. If, on the average, the proteins are present in less
than unit quantity in each ribosome, the total protein content would
be overestimated by simply adding up the proteins' molecular weights.
It is clear in the electrophoretic patterns shown above that the pro¬
teins are not all present In equal quantities, particularly in the

L17
small subunit, but there i: no direct evidence whether the average
spot intensity in these patterns represents one copy per ribosome or
less than one. However the results which have been obtained with
other ribosomes indicate that ribosomal proteins are more likely to
be deficient than to be found in greater than unit quantities when
ribosomes are prepared according to the methods used in the present
experiments.
Although _E. coli ribosomes prepared under very gentle mechanical
and ionic conditions appear to contain unit amounts of nearly all of
the ribosomal proteins, the more usual procedures (particularly the
salt-washing steps used to remove non-ribosomal proteins) remove varying
quantities of many of tin.* proteins (Hardy, 1975). As a result, several
ribosomal proteins appeared on two-dimensional gels as spots that
stained only faintly (Kaltschmidt and Wittmann, 1970), quantitative
analyses showed that many of the proteins were present in less than unit
stoichiometry (Voynow and Kurland, 1971; Weber, 1972), and the sum of
the molecular weights of the E. coli small-subunit proteins was found
to exceed the measured total protein content of the subunit by 25%
(Dzionara ejt al., 1970).
It seems likely that this partial removal of ribosomal proteins by
salt treatment is responsible for the finding that the mitochondrial
ribosomes in each preparation appear to be heterogeneous with respect
to their content of the individual ribosomal proteins, and for the
variations observed among separate ribosome preparations. Such effects

118
may also explain Che observation that none of the proteins are found
to be removed by treatment with Buffer E, even though the buoyant
densities of the subunits are increased by this treatment (Section ITB).
The action of Buffer E may be relatively nonspecific, removing some
proteins from all of the ribosomes but not always the same set of pro¬
teins. In this case none of the proteins would be found to be absent
in the electrophoretic analyses, and even reductions in the relative
quantities of some of the proteins mighi be undetectable.
Unfortunately no method has yet been developed for preparing ribo¬
somes free of contaminating proteins but containing stoichiometric
quantities of all the ribosomal proteins. But if this were done, the
results presented above indicate that the measured protein content of
mitochondrial ribosomes would be as given in the bottom line of Table X.
Since functional ribosomes in vivo probably do contain unit copies of
nearly all the proteins (Hardy, 1975), the comparison shown in Table X
corresponds to some extent to the difference between the mitochondrial
ribosome jri vivo and jin vitro. Thus one useful result of the electro¬
phoretic experiments is that they provide information about the nature
and degree of the differences between the ribosomes as isolated and as
they probably exist in vivo. The table, and the electrophoretic
patterns Lhemselves, permit the conclusion that the isolated ribosomes
are lacking considerable quantities of many of the proteins; further, it
appears that most of the difference is localized in the small subunit,
whereas the large subunit is more nearly intact.

V. COMPARISON WITH PROTEINS OF RAT MITOCHONDRIAL RIBOSOMES
The proteins of eukaryotic cytoplasmic ribosomes show a relatively
loro degree of evolutionary divergence. No differences were found among
mammals, birds, or reptiles in the two-dimensionai electrophoretic
pattern of their cytoribosomal proteins hy Delaunay et al. (1973).
Other laboratories (Martini and Gould, 1975; Kuter and Rodgers, 1974)
have been ahie to detect some electrophoretic differences among these
species, but in all cases the number of such differences has been quite
small. Two proteins were found to differ in comparisons of rat, mouse,
and hamster cytoribosomal proteins, and one further protein was altered
in HeLa cell.1. (Kuter and Rodgers, 1 974). However, the mitochondrial
ribosomes of various distantly related organisms show greater differ¬
ences in their physical-chemical properties than do the cytoplasmic
ribosomes of the same organisms (Figure 1). This diversity at the
level of the overall structure of the ribosome suggested that differ¬
ences might be found between more closely related organisms at the
level of the electrophoretic properties of their ribosomal proteins.
To test this possibility the mitoribosomal proteins of another
mammal, the rat, were analysed and compared with the proteins found in
bovine mitoribosomes. The electrophoretic results from one of the two
preparations of rat mitoribosomes which were analyzed as described
above for bovine mitoribosomes are shown in Figures 27 (large subunit)
and 28 (small subunit). The large-subunit pattern was similar in over¬
all appearance to that obtained from bovine mitoribosomes. In

L2Ü
Figure 27. Electrophoretic pattern of proteins from rat mitochondrial
large subunits.

121
Figure 28. Electrophoretic pattern of proteins from rat mitochondrial
small subunits.

122
comparisons of such complex patterns, it is easiest to recognize
correspondences between proteins that have unusual electrophoretic
properties. Thus a very acidic protein (that is, a protein which
migrates more slowly in the first dimension than other proteins of
the same molecular weight) is seen in the rat pattern at a position
near that of the bovine protein ML30, and a faintly-stained protein
corresponding to ML33 is present as well. Spots corresponding to the
very basi* rotein ML13 and the low-molecular-weight proteins ML49
and 50 can also be identified. Some obvious differences between the
two sets of proteins are also evident, such as the very basic rat
protein just to the right of the spot corresponding to ML20, which
was never found in bovine mitoribosomes. A detailed comparison shows
many other differences. The small-subunit pattern showed no obvious
similarity to that of bovine small-subunit proteins, and no individual
rat proteins clearly corresponded to any bovine proteins.
In order to compare the rat and bovine patterns more accurately,
samples of proteins from both sources were mixed together and co¬
elec trophoresed. The bovine sample was a large (stainable) quantity
of protein and the rat sample was a trace quantity (from a third prepa¬
ration of ribosomes) radioaclively labelled by reductive methylation
as described in MATERIALS AND METHODS. The results are shown in
Figures 29 and 30. In these figures only those proteins which were
found in at least two of the three rat mitoribosome preparations are
given; most ot these appeared in all three. Rat proteins which have

12 i
2-D
68-
44-
23.5-
17.2-
14.4-
Figure 29.
proteins.
+ I-D "
>
2
m ♦ J2L SrfBk) *,3)
15 „?)
1161 *U(19)
(22) (
(30)
(W)^28)
(33)
+ m 5*
<*27 &
37
25
^34
«43«Ü
1(45)
0 o
*49)
(50)
«fc
51
Schematic diagram of rat mitochondrial large-subunit

124
2-D
+1 _D _
>.
4-V
68-
44-
23.5-
17.2-
14.4-
•
I2C)
14**
•
»•
©
24-wa
«*• *
© vj34
36® ®38
©<39)
«41
Figure 30. Schematic diagram of rat mitochondrial small-subunit
proteins.

125
electrophoretic mobilities identical to some of the bovine proteins
are marked with the number of the bovine protein, as given in Figure 8
and Figure 19* In addition some of the rat proteins were found in
positions very close to those of some bovine proteins and are desig¬
nated by parenthesized numbers* Some of these near-correspondences
occur among the proteins with unusual electrophoretic properties which
appeared to represent clear similarities between rat and bovine large-
subunit proteins. In these cases (ML13, 30, 33, 49, and 50) it seems
very likely that the rat protein is closely related to that of the cow
even though some difference is detectable.
Bovine mitoribosomal proteins are clearly more similar to the
mitoribosomal proteins of rat than they are to the proteins of bovine
cytoplasmic or bacterial ribosomes. Both kinds of mitoribosomal pro¬
teins have similar molecular weight and charge properties, which are
different from those of cytoplasmic or Eh coli ribosomal proteins
(Section III). It seems likely that the numbers of proteins in both
mitoribosornes are also about the same. Forty-eight proteins are seen
in the rat large subunit as compared with 51 for the bovine large sub¬
unit* The small number of proteins found in rat small subunits (24)
is rather puzzling. However, the quantity of protein used in these
experiments may have been too small to permit the detection of faintly
staining proteins. The similar physical-chemical properties of rat
and bovine mitoribosomal subunits (O’Brien, 1971; O’Brien et al*,
1974; de Vries and Kroon, 1974) indicate that the protein contents of
the two should be nearly the same.

