n...vu.:‘

 

 

 

 

lorm|l

0-5

3.35:
5'.

. u ‘u.
i‘:!1—=‘§E?!:,'
I..- Iggy);

1 I

up-

n

':
...
1'“
zaamfiuzg;

. fix.
my...“
..

“an...
“n. -

a, “d?“ia .
'"‘",§:§éi§
.; i _

my.»

....

3:
I‘M-t0
4*“
.33..
-
222%
3W

 

0.... .
~ un'
«we...

”1:

- .. £17332:
5...“.
a:

.“zr.

”a:

m 0

gar.“
.

nmaamj

 

IIIIIIII

IIIIIIIIIIIIIIIIIIIIII

This is to certify that the

dissertation entitled

MW? flLWW ? 7,4 /L?(

firtwfl‘m— pw—r( 37:44-

.34 17:3 quzr
Hfiffl/ Z ‘ >1») (/J'J/
52M .13

£0 4&1: fiflZKZ‘ W1?

'57, ’7' (“94- 4/4 WM

ngsééntegfi‘ j’o/ "CZ? WW“?

YM%-7;‘L A/I\(C/0 %(

has been accepted towards fulfillment
of the requirements for

,_PLD

degree in _M TELL wit/xv? l

 

 

 

 

 

LIBRARY
Michigan State
Unlverslty

 

 

PLACE IN RETURN BOX to remove thII checkout from your record.
TO AVOID FINES return on or balm duo duo.

 

DATE DUE DATE DUE DATE DUE

 

‘t"? 9 5 rim:

s‘JQ 0 3 :l‘ N

 

 

 

22’:-

4x411

9-5-3 .1 iv:
a 3 ' I...

 

 

 

" J '. '7' .r '
ILI...‘ L- A : 1’3,

 

an (,- 9 $2“-

 

 

 

 

 

 

 

 

 

 

 

 

IT

 

 

 

 

MSU II An Affirmative Action/Equal Opportunlty Inotltution
Walla-9 I

 

MOLECULAR CHARACTERIZATION OF THE PHAGE EXCLUSION AND THE
LOCAL EFFECT CAUSED BY THE INTERACTION OF AN e14 ENCODED
PROTEIN, LIT, WITH THE gol SITE FROM PHAGE T4 GENE 23.

By

Yuen-Tsu Nicco Yu

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Microbiology

1993

ABSTRACT

MOLECULAR CHARACTERIZATION OF THE PHAGE EXCLUSION AND THE
LOCAL EFFECT CAUSED BY THE INTERACTION OF AN e14 ENCODED
PROTEIN, LIT, WITH THE gol SITE FROM PHAGE T4 GENE 23.

By

Yuen-Tsu Nicco Yu

The Lit protein generated by the cryptic prophage e14 can exclude
bacteriophage T4 at a late stage , particularly when the Lit protein is overproduced
by an up-promoter mutation or expressed from a multi-copy plasmid. This
exclusion is due to cleavage of translation elongation factor Tu caused by the
interaction of the Lit protein with a short sequence, gol, within the T4 major head
protein gene 23. The proteolytic reaction occurs in a region highly conserved
among all the prokaryotic EF-Tus and analogous eukaryotic EF-las. The Lit
protein contains a Zn-dependent protease motif and its proteolytic activity requires
the function of the gol region.

In vitro complementation assays performed by mixing extracts of cells in
which the go] region has been transcribed and translated with extracts of cells
overproducing Lit protein successfully cleave EF-Tu (43 kD)and generate the 37
kD fragment. Moreover, the addition of the purified Gol oligopeptide (29 amino
acids long) to the " Lit " cell extracts also cause the cleavage of EF-Tu in vitro,
indicating that the G01 peptide is the activator of the Lit protease activity.

In addition to blocking translation, the interaction between go] and Lit also
converts the gol region into a transcription terminator. Unlike the translation block,
this transcription termination ( we refer to it as the local inhibition) occurs even if

not all of the cellular EF-Tu is cleaved. Interestingly, in the absence of Lit protein,

the gol sequence actually enhances transcription, suggesting that the normal role of

gol is to be an anti-pausing site that enhances the transcription of gene 23.

To My Parents
and

The One Whose Existence I Believe

ACKNOWLEDGMENT

I would first like to thank Dr. Larry Snyder, my mentor who provided the
financial support and intellectual environment for my graduate work and who also
gave me plenty of patience and encouragement along the way. I am also grateful
for the handy suggestions and thoughtful attentions received from my guidance
committee Drs. Lenny Robbins, Richard C. Schwartz, Wendy Champness and
Michael Bagdasarian.

It has been a great fun to work in Giltner Hall with lots of neat people, and I
wish to thank with particular fondness Dr. Shu-Chih Chen, Dr. Julius Jackson, Sue
Dager, David Aceti, Angel Lake and Natalie Moore.

I would like to thank Greg Velicer, my dear friend, for his friendship and
help to proofreading the manuscript with a great dear of effort.

And lastly, there are my parents, Mr. Yang-Chang Yu and Mrs. Ar-Jue
Chiang-Yu , whom I inherited my genes from and who gave me tons of support

and love.

TABLE OF CONTENTS

List of Tables
List of Figures
Introduction

Chapter 1: Literature Review

Overview of Bacteriophage T4
Molecular Basis For Phage Exclusion
The rex Exclusion
The prr Exclusion
Properties Of Translation Elongation Factor Tu
EF-Tu Encoding Genes And Their Regulation
Function And Structure Of EF-Tu
Transcription Antitermination In E.coli

Bibliography

Chapter 2: Cleavage Of Translation Elongation Factor Tu in Phage
Exclusion

Abstract

Materials And Methods

Results

Discussion
Acknowledgments
References

Chapter 3: Activation Of The e14 Encoded Lit Protease By A 25
Amino Acid Oligopeptide Translated From the T4 gol Site

Abstract
Introduction
Materials And Methods
Results
Discussion
References

vi

Chapter 4: Molecular Basis For The local Inhibition Caused By
gal-Lit Interaction And the Role Of The Phage gol Site

Abstract
Introduction
Materials And Methods
Results
Discussion
References

vii

LIST OF TABLES
Chapter 3
1. Bacterial strains, plamids and phage mutants.

Chapter 4

1. Characterization of M13-gol hybrid phages in Lit-containing cell.

2. A transcriptional gol-Cat fusion clone does not make cells
chloramphenicol resistant in the presence of low amounts of
Lit protein.

3. Overproduction of EF-Tu does not overcome the local inhibition.

viii

LIST OF FIGURES

Chapter 2

1. The "imitation infected cell" used in these experiments.
2. A defect in translation by supematants from induced cells

correlates with the absence of a 43 kD polypeptide.
3. Antibody precipitation of EF-Tu in inhibited and uninhibited cells.
4. Multiplication of T4 in cells with an excess of EF-Tu.
5. A comparison of the sequences of amino acids in EF-Tu

(EF -1a in eukaryotes) from the region of cleavage.

Chapter 3

l. Cleavage of EF-Tu in the inhibited S30 extract.

2. Chloramphenicol-treated cell extract did not cleave EF-Tu.

3. Effect of pH and phosphatase on EF-Tu proteolytic activity.

4. Coomassie Blue stained SDS PAGE assay of in vitro cell extracts.
5. The Zinc-binding protease motif of Lit protein.

6. Coomassie Blue stained SDS PAGE of RN ase-treated cell extracts.
7. Activation of the EF-Tu protease by the 601 polypeptide.

8. The amino acid sequences of the wild type gol region and some mutants.

Chapter 4

l. Plasmid constructs of pUC8PZl and p64PZl-JD.
2. Autoradiogram of the probe protected transcripts.
3. Inhibition of downstream gene transcription by the gol-Lit

interaction.
4. Stimulation of gene transcription by the gol site in the absence

of the Lit protein.
5. EF-Tus get cleaved In the presence of a low, wild-type level of the Lit

protein.

ix

Introduction

The use of host-parasite systems derived from bacteriophage T4 and its host
Escherichia coli has enabled scientists to reveal many essential principles that
form the basis of modern molecular biology. Historic examples include the proofs
that deoxyribonucleic acid is the genetic material ( Hersey and Chase 1952 ), that
ribonucleic acid is the template for proteins ( Brenner et al 1961 ), and that each
triplet is sequentially used as a codon for polypeptide chain synthesis ( Crick et al
1961).

Studies in this field have also brought some surprising and interesting
discoveries that revealed diversity in gene regulatory mechanisms and explored
each classic concept in a deeper and broader manner. For instance, the discovery
of T4 introns led researchers to analyze the self-splicing catalytic activity of an
RNA molecule, and the nature of a mobile intron, which often requires a
endonuclease encoded by the intron itself (Bell-Pedersen and Belfort,
l991)(review see Belfort, 1992). This discovery challenged past beliefs that
prokaryotic genomes do not contain introns and that introns do not code for
proteins. Another example is the characterization of phage T4 Gene 60 which
demonstrated an additional mode of regulation at the level of translation ( Weiss
and Dunn 1990 ). This event is called ribosomal programmed hOpping: ribosomes
can somehow bypass a segment of an open reading frame and continue elongation
at a certain codon downstream, the precise action involving a special context
around the hopped gap. These unusual findings received further recognition as
similar mechanisms were gradually discovered in other genes and organisms

(Ramirez, et al., 1992, Marshalland Lemieux, 1991, Muscarella, 1993).

2

Being one of the largest viruses with a coding capacity of almost 200 genes,
the extent to which bacteriophage T4 depends on the apparatus of its host
Escherichia coli is perhaps surprising. Some phage genes are apparently
dispensable for growth in a laboratory strain, but under circumstances imposed
by host mutants, particular genes of the nonessential group become important.
Thus, although T4 normally relies on the host gene products, phage encoded
proteins can compensate for the host mutations. Phage development also requires
repeated interactions between phage and host to maintain the accuracy of
performance. To determine what host genes are involved, and how they interact
with the phage genome to control the cascade of phage gene expression,
mutations in the host are screened for their ability to restrict phage growth without
affecting the host’s own viability; i.e. host mutants for which no plaques are
formed in the bacterial lawn. Presumably, the products of mutated genes that
interact with phage determinants have either lost their positive regulatory roles or
have negative effects during phage development. The participating phage sites or
gene products can be identified by selecting suppressor mutations in the phage that
can overcome the block and form plaques in the test strain (Cooley, et a1. 1979).

Using the genetic approach described above, Charnpness et a! ( 1982 )
isolated E. coli mutants they called 1i_t, because they cause a late inhibition of 14
gene expression ( later changed to lit(con)). The lit(con) mutations were mapped
to 25', between purB and fadR loci on the E. coli K12 genome, and were later
shown to be in e14, a cryptic DNA element. The dominance of lit(con) mutations
allowed Kao and Snyder to clone the lit allele in a multi-copy plasmid. Nucleotide
sequencing data showed that all 5 independently isolated lit(con) mutants have
the same CG-TA transition in the up-promoter region of the putative lit ORF.
This transition, presumably increasing resemblance to a promoter consensus

sequence, causes a higher transcription level of the gene. Subclones containing the

3

wild type lit gene in a multi-copy plasmid are also phenotypically Lit(con).
Overproduction of RNA transcripts and of the Lit protein were confirmed by
northern hybridization experiments and minicell-labeling assays respectively, the
latter done in cells transformed with appropriate plasmid-home lit constructs. The
1.8 kb EcoRV subclone sequenced contains 2 ORFs with a 4 nt gap in between;
their coding capacity, deduced from the nucleotide sequences, is about 34kD and
17 kD respectively. Results of deletion analysis performed by me indicated that
only the 34 kD protein is essential for the lit phenotype. Two lines of evidence
suggested that gpLit is a membrane ~bound molecule; the majority of Lit protein
labelled in maxicells was found in an inner membrane fraction, and a computer
protein hydropathy analysis of pLit implied 2 possible transmembrane segments
with high hydrophobic density (Kao and Snyder 1988). The connection between
the membrane-association property and the Lit(con) phenotype, which was later
proven to relate to translation elongation factor Tu cleavage, remains a mystery.
The e14 element that contains the lit gene was first discovered by W. C.
Hill as a circular DNA molecule that was found to excise after UV irradiation and
reintegrate back at its original position. It is believed that e14 is a defective
prophage descended from a lambdoid ancestor that gradually lost phage identity
except for a tail-fiber like gene similar to that of Mu, P2 and lambda (Barreiro and
Haggard-Ljungquist, 1991. Sandmeier and Arber, 1992.), and a putative integrase
gene related to that of ¢21 (Schneider and Champbell, personal communication). It
also carries a modified cytosine restriction determinant mcrA (Hiom and
Sedgwick, 1991) and sfiC, a cell-division inhibitor gene (Maguin, 1986). The SOS
independent excision and reintegration is accomplished by an enzyme encoded
by the fragment and the specific a_t_t site in the E.coli genome. When hosts acquired
this cryptic prophage, and why it is maintained are not clear. Cells cured of e14

by UV irradiation produce no Lit protein and have no apparent defect. They can

4

only be distinguished from Lit(con) cells because they allow wild type
bacteriophage T4 to propagate and form plaques (Brody, et al. 1988; Hill, et al.
1989).

