Functional studies on the peptidyl transferase center of mammalian mitochondrial ribosomes


Material Information

Functional studies on the peptidyl transferase center of mammalian mitochondrial ribosomes
Physical Description:
xii, 124 leaves : ill. ; 28 cm.
Denslow, Nancy Mary Derrick, 1945-
Publication Date:


Subjects / Keywords:
Mitochondria   ( lcsh )
Ribosomes   ( lcsh )
Proteins -- Synthesis   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis--University of Florida.
Includes bibliographical references (leaves 112-123).
Statement of Responsibility:
by Nancy D. Denslow.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000167693
notis - AAT4084
oclc - 02863908
System ID:

Full Text








Dr. Thomas W. O'Brien provided helpful guidance

and advice throughout the course of this research project.

I would like to thank Dr. Richard Boyce and the members

of my committee: Dr. Rusty Mans, Dr. Melvin Fried;

Dr. James Preston III, and Dr. Peter Cerutti; who through

their criticisms helped me prepare this manuscript. I

also am thankful to David Matthews and Robert Hessler,

graduate students in biochemistry, for their helpful

discussions, and to Warren Clark and Mark Critoph for their

expert technical assistance.

I am grateful to my parents-in-law, David and

Mary Denslow, for their making it possible for me to com-

plete my graduate studies. I am indebted to Mary Denslow

for typing the manuscript. Finally my thanks go to Marvin

and Mary Derrick and David Denslow Jr. for their encourage-

ment. I.would like to dedicate this work to Sandra and



Acknowledgements..................................... ii
List of Tables........................ ................ v
List of Figures..................................... vi
List of Abbreviations and Definitions................ viii
Abstract............................................. x
Chapter 1. Introduction.............................. 1
Physical-Chemical Properties of Mammalian
Mitochondrial Ribosomes ....................... 1
Biogenetic Properties of Mitochondria............ 3
Protein Biosynthesis in Mitochondria...... 5
Mitochondrial DNA.......................... 6
Mitochondrial RNA........................... 7
The Problem..................................... 9
Chapter 2. Materials and Methods .................... 17
Materials........................................ 17
Methods.............. ........................... .. 18
Preparation of Mitochondria................ 18
Preparation of Mitochondrial Ribosomes..... 18
Preparation of Mitochondrial Subribosomal
Particles................................ 21
Preparation of Bovine Liver 80S
Cytoplasmic Ribosomes..................... 22
Isolation of Crude Mitochondrial Factors... 23
Preparation of E. coli Ribosomes and
Crude Factors.......... .................. 24
Preparation of N-acetyl[3H]leucyl-tRNA..... 25
Peptidyl Transferase Assay................. 26
Poly U-Dependent Polymerization of
[ H]Phenylalanine........................ 27
Binding of [3H]GTP to Ribosomes............. 28
Efficiency of Counting..................... 29
Chapter 3. Results.................................. 31
Preparation of Mitochondrial Ribosomes.......... 31
Peptidyl Transferase Center of the Mitochondrial
Ribsomes: Activity........................... 34
Susceptibility of the Mitochondrial Ribosomal
Peptidyl Transferase Activity to Antibiotic
Inhibitors ...... ....................... ...... 42
Chloramphenicol........................... 47
Streptogramin Group: PA114A and
Vernamycin A...................... ...... 49
Lincosamine Group: Lincomycin and
Celesticetin................. .. ....... 52
Macrolide Group: Carbomycin and Tylosin... 55


Controls for the Antibiotic Susceptibility
Studies.............................. ........... 55
Optimal Activity of Mitochondrial
Ribosomes................................. 55
Absence of Bound Impurities to the
Lincomycin Site of Mitochondrial
Ribosomes................................. 58
The Effect of Lincomycin on Salt-Washed
E. coli Ribosomes......................... 62
Reproducibility of Results ................. 62
The Function of Mitochondrial Ribosomes in
Additional Assays..................... ...... 64
Poly U-Dependent Incorporation of
[3H]Phenylalanine........................ 65
Function of Mitochondrial Ribosomes in
the Binding of [3H]GTP..................... 68
Chapter 4. Discussion ................................ 72
Significance of Results........................... 72
Peptidyl Transferase Activity of
Mitochondrial Ribosomes.................. 72
Antibiotic Susceptibility ................. 73
Additional Assays of Mitochondrial
Ribosomal Activities....................... 89
Stability of the Peptidyl Transferase
Locus and [3H]GTP Binding Ability........ 93
Conclusions.................................. 98
New Directions in Research..................... 101
Appendix A........................................... 103
Appendix B........ ............................ ..... 106
Bibliography.............................. ............ 112
VITA........................... ... ...... ..... ... .... 124


1. Activities of Various Ribosome Classes
in the Peptidyl Transferase Reaction.......... 39

2. Peptidyl Transferase Activity of the
Large Subribosomal Particles from
Bovine Mitochondria and E. coli
Prepared in Selected Buffers................. 41

3. Molar Concentration of Antibiotics
Required to Inhibit the Peptidyl
Transferase Reaction by 50%.................... 44

4. Peptidyl Transferase Activities of
Ribosomes in the Presence of Vernamycin
B, Oleandomycin and Gougerotin................ 46

5. The Effect of Preincubating Ribosomes at
400 C on the Peptidyl Transferase Activity.... 59

6. Lincomycin Susceptibility of E. coli
Ribosomes Resuspended in Mitochondrial
Supernatant ................................... 61

7. Comparison of Lincomycin Susceptibility
of Crude and Salt-washed Ribosomes............. 63

8. Activity of Bovine Mitochondrial Ribosomes
in Poly U-Dependent Systems................... 66

9. Binding of [3H]GTP to Ribosomes................. 70

10. Comparison of the Activities of 55S
Mitochondrial Ribosomes Isolated from
Various Organisms in Poly U-Dependent
Reactions............ ...... ................. ... .91

11. Poly Phenylalanine Polymerizing Activity of
Ribosomes in the Presence of NH4+ Ions........ 107


1. Sucrose density gradient centrifugation of
crude ribosomes isolated from bovine liver
mitochondria by procedure 1................... 32

2. Sucrose density gradient analysis of isolated
55S ribosomes .............................. 33

3. Sucrose density gradient centrifugation of
crude ribosomes isolated from bovine liver
mitochondria by procedure 2................... 35

4. Time course of the peptidyl transferase
reaction................................... 37

5. Effect of chloramphenicol on the peptidyl
transferase activity of various ribosomes..... 48

6. Effect of the streptogramin group of
antibiotics on the peptidyl transferase
activity of various ribosomes................. 50

7. The susceptibility of crude and salt washed
E. coli 70S ribosomes to inhibition by
antibiotics of the streptogramin group........ 51

8. Effect of the lincosamine group of antibiotics
on the peptidyl transferase activity of
ribosomes..................................... 53

9. Effect of the macrolide group of antibiotics
on the peptidyl transferase activity of
ribosomes............................... ...... 56

10. Chemical structures of inhibitors............... 76

11. Effect of lincomycin on peptidyl transferase
activity of ribosomes ....................'..... 84

12. Comparison of functional activities and
buoyant densities of mitochondrial ribosomes
prepared under buffer conditions of
decreasing Mg2+/p ratios ...................... 96

13. Synthesis of polyphenylalanine by E. coli
ribosomes in the presence of increasing
concentrations of E. coli factors............ 108

14. Determination of an inhibitor of protein
synthesis in preparations of mitochondrial
elongation factors................. .......... 111




A260 absorbance at a wavelength of 260 nanometers

ATP adenosine-5'-triphosphate

CAP chloramphenicol

cpm counts per minute

dalton the mass of 1 hydrogen atom

DOC sodium deoxycholate

dpm disintegrations per minute

DNA deoxyribonucleic acid

EDTA ethylene diamine tetraacetate

g gravity

GTP guanosine-5'-triphosphate

h hour

M molar (mole/litre)

min minute

mM millimolar (millimole/litre)

mRNA messenger RNA

poly U poly uridylic acid

RNA ribonucleic acid

rRNA ribosomal RNA

S sedimentation coefficient (Svedberg units)

sec second

Tris tris(hydroxymethyl)aminomethane


tRNA transfer RNA

V ionic strength


"native" subribosomal particles subribosomal particles

obtained directly from lysed mitochondria, after sucrose

density gradient centrifugation, as by-products of mito-

chondrial monosome preparations.

"derived" subribosomal particles subribosomal particles

obtained by dissociating mitochondrial 55S monosomes

in a high salt buffer.

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



Nancy D. Denslow

August, 1975

Chairman: Thomas W. O'Brien
Major Department: Biochemistry

Ribosomes isolated from mammalian mitochondria

have unique physical-chemical properties when compared to

prokaryotic and eukaryotic ribosomes. Although their

molecular weight (2.8 million daltons) is close to that

of Escherichia coli ribosomes, their sedimentation co-

efficient is 55S compared to 70S for those from E. coli

and 80S for those from the cytoplasm of eukaryotic cells.

This sedimentation behavior reflects their unusual compo-

sition. They have a greater proportion of protein, 70%,

than the 35 to 55% normally encountered in ribosomes.

The ribosomal RNAs, which are smaller than those found in

other ribosomes, constitute only 30% of the total mass.

It seems likely that a ribosome having these unique

physical properties would also exhibit unique functional


In this work, some functional characteristics of

mitochondrial ribosomes were investigated. The overall

mechanism of protein synthesis appeared to be similar to

that of other ribosomes as assessed in cell free systems

involving poly-uridylic acid-directed phenylalanine poly-

merization, guanosine triphosphate (GTP)-binding, and

peptidyl transferase activity. For protein synthesis,

mitochondrial ribosomes required, in addition to amino

acids, energy (adenosine triphosphate and guanosine tri-

phosphate), template, transfer ribonucleic acid, and

soluble factors. In the poly-uridylic acid-directed

assay for protein synthesis, mitochondrial ribosomes were

able to utilize unpurified E. coli elongation factors. In

the GTP-binding assay, [3H]GTP was observed to bind to

mitochondrial ribosomes with the same degree of efficiency

as it binds to E. coli ribosomes, which suggests the pres-

ence of a bound elongation factor on the isolated mitochon-

drial ribosomes analogous to the bacterial elongation

factor G.

The peptidyl transferase activity of mitochondrial

ribosomes was investigated in depth using the modified

fragment reaction. The peptidyl transferase activity was

localized on the large subunit. The activity is diminished

by 75% when the ribosomes are salt-washed in buffers which

do'not affect the activity of E. coli ribosomes. The

mitochondrial ribosomal activity is therefore more labile

than that of E. coli ribosomes.


Several antibiotic inhibitors were used as speci-

fic probes of the peptidyl transferase activity of mito-

chondrial ribosomes. The antibiotic susceptibility of 55S

ribosomes was compared to that of E. coli ribosomes and

bovine cytoplasmic ribosomes. Mitochondrial ribosomes

appeared to be sensitive to several inhibitors of pro-

karyotic specificity, such as chloramphenicol and the

streptogramins. But they showed significantly lower sus-

ceptibilities to the macrolides and the lincosamines, also

inhibitors of prokaryotic protein synthesis. Ten- to 700-

fold higher concentrations of these latter drugs were re-

quired to inhibit the peptidyl transferase activity of

mitochondrial ribosomes relative to that required for typi-

cal bacterial ribosomes. From earlier studies, it was

known that mitochondrial ribosomes were not sensitive to

inhibitors specific for protein synthesis occurring on

ribosomes obtained from the cytoplasm of eukaryotic

organisms. The peptidyl transferase center of mitochon-

drial ribosomes is therefore discriminated from those of

both prokaryotic and eukaryotic cytoplasmic ribosomes on

the basis of its response to antibiotic probes.



Physical-Chemical Properties of Mammalian
Mitochondrial Ribosomes

Ribosomes of unique physical-chemical properties

were discovered in mammalian mitochondria by O'Brien and

Kalf eight years ago (O'Brien and Kalf, 1967a; O'Brien

and Kalf, 1967b). Intact rat liver mitochondria were

able to incorporate added [3H]leucine into acid-insoluble

protein in the absence of cell sap or added energy in the

form of ATP, Under these conditions, the microsomal sys-

tem does not synthesize proteins. The treated mitochon-

dria were lysed and the material sedimenting at 230,000

x g was collected. Analysis on sucrose density gradients

revealed that the radioactivity was localized on particles

sedimenting at 55S. Subsequently, the ribosomal nature

of these 55S particles was confirmed by experiments in

which the monosome dissociated reversibly into 28S and

39S subunits under conditions of low Mg2+ and high ionic

strength (O'Brien, 1971). Swanson and Dawid (1970) pro-

vided additional confirmation by showing protein synthe-

sizing activity in a poly U dependent system by the

combined subunits of 55S particles of mitochondria from

Xenopus laevis eggs.


The existence of mitochondrial ribosomes with a

sedimentation coefficient close to 55S has been observed

for other mammalian organisms (reviewed in: Borst and

Grivell, 1971; O'Brien and Matthews, 1975). Having also

been found in Xenopus laevis (Swanson and Dawid, 1970),

shark (O'Brien, 1972), and locust (Kleinow et al., 1971),

they probably appear in all multicellular animals.

The low sedimentation coefficient of these ribo-

somes is not due to a low molecular weight as was once

generally assumed (Borst and Grivell, 1971). Indeed, they

have molecular weights of 2.8x106 daltons, close to the

value of 2.7x106 daltons for E. coli ribosomes (Hamilton

and O'Brien, 1974; de Vries and Kroon, 1974). This is due

instead to a very low buoyant density, corresponding to

an abnormally high protein content (Hamilton and O'Brien,

1974; Sacchi et al., 1973; O'Brien et al., 1974; de Vries

and Kroon, 1974). O'Brien has found a buoyant density of

1.40 g/cc for the 55S form as opposed to 1.58 for extra-

mitochondrial ribosomes and 1.63 to 1.65 for most bacterial

ribosomes. This corresponds to a protein content of 70%

which is considerably more than the 35 to 55% found in

other ribosomes.

Examination of the ribosomal protein content by

2 dimensional polyacrylamide gel electrophoresis reveals

that mitochondrial ribosomes contain many more ribosomal

proteins, 90 (David Matthews, personal communication),

than the ribosomes of E. coli or eukaryotic cytoplasm


which contain 55 (Kaltschmidt and Wittmann, 1970; Kurland,

1972) and 70 (Wool and Stbffler, 1974) respectively. The

role that the extra ribosomal proteins play in the function

of mammalian mitochondrial ribosomes is no.t known.