126
At a more detailed level of analysis, the electrophoretic posi¬
tions of the individual proteins show more differences than similar¬
ities between these two mammalian species. A convenient means of
quantitating the degree of similarity between the electrophoretic
patterns of two sets of riboscmal proteins has been suggested by
Delaunay and Schapira (1974), using the formula
degree of similarity, P = ,(a + b) ~ i?._
%(a + b)
where a is the number of proteins in one of the ribosomes, b is the
number in the other ribosome, and n is the total number of electro-
phoretically different proteins found in a mixture of the two samples.
If all of the proteins are electrophoretically distinct from each
other, P = 0. When no differences at all are found, as reported by
Delaunay et al. (1973) for mammals, birds, and reptiles, a value of
P = 1 is obtained. The three protein differences seen between rat and
HeLa cytoribosomal proteins (Kuter and Rodgers, 1974) correspond to
P = 0.95.
When this formula is applied to the proteins of mitochondrial
large subunits, in which only 13 of the 48 rat proteins are electro-
phoreticaliy identical to bovine proteins, it is found that P = 0.26.
A similar calculation for the small subunits, in which there are only
8 electrophoretic identities, gives a value of 0.25. The formula
slightly underestimates the actual degree of similarity between the
small subunits, since even if all 24 of the rat proteins detected were
identical to cow proteins, a value of P = 1 would not be obtained.

127
However if a correction is made for the apparent failure of these
experiments to detect all of the rat small-subunit proteins, the value
of P is still only 0.33.
These results are consistent with the finding of at least seven
mitochondrial large-subunit proteins which differ even between two
species of the frog genus Xenopus (P = 0.83) (Leister and Dawid, 1975).
The evidence thus indicates that the degree of evolutionary divergence
among mitoribosomal proteins is much higher than among the cytoribosomal
proteins of the same organisms. This conclusion extends to more closely-
related species the generalization (see INTRODUCTION) that most of the
physical and chemical properties of ribosomes appear to diverge more
rapidly during evolution for the ribosomes of mitochondria than for
their extramitochondrial counterparts. Some speculations about the
molecular and genetic basis of this higher evolutionary rate will be
discussed below.

DISCUSSION
The results presented above are relevant to questions about the
structure and the evolution of mammalian mitochondrial ribosomes. It
is found that the large complement of protein observed in these ribo¬
somes by physical-chemical measurements can be accounted for by the
number and sizes of the individual ribosomal proteins. The molecular
weights of these proteins are not exceptionally high; rather, it is
the large number of proteins present in these ribosomes which is pri¬
marily responsible for their high protein content. The proteins are
found to be somewhat more acidic th¿in those of bacterial or eukaryotic
cytoplasmic ribosomes, a property which may be related to differences
in the kinds of intermolecular bonding interactions which prevail in
such a protein-rich structure. Finally, the electrophoretic properties
of mammalian mitoribosomal proteins, like many of the other structural
properties of mitochondrial ribosomes in general, show a greater degree
of phylogenetic variation than is found among cytoplasmic ribosomes.
Ninety-three proteins arc found in bovine mitochondrial ribosomes,
as compared with about 70 in mammalian cytoplasmic ribosomes (Wool and
Stoffler, 1974) and 53 in E. coli ribosomes (Wittmann, 1974; Pettersson
et al., 1976). Similar relatively large numbers of ribosomal proteins
have been reported for the mitochondrial ribosomes of other vertebrate
species. Leister and Dawid (1974) found 84 mitoribosomal proteins in
128

129
the frog Xenopus laevis. Recently the proteins of rat mitochondrial
ribosomes were analyzed, and numbers ranging from 70 (van den Bogert
and de Vries, 1976) to 107 (Czempiel et al., 1976) were reported.
The large number of proteins found in bovine mitochondrial ribo¬
somes does not seem to be attributable to contamination of the ribosomes
by non-ribosomal proteins, on the basis of a variety of evidence.
These proteins are reproduciblv found in separate preparations of ribo¬
somes, and they are not removed by treatments that have been found to
remove contaminating proteins from other kinds of ribosomes. Indeed,
only a few of them are removed by even more stringent treatments that
result in complete inactivation of the peptidyl transferase function of
this ribosome. Furthermore, when heterologous ribosomes are added to
the mitochondria and then re-purified by the same method used for the
preparation of mitochondrial ribosomes, they are not contaminated by
adsorbed proteins.
There is a possibility that the present results overestimate the
number of bovine mitoribosomal proteins to some extent, due to a par¬
titioning of some proteins between the two subunits. Although the
total number of proteins found in analyses of the large and small sub¬
units separately is 93, only 81 of these can be observed in mixtures
of the two sets of proteins. That is, 12 of the large-subunit proteins
are electrophoretically indistinguishable from proteins of the small
subunit. It is therefore possible that the two members of all or some
of these overlapping protein pairs are in fact identical , and that

130
each p¿iir should be counted only once in the enumeration of the bovine
mitoribosomal proteins. The evidence in favor of this possibility is
only suggestive, and further characterization of the individual proteins
involved will be required to resolve this question.
The proteins of bovine mitoribosomes are found to have molecular
weights avi raging 21,200 in the large subunit and 22,600 in the small
subunit, values similar to those of mammalian cytoribosomal proteins.
Much higher molecular weights were reported for the mitoribosomal
proteins of Xenopus, with averages of 27,000 and 33,300 for the large
and small subunits respectively (Leister and Dawid, 1974). Molecular
weights were not determined in the published studies of rat mitoribo¬
somal proteins, but the present experiments indicate that these are
about the same size as bovine mitoribosomal proteins. The sum of the
molecular weights of the bovine mitoribosomal proteins is somewhat
greater than the total quantity of protein calculated to be present in
these ribosomes on the basis of their buoyant density, particularly in
the case of the small subunit. A similar result was obtained with
Xenopus mitoribosomes (Leister and Dawid, 1974). The simplest inter¬
pretation of this phenomenon is that a proportion of the ribosomes in
a given preparation are lacking some of the proteins, so that on the
average there is less than one copy of each of these proteins per
ribosome. This interpretation is consistent with results that have
been obtained for E. coli ribosomes.

131
The finding that bovine mitoribosoiiuJ proteins are relatively
acidic, compared to bacterial or mammalian cytoplasmic ribosomal pro¬
teins, is in qualitative agreement with the reported results of van
den Bogert and de Vries (1976) and Czempiel e_t a^L. (1976) on rat mito-
ribosomal proteins. The proteins of Xenopus mitochondrial ribosomes
also appear to be more acidic than those of the cytoplasmic ribosomes,
at least in comparisons between the large subunits (Leister and Dawid,
1974). Quantitatively, there are some differences between the results
of the present experiments and those which have been published for rat
mitoribosomal proteins. The cause of these discrepancies is unclear.
It does not seem to be related to differences between the rat and
bovine proteins: the direct comparisons of these two sets of proteins
by electrophoresis show that they have similar charge properties at
pH 4.3.
A large number of electroph >retic differences are found between
bovine and rat mitoribosomal proteins. This observation is consistent
with comparisons which have been made between species related both more
closely and more distantly than rat and cow. Xenopus laevis and Xenopus
mulleri differed from each other in at least seven of the proteins of
the large subunit alone (Leister and Dawid, 1975). Rat mitoribosomal
proteins showed no similarity with those of the fungus Neurospora (van
den Bogert and de Vries, 1976). No obvious similarities can be seen
when the patterns of bovine or rat mitoribosomes are compared with those
published for Xenopus laevis (Leister and Dawid, 1974), obtained in an