The exclusion phenomenon of T4 caused by the overproduction of gpLit is
due to a global inhibition. in T4 true late gene expression and expression of those
early genes that are normally made continuously in late stage; this was shown in
pulse-chase labeling experiments to monitor phage protein synthesis during
multiplication. The inhibition can be suppressed by a T4 phage mutation that
mapped within a ‘40 nt region at the first quarter of T4 major head protein gene 23,
one of the true late genes (Bergsland et al., 1990). This wild type region was
designated as go_l, since mutations in this region allow phage to grow on lit. To
avoid confusion, whenever I mention the Lit(con) phenotype or go] activity, I
mean the phenotypic properties of this global inhibition. Champness and Snyder
(1984) later found that subclones of the wild type gol region fi'om gene 23 are
unable to transform lit(con) cells because expression of this fi'agment is sufficient
to trigger a global effect similar to the abortive phage infection caused by gpLit.
The use of this transformation assay enabled Bergsland et a1 ( 1990 ) to define the
minimum gol sequence necessary for its activity and to scan mutations that can
abolish the go] effect. Kao and Snyder ( 1988 ) took advantage of the same assay
and discovered that the global inhibition caused by gol-Lit interaction acts at the
translation level. Transcription and translation of the gol+ region, a minimal size
of 75m, is alone responsible for this global tanslational effect in Lit-containing
cells, and no other phage genes are involved. Conclusive evidence for this was
obtained from study of a plasmid clone, pUC84PZl, that contains the gol sequence
fused to the lacZ gene of a pUC vector in the Gene23 reading frame. In this
plasmid, Gol activity is controlled by the IPTG-inducible lacZ promoter. All
cellular protein synthesis was shut off soon after IPTG was added to cells. This

5

IPTG inducible inhibition is dramatically diminished if the construct translates
the go] region out of Gene23's reading frame, for example in the -1 flame of
pUC 12PZl. This observation also suggested involvement of the go] polypeptide.
However, we couldn‘t rule out the possibility that it is the act of translation, rather
than the product, that is required. For example, the translation-coupling ribosomes
may follow the transcription apparatus , prevent nascent gol RNA from folding
into an appropriate secondary structure, and consequently provoke translation
shutdown. '

The partial gol RNA sequence has a striking resemblance to lambda nut
sites, which are utilized by the anti-terminator N protein to read through
downstream terminators. This resemblance consists of a box A sequence and a
boxB-like hairpin structure. Interestingly, the mutations that inactivate gol activity
all map in the nut-like site. FIuthermore, a computer simulation program showed
that the entire gol RNA (PvuII- I 75 nt fragment) can form a stable stem-loop
structure, with a AG of -35.5 Kcal, suggesting that this secondary structure may
exist in vivo (Y.-T. N. Yu and L. Snyder). The intriguing interconnection between
possible gol secondary structures and their functions is discussed in one of the
manuscripts included in this dissertation.

In addition to the severe global inhibition of all cellular translation, the
gpLit protein also causes a local inhibition that affects gene expression at the gol
site. This local inhibition occurs in the presence of the normal, low amount of
gpLit which occurs in the wild type ( usually referred to as lit+) and does not kill
the cells. The local effect was characterized by Bergsland and Snyder by using
plasmid constructs carrying a gol-lacZ fusion. In J M1011it0: a strain that lacks the
gene, (It—complementation can be achieved by the plasmid that translates both gol
and lacZ in frame and exhibits blue color in X-gal agar plates ( X-gal is subtract
for B—galactosidase). But in JMlOllit+: which has the same genetic background

6

but a functional lit gene, the transforrnants are white, indicating that no functional
LacZ product is made. A similar effect was also observed in lit(con) cells
transformed with a -1 fi'ame clOne that alters completely the go] polypeptide but
retains the same gol RNA sequence. Based on these results, Bergsland et al
proposed that gol RNA, but not a Go] peptide, is essential for the local inhibition.

The failure of or—complementation has two possible explanations; the
fusion protein carrying the G01 polypeptide at its N-terminus can not function in
the presence of gpLit; or transcription or translation of the hybrid segment is
prematurely terminated (before getting to the reporter gene. In an attempt to
differentiate between them, Rajiv Gulati (Bergsland et. al., 1991) did 81 nuclease
mapping to examine the gol-lacZ RNA transcripts, and I made a gol-CAT hybrid
plasmid that is transcribed from the lacZ promoter and translates gol in the -1
flame and the chloramphenicol acetyltransferase gene from a separate ribosome
binding site. Preliminary results showed that there were fewer intact transcripts
protected in presence of Lit protein, and the transcription fusion clone transformed
Lit+ cells to chloramphenicol sensitive instead of to chloramphenicol resistant as
in lito. From these observations, we hypothesized that the local effect is due to
premature termination of transcription. However, this model is further adjusted in
one of the manuscripts presented here.

My main interest was to characterize the molecular basis of the global and
local inhibitions caused by the gol-Lit interaction. First, I investigated how
translational inhibition happens at the onset of the gol-Lit interaction; what key
factor required for translation is affected in the gol-Lit mediated global inhibition,
and how the defect influences translation activity. Second, for the local inhibition,
the main questions to be answered were: How does premature transcription

termination occur? What is the significance of gol sequence and structure in the

7

above phenomenon? How is the local inhibition related to the global effect caused
by the same factors?

This thesis includes 4 main chapters. Chapter 1 reviews modes of gene
regulation in bacteriophage T4, mechanisms of the known exclusion systems,
current understanding of translation elongation factor Tu which relates to the
molecular basis for the global inhibition, and published material on some systems
of antiterrnination that possess similarity to the local effect caused by the gol-Lit
interaction. Chapter 2, 3 and 4 are written as manuscripts for publication. Chapter
2 and Chapter 3 present results to suggest that cleavage of elongation factor Tu is,
at least in part, responsible for the global translational deficiency in the gol-Lit
mediated abortive infection system and demonstrate that the activity of an EF-Tu
encoded by lit protease can be assessed in in- vitro experiments. Chapter 4
documents the evidence that helps us comprehend the local inhibition. The
experiments presented addressed the involvement of gol RNA structure and the
cleaved EF-Tu product in the premature termination of transcription during local
inhibition, and the possible anti-pausing effect of the go] sequence in the absence
of gplit.

CHAPTER 1

LITERATURE REVIEW

Overview of Bacteriophage T4

As one of the largest coliphages, bacteriophage T4 contains 160 kb of DNA
and close to 200 genes and is characterized by an icosahedral head filled with
linear double stranded DNA and a tail through which DNA is ejected into its host.
The development of this obligate lytic virus is a well orchestrated process,
including alteration of host apparati for its own use, transcriptional and
translational gene regulation, and delicate protein-protein interactions during
virion assembly. One unusual feature of the phage T4 genome is that it contains
hydroxymethylcytosine dCMP instead of cystosine as in host DNA. This genetic
feature is important for phage multiplication as discussed later.

Infection of E.coli by T4 is initiated by probing specific host surface
receptors (LPS and ompC in E. coli K strains). This mission relies on phage T4 tail
fiber distal protein gp37, together with gp38, which comprise a "host-range
cassette" functional unit (Snyder and Wood, 1989). Several events occur during
the first minute after T4's entry into the host. The biosynthesis of host
macromolecules (DNA, RNA, and proteins) quickly ceases. Then, the host RNA
polymerase (RNAP) begins to transcribe phage early genes from promoters with
Similarity to host consensus sequences. The early phage proteins include
nucleases, which are responsible for the degradation of the cytosine containing
host genome, and regulatory proteins such as RegA, which serves as a translational
repressor for several phage early genes at a later stage, and phage encoded specific
replication and transcription components. Within three minutes post-infection, the
alteration and modification of host RNA polymerase core enzyme through the
ADP ribosylation of alpha units of RNAP by gpalt and gpmod phage proteins and
binding of T4 encoded 11.4 kD (prA) occurs. This confers increased affinity of
RNAP for phage specific sigma factor gpSS, which is required for recognition of
the late promoters (Williams, et al. 1987; Malik and Goldfarb, 1988).

10

In T4 infected cells the theme of transcription regulation can be divided into
pre-replicative and post-replicative periods. During the pre-replicative period,
which spans from the time of infection to the onset of DNA replication about 5
minutes later, genes are transcribed from the early and middle promoters. The
early transcription units contain IE genes (proximal to the promoter) and DE genes
(distal to the promoter) messages in long polycistronic molecules separated by
potential terminator sequences. Transcriptional readthrough at terminator sites
within the T4 intercistronic region is required for synthesis of the promoter distal
gene products. It should be noted that the early and middle transcriptional units
overlap to a great extent. Under circumstances of polarity effects, RNA synthesis
in the DE region could be initiated from a middle promoter within the IE-DE
junction. Unlike the early promoters, the T4 middle promoters contain a unique
sequence centered at ~30 and utilize T4 modified RNA polymerase and a particular
transcriptional factor gpmotA. The binding of gpmotA on a T4 middle promoter to
facilitate the RNAP function was also demonstrated in vitro by Hinton (1991).

The post-replicative genes are mainly the bacteriophage true late genes
which encode proteins for virion components. Their promoters display a totally
different sequence without typical TATA boxes or -35 sequences. They are
selectively recognized by RNAP coenzyrne and a late gene sigma factor encoded
by T4 gene 55. Efi'rcient initiation at these promoters also requires enhancement
by three T4 encoded DNA polymerase accessory proteins,(gp44, gp62, and gp45)
bound to distal mobile enhancer sites. The enhancer exhibits the unique
characteristic that it is not sequence specific but a structural break at the non-
transcribed DNA strand. Activation by these DNA replication proteins also needs
the function of an RNA polymerase bound co-factor protein which is encoded by
T4 gene 33. Herendeen and Geiduschek (1991) have indicated that this

transcriptional activating signal is conununicated between its enhancer and a T4

11

late promoter by a DNA-tracking mechanism. It is noted that concurrent DNA
replication is normally required to activate T4 late transcription. However,
transcription of a plasmid home late gene is independent of replication in vivo and
in vitro suggesting an alternative pathway for the activation. The studies from both
Geiduschek's and Zoguffs laboratories suggest the idea that the competence for
late transcription results from negative torsional stress of the region in the late
promoter, which can be generated by host DNA gyrase on a plasmid or by viral
replication.