The physical-chemical properties of animal mito-

chondrial ribosomes differ remarkably from those of pro-

tists and fungi (reviewed in O'Brien and Matthews, 1975).

Functional phylogenetic differences between mitochondrial

ribosomes of fungal and mammalian origins were proposed

by Linnane and his group (Clark-Walker and Linnane, 1966;

Lamb et al., 1968; Towers et al., 1972; Towers et al.,

1973). Their conclusion was supported by the observation

of differential inhibition caused by certain antibiotic

inhibitors of protein synthesis when intact mitochondria

isolated from yeast and rats were used. These results

were challenged by Kroon and de Vries (1971) who showed

that in many cases the insensitivity was due to impermea-

bility of the mitochondrial membrane to the antibiotics.

Conclusive evidence for possible phylogenetic differences

.among mitochondrial ribosomes rested on the analysis of

the mechanism of protein synthesis of the isolated par-


Biogenetic Properties of Mitochondria

The existence of mitochondria in the cytoplasm of

eukaryotic cells has aroused new interest in the past few

years, especially after the discovery that these organelles


contain their own DNA, RNA and protein synthesizing appa-

ratus (reviewed in Ashwell and Work, 1970). Several

theories have been advanced to explain their presence in

the early eukaryotic type cell, the most popular of which

is the theory of endosymbiosis. This theory suggests

that mitochondria were once symbiotic bacteria which in-

vaded eukaryotic cells and, through evolution, became de-

pendent on the host cell genome for some of their vital

functions while, conversely, the host cell became depen-

dent on products originating from the mitochondria (re-

viewed in Ashwell and Work, 1970). This theory is

supported by investigators, who find many similarities

between the genetic apparatus of mitochondria and that of

typical prokaryotes.

Alternatively, mitochondria may have originated

through "cluster-cloning" (reviewed in Bogorad, 1975).

According to this hypothesis, the cell's genetic. material

was partitioned off into gene clusters. Each cluster and

the protoplasm immediately around it were enclosed by

membranes to form mitochondria, chloroplasts and nuclei.

These compartments would have the capacity to reproduce

themselves. Because of gene segregation in the original

genome, gene products from different clusters may be

needed to complete functional units the electron

transport system in mitochondria. According to this hy-

pothesis, genes would have been separated into the com-

partments early in the formation of eukaryotic cells,


rather than having been transferred from one compartment

to another at some later time.

The plasmid theory proposed by Raff and Mahler

(1972) is a variant of the cluster-clone hypothesis.

According to this theory groups of genes in the form of

plasmids would become detached from the nucleus. Then

membranes would develop around them to form organelles,

which would have the capacity to reproduce themselves.

Protein Biosynthesis in Mitochondria

Of the many proteins found within mitochondria,

there are probably not more than 12 encoded by the mito-

chondrial DNA (reviewed in Schatz and Mason, 1974). In

fungi 8 products have been identified. They are 3 sub-

units of the cytochrome oxidase (Mason and Schatz, 1973),

4 subunits of the oligomycin-sensitive ATPase (Tzagoloff

and Meagher, 1972; Ebner et al., 1973) and 1 subunit of

cytochrome b (Weiss et al., 1971; Weiss, 1972). The mito-

chondrial ribosomal proteins are among those synthesized

in the cytoplasm of eukaryotic cells (Kintzel, 1969b).

These ribosomal proteins are mitochondrial specific and

differ from those found in cytoplasmic ribosomes (Kintzel,

1969a, and David Matthews, personal communication).

The biogenetic system in mitochondria thus exists

primarily for the synthesis of a few proteins which must

be extremely important for the survival of cells. Evi-

dence for a tight coordination between mitochondrial and

cytoplasmic protein synthesis is slowly appearing.

Exactly how they depend on each other is not yet known.

The mitochondrial products seem to be needed for the cor-

rect integration of the cytoplasmically made proteins

into complexes in the inner membrane (Schatz and Mason,


Mitochondrial DNA

All eukaryotes contain specific mitochondrial

DNAs which are not encoded by nuclear DNA. They occur in

small, circular duplex forms, their size being considera-

bly larger in mitochondria from fungi, 15 to 30 p, than in

mammalian mitochondria, 5 p (Borst and Flavell, 1972). In

mammals, mitochondrial DNA has a limited coding capacity.

Hybridization studies show it codes for mitochondrial rRNA

and tRNA and may code for a few mitochondrial proteins as

described above.

At most, 20 mitochondrial tRNAs have been found

to hybridize with the mitochondrial DNA in yeast. In

animals fewer complements have been found; for example, 15

in Xenopus (Dawid, 1972) and 12 in HeLa cells (Aloni and

Attardi, 1971a; Wu et al., 1972). These numbers are

considerably less than the minimum necessary to read the

possible 61 codons even allowing for maximal wobble.

Various possibilities arise: a) not all of the mitochon-

drial tRNAs have been found, b) some of the'tRNAs found

in the mitochondria are coded by nuclear genes and


therefore do not hybridize with the mitochondrial DNA,

or c) not all the codons are used in the mitochondria.

This last possibility is favored by Costantino and

Attardi (1973).

Apparently both strands of DNA are transcribed

for information. Nass and Buck (1970) have demonstrated

that in mitochondria from rat livers, two tRNAs hybrid-

ized to the heavy strand of DNA, and two to the light

strand. Even more convincing are Attardi's genetic

maps showing hybridization of ferritin-labelled tRNAs

from HeLa cells to both strands of the mitochondrial DNA

(Wu et al., 1972).

Mitochondrial RNA

Aloni and Attardi (1971b) have demonstrated that

there is symmetric transcription of mitochondrial DNA.

The RNA products transcribed from the heavy strand of

mitochondrial DNA had a relatively long half life when

compared to those made from the light strand. Symmetric

transcription of DNA was confirmed by several labora-

tories using RNA polymerase from E. coli to synthesize

mitochondrial RNA in vitro (Schafer et al., 1971; Tabak

and Borst, 1970). In these studies, the product made

from the light strand was predominant.

Several investigators have succeeded in isolating

mitochondrial RNA polymerases from a number of organisms

(Gadaleta et al., 1970; Kiintzel and Schafer, 1971;


Scragg, 1971; Tsai et al., 1971; Wintersberger, 1970;

Wu and Dawid, 1972). Recently Scragg (1974) used isolated

yeast RNA polymerase to transcribe yeast mitochondrial

DNA. He translated the RNA product in a cell free system

using E. coli ribosomes and soluble factors. He obtained

several labeled polypeptides but these have not yet been

identified as specific mitochondrial products.

For the most part, RNA polymerases isolated from

mitochondria appear to be sensitive to rifampicin (Gada-

leta et al., 1970; KIntzel and Schafer, 1971; Scragg,

1971). Mitochondrial RNA synthesis can also be specifi-

cally inhibited by low concentrations of ethidium bromide

and actinomycin D. Preliminary work using intact mito-

chondria suggests that ethidium bromide and rifampicin may

also act to inhibit mitochondrial protein synthesis di-

rectly (Dube et al., 1973; Grivell and Metz, 1973; Avad-

hani and Rutman, 1975),

Mitochondrial tRNAs are thought to resemble bac-

terial tRNAs more closely than cytoplasmic tRNAs. As in

bacterial systems, N-formyl-methionyl-tRNA is the initia-

tor tRNA (Smith and Marcker, 1968). There also appear to

be similarities between synthetases, since the bacterial

enzymes can be used to charge mitochondrial tRNAs in

vitro. This is noteworthy since the mitochondrial syn-

thetases seem to be under nuclear control (Davey et al.,



Ribosomal RNAs isolated from animal mitochondria

are smaller, 12S and 16S (Borst and Grivell, 1971), than

ribosomal RNAs found in either prokaryotic or eukaryotic

ribosomes. Using hybridization experiments, Aloni and

Attardi (1971b) showed that ribosomal RNAs isolated from

HeLa mitochondria were encoded by mitochondrial DNA. The

G+C content of mitochondrial rRNA is considerably lower

than that of both eukaryotic and typical bacterial rRNAs,

giving it a lower stability in its secondary structure

(Dawid, 1973).

In mitochondria of Tetrahymena a ribosomal RNA sedi-

menting at 5S has been observed (Chi and Suyama, 1970),

However in animals, 5S RNA appears to be absent. A ribo-

somal RNA sedimenting at 3S was isolated from mitochondria

of hamsters by Dubin et al. (1974). Even though the 3S

RNA did not associate with the mitochondrial large sub-

ribosomal particles, the authors postulated may

correspond to the 5S RNAs found in other ribosomes.

The Problem

When this study was initiated, little had been re-

ported concerning functional activities of mammalian mito-

chondrial ribosomes. It was generally assumed that, like

the mitochondrial ribosomes isolated from fungi, mammalian

mitochondrial ribosomes were indistinguishable from bac-

terial ribosomes in terms of function. Initial studies

were complicated by the very low yield of mitochondrial

ribosomes, their relatively low activities in the assays

used, and the difficulty of obtaining mitochondrial ribo-

somes free of cytoplasmic contaminants.


The unexpected physical-chemical properties of the

mammalian mitochondrial ribosomes and the possibility that

phylogenetic differences existed between protein synthesis

mechanisms within mitochondria of lower and higher eukary-

otes led me to examine some functional activities of the

isolated 55S ribosomes.

This research was directed towards determining

the role of mitochondrial ribosomes in protein synthesis.

To this end I decided to look at one specific but crucial

reaction, that of peptide bond formation on mitochondrial

ribosomes. First, the peptidyl transferase activity was

localized on the mitochondrial ribosome. Specific anti-

biotic probes were used to characterize this site with

respect to the analogous site on other ribosomes. The

antibiotics chosen were those known to act on the pep-

tidyl transferase site of prokaryotic ribosomes. Some

were known to inhibit protein synthesis directly within

mitochondria (for example chloramphenicol), whereas the

effect of others (for example carbomycin and lincomycin)

was open to interpretation. In my studies I describe the

sensitivities of isolated mammalian mitochondrial ribosomes

with respect to these drugs.

Under normal conditions of protein synthesis, for-

mation of peptide bonds by peptidyl transferase is pro-

moted by complex interactions occurring among the two

ribosomal subunits, mRNA, elongation factors and tRNA.

There are two substrate binding sites on the ribosome:


an "A" site for the binding of amino acyl-tRNA and a "P"

site for the binding of peptidyl-tRNA. Binding of sub-

strate to the "A" site of ribosomes occurs readily and

specifically. The binding of peptidyl-tRNA to the "P"

site is weak in the absence of stabilizing interactions

among the various components of the functioning ribosome

complex. In the presence of alcohol (methanol or ethanol)

the binding of peptidyl-tRNA to the "P" site on the large

subunit is promoted in the absence of the many otherwise

required components. Alcohol can also stimulate the bind-

ing of a N-acetyl-aminoacyl-CCA fragment, produced by

RNase Ti digestion of N-acetyl-aminoacyl-tRNA, to the "P"

site. With the two sites occupied, peptidyl transferase

can catalyze formation of a peptide bond. The mechanism

of action in the presence of ethanol is thought to be the

same as that occurring in vivo in protein synthesis

(Monro et al., 1968). Furthermore, the sensitivity to

specific antibiotic inhibitors is retained (Monro et al.,


The "fragment reaction" of Monro (1967), which'

measures peptide bond formation between puromycin and the

N-acetyl-aminoacyl-tRNA fragment, takes advantage of this

alcoholic medium in order to examine the fine structure

of the peptidyl transferase locus. The large subribosomal

particle containing the intact peptidyl transferase locus

is the minimal ribosomal structure required. Using this

assay it was established that energy and soluble factors


are not required for the formation of peptide bonds

(Monro et al., 1969).

In addition to the predominant peptide bond

forming reaction occurring on ribosomes in the presence

of alcohol, the peptidyl transferase also catalyzes the

transesterification of the amino acid (Scolnick et al.,

1970). The rate of this reaction is one order of magni-

tude smaller than that of the peptide'bond forming re-

action and it appears to be dependent on the nature of

the substrate bound to the "A" site. The peptide bond

forming reaction is favored when the substrate binding to

the "A" site is either an amino acylated tRNA or puromycin,

whereas the transesterification reaction will occur if

the substrate is an uncharged tRNA (Scolnick et al., 1970).

Hydrolysis of the amino acyl-tRNA occurring at room tem-

perature during the course of the reaction, therefore, will

influence the formation of the ester product. However,

the treatment of the reaction mixture with alkali (see

"methods," Miskin et al., 1970) eliminates the ester

without affecting the N-acetyl[3H]leucyl-puromycin product.

The radioactive product extracted into ethyl acetate under

these conditions measures only the peptide bond forming

activity of the peptidyl transferase.

The fragment reaction is particularly well suited

for studies correlating structure and function of ribo-

somes because it requires a minimal ribosomal structure

and a limited number of intermediate steps immediately


surrounding the peptidyl transferase activity. The pep-

tidyl transferase center of ribosomes has been analyzed

extensively by the use of antibiotic probes, i.e. in-

hibitors which bind specifically to proteins in or around

the active center (reviewed in Pestka, 1971; Vazquez,

1974). These studies have yielded information concerning

the structure-function relationship of ribosomal proteins

in the locus. For example, some inhibitors may affect

the binding of aminoacyl-tRNA to the "A" site whereas

others will affect the binding of peptidyl-tRNA to the

"P" site and still others may inhibit the peptidyl trans-

ferase itself. Antibiotics specific for the peptidyl

transferase center of prokaryotic ribosomes were chosen

for this study because of the prior observation that pro-

tein synthesis in intact mitochondria resembles that of


Chloramphenicol, which specifically inhibits the

peptidyl transferase activity of prokaryotic ribosomes

(Pestka, 1971), is one of a few classical antibiotics

used to discriminate the protein synthesizing systems of

prokaryotic and eukaryotic organisms. On the basis of

its chloramphenicol sensitivity, mitochondrial protein

synthesis was determined to be of the bacterial type

(Rendi, 1959; Clark-Walker and Linnane, 1966; de Vries

et al., 1971; Kroon and de Vries, 1971). The early stud-

ies, however, were done with intact mitochondria. In-

vestigators were not able to rule out a direct inhibitory


effect of chloramphenicol on oxidative phosphorylation

with consequently indirect effects on protein synthesis.