132
essentially identical electrophoretic system. Indeed even the molec¬
ular weights of the frog proteins are quite different from those of
the mammalian proteins.
The significance of these results lies in their contribution to
studies on the structure of ribosomes and on the evolution of mito¬
chondrial genes. First, it has been suggested that the low buoyant
density oí anima] mitochondrial ribosomes might be due to the presence
of adherent membrane fragments (Borst and Grivell, 1971). The present
results d>. iot support the hypothesis that a third component of lower
density than protein (i.e., lipid) makes any significant contribution
to the low buoyant density of these ribosomes: the quantity of protein
alone required to account for the observed density does not exceed the
quantity which actually appears to be present by electrophoretic anal¬
ysis .
The remarkably large number (and total quantity) of proteins in
these ribosomes raises some interesting questions about the similar¬
ities and differences that must exist between these particles and other
structural kinds of ribosomes with respect to their biosynthesis,
assembly, and detailed functional activities. For example, it seems
possible that these ribosomes, in which the ratio of RNA to protein is
only 1:2, are held together predominantly by ditferent kinds of inter-
molecular bonding interactions than are found in E. coli ribosomes,
which have an RNA:protein ratio of about 1:0.6. Clearly the structure
of the mitoribosome must involve more prot uin-protein interactions and

133
fewer protein-RNA interactions than that of the _E. coli ribosome.
Proteins which bind directly to polynucleotides, such as histones and
many ribosomal proteins, frequently have exceedingly high isoelectric
points; presumably a large part of the energy of binding in these cases
is due to electrostatic attraction (and lack of electrostatic repulsion)
between charged residues of the proteins and the phosphate backbone
of the nucleic acid. Ionii bonding between two proteins, on the other
hand, requires the Intel iction of both acidic and basic amino acids in
the proteins. Thus the high protein content of animal mitochondrial
ribosomes may be related to the lower basicity which Is observed In
their proteins. It may be predicted that other very protein-rich
ribosomes, like the mitochondrial ribosomes of Tetrahymena (Chi and
Suyama, 1970) and Candida utilis (Vignais et_ a]L. , 1972), will likewise
contain relatively acidic ribosomal proteins.
Since hydrophobic interactions are probably more important in
stabilizing protein-protein associations than protein-RNA binding, it
seems reasonable that this kind of interaction plays a larger role in
the fundamental structural organization of very protein-rich ribosomes.
One characteristic of hydrophobic interactions is that they are resist¬
ant to disruption by high salt concentrations, whereas electrostatic
interactions are highly susceptible. Thus it Is of interest that most
of the proteins of bovine mitochondrial ribosomes are found to be much
more resistant to salt-stripping than those of the cytoplasmic ribosomes
(O'Brien et al., 1976). These mitoribosomes thus share some properties

í Í4
(though to a lesser extent) with the ribosomes of the extreme halophile
Halobacterium cutirubrum: the latter are structurally and functionally
stable at the normal intracellular salt concentration of 4 molar, and
their proteins have isoelectric points much lower than even those of
animal mitoribosomes (Bayley, 1966).
The large number of electrophoretic differences observed between
the mitoribosomal proteins of two mammalian species complements the
previous reports of even larger differences in the structural properties
of mitoribosomes from more distantly related organisms. Taken together,
these observations clearly indicate a relatively high rate of evolution¬
ary divergence among these ribosomes, and invite a consideration of the
possible cellular, genetic, or molecular bases of this high evolutionary
rate.
Perhaps the simplest explanation for this phenomenon is the possi¬
bility that mitochondrial ribosomes are not subject to very strong
selective pressures in nature. ¿Relative to the cytoplasmic ribosomes,
mitoribosomes make only a very few different proteins. Although these
proteins are major (and essential) components of the mitochondrion,
their total quantity is also much less than the total quantity of
cellular proteins synthesized on cytoplasmic ribosomes. Only a small
number of ribosomes in each mitochondrion are required to keep up with
the work load of protein synthesis for mitochondrial growth and turn¬
over. Thus ±l would seem that a diminished efficiency due to the
accumulation of slightly disadvantageous mutations might be more

135
tolerable in the mitochondrial ribosomes; the lower efficiency could
be compensated by increasing the number of ribosomes, at only a modest
expense in terms of the metabolic energy required to synthesize the
extra ribosomes.
In this regard it is necessary to consider also the ribosomes of
chloroplasts, since the role of these ribosomes in cellular protein
synthesis is similar to that of mitochondrial ribosomes. It is note¬
worthy that 'he structural properties of these ribosomes indicate
little evolutionary divergence among distantly related chloroplast-
containing protists and plants, or even between these organisms and
prokaryotes. This observation argues against the significance of a low
level of selective pressure as a contributing factor in the high evo¬
lutionary rate of mitochondrial ribosomes. However the argument is
weakened by the fact that chloroplast ribosomes are responsible for
a somewhat larger proportion of the total protein synthesis in the cell
than mitochondrial ribosomes are: one subunit of the very abundant pro¬
tein ribulose-diphosphate carboxylase is made on chloroplast ribosomes.
Moreover, the number of chloroplast ribosomes per cell is greater than
that of mitoribosomes.
It is also possible that genetic mechanisms exist to effect the
more rapid establishment of selerLively neutral or advantageous mutations
in mitochondrial ribosomes. No .! i Ferences can be visualized in the
genetic processes that operate directly on mitoribosomal versus cyto-
ribosomal proteins, since both groups of proteins appear to be coded in

136
nuclear genes. But a rapid evolutionary rate in mitochondrial rRNA
might force or encourage a complementary high rate in the proteins
which must interact with this rRNA to form functional ribosomes. The
mitochondrial rRNA is copied from mitochondrial genes, and several
fundamental differences in the genetics of nuclear and mitochondrial
DNA are known. Furthermore there is considerable experimental evidence
demonstratin'1; rapid evolution;.1 - divergence not only of mitochondrial
rRNA genes, but of the rest of the mitochondrial DNA as well.
Most oi this evidence is based on DNA-DNA or DNA-RNA hybridization
studies, usually including thermal denaturatxon analyses to evaluate
the level of correspondence between the heterologous molecules hybrid¬
ized. Such experiments measure differences in the primary sequence of
the nucleic acids, and should be roughly comparable to electrophoretic
comparisons of proteins with respect to their sensitivity in detecting
interspecies variations. Unfortunately a direct comparison of the
degree of variation found in mitochondrial DNA with that of nuclear DNA
has only been reported in one instance. C.root et al. (1975) compared
the nuclear and mitochondrial DNAs of the yeasts Saccharomyces
carlsbergensis, Kluyveromyces lactis, and Candida utilis, finding that
both mitochondrial rRNA and total mitochondrial DNA showed greater
differences among these species than did the cytoplasmic rRNA and
nuclear DNA.
A number of indirect comparisons can also be made. Sinclair and
Brown (1971) found detectable homologies between the cytoplasmic rRNA

137
of Xenopus laevis and that of many distantly related organisms, includ¬
ing yeast; but no correspondence in base sequence was detectable between
the mitochondrial DNAs of X. laevis and yeast (Dawid and Wolstenholme,
1968). In comparisons of rat, mouse, ¡minea pig, monkey and chicken,
Jakovcic et_ uK (1975a and b) found a lower degree of homology among
mitochondria] tRMA genes than had been reported for nuclear genes coding
for rRNA or proteins; most of the rest of the mitochondrial DNA showed
even less homology.
No significant differences were detected between the nuclear rRNA
genes of X. laevis and X. mulleri, although the spacer DMA interspersed
among these rRNA genes showed extensive non-homology' (Brown et al.,
1972). Mitochondrial rRN.'. genes, on the other hand, were distinguish¬
able in hybridization comparisons between these two species, and most
of the remaining mitochondrial DMA showed a significantly higher degree
of evolutionary divergence (Dawid, 1972). indeed, the high level of
non-homology found between these two closely-related species in the
bulk of the mitochondrial DMA sequences led Dawid (1972) to propose
that these sequences represent nothing but spacer DNA, analogous to the
spacer regions found in nuclear ribosomal DNA. But the weight of sev¬
eral other kinds of evidence contradicts this proposal, indicating that
most of this mitochondrial DNA represents the structural genes for
enzyme proteins (Schatz and Mason, 1974). Thus it must be concluded
that all kinds of mitochondrial genes can show considerable divergence
between closely related species. All these results are in contrast to