In addition to the transcriptional mechanism that controls phage gene
expression, some T4 genetic clusters are also subject to regulation at the post-
transcriptional level. The RNA processing, control of translation, and cleavage
modification of the protein complex ultimately determines the degree of biological
activity for individual phage-encoded enzymes and structural proteins. Multiple
enzymatic activities that catalyze the cutting and trimming of precursor RNA
molecules, chain extension and base modification are essential for maturation of
T4 tRNA species. The majority of the enzymes involved are host encoded (RNase
P, RNase 3, RNase PEP, tRNA nucleotidyltranferase, etc.). Those enzymes
perform similar functions in uninfected E. coli.

Ribonucleolytic activity was also found to be involved in the function of
messenger RNA. Cleavage within the Shine-Delgamo sequence of phage T4 motA
and ORF 2 by the product of the T4 regB gene was proposed to regulate
expression of messenger RNA transcripts. Gold and coworkers (Ruckrnan, et al.
1989) have identified additional processing sites within other T4 ribosome binding
regions, including two sites in the polycistronic frd transcript, which encodes
difyhydrofolate reductase. Mutations that impaired the messenger RNA processing
resulted in overproduction of frd protein. Introduction of cloned copies of regB

into chromosomes of the mutant phage can restore messenger RNA processing

12

capacity. All known processing sites of mRNA lie within a similar sequence
suggesting a common processing pathway.

Two types of specific translational control are being extensively studied in
phage T4, repression of translation of several phage early genes by the product of
phage encoded regA and the mechanism of autogenous regulation. Karam and
colleagues (Robert, et al. 1987) have demonstrated that T4 regA protein acts as a
translation repressor to regulate synthesis of a set of phage induced early proteins.
Those proteins include nucleotide synthesis enzymes (gpcd, gpl, gp42, and gp56),
DNA polymerase accessory proteins (gp45, gp44, and gp62), rILA, rIIB and regA
itself. The purified regA protein was shown to bind specifically to target mRNA
near the initiation codon AUG, and therefore to exclude ribosomes from binding
(Unnithan, et al. 1990). Unlike other translational repressor regulated genes, the
regA binding domain of all the unlinked transcriptional clusters did not possess
sequences in common. Further analysis of mRNA binding by filtration and
nuclease protection assays suggested that AUG initiator is necessary but not
sufficient to determine regA recognition. A proper distance between ribosome
binding sites and the AUG codon also serves as an important factor for regA
binding (Liang et al. 1988).

The biosynthesis of DNA polymerase gp43 in bacteriophage T4 is
autogenously regulated at the translational level. gp43 represses its own translation
by binding to its RNA transcript 5' to the AUG codon at a 36-40 at region that
includes the Shine-Delgarno sequence and the putative RNA stem loop structure
(Andrake, et al. 1991). Many mutations that either disrupt the stem or substitute
the base in the loop segment diminished binding of purified gp43 in vitro. Tuerk
and Gold (1990) have shown that gene 43 mutations that derepress gp43 synthesis
do not necessarily affect the rate of replication indicating that the replicative and

autoregulatory functions of gp43 are independent

13

The production of ssDNA binding protein gp32, which is utilized in
replication, repair and recombination, is also subject to a similar mode of
autogenous repression (Krassa, et al. 1991; McPheeters, et al. 1988). In vitro
studies suggest that autoregulation of gp32 occurs by its own product specifically
binding to its mRNA at a highly structured pseudo-knot near the 5' end. Nucleation
of gp32 binding through the pseudo-knot is thought to be crucial for cooperative
binding of gp 32 to a largely unstructured region that overlaps the initiation codon,
therefore blocking formation of translation initiation complex. gp32 contains a
zinc binding subdomain to which homology was found in a variety of retroviruses
and plant viruses. Shamoo and Konigsberg (1991) have demonstrated that gp32
loses its autogenous regulation ability upon removal of the zinc binding motif, but
retains the ability to bind ss DNA. Together with the evidence that deletion of the
pseudo-knot from gp32 mRN A abolished autogenous repression, the authors
proposed that the zinc binding motif of gp32 plays an important role in directing
the binding of gp32 at the RNA pseudo-knot region.

Post-translational control of protein function was studied in depth with the
proteolytic processing of protein precursors in phage T4. Examples are the
maturation of proteins that compose the phage virion. Assembly of the phage
virion occurs on the host inner membrane about 15 minutes post-infection. The
majority of constituents are produced by T4 late genes. The head and tail are
assembled in two separate pathways. The process of assembly was summarized
both in Kao's and Bergsland's theses. However, I would like to emphasize our
current understanding of one of the capsid components, ng3, which is intimately
related to my research.

ngB is the major head protein of phage capsid. Each mature head contains
more than 1000 copies of ng3 and each cell produces about 200 progeny.

Therefore, one can imagine that codon usage in gene 23 has to be optimal to keep

14

up with the demand. In SDS polyacrylanride gels, gp23 electrophoreses at two
different positions, 53 kD and 48 kD. The 53 kD precursor form of the ng3 is
assembled into the shell (the outermost layer of prohead). After the formation of
the prohead, a protease (T4 PPaSe) encoded by gene 21 recognizes a unique
secondary conformation of the prohead precursor as the cleavage signal and
cleaves nears the N terminal end (Hinterrnan, 1992). This process results in
rearrangement of the prohead and expansion of the inner head cavity (Steven, et a1.
1992). The phage genome that is replicated in multi-chromosome concatemers is
then measured and cleaved by an ATP dependent DNA nuclease derived fiom a
variant of ng3 as one headful length in the mature head (Xue and Black, 1990).

It is an interesting notion that ng3 possesses both enzymatic and structural
functions and that the role of the protein is determined by the variant derived by
the proteolytic process. Rao and Black (1985) have observed that the DNA
packaging enzyme (also called capsizyme gp23") results fiom the N terminal
processing found in the mature capsid protein ng3“ as well as truncation at the C
terminal end. Lack of the consensus ng3 cleavage site at the C terminus suggests
that the cleavage is unlikely to be due to the ngl proteinase. The observation that
amino acid substitutions affecting the C terminal cleavage exhibited a new gene 23
mutant phenotype, defective DNA filled heads supports the idea that the cleavage
product of ng3 directly acts in the DNA packaging mechanism.

The temporal interconnection of prophage maturation and DNA packaging
by a single polypeptide, ng3, is an intriguing phenomenon. Many questions arise
such as: How are the alternative proteolytic processes involving two different
proteases determined? Does the location of the gp23 molecule on the head
structure influence the decision? Is it possible that gp 23 near the bottom entrance
of the head may possess a distinct conformation that can be recognized by the C-

15

terminus specific protease? Identification and characterization of the protease

acting on the ng3 C terminus would aid in understanding these puzzles.

Molecular Basis for Phage Exclusion

Extrachromosomal DNA elements, such as plasmids and integrated
prophages, are able to protect their hosts from infecting phages in a number of
ways. Basically, these defense systems can be classified into two major categories:
superinfection exclusions and abortive infections. The former deals with the
phenomenon that includes: conferring inhibition of early gene transcription and
replication in superinfecting phage carrying a homologous immunity region (Lu,
1989); preventing successful injections of viral DNA from the adsorbed phages
into the cytoplasm (Kliem, 1989); and altering surface receptors, thus hindering
phage adsorption. Overall, the consequence of this type of exclusion allows hosts
to survive viral attacks. Less well understood are those systems of abortive
infection that are characterized by an initial event that usually takes place in a
productive infection, followed by the occurrence of cellular dysfunction, that
consequently kills the host and prevents the production of phage progeny and
their spread into the bacterial population.

The employment of plasmid clones carrying genes that can provoke a defect
similar to abortive infection enables researchers to probe the molecular basis for
each system. Reviews of our current understanding of some exclusion systems
have recently been published (Snyder and Kaufmann, 1993; Molineux, 1991).
Here, I would like to concentrate on two notable exclusion systems of
bacteriophage T4, with emphasis on their molecular themes: the exclusion of T4
rII mutants by the rex genes of E. coli lambda lysogens and by the prr gene
product of a cryptic DNA prophage identified in the clinical isolate CT196. Both

mediate physiological defects but in different ways: the former interferes with

16

membrane potential and the latter attacks a specific tRN A, thus impairing
translation.

The rex exclusion: The phenomenon of restricting T4 rII mutants by an E.
coli lambda lysogen, first discovered by Benzer in 1955, was later developed into
a powerful genetic system for mapping mutations, uncovering the nature of the
genetic code, and a variety of other important contributions. Compared to the glory
of its applications, the phenomenon itself did not receive very much attention until
the 1980s (Toothman and Herskowitz, 1980). This abortive infection requires the
synthesis of two lambda prophage gene products, RexA and RexB. The RexA
protein, being on average hydrophilic, is thought to be a cytoplasmic protein,
whereas RexB is extremely hydrophobic and is associated with the inner
membrane (Parma et al. 1992). The structural similarity between RexB and some
ion channels, and the dependence of restriction on the presence of monovalent
cations, led to the hypothesis that: RexB may form an ion channel that selectively
transports monovalent cations (Sekiguchi, 1966). The influx of ions might
depolarize the cytoplasmic membrane, causing loss of internal ATP, termination of
cellular metabolism, and blockage of phage development.

On the basis of the observation of requirement of RexA for the exclusion,
Parma et al have proposed a two component regulatory model where RexA is a
cytoplasmic sensor that detects the status of infecting phage and then passes the
signal to membrane bound RexB, destroying the membrane potential. The ratio of
RexA to RexB appears to be important for the loss of membrane potential. Snyder
and McWilliams (1989) have observed that overproduction of RexA over RexB
resulted in a defect in uninfected cells similar to that seen after rH mutant infection
of lambda lysogens. This effect does not exist in the situations of overproduction
of both RexA and RexB or of RexA alone. Moreover, overproducing RexB over
Rex A, does not abort the lytic growth of rII mutants (Parma 1992). According to

l7

the model, this evidence. suggests that it takes more than 1 molecule of RexA
sensor to activate the function of RexB, to mediate the exclusion pathway, in
response to phage infection.

This severe physiological defect can be prevented by the presence of both
rIIA and rIIB gene products of phage T4. The rIIA and rIIB functions are normally
non-essential, although it has been suggested that they may play a role in DNA
replication (Manoil et. al. , 1977). The relationship between rex and rII is
quantitative; and plasmid clones of rex A and B not only restrict rII mutants, but
also T4 wild type and other unrelated phages as well (Shinedling, et al. 1987).
Mutations that can substantially alleviate exclusion of T4 by rex map in the T4
motA gene that activates transcription of T4 middle gene promoters. The delayed
T4 DNA replication of motA' mutants was thought to promote the escape from
abortive infection, but the detailed mechanism is not clear.

If the two component model proposed by Parrna er 01 holds, then one major
question is how the signal which is apparently manifested in the absence of rII
proteins potentiates the sensor protein RexA to activate the putative ion channel
RexB. Several suggestions have been made. As described above, an excess of
RexA to RexB can provoke a phenotype similar to abortive infection in uninfected
cells, so one hypothesis focuses on the absence of rII protein somehow
unbalancing the ratio of RexA to RexB and consequently leading to abortive
infection. Alternatively, RexA might be potentiated by some sort of modification
to activate RexB in response to the infection by rII mutants.

With the significant progress toward unraveling the underlying
mechanisms of abortive infection, hopefully it will not be long before we reach the
stage of full understanding. Furthermore, the knowledge might lead to interesting

new insights of host-parasite interactions.

18

The prr exclusion: T4 mutants lacking polynucleotide kinase and RNA
ligase can't form plaques on E.coli strains derived from CT196. The restriction
depends on the function of a cryptic DNA element, prr, located at 29 min. on the
bacterial chromosome (Snyder and Kaufmann, et al. 1987). If prr is transduced
into other strains, it will confer the same phenotype. Underlying prr restriction is
the specific manifestation of a T4 induced RNA ribonuclease, which cleaves
preexisting host lysine tRN A 5' to the wobble position of the anticodon
loop(Kaufmann, 1985).. The reaction generates 2':3'-phosphorus and 5'-OH
termini (Amistsur, et al. 1987). The damaged tRNAs can be repaired by T4
polynucleotide kinase and RNA ligase in Wild-type phage but not in pnk- and di-
mutants (Abdul, 1984; Kaufmann, 1985). Translational inhibition by depleting
tRNAbLs accounts for the abortive infection in prr-containing cells.