Using isolated mitochondrial ribosomes from rat livers,

de Vries et al. (1971) showed a direct effect of chloram-

phenicol on their peptidyl transferase activity. The

susceptibility of bovine liver mitochondrial ribosomes to

chloramphenicol has been examined and compared to a

standard inhibition profile obtained for E. coli ribosomes.

Members of the streptogramin group of antibiotics

have been shown to interact directly with the 50S subunit

of prokaryotic ribosomes by inhibiting the binding of the

amino acyl end of amino acyl-tRNA to the "A" site (Ennis,

1966; Mao and Putterman, 1969; Vazquez, 1966b). Further-

more, mikamycin, a member of this group, has a direct

effect on protein synthesis occurring within intact yeast

mitochondria (Towers et al., 1972). The susceptibility

of bovine mitochondrial ribosomes to PA114A and vernamycin

A was investigated in this study.

Considerable controversy exists in the literature

concerning the action of the lincosamines and macrolides

within mitochondria of higher organisms (Towers et al.,

1972; Firkin and Linnane, 1969; Kroon, 1969; Williams and

Birt, 1972). Unlike mitochondria from fungi, the intact

mammalian mitochondria respond differentially to these

antibiotics. For example, protein synthesis within intact

organelles is resistant to erythromycin (Firkin and

Linnane, 1969; Kroon, 1969) but highly sensitive to the


action of carbomycin and spiramycin (Towers et al., 1972),

all members of the macrolide group. The intact organelles

are also insensitive to lincomycin (Kroon and de Vries,

1971). A permeability barrier for erythromycin and linco-

mycin at the level of the mitochondrial membrane has been

observed in mammalian mitochondria (Kroon and de Vries,

1971). This conclusion came as a result of experiments

in which mitochondria were ruptured by sonication or by

treatment in a hypotonic medium. Protein synthesis was

then sensitive to the action of these drugs (Kroon and

de Vries, 1971). The observed high sensitivity to carbo-

mycin and spiramycin may be due to a 30-fold concentration

(de Vries et al., 1973) of these drugs within intact

organelles. Studies with isolated mammalian mitochondrial

ribosomes were necessary to determine whether mitochon-

drial ribosomes are indeed susceptible to these antibio-


Poly U dependent polyphenylalanine synthesizing

systems have been used as model reactions for other ribo-

.somes by many investigators. These systems are not as

restrictive as the in vivo system since initiation factors

and initiator tRNA are not required. Nevertheless, the

assays depend on functional and complete ribosomes. Both

ribosomal subunits are required as well as elongation

factors, EF-T and EF-G, phenylalanyl-tRNA, energy in the

form of GTP and ATP, and the correct ionic conditions.

In my studies, a poly U-directed system developed by


Hosokawa et al. (1966) was used to ascertain the func-

tional integrity of beef liver mitochondrial ribosomes.

It was, furthermore, a convenient assay for determining

whether bacterial elongation factors and tRNA can support

the mitochondrial ribosomal activity.

A partial reaction of protein synthesis can be

used to examine the GTP binding ability of ribosomes. The

binding and subsequent hydrolysis of GTP is necessary for

three different steps in protein synthesis: (a) binding

of the initiator tRNA to the peptidyl site mediated by the

initiation factor IF-2; (b) binding of all incoming

aa-tRNAs to the amino acyl site mediated by the elongation

factor EF-T; and (c) translocation of the peptidyl-tRNA

from the "A" site to the "P" site mediated by the elonga-

tion factor EF-G. A large body of evidence suggests that

there is a common binding site on the large subunit for

IF-2, EF-T and EF-G (Hamel et al., 1972; Hamel and

Nakamoto, 1972; Highland et al., 1971; Richman and Bodley,

1972) and that this site is composed mainly of the

r-proteins L7 and L12 (Highland et al., 1974; Highland

et al., 1973). I examined mitochondrial ribosomes for

their ability to bind [3H]GTP as an additional measure of

their competence in the partial reactions of protein




Chloramphenicol was obtained from Sigma Chemical

Company. Lincomycin-HC1 (U-19149A, lot XS-610) and Celes-

ticetin (U-4819, lot 16376) were the generous gifts of

Dr. George B. Whitfield from Upjohn and Company. PA114A

(lot 1454-172A) and Carbomycin (lot 42464) were the gen-

erous gifts of Dr. Nathan Belcher from Pfizer Company.

Tylosin tartrate (lot 4 PE77) was donated by Dr. Robert

Hosley from the Lilly Research Laboratories. Vernamycin A

(SQ16515, Batch 5A) was given by Miss Barbara Stearns from

Squibb. These antibiotics are prepared by the biomedical

research groups in each company and are distributed in

small quantities (5-100 mg) to interested investigators.

They do not appear to lose activity upon storage for 3 to

4 years and no impurities affecting their activity have

been reported. Puromycin diHCl was purchased from Nutri-

tional Biochemical Company, and E. coli de-aminoacylated

tRNA (stripped tRNA) from General Biochemicals. [4,5-3H]L-

leucine, 55 Ci/mmole; [3H]L-phenylalanine, 7 Ci/mmole;

and [8-3H]GTP tetrasodium, packed in 30% ETOH, 12 Ci/mmole

were obtained from Schwarz-Mann. Permablend,' containing

91% PPO and 9% Dimethyl POPOP, was purchased from Packard.




Preparation of Mitochondria

Bovine livers of freshly killed animals were ob-

tained at the slaughterhouse. Alternatively, rat livers

were obtained from young Sprague-Dawley rats, fasted for

18 hr and killed by decapitation. The livers were cooled

promptly by placing them in ice. All preparative pro-

cedures were carried out at 40 C.

The livers were sliced, ground and homogenized in

either a potter Elvehjen glass homogenizer or a Super

Dispax flow-through homogenizer, Model SD-45K, Tekmar Co.,

in 6 volumes of isolation medium A (0.34 M sucrose, 5 mM

Tris-HCl, pH 7.4) or isolation medium B (0.25 M sucrose,

1 mM EDTA, 5 mM Tris-HCl, pH 7.6). Whole cells and cell

debris were sedimented by centrifugation at 970 x g for

10 min in a Sorvall GSA rotor. Mitochondria were resus-

pended three times in isolation medium and sedimented by

differential centrifugation for 10 min at 5200 x g in the

GSA rotor. This procedure has been described previously

by O'Brien (1971).

Preparation of Mitochondrial Ribosomes

During the course of this study the method of ob-

taining ribosomes from mitochondria was varied. Two pro-

cedures are described. The second procedure reduces the

contamination of mitochondrial ribosomes by cytoplasmic 80S


Procedure 1

Three-times-washed mitochondria were resuspended

at 20 mg protein/ml standard buffer (20 mM MgC12, 100

mM KC1; 20 mM triethanolamine, pH 7.6, 5 mM 2-mercapto-

ethanol). Heparin and oligonucleotides, prepared by

partial digestion of yeast tRNA with NaOH (Spencer and

Poole, 1965), were added to 0.05 mg/ml and 0.2 mg/ml

respectively in order to diminish the-ribonuclease degra-

dation of mitochondrial ribosomes. Mitochondria were

lysed by the addition of Triton X-100 to 2% and DOC to

0.5%. This lysate was centrifuged for 10 min at 60,000

x g in a Spinco type 30 rotor to remove membrane fragments.

Crude mitochondrial ribosomes were obtained from this super-

natant by centrifuging through a 2 ml layer of standard

T buffer containing 24% sucrose for 3 hours at 230,000 x g

in a Spinco type 65 rotor.

The pellets were rinsed gently with T buffer and

finally resuspended in a small volume of the same buffer

containing 50 ug/ml heparin and 550 pg/ml puromycin. The

mixture was incubated for 5 min at 370 C and returned to

the ice bath promptly after incubation. The ribosome solu-

tion was slightly turbid after this treatment and was

clarified by centrifugation at 14,000 x g for 10 min in a

Spinco type 65 rotor. The volume of the resulting super-

natant was adjusted to the appropriate level with T buffer

and layered directly onto 10 to 30% sucrose density gradi-

ents to purify the mitochondrial ribosomes further,

Procedure 2

In a modified version of the above procedure,

three-times-washed mitochondria were treated with 0.5%

digitonin (de Vries et al., 1971) in order to remove the

outer mitochondrial membranes and thus diminish the con-

tamination by cytoplasmic 80S ribosomes. Heparin and

oligonucleotides were omitted from this preparation. Prior

to lysis, mitochondria were resuspended in DVT buffer (10 mM

Mg2C1; 0.1 M KC1; 5 mM 2-mercaptoethanol, 10 mM Tris-HC1,

pH 7.6; 0.1 mM EDTA) and lysed by the addition of Triton

X-100 to 2%. DOC was omitted. The lysate was treated

with puromycin and mitochondrial ribosomes were obtained as

described for method 1.

Sucrose Density Gradients

Linear 10-30% sucrose density gradients were pre-

pared three at a time with a Beckman density gradient

former. All solutions contained RNase-free sucrose. Gra-

dients were centrifuged for the appropriate time and speed

in the Beckman SW 27 rotor at 40 C. The gradients were

fractionated by pumping a 34% sucrose solution into the

bottom of the centrifuge tube thereby displacing the solu-

tion at the top. The Gilford model 2400 spectrophoto-

meter was used to monitor the gradients at 260 nm as they

percolated through a modified Gilford flow-through cell.

E. coli ribosomes and subribosomal particles, 'analyzed on

separate gradients, were used as sedimentation coefficient



Fractions of 0.5 ml volumes were collected at 20

sec intervals. Contents of tubes from each sedimenting

species were pooled. and concentrated by centrifugation for

3 to 6 hrs at 230,000 x g in the Spinco type 65 rotor.

These ribosomes were used immediately in functional studies.

Preparation of Mitochondrial Subribosomal Particles

Mitochondrial subribosomal particles can be ob-

tained in two ways. "Native" subunits are by-products of

the monosome preparation, obtained directly from the su-

crose density gradients in T buffer. "Derived" subunits

are prepared by dissociating 55S monosomes in suitable

buffers. Both "native" and "derived" particles were some-

times used directly in functional studies. Alternatively,

they were washed with buffers of varying Mg2+/KCl ratios

known to remove defined groups of ribosomal proteins

(David Matthews, personal communication) prior to being

studied. Ribosome cores thus produced can be separated

from the released proteins by sucrose density centrifuga-

tion or by sedimentation through a 4 ml layer of the same

wash buffer containing 10% sucrose. Sucrose density

gradient centrifugation proved to be the more useful and

reproducible method.

The wash buffers used contained 20 mM triethanola-

mine, pH 7.5, 5 mM 2-mercaptoethanol and varying quantities

of MgC12 and KC1,

Buffer S: 10 mM MgC12, 0.1 M KC1;


Buffer BTR: 10 mM MgC12, 0.5 M KC1;

Buffer 0: 5 mM MgC12, 0.3 M KC1;

Buffer Z: 5 mM MgC12, 0.5 M KC1;

Buffer Y: 5 mM MgC12, 1 M KC1;

Buffer A: 1 mM MgC12, 1 M KC1.

Preparation'of Bovine Liver 80S Cytoplasmic Ribosomes

Bovine liver cytoplasmic 80S ribosomes were ob-

tained as by-products of mitochondrial ribosome prepa-

rations when "procedure 1" was employed. They were

subjected to the same preparative manipulations as mito-

chondrial ribosomes and thus served as useful internal


Alternatively, 80S cytoplasmic ribosomes were

prepared directly from the microsomal fraction of the beef

liver post-mitochondrial supernatant. Microsomes were

resuspended in buffer I C20 mM triethanolamine, pH 7.5;

5 mM MgCl2, 0.1 M KC1; and 5 mM 2-mercaptoethanol) and

lysed with 2% Triton X-100 in order to release their ribo-

.somes. Membranous material was removed by centrifugation

at 60,000 x g in a Spinco type 30 rotor for 10 min.

Ribosomes were removed from the resulting super-

natant by centrifuging them through 1/4 volume of the

same buffer containing 34.5% sucrose for 12 hrs at 63,000

x g in a Spinco type 60 rotor. Subsequently, the cyto-

plasmic ribosomes were incubated with puromycin (550

pg/ml) for 5 min at 370 C to free the nascent proteins

from the ribosomes. Following incubation the mixture was

cooled quickly on ice. The sample was layered onto 10-30%

sucrose density gradients prepared as described. Gradi-

ents were centrifuged for the appropriate time in a Beck-

man SW 27 rotor. After fractionation the material

corresponding to the 80S peak was pooled. Ribosomes were

concentrated by centrifugation and used directly in

activity studies.

Isolation of Crude Mitochondrial Factors

Bovine or rat liver mitochondria were resuspended

in two volumes of KTM (40 mM KC1; 20 mM Tris-HCl, pH 7.5;

15 mM MgC12; 6 mM 2-mercaptoethanol) according to the

method of Leister and Dawid (1974). The organelles were

ruptured by sonication in a Bronwill Biosonik 2 Sonicator

using the small probe (4 mm diameter) at maximum intensity

of 150 watts. Four 15 sec bursts were given at 20 sec

intervals. The lysed mixture was centrifuged at 14,000

x g for 10 min in a Spinco type .65 rotor to remove mem-

branous material. The supernatant was adjusted to 0.5 M

KC1 and centrifuged once more at 230,000 x g for 4 hrs in

a Spinco type 65 rotor. The upper 3/4 of the resulting

supernatant was removed and concentrated 20-fold with dry

Sephadex G-25, 3 g/10 ml of supernatant (Leister and

Dawid, 1974). Dry Sephadex G-25 was added to the super-

natant and was allowed to swell. This mixture was placed

within the barrels of 5 ml syringes containing nylon net


and luer lok fittings over their narrow openings. The

syringe assemblies were placed within centrifuge tubes

and centrifugation was carried out at 1/2 maximum velo-

city in an IEC desk top centrifuge for 2 min at 40 C. The

concentrated crude factors were collected in the centrifuge

tubes. Subsequently, the sample was passed through a

column of Sephadex G-25. The void volume, containing the

factors, was collected and used immediately in poly U-

dependent polymerization of phenylalanine assays.