138
those obtained in chloroplast studies, in which hybridization experi¬
ments revealed little difference in the base sequences of chloroplast
rRNA from various raonocotyledonous and dicotyledonous plants (Thomas
and Tewari, 1974).
What kinds of genetic mechanisms are known that might account for
the high evolutionary rate of mitochondrial DNA? Among the known
differences in the genetic processes acting on the mitochondria and
nuclei of eukaryotic cells is the phenomenon of vegetative segregation
of mitochondrial genes. Yeast zygotes "heteroplasmic" for a mitochon¬
drial gene (that is, containing both mutant and wild-type mitochondrial
DNA molecules) give rise to some homoplasmic progeny within one or two
mitotic cycles, and all descendants are homoplasmic within 10-20 cell-
divisions (Birky, 1976). Groot et al. (1975) have pointed out that
this process would lead to the rapid phenotypic expression of mitochon¬
drial mutations, thereby promoting the establishment of these mutations
if they were positively selective.
The latter authors have also postulated a mechanism for a relatively
rapid establishment of selectively neutral mutations in mitochondrial
genes. The presence of many copies of a gene per cell increases the
probability of mutation in the gene. Multiple copies of genes for both
mitochondrial and cytoplasmic rRNA are found in eukaryotic cells, but
a master-slave correction process acting on the cytoplasmic rRNA genes
has been proposed (Brown j2t al., 1972). If this process does act on
the nuclear but not on the mitochondrial genes, it is argued, the latter

139
would be expected to show a higher mutation rate. Some objections may
be raised to this line of reasoning, however. First, the spacer regions
present in the nuclear rDNA are considered to be subject to the same
correction process as the rRNA genes, because both are part of the same
repeating DNA sequence and both are essentially identical in all of the
copies present in a given cell; but the spacer regions show very rapid
evolutionary divergence, as noted above (Brown et al., 1972). Second,
all of the mitochondrial DNA molecules in a given organism are found to
be identical to each other (DawiJ, 1972; Potter £t_ cfL., 1975), suggesting
that some kind of correction mechanism may operate on the mitochondrial
genes too (Birky, 1976).
Another unique feature of mitochondrial genetics is the high rate
of recombination observed between mitochondrial DNA molecules. In yeast
the rate is high enough that no linkage can be observed among four gene
loci, all of which are physically located on the same 5 x 10^ d circular
DNA molecule (Dujon e_t_ al^. , 1974). These results were interpreted to
mean that the mitochondrial genomes undergo many repeated rounds of
pairing and recombination. These events do not even appear to be
limited in time to a short interval in the life cycle of the organism,
such as meiosis. Such high rates of recombination should promote the
spread of mutations through the population, and permit the combination
of separate mutations into a single genome.
Mitochondrial recombination can also occur in mammalian cells:
Coon £t_ al. (1973) have presented biochemical evidence for the forma-

140
tion of recombinant mitochondrial DNA molecules in somatic cell hybrids
of mouse and human origin. But the importance of this phenomenon in
the exchange of mitochondrial genetic information between individuals
in nature is questionable for mammals and most other organisms. In
most species (unlike the case in yeast), mitochondrial genes are inherited
uniparentally, affording little opportunity for recombination between
the BNAs of the two parents. However some recombination can be de¬
tected between the parental chloroplast genomes in Chlamydomonas crosses,
despite the facl: that chloroplast genes in this organism are also uni¬
parentally inherited (Sager, 1972).
The high recombination rate of mitochondrial DNA may be more sig¬
nificant as a possible inducer of mutations. The model of Dujon _et al.
(1974) includes a suggestion that mitochondrial recombination could
account for the properties of the petite mutants which arise spontane¬
ously at a rather high frequency in yeast. Faulty alignment of the DNA
molecules during recombination might also lead to less dramatic mutations
on occasion. In addition, mitochondrial DNA may be more mutable than
nuclear DNA because of differences in repair processes. Yeast mito¬
chondria are reported to lack an excision-repair system for thymidine
dimers induced by ultraviolet radiation (Prakash, 1975). Mammalian
mitochondria lack both this system and the photoreactivation system
(Clayton et^ al., 1974).
There is certainly no shortage of possible mechanisms to account
for the rapid evolutionary divergence of mitochondrial DNA and mito-

141
chondrial ribosomes. The question is which of these possibilities
actually represent significant factors contributing to this divergence.
Answers to this question can only come from further studies on the
cellular and molecular processes acting on cytoplasmic genes. Par¬
ticularly instructive would be a comparison of such processes between
mitochondria and chloroplasts, since the latter show a much lower rate
of evolutionary divergence.
Whatever the causes, the phylogenetic diversity of mitochondrial
ribosomes should prove to be a useful tool for the detailed analysis
of the biosynthesis, assembly, and mechanism of action of ribosomes in
general. Important informal ion about the relationships between struc¬
ture and function in biological systems can frequently be derived from
comparisons between naturally occurring systems which have the same
function but differ structurally. Mitochondrial ribosomes offer a
great variety of structural differences with considerable similarity-
in the details of their functional activity, and thus should be most
valuable for such comparative studies.

BIBLIOGRAPHY
Aaij, C. N., N. Nanninga and P. Borst, 1972. The structure of ribosome¬
like particles from rat liver mitochondria. Biochim. Biophys. Acta
277:140-148.
Agsteribbe, £., R. Datema and A. Kroon, 1974. Mitochondrial polysomes
from Neurospora crassa. In The Biogenesis of Mitochondria, edited by
A. Kroon and C. Saccone, Academic Press, New York, pp. 305-314.
Avadhani, N. G. and D. E. Buetow, 1972. Isolation of active polyribo¬
somes frum the cytoplasm, mitochondria, and chloroplasts of Euglena
gracilis. Biochem. J. 128:353-365.
Avadhani, N. G. and D. E. Buetow, 1974. Mitochondrial and cytoplasmic
ribosomes: Distinguishing characteristics and a requirement for the
homologus ribosomal salt-extractable fraction for protein synthesis.
Biochem. J. 140:73-78.
Bartoov, B., R. S. Mitra and K. B. Freeman, 1970. Ribosomal-type
ribonucleic acid from rodent mitochondria. Biochem. J. 120:455-466.
Bayley, S. T., 1966. Composition of ribosomes of an extremely halophilic
bacterium. J. Mol. Biol. 1j5:420-427.
Bickle, T. A. and R. R. Traut, 1971. Differences in size and number of
80S and 70S ribosomal proteins by dodecyl sulfate gel electrophoresis.
J. Biol. Chem. 246.:6828-6834.
Birky, C. W., 2976. The inheritance of genes in mitochondria and
chloroplasts. Bioscience 26:26-33.
Bodley, J. W., F. J. Zieve, J. Lin and T. S. Zievel, 1970. Studies on
translocation. III. Conditions necessary for the formation and
detection of a stable rihosome-G factor-guanosine diphosphate complex
in the presence of fusidic acid. J. Biol. Chem. 245:5656-5661.
Bonner, J. D. and J. E. Varner, 1965. Plant Biochemistry, Academic
Press, New York.
142