The RNA ribonuclease has been renamed anticodon nuclease and found to
be encoded by the E. coli prr locus. Nucleotide sequences and mutational analysis
of a cloned prr region indicate that the prr region is composed of four tandem
ORFs designated prrA-D, and that only prrC is responsible for anticondon
nuclease activity. Intriguingly enough, the other ORFs prrA, prrB, and per are
homologous to type-1 restriction-modification genes hde, hst and hst,
respectively (Amistsur, et al. 1992; Linder, et al. 1992). The interaction between
prr and hsd may couple and mutually enhance protection at the DNA and RNA
levels. _

Activity of prrC-encoded nuclease can only be induced upon phage
infection, implying the association Of phage genetic elements with the enzymatic
activity. Suppresser mutations that can inhibit prr restriction were isolated and
mapped in the rII to gene52 of T4 DNA, a region called stp. The suppressors
defined an open reading frame of 29 codons (Chapman, et al. 1988). This small
polypeptide appears to stimulate the prrC-encoded anticodon nuclease. Amistsur

l9

and Kaufinann (1989) have successfully demonstrated cleavage of lysine tRNA in
an in vitro assay by adding synthetic Stp polypeptide to an extract made from prr-
containing cells. However, Stp with a missense mutation will not activate the
cleavage. The result indicates that Stp is a very specific activator for prrC
enzymatic fturction. . .

Studies emerging from a subclone containing only prrC indicated that prrC-
encoded product appears to cleave cellular tRNA without phage infection. To
ensure cellular safety, the expression of prrC definitely has to be tightly regulated
in uninfected cells. The safety policy seems to be endowed by the flanking HSD
sequences, for a clone carrying the whole prr region retains the latency of prrC
function, that can only be activated by the Stp peptide. From these observations,
Kaufmann and coworkers (Levitz, et al. 1990) proposed that the anticodon
nuclease latency might be due to the masking of the core enzyme PrrC by the
flanking HSD elements, and the Prr ribonuclease could be activated indirectly by
the binding of Stp polypeptide, to remove the putative masking agents. Recent
evidence showed that activation of PrrC nuclease also depends on an endogenous
DNA besides Stp, ATP and GTP. Stp function can be substituted by a small, heat-
stable E. coli factor which only exists in non-prrC containing cells. This implies
that PrrC might function in circumstances other than phage infection, and we
might predict these systems may be controlled by different conditions.

The current model of Prr-Stp mediated abortive infection bears some
features similar to our gol-Lit exclusion system: Firstly, they both involve a small
polypeptide that functions as an activator of a particular enzyme. The Go]
polypeptide of about 25 amino acids long activates the EF-Tu protease, Lit.
Secondly, the molecular mechanism in both cases lead to depletion of translational
components; 'EF-Tu in gol-Lit and tRNAlys in Stp-PrrC mediated exclusion. It is

possible that interaction of the host and phage determinants works in a similar

20

fashion in both systems. Hopefully, the forthcoming investigation of one will have

input into understanding the other in detail and vice- versa.

Properties of Translation Elongation Factor Tu

In the course of my research, one of the major discoveries regarding gol-Lit
mediated exclusion is the cleavage of elongation factor Tu. Knowledge of EF-Tu
is therefore important for understanding the physiological basis of the abortive
infection.

EF-Tu encoding genes and their regulation: EF-Tu, as the most abundant
protein in E. coli, is encoded by two nearly identical, unlinked, genes, MA and
tu_fB_(Jacobson and Rosenbusch, 1976; Weiljland, et al. 1992; Young and Zurano,
1981). The only difference in their products is the residue at the C terminus: Gly
and Ser, respectively. The majority of cellular EF-Tu is made from the tqu gene,
which is distal to, but cotranscribed with, two ribosomal proteins (rpsL, rpsG) and
EF-G (fus) as part of the str operon. The str operon is located at 72 min. on the
genetic map (Shibuya, et al. 1979). The expression of tqu does not appear to be
subject to feedback control by 87, the autogenous regulatory protein identified for
the str operon (Zengel, et al. 1984). Moreover, an interesting feature of the tqu
gene is that the translational efficiency of tqu mRNA is significantly higher than
that of other gene transcripts in the same operon. The high molar concentration of
EF-Tu relative to EF-G is attributable to this difference.

The tufB gene, located at.88 min:, is the last gene of the thrU operon that
also contains four tRNA genes and is under the additional contrOl of a vicinal
weak secondary promoter. The thrU operon is particularly interesting since its
transcript functions both as structural RNA and as a messenger RNA. This
arrangement suggests that the polycistronic transcript undergoes extensive

processing to generate the four tRNAs and the tufB message. The synthesis of EF-

21

Tu fiom this region in vitro is autogenously regulated by the intracellular content
of EF-Tu B (van der Meide, et al. 1983). More recently, Nilsson reported that
transcription of the thrU operon is stimulated by a transactivation mechanism
during rapid growth, and. contains a cis-activating sequence, designated UAS
(Upstream Activating Sequence) (Nilsson, 1990; Bosch et al.. 1990). This
stimulatory effect depends on a UAS-binding element with characteristic similarity
to FIS ( E actor for Inversion Stimulation) protein. In early log phase, the cellular
level of FIS is maximal and drops about 70 fold when cells go from exponential
growth to stationary phase. These results indicate that it is the PIS-dependent
transactivating mode that, enables the cell to respond properly and efficiently to
environmental signals. The copurification of FIS with EF-Tu in my work raises the
question whether conununicatiOn between these two factors determines the
expression of tufB. Interestingly, this UAS-FIS regulatory mechanism also works
on other operons including the tRNA(ter) operon and the rRNA (nnB) operon.
For the latter, a regulatory role of EF-Tu is also suggested by my work.

Although there is no functional difference between Tqu and TufB,
inactivation of tqu by virtue of Mu insertion makes the cell sick whereas
inactivation of tum has little effect on cell growth. Moreover, unlike tqu, tufB
expression is under control of growth rate and the stringent response, as are most
rRNA and tRNA operons. Conceivably, these differences are due to specific
transcriptional modulation of their promoter activity. The F IS-dependent
transactivation mechanism may be responsible for this.

Function and structure of EF-Tu: EF-Tu is responsible for bringing
amino-acid charged tRNAs to the ribosome during the peptide elongation cycle,
where interaction of EF-Tu with its targets, tRNAs and ribosome, is promoted by
binding of an allosteric effector GTP. The codon-anticodon interaction triggers
GTP hydrolysis and causes the release of GDP bound EF-Tu from the ribosome.

22

The dissociation allows formation of a new peptide bond between the N-terminus
of the aa-tRNA in the A-site and the C terminus of the peptidyl-tRNA in the P site.
GDP-EF-Tu is in an ”off“ state. To regenerate functional EF-Tu, a GDP/GTP
switch is catalyzed by an exchange factor EF-Ts, and the resulting GTP-EF-Tu
repeats the cycle. "

In addition to elongating peptide chains, EF-Tu is one of the host-encoded
components in RNA phage QB replicase(Blumenthal and Garrnichael 1979). Its
possible role in rRNA synthesis has also been suggested but not yet defined (
Travers et al.,1983; Haseltine, 1972). In fact, the observation that EF-Tu is
methylated in response to starvation conditions (Young and Bemlohr 1991), which
can concurrently provoke stringent control in ribosome production, indirectly
echoes a possible regulatory function in transcription. It has been reported that the
combination of an EF—TuA/B mutant can suppress a number of nonsense and
fi'ameshift mutations, implying a proof-reading ability in EF-Tu (Vijgenboom et
al., 1985).

The single polypeptide of EF-Tu, 393 amino acids long, can fold into three
discrete domains (Arai, et al. 1980). The first 200 residues, composed of a six-
stranded B sheet surrounded by six or helices, is the GTPase domain. This is the
most thoroughly characterized portion of the entire sequences (it is also the first G-
domain crystal structure identified. ) and EF-Tu is the best understood guanine
nucleotide-binding protein. Domain 2 (aa190-aa297) contains a six-stranded
antiparallel B sheet, forming a hydrophobic pocket . Domain 3, located at the C-
terminal region, is a small B barrel ( Jtu'nak, F. 1985). The middle and C-terminal
domains are needed for enhanced affinity that is needed for productive interaction
with the tRNA, EF-Ts, and ribosomal components. Extensive studies of EF-Tu's
its G-domain reveal striking similarity of sequence and three-dimensional structure

to that of the eukaryotic oncogene ras p21 protein (Parmeggiani, 1987; Valencia,

23

et al. 1991). Several point mutations introduced in EF-Tu show the near-identity of
the GTPase mechanism with p21. These structural and frmctional analogies in the
G-domain have also been extended in part to other G—binding proteins, including
adenylate kinase (Chen et al. 1990) and chemotaxis protein CheY (Chen et al.
1990), by comparison of crystal structures.

' The two post-translational modifications known to happen to EF-Tu are
acetylation at the N-terminal serine and methylation at lysine residue 56 (Ames,
1979). Interesting enough, the methylation site is only 4 amino acids upstream of
the gol-Lit initiated EF-Tu cleavage site (residue 59-60) which is identified in my
research (Yu and snyder, 1993). This region also contains two trypsin sensitive
sites (residues 40 and 58) and usually disappears in crystallized EF-Tu. Its
accessibility to a variety of enzymes suggests that the region may be on the
surface. The result of a computer search of all the EF-Tus and the analagous EF-
lors of eukaryotes showed that the 5 amino acids ( R. G. I. T. 1), covering the Lit
and trypsin cleavage sites, at position 5862 of E. Coli EF-Tu, are totally
conserved during evolution. From these observations, we speculate that this
conserved region may play a regulatory role. Perhaps methylated EF-Tu and the
cleaved EF-Tu fragment resulting from different environmental signals (starvation
for methylation, and phage infection for cleavage), have similar effects on the
cellular regulatory networks. With this hope, my research may provide a new
avenue to fully understand the role of this most prominent protein.

Transcription Antitermination in E. coli

g Transcription antiterrnination, which allows readthrough of otherwise
emcient terrninators by a regulatory protein, provides an important mechanism for
regulating gene expression. "Action-at-a-distance" is the hallmark of
antiterminatOrs, such as phage A gene N and Q products (Friedman and Gottesman

1983; Roberts 1988) and, probably, the Tat protein of human immunodeficiency

24

virus HIV (Cullen 1990, 1993; Gunnery, et al. 1990). Unlike A. Q protein which,
by binding to DNA in the -35 to ~10 region, modifies RNA polymerase at an early
elongation pause site (Yamell and Roberts, 1992), the A pN and the HIV Tat
protein probably recognize a particular site on nascent RNA transcripts, designed
as nut for pN and tar for Tat antitermirrator ( Nodwell and Greenblatt, 1991;
Cunnery et al. 1990 ).

The studies emerging from genetic analysis and in vitro reconstitution
assays indicate that, in bacteriophage A, processive transcription antiterrrrination in
early genes requires the direct interaction of N and a cellular cofactor NusA. The
interaction enhances the recruitment of three other host factors NusB, NusG to
subsequently modify RN AP to a terminator-resistant form (Mason and Greenblatt
1991; Linn and Greenblatt, 1992; Roberts 1993). The late-gene antitermination
system in contrast needs only the Q protein, the A qut site and E. coli NusA (Barik
and Das 1990).