Preparation of E. coli Ribosomes and Crude Factors

The procedure of Nirenberg (1963) was used to

prepare E. coli ribosomes and crude factors. E. coli

K-12, strain Hfr DIO RNase- or strain 1200 F'end A 1100

rnsA Su-, cells were grown to early exponential phase in

nutrient broth at 370 C. After harvesting in the cold,

the cells were resuspended in ice cold standard buffer

(10 mM Tris-HCl, pH 7.8; 60 mM KC1, 14 mM Mg(OAc)2; 6 mM

2-mercaptoethanol). They were ruptured by sonication in

a Bronwill Biosonik Sonicator using the small probe at

maximum intensity. Four 15 sec bursts were given at 20

sec intervals. Cell debris was removed by centrifugation

in a Spinco type 65 rotor for 10 min at 14,000 x g, The

resulting supernatant was centrifuged at 230,000 x g for

2 hr in the Spinco type 65 rotor in order to sediment

all the ribosomes. The upper 2/3 of the new supernatant

was removed and passed through a Sephadex G-50 column.


The void volume, containing the factors, was collected and

stored frozen in small aliquots at -700 C. The ribosome

pellets were washed.either in standard buffer or in a high

salt buffer W (10 mM Tris-HCl, pH 7.5; 10 mM MgC12; 5 mM

2-mercaptoethanol; 1 M NH4C1). Ribosomes prepared in this

manner were resuspended in the buffer of choice and stored

frozen at -700 C.

Preparation of N-acetyl['H]leucyl-tRNA

Crude amino acyl tRNA synthetase was prepared

from E. coli by the method of Nishizuka et al. (1968).

Exponential phase E. coli cells were resuspended into

10-15 ml of cold buffer B (10 mM Tris-HCl, pH 7.4; 10 mM

MgC12; 2 mM 2-mercaptoethanol) and lysed by sonication as

described. Cell debris was removed by centrifugation at

14,000 x g for 10 min in the Spinco type 65 rotor. The

top third of the 230,000 x g supernatant obtained by cen-

trifugation for 2 hr in the Spinco type 65 rotor was

dialyzed overnight in 100 volumes of a buffer containing

10 mM Tris-HCl, pH 7.4 and 2 mM 2-mercaptoethanol. Ali-

quots were stored frozen at -700 C in a Revco freezer.

Stripped E. coliB tRNA was charged with [3H]leucine

by the method of Nishizuka et al. (1968). The reaction

mixture contained: 100 mM Tris-HCl, pH 7.2; 25 mM MgCl2;

4 mM ATP; 10 mM glutathione; 10 mg stripped tRNA; 0.80 mg

protein crude synthetase; 0.5 mCi[4,5-3H]leucine, 55

Ci/mmole in a volume of 10 ml. The reaction was started


by the addition of synthetase. Incubation was for 30 min

at 370 C. The reaction was stopped by the addition of an

equal volume of glass distilled phenol. The phenol ex-

traction was performed to remove all protein and thereby

purify the leucyl-tRNA. The tRNA fraction was precipi-

tated from the aqueous phase, adjusted to 2% KOAc, with

two volumes of cold ethanol. It was washed 2 times with

cold 70% ethanol, and once with ether to remove all traces

of phenol. After drying at 0 C under vacuum, the sample

was resuspended into a small volume of water.

[3H]leucyl-tRNA was acetylated by the method of

Haenni and Chapeville (1966). [3H]leucyl-tRNA was dis-

solved in a small volume of 0.2 sodium acetate buffer,

pH 5.0. Three 40 il aliquots of acetic anhydride were

added at 40 min intervals to the mixture maintained at

0 C. The product, acetyl-[3H]leucyl-tRNA, was precipi-

tated with cold ethanol, washed several times and dried

under vacuum. It was then resuspended in a small volume

of H20 and stored frozen at -700 C. This method of acety-

lation is 100% efficient (Haenni and Chapeville, 1966).

Peptidyl Transferase Assay

The peptidyl transferase activity of ribosomes

was measured by the "fragment reaction" of Monro (1971)

with a few modifications. N-Acetyl[3H]leucyl-tRNA (de Vries

et al., 1971) was reacted instead of a leucyl-CCA terminal

fragment obtained by treatment of the molecule with RNase Ti


(Monro et al., 1968), The reaction mixture contained:

36 mM Tris-HCl, pH 7.5; 267 mM KC1; 36 mM Mg(OAc)2; 0,66

mM puromycin; 1 A260 units of ribosomes; 83 nM Ac[3H]leu-

tRNA (6,000 to 10,000 cpm) and 33% v/v ethanol in a volume

of 0.15 ml. The reactions were incubated at 250 C and

were terminated by the addition of KOH to a concentration

of 0.6 N. The mixtures were then incubated at 400 C for

3 min to break down any [3H]leucyl ethyl ester by-products,

and then were neutralized by the addition of 1 ml 1 M

sodium phosphate, pH 7.0 (Miskin et al., 1970). The

N-acetyl[3H]leucyl-puromycin product was extracted into

1.5 ml ethyl acetate by shaking the reaction tubes for

10 sec on a vortex mixer and centrifuging in a desk top

centrifuge at low speed for 1 min to separate the phases,

1 ml of the ethyl acetate extract was mixed with 10 ml

triton-toluene-permablend (50% triton, 0.5% PPO and 0.05%

POPOP) liquid scintillation mix and counted as above.

Counting efficiencies were 25%.

Poly U-Dependent Polymerization of [3H]Phenylalanine

Ribosomal activities were determined using the

poly U-dependent protein synthesizing system of Hosokawa

et al. (1966). The reaction mixture contained: 10 mM

Tris-HCl, pH 7.8; 50 mM KC1; 20 mM Mg(OAc)2; 6 mM

2-mercaptoethanol; 1 mM DTT; 0.032 mM GTP; 5 mMI potassium

phosphoenol pyruvate; 1 mM ATP; 0.64 mg/ml poly U; 25 mM

tyrosine; 50 mM each of 18 amino acids in a mixture


excluding tyrosine and phenylalanine; 5.4 pM [3H]phenyl-

alanine, 1.82 Ci/mmole; 0.5 mg/ml tRNA; 0.1 mg/ml pyruvate

kinase lyophilizedd or as an (NH4)2S04 suspension); 0.1 to

1.0 mg/ml crude factors; 0.5 to 2.0 A260 units of ribo-

somes in a total volume of 0.25 ml. The reaction was

started by the addition of factors, except when ribosomes

were preincubated in the above reaction mixture in the ab-

sence of poly U for 5 min at 300 C to allow for termination

of protein synthesis on any endogenous mRNA. In these

latter cases, the reaction was started by the addition of

poly U. Reaction tubes were incubated at either 300 or 370

C. At specified times 50 pl aliquots were placed onto

Whatman 3 MM filter papers and processed by the method of

Mans and Novelli (1960). [3H]phenylalanine incorporated

into hot acid-insoluble protein was measured by counting

the filter papers in toluene-permablend (0.5% PPO and 0.05%

POPOP) liquid scintillation mix in a Beckman LS 330 liquid

scintillation spectrometer. The samples were counted

for 10 min each at an efficiency of 4%.

Binding of ['H]GTP to Ribosomes

The millipore filtration technique of Bodley et

al. (1970) was used to measure the binding.of ['H]GTP to

Sribosomes. The reaction mixture contained: 10 mM Tris-

HCl, pH 7.4; 10 mM NH4C1; 20 mM\ Mg(OAc)2; 5 mM 2-mercapto-

ethanol; 5 pmoles of ribosomes:* 42 pmoles [8.-3H]GTP

* See Appendix A for calculations.


(0.5 pCi) in a volume of 50 pl. The reactions were

started by the addition of [3H]GTP. Following incubation

for 5 min at 00 C, the reactions were terminated by the

addition of 3 ml cold buffer containing 10 mM Tris-HCl,

pH 7.4; 10 mM Mg(OAc)2; 10 mM NH4Cl. This mixture was

immediately filtered through a millipore filter and

washed 3 times with 3 ml aliquots of the same buffer.

This washing procedure was done very quickly as emphasized

by Bodley et al. (1970) to minimize disaggregation of the

metastable complex. The filters were then air dried,

overlaid with toluene-permablend liquid scintillation mix

and counted as above. The efficiency of counting was 16%.

Efficiency of Counting

To correct for self-absorption and quenching mock

reaction mixtures for each assay condition were prepared

and counted directly in the usual scintillation cocktail

mix. For the poly U directed assay, this involved pipet-

ting 50 pl samples containing 5.49x106 dpm of [3H] phenyl-

.alanine, 1.85 Ci/mmole directly onto Whatman 3 MM filter

paper discs. Increasing quantities of E. coli and mito-

chondrial factors (0-100 pl) at concentrations normally

used in the reactions were layered on top. For the

[3H]GTP binding reaction, 10 pl of [3H]GTP containing
l.1x166 dpm were placed on a millipore filter and overlaid

with 5 to 10 pl of E. coli ribosomes whose concentrations

equaled those used in the reactions. The filters were


air dried before counting in toluene-permablend (0.5% PPO

and 0.05% POPOP).

And, finally, the [3H] efficiency for the "frag-

ment reaction" was determined by adding 10 pl (8.804 mg)

of a standard [3H] toluene solution containing 2.6x106

dpm/g directly to 10 ml triton-toluene-permablend (50%

triton, 0.5% PPO and 0.05% POPOP). One ml of ethyl

acetate was added as well to control for quenching due

to this solvent and the sample was counted.

Counting efficiencies for each assay were deter-

mined by dividing the cpm experimentally obtained by the

dpm added and multiplying by 100.


Preparation of Mitochondrial Ribosomes

Mitochondria are studded with cytoplasmic ribosomes

in vivo. To eliminate a large portion of the cytoplasmic

contaminants, the mitochondria were washed several times in

isolation medium. Two procedures were used to prepare

mitochondrial ribosomes and subribosomal particles as des-

cribed in "methods." In initial studies (procedure 1) the

mitochondria were not treated with digitonin and therefore

contained intact outer membranes. Lysis of mitochondria

with Triton X-100 released not only mitochondrial ribosomes,

but 80S cytoplasmic ribosomes as well. Mitochondrial 55S

ribosomes were readily separated from 80S cytoplasmic ribo-

somes by centrifugation in sucrose density gradients

(Figure 1). However, there was a small but variable (5-15%)

contamination of 55S ribosomes by 60S cytoplasmic subribo-

somal particles. This contamination was analyzed by cen-

trifuging pooled 55S ribosomes in sucrose density gradients

under dissociating conditions (Figure 2). Under these con-

ditions, the derived mitochondrial subribosomal particles

can be distinguished from the residual 60S cytoplasmic
subribosomal particles also on the basis of characteristic

differences in buoyant density (O'Brien et al., 1974).

c 28S 39S 55S 80S


S1.0 -

10 20 30

Figure 1. Sucrose density gradient centrifugation of crude
ribosomes isolated from bovine liver mitochondria (1.1 g
of protein) by procedure 1 as described under "methods."
The crude ribosomal fraction was layered onto linear 10-30%
sucrose gradients in buffer T (20 mM MgC12, 0.1 M KC1,
Centrifugation was for 5 hr at 27,000 rpm in the Beckman
SW 27 rotor. The shaded areas represent the fractions
which were pooled.

"which were pooled.

E i II- I I I
_ 27S 38S 60S


o0 .5


10 20 30
Figure 2. Sucrose density gradient analysis of isolated
55S ribosomes prepared by sucrose density centrifugation
(see Figure 1). The 55S ribosomes were dissociated in
buffer Z (5 mM MgC12, 0.5 M KC1, 5 mM 2-mercaptoethanol
and 20 mM triethanolamine, pH 7.5) and layered onto linear
10-30% sucrose density gradients in buffer Z. Centrifu-
gation was for 13.5 hr at 20,000 rpm in the SW 27 rotor.


In later experiments procedure 2 was used to iso-

late mitochondrial ribosomes. As described in "methods,"

mitochondria were washed with digitonin to remove the outer

membranes. This procedure has the additional advantage of

removing lysosomes from the mitochondrial fraction.

Figure 3 illustrates a typical sucrose density

profile obtained for mitochondrial ribosomes by this

method. As can be observed the peak of contaminating 80S

cytoplasmic ribosomes was greatly diminished by this pro-

cedure. Mitochondrial ribosomes prepared by both methods

were equally active in the assays and resulted in identi-

cal inhibition patterns with antibiotics specific for the

peptidyl transferase locus.

Peptidyl Transferase Center of the Mitochondrial
Ribosome: Activity

The peptidyl transferase activity of mitochondrial

ribosomes was examined independently of other ribosomal

activities by a modified "fragment reaction" (de Vries

et al., 1971; Monro, 1971; Miskin et al., 1970) as des-

cribed in "methods." The time course for peptidyl trans-

ferase activity of mitochondrial monosomes and subriboso-

mal particles is compared to that of E. coli 70S ribosomes

in the experiment depicted in Figure 4. Mitochondrial 55S

ribosomes appear to be as active as E. coli 70S ribosomes.

In fact, mitochondrial 55S activity was not significantly

different from that of 70S ribosomes as determined by

c 28S 39S 55S 80S

0 ..

< 0 ___ i ____ i i
10 20 30

Figure 3. Sucrose density gradient centrifugation of crude
ribosomes isolated from bovine liver mitochondria (4.2 g of
protein) by procedure 2 as described under "methods." The
crude ribosomal fraction was layered onto linear 10-30%
sucrose gradients in buffer DVT (10 mM MgCl2, 0.1 M KC1,
5 mM 2-mercaptoethanol, 0.1 mM EDTA and 10 mM Tris-HCl,
pH 7.6). Centrifugation was for 15 hr at 20,000 rpm in the
,Beckman SW 27 rotor. The shaded areas represent the pooled

Figure 4. Time course of the peptidyl transferase reac-
tion. The reaction conditions are described in "methods."
Assay mixtures contained 9,245 cpm of N-acetyl[3H]leucyl-
tRNA and ribosomes as indicated: A---A, 1.32 A260 units
of E. coli 70S ribosomes (30 pmoles); e----, 1.09 A260
units of mitochondrial 55S ribosomes (35 pmoles); o---o,
1.23 A260 units of mitochondrial 39S subribosomal particles
(68 pmoles); or D----, 1.03 A260 units of mitochondrial
28S subribosomal particles (87 pmoles). Each reaction
mixture was incubated at 22 for the prescribed time and
terminated by the addition of KOH to 0.63 N as described
in "methods."