143
Bonner, W. M. and R. A. Laskey, 1974. A film detection method for
tritium-labelled proteins and nucleic acids in polyacrylamide gels.
Euro. J. Biochem. 46:83-88.
Borst, P. and L. Grivell , 1971. Mitochondrial ribosomes. FEBS Letters
13:73-88.
Bourque, D. P., J. E. Boynton and N. W. Gillham, 1971. Studies on the
structure and cellular location of various ribosomes and ribosomal
RNA species in the green alga Chlamydomonas reinhardi. J. Cell Sci.
8^:153-183.
Brouwer, J. and R. J. Planta, 1975. The origin of high molecular weight
proteins in ribosomal preparations of Bacillus licheniformis. FEBS
Letters i:73-75.
Brown, D. D., P. C. Wensink and E. Jordan, 1972. A comparison of the
ribosomal DNA's of Xenopus laevis and Xenopus mulleri: Lhe evolution
of tandem genes. J. Mol. Biol. 63:57-73.
Brunk, C. F. and V. Leick, 1969. Rapid equilibrium isopycnic CsCl
gradients. Biochim. Biophys. Acta 179:136-144.
Cammarano, P., A. Eelsani, A. Romeo and F. M. Alberghina, 1973. Parti¬
cle weights of active ribosomal subunits from Neurospora crassa.
Biochim. Biophys. Acta 308:404-411.
Chi, J. C. H. and Y. Suyama, 1970. Comparative studies on mitochondrial
and cytoplasmic ribosomes of Tetrahymena pyriforáis. J. Mol. Biol.
53^:531-556.
Chrambach, A., P. Doerr, G. R. Finlayson, L. E. M. Miles, R. Sherins
and D. Roland, 1973. Instability of pH gradients formed by isoelectric
focusing in polyacrylamide gel. Ann. N.Y. Acad. Sci. 209:44-64.
Clayton, D. A., J. N. Doda and E. C. Friedberg, 1974. The absence of a
pyrimidine dimer repair mechanism in mammalian mitochondria. Proc.
Natl. Acad. Sci. USA 71:2777-2781.
Coon, H. G., I. Horak and I. B. Dawid, 1973. Propagation of both
parental mitochondrial DNAs in rat-human and mouse-human hybrid cells.
J. Mol. Biol. 8^:285-298.
Cooper, C. and C. Avers, 1974. Evidence of involvement of mitochondrial
polysomes and messenger RNA in synthesis of organelle proteins. In
The Biogenesis of Mitochondria, edited by A. Kroon and C. Saccone,
Academic Press, New York, pp. 289-303.

144
Coote, J. L., T. H. Rabbitts and T. S. Work, 1971. The mitochondrial
ribosomes of baby hamster kidney cells. Biochem. J. 123:279-281.
Creusot, F., J. Delaunay and G, Schapira, 1975. Eukaryotic ribosomal
proteins. Molecular weights of proteins from rabbit liver subunits.
Biochimie 57:167-173.
Czempiel, W., J. Klose and R. Bass, 1976. Mammalian mitochondrial
ribosomes: Characterization of ribosomal proteins by two-dimensional
gel electrophoresis. FEBS Letters 62:259-262.
Darnell, J. E., 1968. Ribonucleic acid from animal cells. Bacterio!.
Rev. 12: '.62-290.
Dawid, I. 8., 1972. Evolution of mitochondrial DNA sequences in Xenopus
Dev. Biol. 29:139-151.
Dawid, I. B., D. D. Brown and R. H. Reeder , 1970. Composition and
structure of chromosomal and amplified ribosomal DNA's of Xenopus
laevis. J. Mol. Biol. 51:341-360.
Dawid, I. B. and J. W. Chase, 1972. Mitochondrial RNA in Xenopus laevis
IT. Molecular weights and other physical properties of mitochondri. 1
ribosomal and 4S RNA. J. Mol. Biol. 63:217-231.
Dawid, I. B, and D. R. Wolstenholme,1968. Renaturation and hybridiza¬
tion studies of mitochondrial DNA. Riophys. J. 8:65-81.
Delaunay, J., F. Creusot and G. Schapira, 1973. Evolution of ribosomal
proteins. Euro. J. Biochem. 39:305-312.
Delaunay, J., F. Creusot and G. Schapira, 1974. Characterization of the
proteins extracted with KC1 from eukaryotic ribosomes. Biochimie
_56:189-191.
Delaunay, J. and G. Schapira, 1974. Pliylogenic distance between pro¬
karyotes and eukaryotes as evaluated by ribosomal proteins. FEBS
Letters 40:97-100.
Denslow, N. D. and T. W. O'Brien, 1974. Susceptibility of 55S mito¬
chondrial ribosomes to antibiotics inhibitory to prokaryotic ribosomes
lincomycin, chloramphenicol, PA114A. Biochem. Biophys. Res. Comm.
57:9-16.
de Vries, H., E. Agsteribbe and A. M. Kroon, 1971. The "Fragment
Reaction": A tool for the discrimination between cytoplasmic and
mitochondrial ribosomes. Biochim. Biophys. Acta 246:111-122.

145
de Vries, H. and A. Kroon, 1974. Physiochemical and functional char¬
acterization of the 55S ribosomes from rat liver mitochondria. In
The Biogenesis of Mitochondria, edited by A. Kroon, and C. Saccone,
Academic Press, New York, pp. 357-365.
de Vries, H. and R. van der Koogh-Schuuring, 1973. Physiochemical
characteristics of isolated 55S mitochondrial ribosomes from rat
liver. Biochem. Biophys. Res. Comm. 54:308-314.
Dubin, D. T., 1974. Methylated nucleotide content of mitochondrial
ribosomal RNA from hamster cells. J. Mol. Biol. 84^257-273.
Dujon, B., P. P. SJonimski and L, Weill, 1974. MiLochondrial genetics
IX: A model for recombination and segregation of mitochondrial
genomes in Saccharomyces cerevisiae. Genetics 78:415-437.
Dzionara, E« Kaltschmidt and H. G. Wittmann, 1970. Ribosomal
proteins, XIII. Molecular weights of isolated ribosomal proteins of
Escherichia coli. Proc. Natl. Acad. Sci. USA 67:1909-1913.
Edelman, M., I. M, Verma and U. Z. Littauer, 1970. Mitochondrial
ribosomal RNA from Aspergillus nidulans: C’haracterization of a
novel molecular species. J. Mol. Biol. 49:67-83.
Fazekas de St. Groth, S., R. G. Webster and A. Datyner, 1963. Two new
staining procedures for quantitative estimation of proteins on electro¬
phoretic strips. Biochem. Biophys. Acta 71:377-391.
Geisser, M., G. W. Tischendorf, C. Stuffier and H. G. Wittmann, 1973.
Immunological and electrophoretic comparison of ribosomal proteins
from eight species belonging to Enterobacteriaceae. Mol. Gen. Genet.
127:111-128.
Gerding, J. J. T., A. Koppers, P. Hagel and H. Bloemendal, 1971.
Cyanate formation in solutions of urea. II. Effect on the eye lens
protein a-crystallin. Biochim. Biophys. Acta 243:374-379.
Gesteland, R. F. and T. Stahelin, 1967. Electrophoretic analysis of
proteins from normal and cesium chloride-treated Escherichia coli
ribosomes. J. Mol. Biol. 24:149-155.
Gressner, A. M. and I. G. Wool, 1974. The phosphorylation of liver
ribosomal proteins in vitro. J. Biol. Chem. 249:6917-6925.
Grivell, L. A., P. Netter, P. Borst and P. P. Slonimski, 1973. Mito¬
chondrial antibiotic resistance in yeast ribosomal mutant.- resistant
to chloramphenicol, erythromycin, and spiramycin. Biochim. Biophys.
Acta 312:358-367.