The nut site that is targeted by pN has a distinct structure (Lazinski, et al.
1989). It is comprised of characterized boxA sequences and a region of dyad
symmehy (boxB) with an undefined boundary between them. Cloning experiments
by deCrombrugghe and coworkers (1979) demonstrated that the nut site is
functionally separable from the cognate promoter. Furthermore, the nut site allows
pN to suppress multiple tenninators located thousands of base pairs away. The
genetic dissection of nut elements by several mutations and hybrid nut sites
between close relatives of A, such as phages p21 and p22 , indicate that BoxB is
essential and is the locus of specificity, while BoxA promotes the specificity by
facilitating cognate pN binding to BoxB. A remarkable discovery regarding nut
function is that, the same elements can be converted fi'om an antiterminator into a
strong terminator by a phage HK022 protein (Nun, Oberto and Weisberg, 1989).

The similarities between Nun-mediated termination and N-mediated

25

antiterrrrination suggest that the two processes have biochemical steps in common
(Robert, et a1 1987).

Several observations indicate that transcription antitermination also
regulates ribosomal (rm) gene expression. The detailed mechanism is not clear, but
a number of features are shared by the rm and A systems (Suzanne et a1 1984).
First, a version of the boxA and boxB elements, although reversed, is found in the
highly conserved bacterial rm. operons (Moorage 1986). However, unlike A nut
sites, the BoxA sequences are both necessary and sufficient for antitermination.
Secondly, the productive antitermination also requires ribosomal protein S10 and
cellular Nus proteins, such as NusB (Squires et al 1993). No cellular analog of N
protein has been identified thus far. In vitro reconstitution using purified proteins
and templates containing rrn nut-like sites failed to completely read through
downstream tenninators, suggesting the requirement of additional not-yet-
characterized cellular factors. A

The gol site that is essential for cleavage of EF-Tu in Lit-containing cells,
not only looks like the antitermination nut site, but, from my work, may actually
function as one by enhancing transcription through the region in the absence of Lit
protein. In local inhibition, the. gol sequences can then be converted to a
transcriptional terminator presumably by the cleaved form of EF-Tu or by other
un-identified factors. Further investigation of this gol-rnediated effect may provide
new insights toward a full understanding of the molecular nature of transcription

antitermination.

26

Bibiogrophy

Abdul, J. M. and L. Srtrglder. 1984. Genetic and physiological studies of an
Escherichia coli locus at restrrcts polynucleotide ase and RNA ligase-
deficient mutants of bacteriophage T4. J. Vrol. 51:522-529.

Albright L.M., G.A. Kassavetis, and E.P..' Geiduschek. 1988. Bacteriophage T4 late
transcrr tron from plasnud templates rs enhanced by negative supercorlmg. J.
Caterio. 17021279- 289.

Ames, G. F.-L. and K. Nikaido. 1979. J. Biol.Chem 254:9947-50.

Amistsur, M., R Levitz, and G. Kaufinann. 1987 anticodon nuclease,
gozlyrgngrcégtatgde kinase and RNA ligase reprocess the host lysrne tRNA. EMBO

Amistsur, M., I. Morad, and G. Kaufinann. 1989. in vitro reconstitution of
anticodon nuclease from components encoded by phage T4 and Escherichia coli
CTr5X. EMBO 6:2411-15.

Amistsur, M.,I. Morad,D. Chapman-Shimshoni and G. Kaufmann. 1992. Hsd
restriction-modification proteins partake 1n latent anticodon nuclease.
EMBOl 1:3129-34.

Andrake, M., N. Guild, T. HSu, L.Gold, C. tuerk,and J. Karam. 1988. DNA
ggl 94e2ra4sg of bacteriophage T4 is an autogenous translational repressor. PNAS

Andrake, M.,and J. Karam. 1991. Mutational analysis of the mRNA operator for
T4 DNA polymerase. Genetics. 128:202-13.

Arai. K. et al. 1980. Primary structure of elongation factor Tu from Escherichia
coli. PNAS 77:1326-30.

Bell-Pedersen, D., S. M. Quirk, M. Bryk, and M. Belfort. 1991. I-Tevl, The
endonuclease encoded by the mobile td intron, recognizes binding and cleavage
domains on its DNA target. PNAS 88:7719-23.

Benzer, S. 1955. Fine structure of a genetic region in bacteriophage . PNAS
41:344-354.

Bergsland, K. J. et al. 1990. A site in the T4 bacteriophage major head rotein
gene that can promote the mhrbitron of all translation In Escherrchra colr. . Mol.
iol. 213:477-496.

Biebricher C. and R. Luce. 1992. In vitro recombination and terminal elongation
of RNA by QB replrcase. EMBO 11:5129-35.

Blumenthal, T. 1979. RNA replication :Function and structure of QB replicase.
Ann. Rev. Biochem. 48:525-48.

27

Bosch, L., L. Nilssson, E. Vij enboom and H. Verbeek. 1990. FIS-de ndent
trans-activation of tRNA and r A operons of Escherichia coli. Biolc em. et
Biophys. acta. 1050:293-301.

Brenner, S., L.Bamett, F.H.C. Crick and A. Orgel. 1961. The theory of
mutagenesis. J. Mol. Biol. 3:121-124.

Brody, H, A. Greener and CW, Hill. 1985. Exclusion and reinte tion of the
Escherichia coli chromosomal element e14. J. Bacteriol. 161:1112—1 17.

Brody, H and CW. Hill. 1988. Attachment site of the genetic element e14. J.
Bacteriol. 170:2040-44.

Caldentey, J. and E. Kellenber er. 1986. Assembly and disassembly of
Bacteriophage T4 polyheads. J. Mo . Biol. 188239-48. -

Campbell A., S. J. Schneider, B. Song. 1992. Lambdoid phages as elements of
bacterial genomes. Genetica 86:259-67.

Chapman, D. et al. 1988. Nucleotide and deduced amino acid sequence of stp: The
bacteriophage T4 anticodon nuclease gene. J. Mol. Biol. 199:373-77.

Champness, W. and L. Snyder. 1982. The gol site: a cis-acting bacteriophage T4
lie 5 3a9tgrzogegron that can affect expressron of all the T4 late genes. J. Mol. Biol.

Chamgness, W. and L. Snyder. 1984. Bacteriopha e T4 gol site: sequence analysis
and e ects of the srte on plasmid transformation. . Viol. 502555-56 .

Chen, J. M. et al 1990. Conformatio‘ns of the central transforming region (Ile 55-
Met 67) of the p21 grotein and their relationship to activation of the protein. Int. J.
Pept.Protein Res. 3 2247-54.

Chen, J.M. et _al. 1990. Com arisonsbetween the tlrree-dirnensional structures of
the chemotactrc protein Che and the normal Gly 12-p21 protein. Int. J. Pept.
Protein Res. 36: 1-6.

Crick, F. H. C., L. Barnett, S. Brenner and R. J. Watts-Tobin. 1961. General
nature of the genetic code for protein. Nature (London). 192: 1227-1232.

Cullen, B. 1990. The HIV-1 Tat protein: An RNA sequence-specific processivity
factor? Cell 63:655-57.

Cullen, B. 1993. Does HIV-1 Tat induce a change in viral initiation rights? Cell
73:417-420.

Cunnery,S. et al. 1990. Tat-responsive region RNA of human immunodeficiency
vrrus 1 can revent activation of the double-stranded RNA -activated protein
kinase. PNA 87: 8687-8691.

de Crombru e, B. et al. 1979. Specificity of the bacterio hage lambda N gene
product : Nut sequences are necessary and sufficrent or antrterrnrnatron by
pN. Cell 8:1145-115 .

28

Duckworth, D.- H., J. Glenn and D. J. Maccoquodale. 1981. Inhibition of
bacteriophage rephcation by extrachromosomal genetic element. Mrcrobiol.
Review. 45:52-71.

Friedman, D. I. and GOttesman. 1983. In Lambda II. RW. Hendrix, J. W. roberts,
F.W. Stalrl. and R. Weisberd. CSH Press. pp21-51.

glipcsrding, L. R 1992. Virus proteins that counteract host immune defenses. Cell

Haseltine, WA. 1972. In vitro transcription of Escherichis coli ribosomal RNA
genes. Nature (London). 2352329. '

Herendeen D.R. et al. 1989. Enhancement of bacteriophage T4 late transcription
by components of the T4 DNA replication apparatus. Scrence 245:952-958.

Herendeen, DR. et al. 1990. An RNA polyrnerase-bindin protein that is re uired
for commumcation between an enhancer and a promoter. crence 248:573-57 .

Herendeen, D.R., G.A. Kassavetis, and ER Geiduschek. 1991. A transcriptional
enhancer whose function imposes a requirement that proteins track along DNA.
Science 256: 1298-1303.

Hersey, A. D. and M. Chase. 1952. Independent functions of viral proteins and
nuclerc acrd rn growth of bacteriophage. J. Gen. Physiol. 36:39-56.

Hersey, J. W. B. 1989. Protein phosphorylation controls translation rates. J. Biol.
Chem. 264:20823-26.

Hill, C.W., J.A. Grag and H. Brod . 1989. Use of the isocitrate dehydro enase
zgrgrgttaal gene for attachment of e 4 in Escherichia coli K-12. J. Bacterio . 171:

Hintermann, E. and A. Kuhn. 1992. Bacteriophage T4 ene 21 encodes two
proterns essential for phage maturation. Virology. 189:474-8 . '

Hinton, D.M. 1991. Transcription from a bacteriophage T4 middle promoter using
gzmotA protein and phage-modified RNA polymerase. J. Biol. Chem. 266: 18034-

Hiorrr, K. and S.G. Sed 'ck. 1991. Cloning and structural characterization of the
mcrA locus of Escheric 'a coli. J. Bacteriol. 173:7368-73.

Hobson, A.H. et al. 1993. Activation of a bacterial lipase by its chaperone. PNAS
90: 5682-5686.

Hsu, T.‘, and JD. Karam. 1990. Transcri tional mapping of a DNa replication ene
cluster .In bacterro hage T4. Sites or mitiatron, termmatron and A
processrng. J. Biol. hem. 265:5303-16.

Hughes, M.B. et al. 1988. Genetic mapping of the arnino-terrninal domain of
bacteriophage T4 DNA polymerase. Genetics 119:743-749.

29

Ishimoto, L.K., et al. 1988. The structure of three bacteriophage T4 gene required
for tail-tube assembly. Virology. 164: 81-90.

Jacobson G. and J .P. Rosenbusch. 1976. Abundance and membrane association of
elongation factor Tu in E. coli. Nature (London). 261 :23.

Jarvis T.C.,J. W. Ne ort and RH. von-Hippel.l99l A stimulation of the
processivity of the DN potlymerase of bacteriophage T4 by the {01 erase
accessory proteins. The role 0 ATP hydrolysrs. J. 8101. Chem. 266: 18 0.

Johnson, K. B. Condie, D.T. Mooney and A.H.Doermann. 1992. Mutations that
eliminate the requirement for the vertex protern In bactenolphage T4 capsrd
assembly. J. Mol. Biol. 224:601-11.

Jumak, F. 1985. Structure of the GDP domain of EF-Tu and location of the amino
acids homologous to ras oncogene proteins. Science 230:32.

Kao, C. and L. Snyder. 1988. The lit gene product which blocks bacteriophage T4
late gene expression is a membrane protein encoded by a cryptrc DNA element,
e14. . Bacteriol. 170: 2056-2062.

Kao, C. E. Gumbs and L. Snyder. 1987. Cloning and characterization of the
Escherichia coli lit gene, which blocks bacteriophage T4 late gene expressron. J.
Bacteriol. 16921232-1238.

Kaufinann, G. et al. 1985. Host transfer RNA cleavage and reunion in T4-infected
Escherichia coli CTr5X. Nucleic acids Res. 13: 1333-41.