Time (min)
















t-statistic analysis of data from several experiments

(Table 1).

The generally lower activities observed in these

studies for cytoplasmic ribosomes have also been reported

by other investigators (Neth et al., 1970; Thompson and

Moldave, 1974). The specific activity of the 60S particles

is only 10 to 15% that of 55S mitochondrial ribosomes.

Having lower specific activities for cytoplasmic ribosomes

worked to my advantage since early preparations of mito-

chondrial 55S ribosomes were slightly contaminated (5 to

15%) by 60S monosomes. It was calculated that only 2 to

3% of the observed activity for 55S mitochondrial ribosomes

could be attributed to the 60S cytoplasmic ribosomes (see

Appendix A for calculations).

To localize the peptidyl transferase activity on

one of the mitochondrial ribosomal subunits, 28S and 39S

particles were examined in the fragment reaction. Each

subunit was free of any contamination by the other subunit

or monosomes. In this manner we found the activity con-

fined to the large 39S subunit (Table 1 and Figure 4).

The 28S subribosomal particles were completely inactive.

This enzymatic activity is located analogously on the

large subunits of both prokaryotic and eukaryotic ribo-

somes (Thompson and Moldave, 1974; Neth et al., 1970;

Monro, 1971). Generally, native 39S subribosomal parti-

cles were only 20% as active as the 55S monoribosomes. By

comparison, the activity of the E. coli large subunit is



Ribosome Ac[3H]Leu-puromycin SD n
cpm/pmole ribosome

Mit 55S 143 32 6

39S 28 15 11

28S 1 1

Cyt 80S 7 3 6

60S 10 1

E. coli 70S 159 76 21

50S 148 53 3

The reaction conditions are described in "methods." The re-
action vessels were incubated for 10 min. Assay mixtures
contained 9,245 cpm of N-acetyl [3H]leucyl-tRNA and ribosomes
as indicated: 1.0 to 1.6 A260 units (23 to 37 pmoles) E.
coli 50S; 0.5 to 1.1 A260 units (16 to 35 pmoles) mitocnon-
drial 39S; 1.0 A2s0 unit (87 pmoles) mitochondrial 28S;
6 to 14 A260 units (102 to 238 pmoles) cytoplasmic 80S, and
6 AZ,6 units (156 pmoles) cytoplasmic 60S ribosomes. The
activities reported represent the mean of n experiments.
Standard deviations, SD, were calculated for each series.
The comparison of mitochondrial 55S activities with E. coli
70S activities yielded a t-statistic, with 25 degrees of
freedom, equal to 0.48, which is not significantly differ-
ent from zero.

93% of that shown by the monosome, and the cytoplasmic

large subunit appears to be slightly more active than its

monosome (Table 1).. Our observations for the activities

of prokaryotic and eukaryotic cytoplasmic ribosomal and

subribosomal particles confirm published results (Monro,

1967; Neth et al., 1970; Thompson and Moldave, 1974). In

the modified fragment reaction, the entire amino acyl-tRNA

molecule is used rather than a RNase T1 digest resulting

in the terminal N-acetyl-leucyl-ACC fragment. Under these

conditions in the mitochondrial system it is possible that

the 28S subunit is required to lend binding stability to

the complex. This would explain the low activities ob-

tained with the isolated 39S subribosomal particles.

The stability of the peptidyl transferase locus of

mitochondrial ribosomes was investigated by washing "na-

tive" subribosomal particles in buffers of varying ionic

strength and MgCl2 concentrations. Depending on the KC1

and MgC1, concentrations used, discrete groups of ribosomal

proteins can be removed from the 39S particles. The re-

sulting "core" particles are purified by sedimentation

through sucrose density gradients prior to activity de-

terminations in the fragment reaction (Table 2).
Washing 39S particles in T buffer (20 mM trietha-

nolamine, pH 7.6, 20 mM MgCl2, 0.1 M KC1, 5 mM 2-mer-

captoethanol) did not diminish their activity in the

"fragment reaction," thereby indicating that the peptidyl

transferase locus is an integral part of the mitochondrial



Buffer MgC12/KC1 Relative Activity
(mM/M) Mit 39S E. coli 50S

T 20/.1 100

S 10/.1 100

BTR 10/.5 71

0 5/.3 60

Z 5/.5 23 119

Y 5/1 4

A 1/1 0 -

Average of 2 or 3 experiments. The buffers contained
20 mM triethanolamine, pH 7.5, and 5 mM 2-mercaptoethanol
in addition to the MgC12 and KC1 concentrations indicated
in the table. Native 39S particles obtained from sucrose
density gradients in T buffer (see Figure 1) were pooled
and pelleted by centrifugation. They were then resuspen-
ded in the indicated buffers, and centrifuged once more
through sucrose density gradients of the same buffer com-
position. E. coli 50S subunits were prepared by sucrose
density centrifugation of crude ribosomes in S buffer or
Z buffer. The subribosomal particles (55 pmoles 39S and
40 pmoles 50S) were finally resuspended directly in the
reaction mixture for assay of their peptidyl transferase
activity as described in "methods."


ribosome rather than a loosely attached function. This

was an expected result since mitochondrial ribosome struc-

ture is stable under these ionic conditions (O'Brien,


The activity of 39S particles in T buffer was

chosen as the 100% activity standard for mitochondrial

ribosomes in Table 2. As the ratio of Mg2+/ionic strength

of the buffers decreases, the mitochondrial subribosomal

particles gradually lose peptidyl transferase activity.

Buffer A (20 mM triethanolamine, pH 7.5, 5 mM 2-mercapto-'

ethanol, 5 mM MgC12 and 0.5 M KC1) is sufficient to reduce

the peptidyl transferase activity by more than 75%, and

buffer A (20 mM triethanolamine, pH 7.5, 5 mM 2-mercapto-

ethanol, 1 mM MgC12, and 1 M KC1) abolishes this activity.

E. coli 50S ribosomes prepared by sucrose density

gradient centrifugation in buffer S (20 mM triethanolamine,

pH 7.5, 5 mM 2-mercaptoethanol, 10 mM MgCl2, and 0.1 M KC1)

were used as controls. Buffer A has no detrimental effect

on the peptidyl transferase activity of these ribosomes.

It appears that the peptidyl transferase activity of mito-

chondrial ribosomes is easier to disrupt than that of

E. coli ribosomes.

Susceptibility of Mitochondrial Ribosomal Peptidyl
Transferase Activity to Antibiotic Inhibitors

Preliminary observations indicated that protein

synthesis within intact mitochondria was sensitive to

inhibitors of prokaryotic specificity and not to those of


eukaryotic specificity. For this reason several inhibi-

tors which interact specifically with the peptidyl trans-

ferase locus of bacterial ribosomes were selected to

investigate this locus in mitochondrial ribosomes. In

general the peptidyl transferase activity of mitochondrial

ribosomes could be inhibited by the antibiotics tested.

For some of the drugs, iM concentrations were sufficient

to inhibit the reaction by 50% whereas for others mM con-

centrations were required to achieve the same result.
A comparison of the mitochondrial ribosome res-

ponse with that of other ribosomes was essential to

determine whether the observed inhibition of peptidyl

transferase was significant for all the drugs tested.

The susceptibility of the peptidyl transferase activity

of E. coli ribosomes in the presence of each antibiotic

was set as the standard for sensitivity. The response of

bovine liver cytoplasmic 80S ribosomes was the standard

for resistance. The peptidyl transferase locus of mito-

chondrial ribosomes was characterized with respect to the

action of each inhibitor.

The molar concentrations of vernamycin A, PA114A,

and chloramphenicol required to inhibit the peptidyl

transferase reactions of mitochondrial ribosomes and

E. coli ribosomes by 50% are of the same order of magni-

tude (Table 3). But the response of mitochondrial ribo-

somes to lincomycin, celesticetin, carbomycin and tylosin

is significantly different from that of E. coli ribosomes.




Ribosomal Particle

E. coli 70S Mitochondrial Mitochondrial
Antibiotic Washe- Crude 39S 55S

Vernamycin A 7x10"- 1x10-7 3x10-7 6x10-7

PA114A 9x10-7 2x10-7 7x10~7 9x10"7

CAP 4x10- 4x10-5 1x10-4

Lincomycin 10-s 2x104 10-3

Celesticetin 2x10-4 2x10-3

Carbanycin 10-6 5x10- 7x10-4

Tylosin 3x10- 10-2

The molar concentrations required to inhibit the reactions
of E. coli 70S and mitochondrial 39S and 55S ribosomes by
50% were estimated from the antibiotic inhibition profiles
(Figures 5 to 9). Untreated E. coli and mitochondrial
ribosomes served as the controls for 100% activity. The
fragment reaction conditions are as in Table 1.

The mitochondrial ribosome reaction is inhibited by 50%

when 10- to 700-fold higher molar concentrations of these

drugs are used. The difference between the responses

of ribosomes from mitochondria and bacteria becomes more

evident when one examines their inhibition profiles over a

wide range of antibiotic concentrations as described below.

To emphasize the specificity of action of the

inhibitors on the peptidyl transferase activity as mea-

sured by the modified fragment reaction, two antibiotics,

vernamycin B and oleandomycin, that exert an inhibitory

action at sites other than peptidyl transferase, were

used as controls. The lack of inhibition of the peptidyl

transferase reaction by these antibiotics is illustrated

in Table 4. Vernamycin B is known to enhance the binding

of vernamycin A to ribosomes and thus synergistically in-

crease the inhibitory action of vernamycin A (Pestka,

1971). Oleandomycin, a member of the macrolide group, is

known to inhibit translocation but not peptide bond forma-

tion. In fact, high concentrations (1 mM of oleandomycin)

actually stimulate the binding of CACCA-acetyl-leucine to

ribosomes by 25% (Celma et al., 1970). In accord with this

observation, I found that high concentrations of oleando-

mycin also stimulated the peptidyl transferase activity of

bacterial, mitochondrial or cytoplasmic ribosomes by

10 to 20% (Table 4).

In addition, an antibiotic gougerotin, known to

affect the peptidyl transferase activity of both prokaryotic




Antibiotic Concentration 70S 55S 80S

(% of control)

Vernamycin B 0.2 98 112

Oleandomycin 1 120 117 111

Gougerotin 0.01 90 -
0.1 60 96 113
1.0 40 60 60

The peptidyl transferase activity of E. coli 70S (1.5 A260
units), mitochondrial 55S (1 A260 unit), and microsomal
80S (8 A260 units) was assessed by the fragment reaction
as described in "methods." Untreated ribosomes served as
the 100% activity controls.


and eukaryotic ribosomes was used. Mitochondrial ribosomes

were found to be susceptible to the action of gougerotin

as well.


Chloramphenicol specifically inhibits the peptidyl

transferase activity of prokaryotic ribosomes (Pestka,

1971). In this study it is shown that chloramphenicol

also exerts a direct inhibitory action on the peptidyl

transferase activity of mitochondrial ribosomes. A similar

response by rat liver mitochondrial ribosomes has been

shown by de Vries et al. (1971). Bovine liver mitochon-

drial ribosomes are 50% inhibited by 0.1 mM chlorampheni-

col, concentrations which are 3-fold greater than those

required to inhibit bacterial ribosomes to the same extent

(Table 3). The similarity of the responses obtained for

mitochondrial and bacterial ribosomes becomes apparent

when one examines their antibiotic susceptibility profiles

(Figure 5). The profile for mitochondrial 39S particles

closely resembles that for 70S ribosomes. These results

may indicate that the site for chloramphenicol binding is

partially obstructed by the 28S subunit in the 55S

ribosome, thereby giving the appearance of a slightly re-

duced susceptibility.



0 100


0c 5 0

10 105 1010
I Oo IO4 Io 1

Antibiotic concentration (M)

Figure 5. Effect of chloramphenicol on the peptidyl trans-
ferase activity of various ribosomes. Mitochondrial 55S,
---e; mitochondrial 39S, o---o; E. coli 70S, A-- A;
and microsomal 80S, I---*, ribosomes were used. The pep-
tidyl transferase reaction was performed as described in
"methods." The reaction vessels contained 0.5 to 1.5 A260
units of mitochondrial or bacterial ribosomes, or 5 to 10
A260 units of microsomal ribosomes. The vessels were in-
cubated for 10 min at 220 C and terminated by the addition
of KOH to 0.63 N. The activities of ribosomes in the pres-
ence of inhibitors were always compared to the activities
of untreated controls examined concomitantly. The values
recorded are averages of 2 to 4 experiments.


Streptogramin Group: PA114A and Vernamycin A

PA114A and vernamycin A, which are.closely related

in structure, bind directly to the peptidyl transferase

locus of prokaryotic ribosomes (Vazquez, 1966a; Pestka,

1971; Ennis, 1965 and 1971). The peptidyl transferase

locus of bovine 55S mitochondrial ribosomes is susceptible

to the inhibitory action of PA114A and vernamycin A

(Figure 6). The 39S large subribosomal particles are

slightly more sensitive to these drugs than are the mono-

somes. The presence of 28S subunits in 55S monosomes may

obstruct the binding sites for this class of antibiotics.

The 55S inhibition profile resembles the 70S pro-

file more closely than that of 80S ribosomes. Low concen-

trations of PA114A and vernamycin A do not inhibit the

peptidyl transferase activity of mitochondrial ribosomes

to the same extent as that of crude 70S ribosomes. The

sensitivity difference observed at low concentrations

vanishes when the inhibition pattern of 55S ribosomes is

compared to that of washed 70S ribosomes (Figure 7).

When this is done, the inhibition patterns differ at high

antibiotic concentrations. Since 55S mitochondrial ribo-

somes are not washed with 1 M NH4C1 during their prepara-

tion, it is more appropriate to compare them to the crude

bacterial ribosomes as in Figure 6.