146
Grivell, L. A., L. Reijnders, and P. Borst, 1971. Isolation of yeast
mitochondrial ribosomes highl' active in protein synthesis. Biochiin.
Biophys. Acta 247 : 91-103.
Groot, G. S. P., R. A. Flavell and J. P. M. Sanders, 1975. Sequence
homology of nuclear and mitochondrial DMAs of different yeasts.
Biochim. Biophys. Acta 378:186-194.
Hamel, E.( M. Koka and T. Nakomoto, 1972. Requirement of an Escherichia
col i 50S ribosomal protein component for effective interaction of the
ribosome with T and G factors and with guanosine triphosphate. J.
Biol. Chem. 247:805-814.
Hamilton, M. G., 1971. Isodensity equilibrium centrifugation of ribo-
S‘..:ial p.irticles; tiie calculation of the protein content of ribosomes
and other ribonucleoproteins from buoyant density measurements.
Methods. Knzymol. 70:Part G 512-521.
Hamilton, M. G. and T. W. O’Brien, 1974. Ultracentrifugal, character¬
ization of the mitochondrial ribosome and subribosomal particles of
bovine liver: molecular size and composition. Biochemistry 13:5400-
5403.
Hamilton, M. C. and M. L. Petermann, 1959. Ultracentrifugal studies on
ribonucleuprotein from r.it liver microsomes. J. Biol. Chem. 234:
1441-1448. ~~~
Hardy, S. J. S., 1975. Tiie stoichiometry of the ribosomal proteins of
Escherichia coll. Mol. Gen. Genet. 140:253-274.
Hardy, S. J. S. and C. G. Kurland, 1966. The relationship between
Poly A polymerase and the r ibosumes. Biochemistry 5:3676-3684.
Hartley, M. R. and R. J. Ellis, 1973. Ribonucleic acid synthesis in
ehloroplasts. Biochem. J. 1234 :249-262.
Hill, W. E., G. P. Rossetti and K. E. van Holde, 1969. Physical studies
of ribosomes from Escherichia coli. J. Mol. Biol. 44:263-277.
Hosokawa, K., R. K. Fujimura and M. Nomura, 1966. Reconstitution of
functionally active ribosomes from inactive subparticles and proteins.
Proc. Natl. Acad. Sci. USA 55:198-204.
Howard, G. A., J. A. Traugh, E. A. Croser and R. R. Traut, 1975, Ribo¬
somal proteins from rabbit reticulocytes: Number and molecular weights
of proteins from ribosomal subunits. .1. Mol. Biol. 93:391-404.
Huynh-van-Tan, M. Gavrilovic and G. Schapira, 1974. Eukaryotic ribo¬
somal proteins: The number of proteins in tiie subunits and their
isoelectric points. FEBS Letters 45:299-303.

147
Jakovcic, S., J. Casey and M. Rabinowitz, 1975a. Sequence homology of
the mitochondrial leucyl-tRNA cistron in different organisms, bio¬
chemistry 14q 20 57-2042.
Jakovcic, K., J. Casey and M. Rabinowitz, 1975b. Sequence homology
between mitochondrial DNA of different eukaryotes. Biochemistry
14:243-250.
Johnson, J. D. and J. Horowitz, 1971. Characterization of ribosomes
and RNAs from Mycoplasma hominis. Biochira. Biophys. Acta 247:262-279.
Kaltschmidt, E., 1971. Ribosomal Proteins. XIV. Isoelectric points
of ribosomal proteins of E. coli as determined by two-dimensional
polyacrylamide gel electrophoresis. Anal. Biochem. 43:25-31.
Kaltschmidt, E. and H. G. liittman, 1970. Ribosomal proteins. XII.
Number ot proteins in small and large ribosomal subunits of E. coli
as determined by two-dimensional gel electrophoresis. Proc. Natl.
Acad. Sci. USA 67:1276-1280.
Kirby, K. S., 1965. Isolation and characterization of ribosomal ribo¬
nucleic acid. Biochenn J. 96:266-269.
Kleinow, W., 1974. RNA from mitochondrial ribosomes of Locusta migra-
toria. In The Biogenesis of Mitochondria, edited by A. M. Kroon and
C. Saccone, Academic Press, New York, pp. 337-346.
Kleinow, W., W. Neupert and T. Bucher, 1971. Small sized ribosomes
from mitochondria of Locusta migratoria. FEBS Letters 12:129-133.
Kleinow, W., W. Neupert and F. Miller, 1974. Election microscope study
of mitochondrial 60S and cytoplasmic 80S ribosomes from locusta
migratoria. J. Cell Biol. 62:860-875.
Klose, J., 1975. Protein mapping by combined isoelectric focusing and
electrophoresis of mouse tissues. Humangenetik 26:231-243.
Krawiec, S. and J. M. Eisenstadt, 1970. Ribonucleic acids from the
mitochondria of bleached Euglena gracilis. Biochim. Biophys. Acta
217:132-141.
Küntzel, II. and H. Noll, 1967. Mitochondrial and cytoplasmic poly¬
somes from Neurospora crassa. Nature (london) 215:1340-1345.
Kurland, C., 1960. Molecular characterization of ribonucleic acid
from Escherichia coli ribosomes. I. Isolation and molecular weights.
J. Mol. Biol. 2:83-91.
Kurland, C. G., 1966. The requirements for specific sRNA binding by
ribosomes. J. Mol. Biol. 18:90-108.

148
Kuter, D. J. and A. Rodgers, 1974* Ribosomal protein differences
between animal cells. Exptl. Cell Res. 87:186-194.
3
Laskey, R. and A. Mills, 1975. Quantitative film detection of H
and -^C in polyacrylamide gels by fluorography. Eur. J. Biochem.
56:335-341.
Leaver, C. J. and M. A. Harmey, 1973. Ribosomal RNAs of higher plant
mitochondria. In Ribosomes and RNA Metabolism, edited by J. Zelinka
and J. Balan, Publishing House of the Slovak Academy of Sciences,
pp. 407-417.
Leister, T>. E. and I. B. David, 1974. Physical properties and protein
constituents of cytoplasmic and mitochondrial ribosomes of Xenopus
laevis. J. Biol. Chem. 249:5108-5118.
Leister, D. E. and T. B. David, 1975. Mitochondrial ribosomal proteins
in Xenopus laevis/X. mulleri interspecific: hybrids. J. Mol. Biol.
96:119-123.
Lin, A. and I. G. Wool, 1974. The molecular weights of rat liver ribo¬
somal proteins determined by "three-dimensional" polyacrylamide gel
electrophoresis. Mol. Gen. Genet. 134:1-6.
Loening, U. E., 1968. Molecular weights of ribosomal RNA in relation
to evolution. J. Mol. Biol. 38:355-365.
Locuing, U. E,, K. W. Jones and M. L. Birnstiel, 1969. Properties of
the ribosomal RNA precursor in Xenopus laevis; comparison to the.
precursor in mammals and in plants. J. Mol. Biol. 45:353-366.
Loewenstein, J., H. R. Scholte and E. M. Wit-Peeters, 1970. A rapid
and simple procedure to deplete rat-liver mitochondria of lysosomal
activity. Bíochim. Biophys. Acta 223:432-436.
Lowry, 0. H., N. J. Rosebrough, A. L. Farr and R. J. Randall, 1951.
Protein measurement with folin phenol reagent. J. Biol. Chem.
193:265-275.
Lyttleton, J. W., 1962. Isolation of ribosomes from spinach chloro-
plasts. Exp. Cell Res. 26:312-317.
McConkey, E. H., 1974. Composition of mammalian ribosomal subunits:
A re-evaluation. Proc. Natl. Acad. Sci. USA 71:1379-1383.