Kaufmann , G. et al. 1986. Phage and host genetic determinants of the specific
lat/Infilccitaloq 1%) fgegréages In bactenophage T4-rnfected Escherichia coli CTr5X. J.
o . ro . 2 - .

Kliem, M. and B. Dreiseikehnann. 1989. The su erimmruri ene sim of
bacteriophage Pl causes superinfection exclusion. Viro ogy 171:3 0- 55.

Krassa,K.B., L.S.Green and L. Gold. 1991. Protein- rotein interaction. with the
acrdrc COOH terminus of the Single-stranded D A brndmg protern of the
bacteriophage T4. PNAS 88:4010-14.

Lazinski, D., E. Grzadzielska and . DAS. 1989. Sequence-specific recognition of
RNA hairpins by bacteriophage antiterrninators requires a conserved argirrine-rich
motif. Cell 572207-18.

Levitz. R. 1990. The o tional E. coli rr locus encoded a latent form of phage T4-
infected anticodon nuc ease. EMBO :1383-89.

Li, Suzanne C., Catherine L. Squires, and Craig Squires. 1.984. Antitermination. of
E. colr rRNA transcription 15 caused by a control regron segment contarmng
lambda nut-lrke sequences. Cell 38:851-860.

 

Lian Y., et al. 1988. Autogenous regulation of the regA gene of bacteriophage
T4: erepressron of translation. Genetics 119:743-49.

3O

Linder, P., et al. 1990. An. anticodon nuclease gene inserted into a hsd region
encoding a type 1 DNA restriction system, Nucl. Acrds Res. 1827170.

L'Italien, J. J. and R A. Laursen. 1979. FEBS lett. 1072359-362.

Linn.T and J -. Greenblatt. 1992. The NusA and NuaG proteins of Escherichia coli
Increase In vrtro readthrough frequence of a transcriptional attentuator preceding
the gene for the beta subumt of RNA polymrase. J. Biol.Chem. 267: 1449-54.

Lu, M.J. and U. Hennin . 1989. The immunitgl (irnm) gene of escherichai coli
bactenophage T4. Joum of Virology 63:3472- .

Mangel, W. F.., W. J. Mcgrath, D. L. Toledo and C. W. Anderson. 1992. Viral
DNA and a vrral peptide can act as cofactors of adenovrrus vrrion proteinase

activity. Nature

Magnin, E., H. Brody , C.W.Hill and R D'Ari. 1986. SOS-associated division
mhrbitron gene sfiC rs part of excrsable element e14 In Escherichia coli. J.
Bacteriol. 1682464-466. '

Malik, S., M. Dirnitroy and A. Goldfarb. 1985. Initiation of transcription by
actor. .

bacterio ha e T4-modi red RNa ol erase inde ndentl of host si a
Mol.Bigl.1885:83-91. p ym 93 Y gm

Malik, .S. and A. Goldfarb. 1988. Late sigma factor of bacteriophage T4.
Sgrnrialqenlalnsdlpropertres of RNA polymerase-promoter complexes. J. Brol. Chem.

 

Mason, S. W. and J. Greenblatt. 1986. Host factor requirements for processive
antrterrnrnatron of transcription and sugpressron of pausrng by the N protern of
bacteriophage lambda. J. Brol. Chem. 2 219418-26. '

Mason, S. W. and J. Greenblatt. 1991. Assembl of transcri tion elongation
complexes contauung the N rotern of pha e lam da and the scherrchra colr
elongatron factors NusA, Nus , NusG, and S 0. Genes. Dev. 5: 1504-12.

Mason, _S. .W., J, Li. and. J .. Greenblatt. .1992. Direct interaction between two
Eschererchra coh transcrr tron antrterrnrnatron factors, NusB and ribosomal
protein S10. J. Mol. Biol. 2 3:55-66.

Marshall P., C. Lemieux. .1991. Cleavage attem of the homing endonuclease
encoded by the fifth rntron In the chloroplast ar e subunit rRNA-encoding gene of
chlomydomonose eugametos. Gene 1042241-24 .

McPheeters D.S., G. stromo and L. Gold. 1988. Autogenous regulatory site on the
bactenophage T4 gene 32 messenger RNA. J. Mol. Biol. 201:517-35.

Me anzhinov, V..V. et al. 1990. A (proposed structure of bacteriophage T4 gene
pro uct 22- A major prohead scaffol rng core protern. J. Struct. Bro . l 42 24- 1.

Miller E.S. et al. 1987. Translationalreg ession: Biological activity of plasmid-
encoded bacteriophage T4 RegA protern. . Mol. Biol. 194:397-410.

31

Molineux, I. J. 1991. Host-parasite interaction: recent developments in the genetics
of abortive phage Infection. The New Biologist 32230-236.

Morgan, E. A. 1986. J. Bacteriol. 168:1-5.

Muscarella, D.; V..M. Vogt. 1993. A mobile group I intron fiom Physarum
Bollycephalum can Insert Itself and Induce mtomutatrons In the nuclear ribosomal
A of Saccharomyces cerevisiae. Mol. ell.Brol. 132 1023-1033

Nodwell, J. R. and J. Greenblatt. 1991. The nut site of bacteriophage lambda is
made of RNA and is bound b transcri tion antitermination factors on the surface
of RNA polymerase. Genes. ev. 5221 2-51. -

Newman, E. B., R, D'Ari and R. T. Lin. 1992. The Leucine-Lrp regulon in E. coli:
A global response in search of a raison d'Etre. Cell 682617-19.

Nilsson, L. et al. 1990. The role of FIS in trans activation of stable RNA operons
of E.coli. EMBO 92727-734.

Oberto, J and R. A. Weisberg. 1989. Structure and function of the nun gene and
the immunity region of lambdoid phage HK022. J. Mol. biol. 207 2675-94.

Parma, D. H. et al. 1992. The rex system of bacteriophage A2 tolerance and
altruistic cell death. Genes and Devel. 6: 197-510.

Parmeggiani, A. et al. 1987. Pro erties of a genetically engineered G domain of
elongatron factor Tu. PNAS 8423 41-45.

Ramirez, M. R C. Wek, L. R. Vazguez de Aldann, B. M. Jackson, B. Freeman,

and A. G. Hmnebusch. 1992. Mutations activatin the yeast elf-2 al ha kinase

3.052122 igtsrgrltier; of alleles altering the domain re ated to his-tRNA. 01. Cell.
10 . 2 - .

Rao, VB and L.W. Black. 1985. Evidence that a phage T4 DNA ackaging
enzyme IS a processed form of the major capsid gene product. Cell 42296 -77.

Robert,B., et al. 1987. Bacteriophage T4 regA protein binds to mRNAs and
prevents translation initiation. PN S 427822-26.

Robert, J. et al. 1987. The remarkable specificity of a new transcription

termination factor su ests that the mechanisms of termination and antiternrination
are srmrlar. Cell 51:4 -92.

Roberts, J. W. 1993. RNA and rotein elements of E. coli and A transcription
antrterrmnatron complexes. Cell 7 2653-55.

Ruckrnan J., D. Panna, C. Tuerk, D.H. Hall and L. Gold. 1989. Identification of a
T4 gene required for bactenophage mRNA processrng. New Brol. 1:54-65

Sandmeier, H.,' S. Iida, and W. Arber. 1992. DNA inversion re ions Min of
plasmid 15B and Cin of bacteriophage P1: evolution of bacteriop age tail fiber
genes. J. acteriol. 17423936-44.

32

Sekiguchi,M. 1966. Studies on 'the hysiological defect in rII mutants of
bacteriophage T4. J. Mol. Biol. 162503- 22.

Shamoo, Y., K.R.Webster, K.R Williams and W.H.Koni sberg. 1991. A
retrovrnts-like zinc domain is- essential for translation repression of
bacteriophage T4 gene '32. J Biol.Chem. 26627967-70.

Shibuya, M. H. Nashimoto and Y. Kaziro. 1979. Cloning of an EcoRl fragment
carrying E. coli‘tqu gene. Molec. gen. Genet. 170:231-34.

Singer, B.S., Shinedling, S.T., Gold, L. (1983) The rII enes: a histo and a
gospectus. In Bacteriophage T4. p. 327-333. Ber et P, utter E, Ma ews C,
osrg M (eds) American society or Microbiology. ashmgton, D.C.

Shinedindg, S., D. Parrna and L. Gold. 1987. Wild type bacteriophage T4 is
restricte by the lambda rex genes. J. Virol. 6123790-94.

Snyder, L.' 1983. T4 eelynucleotide kinase and RNA ligase, p.351-55. The
bacteriophage T4. ASM ess, Washington, DC.

Snyder, L. and K. McWilliams. 1989. Gene 81217-24.

Snyder, Land G. Kaufrnann. 1993. T4 Phage exclusion mechanisms in the
Molecular biology of Bacteriophage T4, Section II: Phage -host interactions. ASM.

Snyder, M. and WE. Wood. 1989. Genetic definition of two functional elements
In a bacterrophage T4 host- range " cassette". Genetics 1222471-479.

Squires, C. L. et al. 1993. Ribosomal RNA antitermination in vitro: Requirement
for Nus factors and one or more unidentified cellular components. Cell 90:970-74.

Steven, A.C. et al-._ 1992. Conformational changes of a viral capsid protein.
Thennodynarmc rationale for (proteolytic regulatron of bacterro hage T4 ca srd
er; 113760184 co-operativity, an super-stabilrzation by soc bin ing. J .Mol. iol.

Suzanne, C. et al. 1984. Antitermination of E. coli rRNA transcri tion is caused by
a control regron segment containmg lambda nut-like sequences. ell 38:851-860.

Toothman, P. and I. Herskowitz. 1980. Rex-dependent exclusion of
lambdordphage. I. Prophage requirments for exclusion. Virology. 1022 1339-50.

TraversA. A. R. Karnen and R. Schleif. 1970. Factor necessary for ribosomal RNA
synthesrs. Nature (London) 2282748.

Tuerk, C, P. Gauss, C. Thermes, D.R. Groebe, M. Gayle, N. Guild, G. Stromo, Y.
D'Aubenton Carafa and L.Gold. 1988. CUUCGG hairpianxtraordinarily stable
1;ng 6s‘ecgérdary structures associated with various bioc emical processes. PNAS

Tuerk, C., S. Edgy, D. Parma and. L. Gold. 1990. Autogenous translational
pigerator recogruze by bacterrophage T4 DNA polymerase. J. Mol.Brol. 2132749-

33

Unnithan,S. et al. 1990. “Binding of the bacteriophage T4 re protein to mRNA
target: an Initiator AUG rs required. Nucleic acrd research 18: 083-92.

Valencia, A. et al. 1991. GTPase domains of ras p21 oncogene protein and
elon ation factor Tu 2 Anal sis of tlrree-dirnensional structures, sequence families
and ctional sites. PNAS 825443-5447.

van der Meide, P.H. et al. 1983. tuf Gene dosage effects of the intracellular
concentration of EF-TuB. Eur. J. Biochem. 130:409-17.

Vijgenboorrifl E. and L. Bosch. 1989. Translational fi'ameshifts induced by mutant
5 ecies if e polypeptide chain elongation factor Tu of Escherichia coli. J.
acteriol. 2642130 - 7.

Webster, A., R. T: Ha and G. Kemp. 1993. The adenovirus protease is activated
by a Viral-coded drsul rde-lmked peptide. Cell 72: 97-104.

Weijland A. et al. 1992. Elongation factor Tu: a molecular switch in protein
brosynthesrs. Molecular Microbiology 6: 683-88.

Weiss, R. B., W. M. Huang and D. M. Dunn. 1990. A nascent peptide is re uired
fierérrbosomal bypass of the coding gap in Bacteriophage T4 gene 60. Cell 6 2117-

Wickens, M. P. and J.E. Dahlberg. 1987. RNA-protein interaction. Cell 512339—42.