Much scatter was observed in the antibiotic

susceptibility profile of 70S ribosomes treated with

PA114A or vernamycin A when different preparations of

100 ----

c 50-




10-7 10-6 10-5 10-4

Antibiotic concentration (M)
Figure 6. Effect of the streptogramin group of antibio-
tics, (A) PA114A, and (B) vernamycin A, on the peptidyl
transferase activity of various ribosomes: e-- mito-
chondrial 55S; o-- o mitochondrial 39S; A-- A crude E.
coli 70S; and microsomal 80S. Conditions are as
in Figure 5.


o O
? B
= 100-

N4 50

10-7 10-6 10-5 I0-4

Antibiotic concentration (M)
Figure 7. The susceptibility of crude and salt washed E.
coli 70S ribosomes to inhibition by antibiotics of the -
streptogramin group, (A) PA114A, and (B) vernamycin A.
Crude 70S E. coli ribosomes, A---A were prepared as des-
cribed in "methods." These ribosomes were washed once,
A---A, or four times, A---A, with a high salt buffer
(1 M NHC1I, 10 mM MgCl 5 mM 2-mercaptoethanol, and 10 mM
STris-HC1, pH 7.5). Conditions are as in Figure 5.

ribosomes were used. It was postulated that the varia-

bility observed could be correlated with the effective-

ness of the 1 M NH4C1 wash on each ribosome preparation.

To test this, crude and washed E. coli ribosomes were

examined. The susceptibility of the crude 70S ribosomes

to the streptogramins was gradually lost as the ribosomal

particles were washed successive times with 1 M NH4C1

(Figure 7). It seems likely that the strong binding of

PAll4A and vernamycin A requires proteins that are dis-

lodged relatively easily from the ribosome. However,

the antibiotic susceptibility profiles of crude and salt-

washed 70S ribosomes treated with other antibiotics, for

example lincomycin, are identical as discussed below.

Apparently binding sites for the other antibiotics are

held more tightly by 70S ribosomes.

It should be stressed that the inhibitory pat-

terns for mitochondrial ribosomes treated with either

PAll4A or vernamycin A were reproducible. The binding

site for the streptogramins may be more integrated into

mitochondrial ribosomes than is the corresponding site in

bacterial ribosomes,

Lincosamine Group: Lincomycin and Celesticetin

The peptidyl transferase center of mitochondrial

ribosomes is sensitive to the lincosamines. Their suscep-

tibility is clearly delineated in the antibiotic inhibition

profile obtained at various concentrations of antibiotics



0 -A
4- 100
o0 0


10-6 10-5 10-4 10-3 10-2

Antibiotic concentration (M)
Figure 8. Effect of the lincosamine group of antibiotics,
(A) lincomycin, and (B) celesticetin, on the peptidyl
transferase activity of ribosomes: *-- *, mitochondrial
55S; o- o, mitochondrial 39S; A-- A, E. coli 70S;
A- A, E. coli 50S; and0----- microsomal 80S. Condi-
tions are as in Figure 5.


(Figure 8). By comparing to the E. coli 70S profile, one

can notice that higher concentrations of antibiotics are

required to inhibit the mitochondrial ribosomal system.

In fact, 1 mM lincomycin and 2 mM celesticetin are re-

quired for 50% inhibition of the mitochondrial ribosomes

in comparison to 0.01 mM lincomycin and 0.2 mM celesti-

cetin needed for the same level of inhibition of E. coli

ribosomes (Table 3).

When tested separately from the 55S monosome, the

39S large subribosomal particle of mammalian mitochondria

appears to be more sensitive to lincomycin inhibition.

The 39S particle requires 0.2 mM lincomycin to achieve 50%

inhibition. This concentration is significantly higher

than that required to inhibit bacterial ribosomes to the

same extent. One can conclude that mitochondrial ribosomes

are less sensitive than typical bacterial ribosomes to the

action of the lincosamines.

The large 50S bacterial subunit is more suscep-

tible to the effect of lincomycin than is the 70S mono-

some. This enhanced sensitivity is similar to the effect

observed for 39S subribosomal particles as described above.

It is possible that the presence of the small subunit in

the monosome makes it more difficult for the lincomycin

to bind to its site on the large subunit.

Macrolide Group: Carbomycin and Tylosin

It is well known that the macrolides exert an in-

hibitory action on protein synthesis by binding close to

the site of peptidyl transferase on prokaryotic ribosomes

(Pestka, 1971). Mammalian mitochondrial ribosomes have

markedly reduced susceptibilities to both carbomycin and

tylosin (Figure 9). Concentrations of 10's M carbomycin

which inhibit the E. coli ribosomes by 90% have no effect

on mitochondrial ribosomes. At this same concentration

tylosin similarly does not affect mitochondrial ribosomes

whereas it inhibits E. coli ribosomes by 50%. Higher con-

centrations of tylosin (300-fold) and carbomycin (700-fold)

are required to inhibit the mitochondrial ribosomal re-

action by 50% when compared to E. coli ribosomes (Table 3).

The effect of the macrolides on the activity of 39S sub-

units parallels the observations made for 55S monosomes.

Clearly, if mitochondrial ribosomes evolved from prokary-

otic ribosomes, they have become less sensitive to these


Controls for the Antibiotic Susceptibility Studies

Optimal Activity of Mitochondrial Ribosomes

Several investigators have proposed that ribosomes

can exist in more than one state and that an equilibrium

between the states exists in vivo (Chuang and Simpson,

1971; Schreier and Noll, 1971). The conformational


I-. O ---


0 100


107 10 6 10-5 1014 IO O10

Antibiotic concentration (M)
Figure 9. Effect of the macrolide group of antibiotics,
(A) carbomycin, and (B) tylosin, on the peptidyl trans-
ferase activity of ribosomes: e-----, mitochondrial 55S;
o-- o, mitochondrial 39S; A- A, E. coli 70S; and M--,
microsomal 80S. Conditions are as in Figure 5.


transitions of ribosomes going from one state to another

may be the basis of their molecular function in vivo

(Nishizuka and Lipmann, 1966; Spirin, 1969). These con-

clusions are supported by studies in which the ionic en-

vironment of ribosomes was varied (Miskin et al., 1970;

Zamir et al., 1971; Zamir et al., 1973; Ginzburg et al.,

1973). When ribosomes are isolated in a medium depleted

of monovalent cations and low in Mg2+, they become in-

active in cell free assays. Activity can be fully restored

by heating the ribosome to 400 C in a medium containing

the appropriate ionic conditions. Buffer conditions that

maintain an intermediate level of activity have been used

to establish this equilibrium in vitro. Under these con-

ditions, 10-6 to 10-5 M concentrations of certain anti-

biotics, including those that affect peptidyl transferase

activity, are capable of shifting the ribosomal equili-

brium in favor of the active state (Miskin and Zamir, 1974).

In view of the diminished susceptibilities ob-

tained to the lincosamines and the macrolides, it was

important to demonstrate that the isolated mitochondrial

ribosomes existed predominantly in an active form. If

this were not the case, the addition of low concentrations

of antibiotics could shift the equilibrium so that a

higher percentage of active ribosomes would appear in the

population and thus an apparent loss of susceptibility to

the drugs would be observed. Mitochondrial ribosomes were

isolated in a buffer containing 10 mM MgCl2 and 100 mM KC1

(see "methods"). I did not expect to find a large per-

centage of inactive mitochondrial ribosomes in the prepara-

tions since Miskin and Zamir (1974) have shown that E.

coli ribosomes are predominantly active when isolated in

buffers containing at least 10 mM Mg2+ and 50 mM KC1.

In order to determine if there were any "inactive"

mitochondrial ribosomes, incubation conditions shown by

Miskin and Zamir (1974) to completely activate E. coli

ribosomes (1 mM Mg2+, 0.1 M NH4C1, 10 mM Tris, pH 7.5)

were used (Table 5). As can be seen 70S control ribosome

activity was only slightly increased by these conditions.

Mitochondrial ribosomes on the other hand were not acti-

vated. The decrease in activity observed is probably due

to nuclease activity operating at the high temperature

(Appendix B and de Vries et al., 1971). The activity of

the non-preincubated mitochondrial ribosomes compares well

to that of E. coli ribosomes; therefore it is unlikely

that a large percentage of these ribosomes are inacti-

vated as they are isolated. Furthermore, the ionic

conditions of the fragment reaction and the temperature

(220 C) at which the reaction vessels are incubated are

sufficient to activate ribosomes previously inactivated

(Miskin and Zamir, 1974).

Absence of Bound Impurities to the Lincomycin
Site of Mitochondrial Ribosomes

The diminished susceptibility of mitochondrial

ribosomes to lincomycin could be due to the binding of a



Ribosome Preincubation Activity (cpm)A260 % of Un-incubated

E. coli 70S none 522 100
70S 400 C, 5 min 574 110

Mit 55S none 546 100
55S 400 C, 5 min 234 43

55S mitochondrial ribosomes and control 70S E. coli ribosomes
were resuspended in a buffer containing 1 mM MgC12, 0.1 M
NH4C1, 10 mM Tris-HCl, pH 7.5. Half of each sample was pre-
incubated at 400 C for 5 min. The buffer was adjusted and
fragment reactions were performed as described in "methods."


contaminant to or near the lincomycin site thereby par-

tially occluding this site. Two control experiments were

performed to examine this possibility.

(1) If a contaminant were bound to mitochondrial

ribosomes it also would be present in the high speed super-

natant obtained during the preparation of these ribosomes.

It was postulated that if the contaminant were nonspecific

it might bind to E. coli ribosomes as well. 70S E. coli

ribosomes were resuspended in the mitochondrial high speed

supernatant in order to scavenge any such contaminants.

After 30 minutes the 70S ribosomes were concentrated by

centrifugation and used directly in the fragment reaction.

The susceptibility of E. coli ribosomes to lincomycin was

not altered by this treatment (Table 6). This experiment

does not rule out specific binding of the contaminant to

mitochondrial ribosomes.

(2) Mitochondrial ribosomes were examined di-

rectly to determine whether a loosely bound contaminant

was responsible for the diminished inhibitory action of

lincomycin. 39S mitochondrial ribosomes were washed with

buffer Z (20 mM triethanolamine, pH 7.5; 5 mM MgC12;

0.5 M KC1; 5 mM 2-mercaptoethanol) and then centrifuged

through a 4 ml layer of buffer Z containing 20% sucrose

in order to separate the washed ribosomes from any dis-

lodged proteins. This treatment is capable of removing

some ribosomal proteins from mitochondrial ribosomes

(David Matthews, unpublished observation). The resulting



Lincomycin Concentration
Particles Treatment 103M 10-4M 10-sM

% of Control
E. coli 70S Mit. Supt. 2 19 57

70S --- 2 16 59

2 A260 units of E. coli ribosomes were resuspended into
6 ml of postribosomal mitochondrial supernatant. The mix-
ture was stirred for 30 min. The 70S ribosomal particles
were concentrated by centrifugation at 230,000 x g for
3 hrs in a Spinco type 65 rotor. They were resuspended
directly into the fragment reaction incubation mixture.
The control 70S ribosomes were the standard 1 M NH4C1
washed ribosomes used in all the inhibition studies. The
reaction conditions are described in "methods."


"core" particles were examined for their susceptibility

to lincomycin (Table 7). A slight increase in suscepti-

bility to this drug is observed for the core particles.

This change is not significant when compared to the anti-

biotic profile of E. coli ribosomes. It appears that the

diminished susceptibility of mitochondrial ribosomes to

lincomycin is not due to a bound contaminant but that this

reflects an intrinsic property of mitochondrial peptidyl


The Effect of Lincomycin on Salt-Washed
E. coli Ribosomes

Crude (buffer S) and salt-washed (buffer Z) E.

coli ribosomes are inhibited to the same extent by several

concentrations of lincomycin (Table 7). An identical

effect was obtained when E. coli ribosomes washed with

1 M NH4C1 were used (Figure 8). It appears that the bind-

ing site for lincomycin is more tightly held by E. coli

ribosomes than that for the streptogramins (Figure 7).

Reproducibility of Results

When this study was initiated low yields (5 to 10

A260 units) of mitochondrial ribosomes were obtained from
each preparation. Antibiotic inhibition profiles were

constructed by averaging 2 to 3 experimental points. The

effect of antibiotics on ribosomes prepared on different

days was reproducible. The maximum error did not exceed

10% of the untreated control for any antibiotic


Ribosomal Wash Lincomycin Concentration
Particles Condition
10-3 10-4 3.3x10-5 10-5 10-6

% of Untreated Control

Mitochondrial 55S DVT 50 85 92 110 -
39S DVT 20 70 90 97
39S Z 60 -

E. coli 70S S 0 10 50 81
70S Z 1 8 46 86
50S S 0 7 45 -
50S Z 0 6 32 -

Mitochondrial 555 and 39S ribosomes were prepared through
sucrose density gradients in DVT buffer (10 mM MgC12, 0.1
M KC1; 10 mM Tris-HC1, pH 7.6, 5 mM 2-mercaptoethanol,
0.1 mM EDTA). 0.8 A260 units of each were examined for
sensitivity to lincomycin. 5 A260 units of 55S ribosomes
were washed and dissociated in Z buffer (5 mM MgC12, 0.5
M KC1, 20 mM triethanolamine, pH 7.5, and 5 mM 2-mercapto-
ethanol) and centrifuged through sucrose density gradients
in that condition. The material sedimenting at 39S was
collected and examined. 1.2 A260 units per reaction mixture
was used. Crude E. coli ribosomes were centrifuged through
sucrose density gradients in S buffer (10 mM MgC12, 0.1 M
KC1, 20 mM triethanolamine, pH 7.6 and 5 mM 2-mercapto-
ethanol) or Z buffer. Material sedimenting at 50S and 70S
in each gradient was collected. 0.6 to 1.2 A260 units of
ribosomes were used. The conditions for the fragment
reaction are described in "methods."


concentration tested and often varied only by 2 to 3%.

Significant differences between responses to the anti-

biotics by the various ribosomes are those which are

greater than 10% of the controls.

The presence of cytoplasmic 60S ribosomes in

preparations of 55S mitochondrial ribosomes isolated by

procedure 1 did not affect the measured activity by more

than 2 to 3%. This contamination did not significantly

alter the inhibition patterns observed. Indeed, all the

antibiotic inhibition patterns for mitochondrial ribo-

somes were confirmed with ribosomes prepared by procedure

2 (see "methods").