149
Maizel, J. W., 1971. Polyacrylamide gel electrophoresis of viral
proteins. In Methods in Virology, Vol 5. Chap. 5., edited by
K. Maramorosch and li. Kaprowski, Academic Press, New York.
Malkin, L. J., 1971. Minor species of ribonucleic acid associated
with rat liver mitochondria. Biochemistry 10:4752-4756.
Mans, R. J. and G. D. Novelli, 1960. A con.anient, rapid and sensi¬
tive method for measuring the incorporation of radioactive amino
acids into protein. Biochem. Bicphys. Res. Comm. 3^:540-543.
Martini, 0. H. W. and H. J. Gould, 1971. Enumeration of rabbit
reticulocyte ribosomal proteins. J. Mol. Biol. 62:403-405.
Martini, O. H. W. and H. J. Gould, 1975. Molecular weight distribution
of ribosomal proteins from several vertebrate species. Mol. Gen.
Genet. 142:317-331.
Morimoto, H. and H. 0. Ilalvorson, 1971. Characterization of mito¬
chondrial ribosomes from yeast. Proc. Natl. Acad. Sci. USA
68:324-328.
Morimoto, H., A. H. Scragg, J. Nekhoroceff, V. Villa and H. 0. Halvorson,
1971. Comparison of the protein synthesizing systems from mitochondria
and cytoplasm of yeast. In Autonomy and Biogenesis of Mltochondria
and Chloroplasts, edited by N. K, Boardman, A. W. Linnane, and R. M.
Smillie, American Elsevier, New York, pp. 282-292.
Morr i on, C. A., R. A. Garrett, H. Zeichhardt and G. Stdffler, 1973.
Proteins occurring at, or near, the subunit interface of JE. coli ribo¬
somes. Molec. Gen. Genet. 127:359-368.
Neupert, W., W. Sebald, A. J. Schwab, A. Pfaller and T. Bucher, 1969.
Puromycin sensitivity of ribosomal label after incorporation of Re¬
labeled amino acids into isolated mitochondria from Neurospora crassa.
Eur. J. Biochem. 10:585-588.
O'Brien, T. W., 1971. The general occurrence of 55S ribosomes In
mammalian liver mitochonli ia. J. Biol. Chem. 246:3409-3417.
O'Brien, T. W., 1972. Occurrence of 55S miniribosomes in mitochondria
of the shark. J. Cell Biol. 55:191a.
O'Brien, T. W., N. D. Denslow and G. R. Martin, 1974. The structure,
composition, and function of 55S mitochondrial ribosomes. In The
Biogenesis of Mitochondria, edited by A. Kroon and C. Saccone,
Academic Press, New York, pp. 347-356.

150
O’Brien, T. W. and G. P. Kalf, 1967. Ribosomes from rat liver
mitochondria. II. Partial characterization. J. Biol. Chem.
242:2180-2185.
O’Brien, T. U. and D. E. Matthews, 1976. Mitochondrial ribosomes.
In Handbook of Genetics: Vol. 5-Molecular Genetics, edited by
R. C. King, Plenum Press. New York, pp. 535-580.
O'Brien, T. W., D. E. Matthews and N. D. Denslow, 1976. Protein compo¬
sition of the bovine mitochondrial ribosome. In The Genetics and
Biogenesis of Chloroplasts and Mitochondria, edited by T. Bucher,
W. Neupert, W. Sebald, and S. Werner. North Holland, Amsterdam,
in press.
O’Farrell, P. H., 1975. High resolution two-dimensional electro¬
phoresis of proteins. J. Biol. Chem. 250:4007-4021.
Olsnes, 1971. The isolation of polysomes from rat liver: contam¬
ination by membrane proteins with high affinity to ribosomes.
Biochim. Biophys. Acta 232:705-716.
Perlman, S. and S. Penman, 1970. Protein-synthesizing structures
associated with mitochondria. i.Mure (London) 227: J 33-137 .
Pettersson, T., S. J. S. Hardy and A. Liljas, 1976. The ribosomal
protein L8 is a complex of L7/12 and LIO. FEBS Letters 64:135-138.
Potter, S. S., J. E. Newbold, C. A. Hutchinson III, and M. H. Edgell,
1975. Specific cleavage analysis of mammalian mitochondrial DNA.
Proc. Natl. Acad. Sci. USA 7_2:4496-4500.
Prakash, L., 1975. Repair of pyrimidine dimers in nuclear and mito¬
chondrial DNA of yeast irradiated with low doses of ultraviolet
light. J. Mol. Biol. 98:781-795.
Pring, D. R., 1974. Maize mitochondria: Purification and characteri¬
zation of ribosomes and ribosomal ribonucleic acid. Plant Physiol.
53:677-683.
Pring, D. R. and D. W. Thornbury, 1975. Molecular weights of maize
mitochondrial and cytoplasmic ribosomal RNAs under denaturing condi¬
tions. Biochim. Biophys. Acta 383:140-146.
Rabbitts, T. H. and T. S. Work, 1971. The mitochondrial ribosome and
ribosomal RNA of the chick. FEES Letters 14:214-218.

151
Rawson, J. R. and E. Stutz, 1969. Isolation and characterization of
Euglena gracilis cytoplasmic and chloroplast ribosomes and their
ribosoma1 RNA components. Biochim. Biophys. Acta 190r368-380.
Reijnders, L., P. Sloof, J. Sival and P. Borst, 1973. Gel electro-
phoresis of RNA under denaturing conditions. Biochim. Biophys.
Acta 324:320-333.
Reisfeld, R. A., U. J. Lewis and D. E. Williams, 1962. Disk electro¬
phoresis of basic proteins and peptides on polyacrylamide gels.
Nature 195:281-282.
Robbersori, Y. Moni, G. Attardi and N. Davidson, 1971. Expression
of the r¡¡, _ochondrial genome in HeLa cells. VT» Size determinat ion
of mitochondrial ribosomal RNA by electron microscopy. J. Mol. Biol.
60:473-4ft4.
Sacchi, A., F. Cerbone, P. Cammarano and U. Ferrini, 1973. Physio-
chemical characterization of ribosome-like (55S) particles from rat
liver mitochondria. Biochim. Biophys. Acta 308:390-403.
Sager, R., 1972. Cytoplasmic Cenes and Organelles. Academic Press,
New York.
Sager, R. and M. G. Hamilton, 1967. Cytoplasmic and chloroplast ribo¬
somes of Chlamydomonas: UltracenLrifugal characterization. Science
157:709-711.
Schatz, G. and T. L. Mason, 1974. The biosynthesis of mitochondrial
proteins. Ann. Rev. Blochcm. 43:51-87.
Schmitt, H., 197U. Characterization of a 72S mitochondrial ribosome
from Saccharomyces cerovisiae. Eur. J. Biochem. 17:278-283.
Schmitt, H., H. Grossfeld, J. S. Beckmann and U. Z. Littauer, 1974.
Biogenesis of mitochondria from Artemi a salina cysts and the tran¬
scription jin vivo of the DNA. 7 : The Biogenes is of Mitochondria,
edited by A. M. Kroon and C. Saceone, Academic Press, New York,
pp. 135-146.
Schnaitman, C. and J. W. Creenawalt,1968. Enzymatic properties of the
inner and outer membranes of rat liver mitochondria. J. Cell Biol.
38:158-175.
Slierton, C. C. and I. G. Wool, 1974, A comparison of the proteins of
rat skelei ii muscle and Liver ribosomes by two-dimensional poly¬
acrylamide gel electrophoresis. J, Biol. Cham. 249:2258-2267.