Williams, K.P., .G.A. Kassavetis, and BF. Geiduschek. 1987. Interactions of the
bacteriophage T4 gene 55 product wrth Escherichia coli RNA polymerase.
Competition with scherichia coli sigma su 3870‘) and release from late T4
transcription complexes following initiation. J. io . hem. 262: 12365-371.

Winter, Robert B. et al..1987. Bacterio hage T4 re6gA protein binds to mRNAs and
prevents translation irritration. PNAS 8 27822-782 .

Xue, M.Q., and L.W. Black. 1990. Role of the major capsid protein of phage T4 in
P1547 egrekagmg from structure-function and srte-directed mutagenesrs studies.

Yarnell W.S. and J.W. Roberts. 1992. The phage 1 gene Q transcri tion
antrternrinator binds DNA in the late gene promoter as it modifies A
polymerase. Cell 6921 181-89.

Young,C.C. and RW. Bemlohr. 1991 Elongation factor Tu is metlr lated in
responds to nutrient deprivation in Escherichia coli. J. Bacteriol. 1732 30 6-3100.

Young, RS. and A. V. Furano. 1981. Regulation of the synthesis of E. coli
elongation factor Tu. Cell 24:695-706.

Yu,. Y.-T. N. and L. Sn der. 1993. Cleavage of E. coli translation elongation factor
Tu In phage exclusion. NAS in press.

Zen e1, J.M., RH. Archer and L. Lindahl. 1984. The nucleotide sequence of the
1133c2 1681101191; colr fus gene codrng for delongatron factor G. NUCICIC Acrds Research

CHAPTER 2

TRANSLATION ELONGATION FACTOR TU CLEAVED BY A PHAGE
EXCLUSION SYSTEM
KEY TERMS (E14/PROTEASE/T4 PHAGE)

Submitted to Proc. Nat. Acad. Sci. USA

YUEN-TSU NICCO YU AND LARRY SNYDERT
DEPARTMENT OF MICROBIOLOGY
MICHIGAN STATE UNIVERSITY
EAST LANSING, MI 48824-1101

To whom reprint requests should be addressed.

34

35

ABSTRACT Bacteriophage T4 multiplies poorly in Escherichia coli strains
carrying the defective prophage, e14. The exclusion is caused by the
interaction of an e14 encoded protein, Lit, with a short RNA or polypeptide
sequence, got, from within the major head protein gene of T4. The interaction
between Lit and go! causes a severe inhibition of all translation, and prevents
the transcription of genes downstream of the gal site in the same transcription
unit. It does not, however, inhibit most transcription: nor does it inhibit
replication, or affect intracellular levels of ATP. Here we show that the
interaction of gal with Lit causes the cleavage of translation elongation factor
Tu (EF-Tu) in a region highly conserved from bacteria to humans. The
depletion of EF-Tu is at least partly responsible for the inhibition of
translation and the phage exclusion. The only other phage exclusion system to
be understood in any detail also attacks a highly conserved cellular
component; suggesting that phage exclusion systems may yield important

reagents for studying cellular processes.

36

Resident prophages, plasmids and transposons often help their host by excluding
infecting phages. Well known examples include the exclusion of phages by the
rex gene products of A prophage and exclusion of T7 and related phages by the ptf
gene product(s) of the F plasmid. In all these exclusions, a nonessential protein or
proteins expressed by the resident element somehow recognizes that the cell has
been infected by a phage and kills the cell; thereby preventing the spread of the
phage to other cells that harbor the DNA element. For a review of phage
exclusion systems see (1).

The DNA element e14, a defective prophage integrated in the isocitrate
dehydrogenase (icd) gene (2), partially excludes Teven phages such as T2, T4 and
T6. The exclusion is due to an e 14 encoded protein Lit that, when overproduced,
promotes a severe inhibition of cellular translation late in Teven phage infection
(3,4). The inhibition requires an interaction between the Lit protein and probably
either the RNA or polypeptide fi'om a short region of only about 75 base pairs in
the major head protein gene of Teven phages (4). The inhibition of translation
occurs when the major head protein gene including this region. begins to be
transcribed and translated late in infection. We call this short region the go!
region, because it was first defined by gal mutations that allow the phage to grow
_o;n lit. In this report, we show that translation elongation factor Tu (BF-Tu) is
cleaved during the phage exclusion. Apparently, the Lit protein is a specific
protease which is activated by the polypeptide or RNA from the T4 gal region to
cleave EF-Tu, thereby causing the inhibition of translation. The EF-Tu is cleaved
very close to the site of EF-Tu methylation in a sequence that has been highly

conserved throughout evolution.

37

MATERIALS AND METHODS

Bacterial Strains and Plasmids. Bacterial strains were derived from
JM101 and W3110 laclq recA, which have been described (4). To construct
isogenic derivatives of these strains, one of which lacks e 14 so has no Lit protein
and the other of which has an excess of Lit protein, we first cloned a kanarnycin
resistance (KnR) cassette between the BamHI sites in the large HindIII fi'agment of
e14 in the plasmid pAG2, obtained from Charles Hill (5). We then crossed this
clone into e14 in the chromosome using a palA(TS) mutation as described
previously (6). The e 14 had a lit(Con) mutation, which is an up-promoter
mutation causing the overproduction of Lit protein (3). A low percentage of the
recombinants still retained the lit(Con) mutation and such a strain was selected and
used as a donor for P1 transduction to move the lit(Con) mutation into W3110
Lach recA' and JM101 selecting for KnR. The e14 prophage can be transduced
into a RecA' recipient because e14 encodes its own Integrase (5). The isogenic
Lit° derivatives were obtained by UV irradiating the LitCon KnR strains and
selecting a KnS derivative which had simultaneously lost the LitCon phenotype so
presumably had been cured of e14. The plasmid pUC84PZ1 described previously
(4) has a 159 hp. DNA fragment including the gal region cloned into a derivative
of pUC8 such that the gal region is transcribed from the lac promoter and
translated in the gene 23 frame from the lacZ ribosome initiation site. The plasmid
pACRV9-KAN was derived from pACRV9 (4) by cloning a KnR cassette into the
EcoRI site in the cat gene of pACYCl84. It has the wild type lit gene cloned in
the tel gene of pACYCl84 so it will be expressed fi'om its own promoter. The
plasmid pTA9 was obtained from David L. Miller and has the 1qu gene (BF-Tu)
cloned in such a way that it will be transcribed from its own promoter as well as

the lac promoter in these cells.

38

Preparation of Cell Extracts and S30 and 8150 Supernatants and in
vitro Translation. The extracts were prepared from cells grown in 1% tryptone,
1.0% NaCl, 0.5% glucose, 50 pg/ml ampicillin, 50 pg/ml kanarnycirr. When the
O-D-625mu reached 0.3, the cells were divided into two 500 ml cultures, and
IPTG was added to 5 mM to one of the cultures for 20 more minutes. The cells
were collected, resuspended in 4 ml resuspension buffer and lysed in a French
Pressure Cell. The S30 and S 150 supematants were prepared and the in vitro
translations performed as described (7) with a few minor modifications. No RNA
and no ribonucleoside triphosphates except GTP and ATP were added to the in
vitro reactions. The reaction rrrixtures contained 200 pl supernatant (100 pl 830 +
100 pl 8150) in a total volume of 300 pl. Aliquots of 70 pl were withdrawn and
precipitated with 3 ml 5% trichloroacetic acid (TCA) in the cold. The precipitates
were heated to 100°C for 5', centrifuged, resuspended in 0.1 ml 2% KOH, and
reprecipitated with 3 ml 5% TCA before collecting on WF/A Whatrnan filters for
counting.

Antibody Precipitations. 10 ml of cells were grown-in M9 medium plus
all 20 amino acids except methionine plus 50 pg/ml arnpicillin to an DD. = 0.4 at
30°C. The proteins were labeled by adding 4 pl of 35S methionine (1200
Ci/mmole) for 10' and chased with unlabeled methionine (10 mM) for 10'.
Depending on the experiment, the cells were either infected with CsCl purified T4
at an M.Q.I. of 10, or IPTG was added to 5 mM to induce transcription and
translation of the go] region. After 30 rrrinutes, the cells were collected by
centrifugation and resuspended in 100 pl H20. They were lysed and the antibody
precipitations performed as. described 8).

Phage One-step Growth Experiment. E. coli JM101 with and without the
Lit protein (see bacterial strains) and with and without the tqu clone in pTA9
were grown at 30°C in 1% tryptone, 1% NaCl, 0.5% glucose, to midlog phase and

39

IPTG was added for 20 minutes before infection. CsCl purified T4 were added at
an M.Q.I. of 10 and 1 ml aliquots were taken to add to ice for SDS-PAGE
electrophoresis. Aliquots were also taken to add to saline plus CHCl3 and diluted
to plate with indicator bacteria to determine the phage yield.

Other Methods: SDS-PAGE electrophoresis was by the method of
Laemmli (9). The cleavage fiagment of EF-Tu was electroeluted onto
nitrocellulose and sequenced by the Michigan State University Macromolecular

Facility using Edrnan Degradation.

40

RESULTS

Extracts of Inhibited Cells Are Inactive for Translation In V‘ttra.
Infection of e 14 Lit prOtein containing cells by T4 causes a severe inhibition of
translation due to the interaction of the RNA or polypeptide from the T4 gal
region with the Lit protein. If extracts of the infected cells are inactive for
translation, it may be possible to determine the cause of the inhibition. To avoid
possible complications due to phage infection, we can mimic the effect of phage
infection using constructs like that shown in figure 1. In this construct, the lit gene
of e14 has been cloned in the multicopy plasmid pACYCl84 such that it will be
expressed from its own promoter. The gal site of T4 has been cloned in the
compatible vector, pUC8, such that it will be transcribed from the lac promoter
and translated in the gene 23 frame fi'om the lacZ ribosome initiation site. When
IPTG is added to cells containing both plasmids, the transcription and translation
of the T4 gal region will commence and the severe inhibition of translation will
ensue (4). To make the extracts as similar as possible, we grew the cells cartooned
in figure 1 and divided the culture in half just before adding IPTG to one of the
two subcultures. After 20 minutes induction, to allow transcription and translation
of the gal region, the cells were concentrated and lysed as described in Methods,
and radioactive methionine was added to measure incorporation into acid insoluble
polypeptides. We observed the extracts of the cells to which IPTG had been
added to be significantly less active for in vitro translation than the parallel culture
(data not shown).

There are many possible causes for the relative inactivity of the extracts of
the inhibited cells. An inhibitor of translation could have been generated.
Alternatively, the ribosomes or one of the soluble factors, e. g. tRNA, initiation or
elongation factors, could be inactivated somehow. To distinguish these

possibilities, we further fractionated the extracts by differential centrifugation. We

41

prepared S30 supematants as described in Methods. These supematants retained
everything required for translation but lost their activity when they were dialde
overnight in the cold, presumably because the endogenous mRNA was degraded.
We also prepared S 150 supematants by centrifuging part of each S30 supernatant,
before dialysis. The S 150 supematants were not dialyzed. They should lack
ribosomes but have all the soluble factors including mRN A required for
translation.

To determine the effect of the supematants on translation, we mixed the
supematants from inhibited and uninhibited cells in the in vitro translation assay
(see Methods). Addition of the S150 supernatant from the uninhibited cells
stimulated incorporation of amino acids by the S30 supernatant fi'om uninhibited
cells (see Fig. 2A). In contrast, addition of S 150 supernatant fi'om inhibited cells
did not stimulate incorporation. We conclude that some difference in the S 150
supematants is responsible for the inhibition of translation.