The Function of Mitochondrial Ribosomes
in Additional Assays

The functional activity of mitochondrial ribosomes

was examined by two additional assays in order to demon-

strate the integrity of the particles used for the peptidyl

transferase assay. These assays included the poly U-

dependent incorporation of [3H]phenylalanine and the

[3H]GTP binding ability of mitochondrial ribosomes. The

poly U-dependent assay was a convenient and standard

method for showing an intact and functional ribosome and

the GTP binding assay was used to demonstrate that mito-

chondrial ribosomes bind GTP in near stoichiometric

amounts in a manner analogous to E. coli ribosomes.


Poly U-Dependent Incorporation of [3H] Phenylalanine

Mitochondrial ribosomes isolated from bovine liver

are able to polymerize phenylalanine when assayed in stan-

dard poly-U directed cell free systems (O'Brien et al.,

1974). As shown in Table 8, the mitochondrial ribosomes

depend on added soluble factors and poly U for maximal

activity. The lower activities obtained for mitochondrial

ribosomes when compared to bacterial ribosomes; 850 pmoles

phenylalanine/mg RNA in 15 min for rat mitochondrial ribo-

somes (Table 8) compared to 8085 pmoles phenylalanine/mg

RNA in 15 min for E. coli ribosomes may be due to the

nature of the mitochondrial system. Nascent proteins

attached to the isolated 55S mitochondrial ribosomes

(O'Brien, 1971) are probably of a "very sticky," hydro-

phobic nature and may not be easily dislodged from the

ribosomes. This property of the nascent proteins may not

allow for proper ribosome run off and reinitiation of

protein synthesis even though the ribosomes are incubated

for 5 min in complete medium before the addition of poly U.

Mitochondrial factors are adequate to support mito-

chondrial protein synthesis (experiments 1 and 3a), but most

homologous factor preparations work less well than those

prepared from E. coli (experiments 2a and 3a). Nevertheless,

activities were obtained with 0.1 to 0.2 mg protein/ml

mitochondrial factors. Although 0.1 mg protein/ml E. coli

factors were sufficient to saturate control assays using

E. coli ribosomes (see Appendix B), maximal activities with



Concen- [3H]Phe incorporated
Exp. Ribosane Factors tration pmoles/mgRNA in 15 min

1 Cow 55S

2(a) Cow 55S

(b) Cow 28S+39S

3(a) Rat 55S

complete Mitochondria

-Poly U
(100 pg/ml'

E. coli

) "n

complete E. coli
-Poly U -
(100 pg/ml)


E. coli

(b) Rat 28S+39S complete Mitochondria















The results are corrected for the minus ribosome controls
at 15 min determined for each experiment as follows: exp 1,
50 pmoles/mgRNA; exp 2, 130 pmoles/mgRNA; and exp 3, 1S0

Bovine and rat liver mitochondrial ribosomes were prepared
through sucrose density gradients in T buffer as described
in "methods." The quantities of ribosomes used in each
experiment were as follows: exp 1, 0.5 A260 units 55S;
exp 2, 0.5 A260 units 55S, 0.2 A260 units 28S, 0.2 A26Q
units 39S; exp 3, 0.5 A260 units 55S, 0.3 A260 units 28S
and 0.3 A260 units 39S. The reaction conditions are des-
cribed in "methods."


mitochondrial ribosomes were only obtained when 0.8 mg

protein/ml E. coli factors were used. The level of activ-

ity obtained with the mitochondrial factors was always

lower than that obtained with bacterial factors. For

example, in experiment 3a the homologous mitochondrial

system incorporated 297 pmoles phenylalanine/mg RNA in 15

min. The value for the heterologous system was 850 pmoles

(Table 8). The low activity of mitochondrial ribosomes in

the presence of homologous factors appears to be due to

the presence of an inhibitor in these factor preparations,

as described in Appendix B.

The polymerization of phenylalanine by 55S mito-

chondrial ribosomes is dependent on the presence of poly U

in the reaction mixture and is inhibited by the addition of

100 pg/ml puromycin. As seen in experiment 2 (Table 8),

the omission of poly U results in lowering .the incorpora-

tion of phenylalanine by 60%, and puromycin inhibits this

reaction by 73%.

Mitochondrial native subribosomal particles also

are active in the poly U dependent system (experiments 2b

and 3b, Table 8). However, they have high activities

(879 pmoles/mgRNA in 15 min) in the absence of poly U.

This endogenous activity is not due to a non-protein

synthesizing side reaction since it can be inhibited by

100 pg/ml puromycin to 152 pmoles phenylalanine/mgRNA in

15 min. This observation can possibly be explained by

having a natural message either bound to one of the


mitochondrial subunits or simply cosedimenting with

either subunit in the gradient. If so, it is probably

bound to the small subunit in initiation complexes as is

observed for other ribosomal systems. Alternatively,

some of the native subunits. normally found may have been

derived from dissociation of ribosome couples early in

the process of protein synthesis, retaining their mRNA.

Generally, mitochondrial ribosomes obtained from

rat liver were more active in the poly U dependent system

than those obtained from bovine liver. This may be due to

the longer preparation time required for the bovine 55S

ribosomes and because the animals used were much older

than the rats.
In summary, 55S mitochondrial ribosomes are active

in a poly U dependent system and absolutely require the

addition of either homologous mitochondrial factors or

heterologous bacterial factors. The recombined subunits

are active as well.

Function of Mitochondrial Ribosomes in the Binding
of [3H]GTP

As another measure of mitochondrial ribosomal

function, we examined their capacity to bind [3H]GTP.

From studies in other ribosomal systems (Nishizuka and

Lipman, 1966; Bodley et al., 1970), it is known that

isolated ribosomes can bind [3H]GTP only if they contain

bound EF-G, the translocase elongation factor. It is,

therefore, of interest to study this partial reaction of


protein synthesis to elucidate the possible involvement

of a translocase type of elongation factor in the function-

ing of mammalian mitochondrial ribosomes.

As seen in Table 9, both mitochondrial monosomes

(55S), and combined 28S and 39S subribosomal particles,

as normally prepared through sucrose density gradients in

DVT-30 buffer containing 30 mM Mg2+ and 0.07 M KC1, bind

[3H]GTP well. This result suggests that the isolated

mitochondrial ribosomes contain bound EF-G.

Under the conditions of this assay, the 55S

ribosomes bind about 3.6 pmoles of [3H]GTP/5 pmoles* of

ribosomes. This value is better than that obtained with

E. coli ribosomes, 2.4 pmoles/5 pmoles ribosomes. Fusidic

acid (3 mM) is required to stabilize the E. coli ribosome-

EF-G-GTP complex. The conditions of the millipore fil-

tration technique are such that only half the complexes

formed in free solution are reproducibly measured (Bodley

et al., 1970). Thus, a maximum of 2.5 pmoles [3H]GTP

bound/5 pmoles active E. coli ribosomes would be detected

by this method. The mitochondrial ribosome-factor-GTP

complex, on the other hand, is not unstable in the absence

of fusidic acid. [3H]GTP binding reactions performed in

the absence of fusidic acid gave the same results for

mitochondrial ribosomes. Therefore, the millipore fil-

tration technique probably detects all the complexes

* See Appendix A for determination of pmoles.



MF. Mg2+ pmoles [3H]GTP
Ribosomal Particles Wash Buffer M Salt bound per
5 pmoles ribosomes


Monoribosome (55S) DVT-30 30/0.07b 3.58

Subribosomal particles DVT-30 30/0.07b 4.78

Subribosomal particles NCB 20/1C 1.08
NCB 20/lc(reconstituted) 3.60

NCC 20/2c 0.06

E. coli

Monoribosome (70S) SGB 10/0.05C 2.40

W 10/1i 0

aAll values corrected for
Salt = KC1.
CSalt = NH4Cl.

the minus ribosome control (0.07 pmoles

Bovine liver mitochondrial monosomes and native subribosomal particles
were prepared through sucrose density gradients in DVT-30 buffer (10
mM Tris-HCl, pH 7.5, 30 mM HgC2l, 0.07 M KC1, 6 mM 2-mercaptoethanol).
Derived subribosomal particles were obtained by treating 55S monosomes
with the following buffers: NCB, 10 nM Tris-HCl, pH 7.6, 20 mM MgCl2,
1 M NH4C1, 5 mM 2-mercaptoethanol; and NCC, 10 mM Tris-HCl, pH 7.6,
20 MM MgC12, 2 M NH4C1, 5 nmM 2-mercaptoethanol. The derived particles
were centrifuged through sucrose density gradients in these buffers.
6.7 pmoles 55S, 3 pmoles native subribosomal particles, and 4.7
poles derived subribosomal particles were used. 6-pmoles E. coli
ribosomes in SGB buffer (10 mM Tris-HCl, pH 7.5, 10 mM MgCl2, 0.05 M
NH4C1, and 5 dM 2-mercaptoethanol) or washed in W buffer (10 niI Tris-
HC1, pH 7.5, 10 mN MgC12, 1 M NH4C1, and 5 fmM 2-mercaptoethanol were
used as controls. All ribosomes were resuspended directly into the
reaction mixture. Reaction conditions are described in 'methods."


formed in free solution. On this basis, isolated mito-

chondrial ribosomes can bind [3H]GTP in near stoichio-

metric amounts.

The ability of E. coli ribosomes to bind [3H]GTP

can be diminished by repeated washes in buffers containing

1 M NH4C1 (Nishizuka and Lipmann, 1966). This procedure

serves to deplete E. coli ribosomes of their bound EF-G.

To determine whether the [3HJGTP binding activity of mito-

chondrial ribosomes was due to the presence of an analogous

factor which could be stripped from the ribosomes by washes

in high salt buffers, mitochondrial ribosomes were prepared

through sucrose density gradients in these buffers. When

mitochondrial subribosomal particles are prepared in the

presence of NCB buffer, containing 1 M NH4C1, their capa-

city to bind [3H]GTP is reduced (77%) to 1.08 pmoles/5

pmoles of ribosomes (Table 9). This activity can be abol-

ished by washing the subribosomal particles in NCC buffer,

containing 2 M NH4C1. These results suggest that mito-

chondrial ribosomes, like other ribosomes, contain a

bound EF-G type of elongation factor. This factor further-

more begins to be stripped from mitochondrial ribosomes

under conditions which are effective in removing EF-G

from E. coli ribosomes (Table 9 and Nishizuka and Lipmann,

1966; Bodley et al., 1970). Significantly, the capacity to

bind [3H]GTP can be reconstituted by adding the removed

factors to the depleted subribosomal particles followed by

dialysis in 10 volumes of DVT-30 buffer to remove the high

salt (Table 9).


Significance of Results

Peptidyl Transferase Activity of Mitochondrial Ribosomes

Because of the complexity of ribosomes, a

detailed understanding of the structure-function rela-

tionships within these macromolecules is slow in coming

into existence. Several approaches have been used to

elucidate many of the functions of the individual ribo-

somal proteins: (a) mutation of ribosomal proteins,

(b) reconstitution of particles missing one protein,

and (c) interaction of inhibitors with specific proteins.

The fragment reaction is a convenient and

direct way to study, specifically, the peptidyl trans-

ferase activity of mitochondrial ribosomes with a

limited interference from other ribosomal activities.

Mitochondrial ribosomes contain a functional peptidyl

transferase center as is demonstrated by their activity

in the fragment reaction. As with all other ribosomes,

the peptidyl transferase activity is an integral function

of the large subunit. Therefore, at a coarse level of

analysis, the peptidyl transferase locus of mitochondrial

ribosomes is similar to that found on other ribosomes.

In the modified fragment reaction, an entire molecule


of amino acyl-tRNA rather than a fragment produced

by RNase T1 digestion is used. In this case, a small

subunit may be required to lend greater stability

to the binding of the entire amino acyl-tRNA. This

requirement may account for the stimulation of peptidyl

transferase'activity observed in the presence of both


Antibiotic Susceptibility

The fragment reaction can be divided into

three steps, as pointed out by Pestka (1969): (a)

binding of the peptidyl-tRNA to the "P" site of the

peptidyl transferase center; (b) binding of puromycin,

an analogue of amino acyl-tRNA, to the "A" site, and

(c) the formation of a peptide bond between the amino

acid moiety of the peptidyl-tRNA and puromycin. Anti-

biotic probes can affect any one of the steps either

by binding directly to proteins within each site or by

causing conformational changes when bound to a distant

site. The binding of antibiotic inhibitors to ribo-

somal proteins is known to be highly specific (reviewed

in: Pestka, 1971; Pongs et al., 1974). Eventually

a detailed map of protein synthesis will be made possible

by probing the substrate binding sites with antibiotics.

Several early studies demonstrated the suscepti-

bility of mitochondrial protein synthesis to'prokaryotic

specific inhibitors but not to inhibitors acting specif-

ically on eukaryotic ribosomes (Lamb et al., 1968;


Freeman, 1970; Kroon and de Vries, 1971; de Vries

et al., 1971). For this reason, several antibiotics

known to have a strong inhibitory effect on bacterial

protein synthesis were chosen to probe the peptidyl

transferase site of mitochondrial ribosomes.

Structures of representative antibiotics used

in this study appear in Figure 10. They have molecular

weights of 300 to 1100 gm/mole. Their size permits

them to bind specifically to ribosomal proteins with-

out greatly perturbing ribosomal structure. Binding

sites on E. coli ribosomes have been described for

each antibiotic (reviewed in: Vazquez, 1974; Pestka,

1971). In particular, the binding site for chloram-

phenicol has been thoroughly investigated (Nierhaus

and Nierhaus, 1973; Pongs et al., 1973; Dietrich et

al., 1974). The ribosomal binding sites for other

antibiotics have been related to the chloramphenicol

binding site by competition experiments (Vazquez,

1966a; Pestka, 1974; Vazquez et al., 1973; Vazquez

and Monro, 1967). These experiments have been the

basis for speculations about the spatial interrelation-

ship of the peptidyl transferase locus of bacterial

ribosomes. It was hoped that a direct comparison of

the antibiotic susceptibilities of mitochondrial and

bacterial ribosomes would lead to the elucidation of

the mitochondrial ribosomal locus for peptidyl trans-


Figure 10. Chemical structures of inhibitors. The struc-
tures for puromycin, chloramphenicol, streptogramin A and
lincomycin were obtained from BUcher and Sies, 1969. The
schematic structures for carbomycin and tylosin were ob-
tained from Gale et al., 1972.