152
Sinclair, J. H. and D. D. Brown, 1971. Retention of common nucleotide
sequences in the ribosomal deoxyribonucleic anil of eukaryotes and
some of their physical characteristics. Biochemistry 10:2761-2769.
Spencer, M. and F. Poole, 1965. On the origin of crystal!zable RNA
frem yeast. J. Mol. Biol. 11:315-326.
Spitnik-Elson, P. and A. Atsmon, 1969. Detachment of ribosomal pro¬
teins by salt. I. Effect of conditions on the amount of protein
detached. J. Mol. Biol. 45:113-124.
Staehelin, T., D. Maglott and R. E, Monro, 1969. On the catalytic
center of peptidyl transfer: A part of the 50S ribosome structure.
Cold Spring Harbor Symposium 34:39-48.
Stutz, E. and A. Boschetti, 1976. Cliloroplast ribosomes. In Handbook
of Genetics: Vol. 5-Molecular Genetics, edited by R. C. King, Plenum
Press, New York, pp. 425-449.
Subramanian, A. R., 1974. Sensitive separation procedure for Escherichia
coli ribosomal proteins and the resolution of high-molecular-weight
components. Eur. J. Biochem. 45:541-546.
Swanson, R. F, and I. B. Dawid, 1970. The mitochondrial ribosome of
Xenopus laevis. Proc. Natl. Acad. Sci. USA 66:117-124.
Tappel, A. L., 1973. Lipid peroxidation damage to cell components.
Fed. Proc. 3^:1870-1874.
Taylor, M. M. and R. Storck, 1964. Uniqueness of bacterial ribosomes.
Proc. Nail. Acad. Sci. USA 52:958-965.
Terhorst, C., W. Müller, R. Laursoii and B. Wittmann-Liebold, 1972.
Amino acid sequence of a 50S ribosomal protein involved in both EFC
and EFT dependent GTP-hydrolysis. FEBS Letters 28:325-328.
Thomas, J. R. and K. K. Tewari, 1974. Conservation of 70S ribosomal
RNA genes in the chloiroplast DNAs of higher plants. Proc. Natl.
Acad. Sci. USA TU 3147-3151.
Tischendorf, G. W., G. Stoffler and H. G. Wi.tmann, 1974. The com¬
parative characterization of the proteins of Escherichia coli 30S
and 50S ribosomal subunits. The identity of S20 and L26. Eur. J.
Biochem., in press,
Tissiéres, A., J. Watson, D. Schlessinger, and B. Hollingworth, 1959.
Ribonueleoprotein particles from Escherichia coli. J. Mol. Biol.
1:221-234.

153
Ts’o, P. 0. P., J. Bonner and J. Vinograd, 1958. Structure and
properties of microsomal nucleoprotein particles from pea seedlings.
Biochim. Biophys. Acta 30:570-582.
van den Bogert, C. and H. de Vries, 1976. The mitochondrial ribosomes
of Neurospora crassa. II. Comparison of the proteins from Neuros-
pora mitochondrial ribosomes with ribosomal proteins from Neurospora
cytoplasm, rat-liver mitochondria and from bacteria. Biochim.
Biophys. Acta, in press.
Vasconcelos, A. C. L. and L. Bogorad, 1971. Proteins of cytoplasmic,
chloropL.r.t and mitochondrial ribosomes of some plants. Biochim.
Biophys. Acta 228:492-502.
Verma, I. M., M. Edeliu.m, M. Herzhog and U. Littauer, 1970. Size
determi¡j.jtion of mitochondrial ribosomal RNA from Aspergillus nidulans
by election microscopy. J. Mol. Biol. 52:137-140.
Vesco, C. and S. Penman, 1969. The cytoplasmic RNA of HeLa cells:
New disci ote species associated with mitochondria. Proc. Natl. Acad.
Sci. USA 62:218-225.
Vignais, P. V., B. J. Stevens, J. Huet and J. Andre, 1972. Mitoribo-
somes from Candida utilis. Morphological, physical and chemical
characterization of the monomer form and its subunits. J. Cell Biol.
54:468-492.
Voynow, P. and C. G. Kurland, 1971. Stoichiometry of the 30S ribosomal
proteins of Escherichia coli. Biochemistry 10:517-524.
Weber, H. J., 1972. Stoichiometric measurements of 30S and 50S ribo¬
somal proteins. Mol. Gen. Genet. 119:233-248.
Wengler, G., G. Wengler and K. Scherrer, 1972. Ribonucleoprotein
particles in HeLa cells: The contamination of the postmitochondrial-
cytoplasraic fraction from HeLa cells with ribonucleoprotein particles
of mitochondrial origin. Eur. J. Biochem. 24:477-484.
Wilson, R. H., J. B. Hanson, and H. H. Mollenhauer, 1968. Ribosome
particles in corn mitochondria. Plant Physiol. 43:1874-1877.
Wittmann, H. G., Ü/74. Purification and identification of Escherichia
coli ribosomal proteins. In Ribosomes, edited by M. Nomura, A.
Tissieres, and P. Lengyel, Cold Spring Harbor Laboratory, Cold Spring
Harbor, pp. 93-114.

154
Wittmann, H. C., O. Stoffler, E. Kaltschmidt, V. Rudloff, H. Janda,
M. Dzionara, D. Bonner, K. Nierhaus, II. Cech, I. Hindennach and
B. Wittmann, 1970. Protein chemical and serological studies on
ribosomes of bacteria, yeast and plants. FEBS Syinp. 21:33-46.
Wittmann, H. G., G. Stoffler, W. Plepersberg, P. Buckel, D. Ruffler
and A. Bock, 1974. Altered S5 and S20 ribosomal proteins in
revertants of an al any!-LRNA-cyn the tase mutant of Kschericia coli .
Mol. Gen. Genet. 134:225-236.
Wool, 1. G. and G. Stoffler, 1974. Structure and function of eukary¬
otic ribo:.nines. In Ribosomes, edited by M. Nomura, A. Tissieres,
and P. Lcngyel, Cold Spring Harbor Laboratory, Cold Spring Harbor,
pp. 417-460.
Yurina, N. P. and M. S. Odintsova, 1974, Buoyant density of chloro-
plast ribos mes in CsCl. Plant. Sci. Lett. 3^:229-234.
Zitomer, R. S. and .1. G. Flaks, 1972. Magnesium dependence and equilib¬
rium of the Escherichia coli ribosomal subunit association. J. Mol.
Biol. 71:263-279.

BIOGRAPHICAL SKETCH
David Earl Matthews was born May 10, 1948, in Birmingham, Alabama.
He was raised in Alabama, Georgia, and Florida, and graduated from Lake¬
land High School, Lakeland, Florida in 1966. His undergraduate education
was at Davidson College and Florida Southern College, with support from
a National Merit Scholarship. In 1972 he received a B.S. degree in
Chemistry, summa cum laude, from Florida Southern College. Shortly there¬
after he was married to PaLty S. Salisbury of St. Petersburg, Florida.
His graduate studies at the University of Florida have been supported by
a National Science Foundation Graduate Fellowship and a Graduate Council
Fellowship.
155

I certify that I have read this study and that, in my opinion,
il conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.
Thomas W. O'Brien, Chairman
Associate Professor of
Biochemistry and Molecular Biology
1 certify that I have read this study and that, in my opinion,
it conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.
Ms£££L
Robert J-. Cohen
Associate Professor of
Biochemistry and Molecular Biology
I certify that I have read this study and that, in my opinion,
it conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.
tC, -t,.
L. O'Neal Ingram j
Assistant Professor of
Microbiology

I certify that I have read this study and that, in my opinion,
it conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.
and Molecular Biology
l certify that I have read this study and that, in my opinion,
it conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.
R. Michael Roberts
Professor of Biochemistry
and Molecular Biology
This dissertation was submitted to the Graduate Faculty of the Depart¬
ment of Biochemistry in the College of Arts and Sciences and to the
Graduate Council, and was accepted as partial fulfillment of the require¬
ments for the degree of Doctor of Philosophy.
December 1 976
Dean, Graduate School

UNIVERSITY OF FLORIDA
3 1262 08553 3049