1 1


D(-) Threo Chloramphenicol


0 ^, }--,

-/ S-CH3

Streptogramin A

amino sugar



arhino sugar


CH \







Our results with isolated E. coli ribosomes

reproduced results reported in the literature (Monro

and Vazquez, 1967; Contreras et al., 1974). Mammalian

mitochondrial ribosomes were inhibited to the same

extent as E. coli ribosomes by moderate, that is 1 to

100 pM, concentrations of chloramphenicol and the

streptogramin group of antibiotics. Significant differ-

ences in their responses to other antibiotics were

observed. Antibiotics of the lincosamine and macrolide

groups were required at concentrations one to two

orders of magnitude higher to inhibit the reaction

by 50% than was necessary for similar inhibition of the

70S ribosomes (Table 3). The inhibition patterns ob-

tained with each of the antibiotic groups will be discussed

and their significance for the understanding of protein

synthesis within mammalian mitochondria delineated in



Chloramphenicol (CAP) acts to prevent the

binding of substrates to the "A" site of the peptidyl

transferase locus of prokaryotic ribosomes (Pestka,

1969; Pestka, 1970). Monoiodoamphenicol, a synthetic

analogue of CAP has been used as an affinity label

to identify protein L16 of the large E. coli subunit

as the CAP binding site (Pongs et al., 1973). Another

approach, that of partial reconstitution of the ribo-


somes, has also been used to identify this binding site

(Nierhaus and Nierhaus, 1973). The specificity of CAP

binding to ribosomes is emphasized by the fact that

only the D-threo configuration will bind (Vazquez,


In this study, it is shown that CAP acts directly

on the mitochondrial ribosomes (Figure 5). Concentrations

of the same order of magnitude required to inhibit

bacterial ribosomes by 50% inhibited mitochondrial

ribosomes to the same extent (Table 3). In this respect

both types of ribosomes seem very similar. As expected,

CAP does not inhibit cytoplasmic 80S ribosomes (Figure

5 and de Vries et al., 1971). This indicates a high

degree of homology between mammalian mitochondrial

ribosomes and bacterial ribosomes.

The CAP binding site of bacterial ribosomes

appears to be highly conserved since no chloramphenicol

resistant mutants with a lesion at the ribosomal level

have been isolated (Benveniste and Davies, 1973; O'Brien

and Matthews, 1975). However, mitochondrial ribosomes

resistant to chloramphenicol have been isolated from

yeast (Grivell et al., 1973) and HeLa cells (Spolsky

and Eisenstadt, 1972). These mutants will be important

for studies elucidating the nuclear-mitochondrial

interactions occurring in mitochondrial biogenesis.


The susceptibility of mitochondrial protein

synthesis to CAP has been known since 1959 (Rendi,

1959). Mitochondria of all eukaryotes are similarly

affected by this drug. In past years, CAP has been the

tool used to discriminate protein products of mito-

chondrial origin from those of cytoplasmic origin

(Clark-Walker and Linnane, 1966). In particular several

subunits of the cytochrome oxidase (Mason and Schatz,

1973; Clark-Walker and Linnane, 1966; Schatz and Mason,

1974; de Vries and Kroon, 1970) and several subunits

of the oligomycin-sensitive ATPase (Tzagoloff and

Meagher, 1972) have been identified as mitochondrial

products. Studies with intact mitochondria show no

membrane barrier to the penetration of the drug.

The possibility that CAP had a direct effect

on oxidative phosphorylation in addition to its effect

on protein synthesis was raised by several investigators.

In these studies, the administration of CAP to intact

mitochondria resulted in a dysfunction of some of the

cytochrome systems. The site of action of CAP within

mitochondria was not resolved until recently (de Vries

et al., 1971). Nijhof and Kroon (1974) argued that

the effect observed on energy production was indirectly

caused by a primary effect on mitochondrial protein

synthesis. Since mitochondrial products are'necessary

for the operation and integration of the electron


transport system within the mitochondria, this argument

can explain the toxic action of chloramphenicol on

tissues that are rapidly regenerating. Daughter cells

in these tissues would contain functionally deficient

mitochondria (Kroon et al., 1973; O'Brien and Matthews,

1975). For example, elevated concentrations of CAP

cause reversible bone-marrow depression in man.

PA114A and Vernamycin A

PA114A and vernamycin A, which are closely related

in structure, are potent inhibitors of protein synthesis

on prokaryotic ribosomes (Vazquez, 1974). They have

been shown to interact directly with the 50S subunit

of prokaryotic ribosomes by inhibiting the binding

of the amino acyl end of amino acyl-tRNA to the "A"

site (Ennis, 1966; Mao and Putterman, 1968; Vazquez,

1966b). Only one binding site per 50S subunit of the

ribosome has been detected for vernamycin A (Ennis,

1971). Optimal binding of vernamycin A (Ennis, 1971)

occurs at concentration of K+ and Mg2+ found in the

fragment reaction.

Mitochondrial ribosomes are quite susceptible

to the inhibitory action of antibiotics from the strepto-

gramin A group (Figure 6). The concentration required

to inhibit the peptidyl transferase reaction by 50%

is of the same order of magnitude for both bacterial

and mitochondrial ribosomes.


The inhibition profile of mitochondrial ribo-

somes treated with these antibiotics is sigmoidal in

shape. This may indicate a conformational change in

the ribosome caused by binding of these antibiotics.

Miskin et al. (1974) have observed conformational

changes in E. coli ribosomes when these were treated

with antibiotics specific for peptidyl transferase.

Conformational changes also have been observed when

vernamycin B interacts with these ribosomes (Ennis,


Competition experiments were utilized to relate

the binding site of the streptogramins to that of CAP.

Although vernamycin A can inhibit CAP binding to ribo-

somes (Vazquez, 1966a), the reverse was not observed

(Ennis, 1971). The ribosomal protein to which the

streptogramins bind has not yet been identified. This

protein probably is not tightly bound to the ribosome

since it is easily removed by 1 M NH4C1 (Figure 7).

The binding sites for CAP and the streptogramins are

probably distinct but overlapping in bacterial ribosomes.

Since the inhibition profiles obtained with both CAP

and the streptogramins for mitochondrial ribosomes

are similar to those obtained for bacterial ribosomes,

it is likely that the CAP and streptogramin binding

sites in mitochondrial ribosomes overlap as well.


Lincomycin and Celesticetin

Antibiotics of the lincosamine group inhibit

substrate binding to the "A" and "P" sites of the

peptidyl transferase center of prokaryotic type ribo-

somes (Chang et al., 1966; Pestka, 1971). Studies

by Vazquez et al. (1973) have indicated only one binding

site for lincomycin on bacterial ribosomes. On the basis

of competition experiments, the lincomycin binding site

is closely related to, but distinct from, the binding

sites for CAP and the macrolides.

Although this group of antibiotics can inhibit

the peptidyl transferase activity of mammalian mitochon-

drial ribosomes, very high concentrations are required to

inhibit the reaction by 50% as compared to the bacterial

system (Table 3). Lincomycin resistance is also observed

in poly U directed systems (Ibrahim et al., 1974). The

abnormally high concentrations of the drug required

to inhibit the reaction indicate that the binding site

of mitochondrial ribosomes is not similar to that of

bacterial ribosomes. Concentrations which inhibit

mitochondrial peptidyl transferase do not inhibit the

activity of 80S microsomal ribosomes. Therefore, the

inhibitory effect is specific and may reflect a loss

of affinity of the ribosomal protein for lincomycin.

The possibiltiy of phylogenetic differences

among mitochondrial ribosomes was raised by Linnane


and coworkers (Firkin and Linnane, 1969; Towers et

al., 1972), who showed marked differences in the concen-

trations of several antibiotics, including lincomycin,

required to inhibit protein synthesis by isolated intact

mitochondria. Kroon and de Vries (1971), however,

attributed these observed differences to the impermea-

bility of mammalian mitochondrial membranes. When

the mitochondria were osmotically shocked, protein

synthesis was inhibited by these drugs. Linnane,

repeating his experiments with osmotically shocked

mitochondria, maintained his argument (Towers et al.,

1973). The similarity of their data is not obvious

until one examines the results of both groups relative

to the lincomycin inhibition profile obtained in this

study (Figure 11; Denslow and O'Brien, 1974). Inhibition

data obtained by Grivell et al. (1971b) for intact

yeast mitochondria are plotted for comparison. One

can conclude that despite suggestions of a membrane

permeability barrier to lincomycin, the isolated mito-

chondrial ribosomes are partially resistant to the

drug. The comparison illustrated in Figure 11 stresses

the importance of using isolated ribosomes in studies

of antibiotic susceptibilities. These results are

significant in understanding the effect of lincomycin

on mammalian cells in vivo. It has been suggested

that the impermeability of mitochondrial membranes to


I I -- I I
o0100 _-________

50- 1t oa

0 0

i0-6 i0-5 0-4 0-3 10-2

Antibiotic concentration (M)
Figure 11. Effect of lincomycin on peptidyl transferase
activity of ribosomes: a comparison of my results with
relevant data from other laboratories. *--- 55S mito-
chondrial; A--- A, 70S E. coli, and.----, 80S cytoplasmic
ribosomes were examined-as described in Figure 5. 0 in-
hibition of protein synthesis in intact yeast mitochondria
(Grivell et al., 1971). A inhibition of protein synthesis
in osmotically- shocked rat liver mitochondria according to
Kroon and de Vries (1970). 0 inhibition of protein syn-
thesis in osmotically shocked rat liver mitochondria accord-
ing to Towers et al. (1973).

lincomycin (Kroon and de Vries, 1971; Kroon et al.,

1973) is responsible for their resistance to this anti-

biotic. Results in. this study indicate that this may

not be the case since very high levels of lincomycin are

needed to inhibit the isolated mitochondrial ribosomes.

Carbomycin and Tylosin

The macrolide antibiotics contain large lactone
rings of 12 to 22 atoms and are thus structurally related

(Pestka, 1971). It is thought that all macrolides

bind to one major site on the ribosome but that the

interactions with this site vary according to the size

of the lactone ring (Pestka, 1971; Mao and Robishaw,

1971). For example, the larger macrolides such as

carbomycin and tylosin inhibit peptide bond formation,

whereas the smaller macrolides such as erythromycin do

not. Even though erythromycin cannot inhihit peptidyl

transferase activity, it can still interfere with the

binding of CAP to ribosomes, thereby indicating that

it binds to a site close to the CAP binding site.

In this study we have demonstrated that the

peptidyl transferase activity of mammalian mitochondrial

ribosomes shows a greatly diminished susceptibility

to carbomycin and tylosin (Figure 9). High antibiotic

concentrations, 0.7 mM carbomycin and 10 mM tylosin,

which begin to inhibit the microsomal 80S ribosomes,

are required to inhibit bovine liver mitochondrial


ribosomes by 50%. This may represent a nonspecific

inhibition. Similar effects were noted in isolated

rat liver mitochondrial ribosomes as well (de Vries

et al., 1973; Kroon et al., 1974). In experiments

recently published by Ibrahim et al. (1974), isolated

rat liver mitochondrial ribosomes examined in a poly

U dependent system required a 0.1 mM concentration of

carbomycin to inhibit the reaction by 50%. The authors

chose to interpret their results as indicating a high

susceptibility of the isolated mitochondrial ribosomes

to carbomycin. However, if their data are compared

to the antibiotic profile in Figure 9, it is clear that

their mitochondrial ribosomes are not much more susceptible

to carbomycin than those reported here.

These results differ sharply from the high

susceptibility of protein synthesis observed when

intact mammalian mitochondria are treated with carbomycin

(Kroon and de Vries, 1971; de Vries, 1973; Ibrahim

et al., 1974). For example, Ibrahim et al. (1974)

found that a 2 pM concentration of carbomycin was

sufficient to inhibit protein synthesis by 50%. De Vries

et al. (1973) found that intact mitochondria are capable

of concentrating carbomycin 30-fold from the medium.

This, in part, explains the sensitivity observed when

intact mitochondria are used.


The lower sensitivity of mammalian mitochondrial

ribosomes to carbomycin and tylosin may reflect an

altered peptidyl transferase site in the ribosomes.

Lower sensitivities towards other members of the macro-

lide group are found as well. Erythromycin was a poor

inhibitor of CAP binding to rat liver mitochondrial

ribosomes (Ibrahim et al., 1974; de Vries et al.,

1973). It appears that the binding sites for CAP and

the macrolides are not linked in mitochondrial ribosomes

as they are in bacterial ribosomes.

Summary of Antibiotic Studies

Mitochondrial ribosomes contain a CAP binding

site which is located within the peptidyl transferase

locus. They also contain a streptogramin binding site.

These sites are known to overlap in E. coli ribosomes.

They are thought to overlap in mitochondrial ribosomes

as well since their peptidyl transferase activity is

inhibited to the same extent as that of bacterial ribo-

somes when antibiotics from these groups are used.

The lincomycin binding site is distinct from

the CAP site. This is supported by the observation

that mitochondrial ribosomes respond to CAP in a manner

analogous to bacterial ribosomes, but require 100-fold

greater concentrations of lincomycin to achieve 50%

inhibitory values. Only a 10-fold difference in 50%

inhibitory concentration is observed when celesticetin

is the drug employed. This indicates that mitochondrial


ribosomes have a binding site for the lincosamines.

The reduced sensitivity of the peptidyl transferase

activity may be due to one or more of these possibilities:

(a) there is a reduced binding affinity for the lincosa-

mines to their site; (b) the binding site is no longer

in the correct spatial position for effective inhibition

of peptidyl transferase activity; or (c) a conformational

change in the ribosome, which may be required for effec-

tive inhibition, is not occurring. This study does not

discriminate among these possibilities.

The high concentration of the macrolides required

to inhibit the mitochondrial ribosomes probably indicates

a decrease in the binding affinity of the drugs for

their specific sites. This conclusion is supported

by the competition studies of de Vries et al. (1973)

and Ibrahim et al. (1974), who showed that erythromycin,

another macrolide, was not as effective in reversing

CAP binding to rat liver mitochondrial ribosomes as it

is to yeast mitochondrial ribosomes.

While no bacteria are known to have become CAP

resistant through ribosomal mutation, point mutations

at this level have been found to induce resistance

to the macrolides. These mutants often show cross

resistance to lincomycin (Vazquez et al., 1973). These

findings may indicate that the macrolide and lincomycin

binding sites are not as intimately involved in the