EF-T u has been cleaved in the inhibited extracts. As part of an effort to
determine the difference between the inhibited and uninhibited extracts, we
electrophoresed the proteins in the S30 and S 150 supematants on SDS
polyacrylarnide gels and stained the proteins. The results are shown in figure 2B.
A major protein band of 43 kDa is missing from the induced supematants, and it is
replaced by a smaller band of about 37 kDa. This is the only reproducible
difference we have detected betWeen the inhibited and uninhibited extracts. In the
experiment shown in Fig. 2B, another polypeptide of about 75 kD appears to be
missing from the supematants of the inhibited extracts. In other inhibited extracts,
this polypeptide is present and is probably a protein that sporadically sediments
during the preparation of the supematants.

From its size and abundance, we suspected that the 43 kDa protein is EF-

Tu, which is being cleaved, and the 37 kDa polypeptide is one of the cleavage

42

products. To prove this, we performed an antibody precipitation experiment. The
results are shown in figure 3. In this experiment, the proteins were radioactively
labeled and the gal region was induced in the presence and absence of Lit protein.
After lysis, the proteins were precipitated with specific antibodies directed against
EF-Tu. As a control, we included cells that contain an overproducing clone of
tqu, a gene for EF-Tu. As expected, the antiserum specifically precipitates the 43
kDa EF-Tu protein and more .of this protein precipitates in cells with the
overproducing clone. In the cells in which the gal region had been induced in the
presence of Lit protein, much less of the 43 kDa protein precipitates, and a second
polypeptide of 37 kDa also precipitates (Lanes 9-11). We conclude that the 43
kDa protein that is being cleaved is EF-Tu, and that the 37 kDa polypeptide is one
of the cleavage products.

The induction of the cloned gal region in the presence Of Lit protein mimics
in many ways the infection of Lit protein containing cells by T4 phage suggesting
that EF-Tu would also be cleaved after T4 infection. To test this, we labeled cells
with S35 methionine and chased with cold methionine before infecting them with
T4 as in Methods. After 30 more minutes we lysed the cells and precipitated the
EF-Tu with anti EF-Tu as before. After infection of the non-Lit protein-containing
cells only one protein, the 43 kDa EF-Tu protein precipitated. However, after
infection of Lit protein-containing cells, the 37 kDa cleavage fragment also
precipitated (data not shown). Therefore, EF-Tu is also cleaved afier T4 infection
of Lit protein containing cells. In contrast EF-Tu was not cleaved after infection
by T4 with a gal mutation. This was expected since gal mutations allow T4 to
multiply in Lit protein containing cells. Presumably, they change the, gal peptide
or RNA so it can no longer activate the cleavage of EF-Tu.

The experiment above was performed with cells that contain abnormally

high levels of Lit protein. It is of interest whether EF-Tu is also cleaved after

43

infection of E. coli K12 containing normal levels of Lit protein. When E. coli
containing an integrated copy of wild type e14 is infected by T4, EF-Tu is also
cleaved (data not shown). However, in this case, only about 50% of the EF-Tu is
ever cleaved, even at late times. Nevertheless, phage production is significantly
delayed (data not shown). Apparently, normal levels of Lit protein can cleave a
significant part of the cellular EF-Tu and inhibit, but not totally block, phage
development. 3

Is depletion of EF-T u solely responsible for the inhibition of T 4 production?
Depletion of EF-Tu through inactivation by cleavage would be sufficient to
explain the inhibition of translation and the block to phage production since EF-Tu
is required for translation. However, this may not be the only contributing factor.
Some other protein may also be cleaved, or the cleavage fragment may be
inhibitory. To begin to investigate these possibilities, we performed the
experiment shown in figure 4. The rationale behind this experiment is as follows.
If cells contain an excess of EF-Tu because they have a clone of the tqu gene in a
multi-copy plasmid, then not all the EF-Tu might be cleaved after T4 infection. If
depletion of EF-Tu due to inactivation by cleavage is solely responsible for the
inhibition of phage production, then, under these conditions of an excess of EF-
Tu, T4 production should be normal. If, however, the cleaved form of EF-Tu is
somehow interfering with phage production, or another protein required for T4
development is also being cleaved, then phage production under these conditions
should not be normal. From figure 4A, it is apparent that cells with the tuf/l clone
have about three times the normal amount of EF-Tu (compare lanes 10-12 to lanes
4-6); and even though more EF-Tu is cleaved after T4 infection, judging by the
amount of cleavage fragment (lanes 7-9), a supemorrnal amount still remains

(compare lanes 7-9 to lanes 1-3). The remaining uncleaved EF-Tu does support

44

some phage production (Fig. 4B) indicating that the depletion of intact EF-Tu is
partially responsible for the block in phage production.

While the experiment shown in Fig. 43 supports the conclusion that
depletion of EF-Tu is partially responsible for the block in T4 production, it also
suggests it. may not be the only contributing factor. The excess of EF-Tu causes
the cells to grow more slowly and delays phage production (Fig. 4B), complicating
the interpretation of the results somewhat. Nevertheless, if depletion of intact EF-
Tu were solely responsible for the inhibition of phage production, we nright expect
production in the presence of excess EF-Tu and Lit protein to be at least as high as
phage production in cells with excess EF-Tu but without Lit protein. That it is not
makes us entertain other possibilities such as that the cleaved form is inhibitory, or
that some other protein besides ,EF-Tu is being cleaved. More experiments are
required to resolve these issues. I

Locating the cleavage site of EF-T u. To locate the site of cleavage of EF-
Tu, we isolated the 37 kDa cleavage fi'agment from gels and had the N terminus
sequenced in our Macromolecular Facility. If the N terminus has been cleaved off,
the N terminal sequence should be different from that of intact EF-Tu and the
sequence should reveal the exact site of cleavage. The N terminal sequence of the
cleavage fragment was determined to be N-ile-thr-ile-asn-thrm (see Figure 5). A
comparison of this sequence with the known sequence of E. coli EF-Tu (10)
revealed the cleavage occurs between glycine 59 and isoleucine 60. This cleavage
site is also consistent with the relative sizes of the intact EF-Tu and the cleavage
product. The entire EF-Tu has 393 anrino acids and a cleavage at arrrino acid 60
would remove about 15% of the protein, leaving a polypeptide of about 37 kDa.
Note that there could be other cleavage sites in the extreme N terminus of the
protein, but the smaller cleavage fragments would not be detected by our method.
DISCUSSION

45

In this paper we have shown that EF-Tu is cleaved following T4 infection
of cells containing the defective prophage, e 14. The cleavage requires the
interaction between the e14 encoded Lit protein and the polypeptide or RNA fi'om
the gal region of T4, a short sequence within gene 23, the major head protein gene.
The cleavage occurs between glycine 59 and isoleucine 60 in a highly conserved
region of EF-Tu. The cleavage apparently inactivates EF-Tu for translation
because translation is completely blocked if all the EF-Tu in the cell is cleaved.
However, depletion of EF-Tu is probably not solely responsible for the block in
T4 phage development since T4 deVelopment is not normal even if supernormal
amounts of EF-Tu remain uncleaved.

At present, we do not know the mechanism of cleavage of EF-Tu. We have
only observed. the cleavage when both the e14 Lit protein and the T4 gal region
are present and the T4 gal region is both transcribed and translated. The Lit
protein contains a motif characteristic of zinc dependent proteases (Yu and
Snyder, to be published); indicating that it is probably the protease. If so, the Lit
protease must be activated by something generated from the transcription and
translation of the gal region afier T4 infection. Note that the gal RNA or peptide
could possibly activate the Lit protease not by binding to it but rather by binding to
EF-Tu, changing its conformation and thereby making it susceptible to cleavage.
Our preliminary searches have revealed no similarities between the go] RNA and
peptide and tRNA, EF-Ts or other proteins known to bind to EF-Tu. It is also
possible that the Lit protein is not the protease, and that another protease is being
activated, or that EF-Tu cleavage is due to autoproteolysis, stimulated somehow by
the gal-Lit interaction. In this respect, it may be relevant that EF-Tu is often
cleaved during purification and/or. crystallization (16,17) giving rise to fragments
of - approximately the same size as those induced by gal-Lit. It is of interest

whether this apparently spontaneous cleavage bears any relationship to the gal-Lit

46

induced cleavage or whether this region is merely unusually sensitive to
proteolysis. We have recently observed the cleavage of EF-Tu in vitro in
experiments in which crude extracts of cells in which the gal region had been
induced were mixed with crude extracts of cells containing Lit protein. This assay
should allow us to determine, unambiguously, the function of the various
components of the gal-Lit induced cleavage of EF-Tu.

There are many possible mechanisms by which the cleaved form of EF-Tu
could afl‘ect cellular functions. It may bind tRNA but not function for translation,
thereby depleting the available pool of tRN A. Or it may enter the ribosome but
not function for translation, thereby clogging up the translatiOn apparatus. Or the
effect of the cleaved farm of EF-Tu may be indirect. For example, the cleaved
form may bind GTP but not recycle it, thereby making it unavailable for other
cellular processes but this seems unlikely considering that there is much more GTP
in the cell than EF-Tu. Also there seems to be little or no general effect of the
cleavage of EF-Tu on transcription (4) which also requires GTP. Finally, the
inhibition of translation may be due to depletion of the active form of EF-Tu. This
latter explanation is most consistent with the observation that overproducing EF-
Tu, so that not all is cleaved, has an ameliorating effect on the inhibition of protein
synthesis and phage production.

The results of in vitro experiments with EF-Tu do not give a clear
indication of what properties to expect of the cleaved form of EF-Tu in. viva. In
the early stages of digestion of the native protein, trypsin cleaves at amino acids
arginine 44 and arginine 58 (14), very close to the gal-Lit induced cleavage. The
larger fi'agments produced by these cleavages do not dissociate although the small
fragment of 14 amino acids may be lost. These results suggest that the fiagments
created by the gal-Lit induced cleavage may not dissociate in viva. On the

question of whether this cleaved form of EF-Tu retains activity, different methods

47

have yielded apparently contradictory results. For example, by their methods,
Wittinghofer et al. (14) fOund that the EF-Tu of E. coli cleaved by limited trypsin
digestion, retains the ability to stimulate polyphenylalanine synthesis in response
to poly U.- However, Gulewicz et al. (15) found that the cleaved form of the
closely related EF-Tu of T hermus thermaphilus has lost this ability although it
retains the ability to form a ternary complex with GTP and tRNA. Such seemingly
contradictory results make it difficult to. predict what activities would be retained
by the gal-Lit cleaved EF-Tu in viva.

At present, it is not known if the interaction of gal and Lit is activating the
cleavage of other proteins besides EF-Tu. The five amino. acid sequences
surrounding the cleavage site of EF-Tu is shared by EF-G and a similar sequence
exists in SelB, the specific EF-Tu for selenocysteine tRNA. In. the absence of any
knowledge of the sequence and structural requirements for the cleavage, it is
diflicult to predict what other E. coli proteins can be cleaved. If the components
of the cleavage reaction can be purified, other proteins can be tested. Until then,
the possibility remains that cleavage of another protein required for translation
could be at least partially responsible for the inhibition of translation.

Even if depletion of intact EF-Tu sufficiently explains the inhibition of
translation, from Figure 4 it apparently is not the only cause of the block in phage
production. When EF-Tu is overproduced, so not all the EF-Tu is cleaved, T4
production is still retarded. This result might have been predicted from our earlier
work which showed that there is another effect of the interaction of the Lit protein
with the gal region. In addition to promoting the inhibition of all translation, the
Lit protein also prevents the transcription of genes serviced by the same promoter,
but downstream of the gal site. This phenomenon, which we call the 13%!
inhibition (4), occurs even when not all the EF-Tu is cleaved. Because it interferes

with the transcription of T4 gene 23, an essential gene, the local inhibition should

48

also cause a delay in phage production. At present, we do not know what
consequence of the gal-Lit interaction is causing the local inhibition. It may be
due to the cleaved form of EF-Tu, or the cleavage of an, as yet unknown, other
protein. Further experiments are needed to distinguish these and other
possibilities.

The site of cleavage of EF-Tu has some interesting features. The cleavage
occurs in what is predicted to be a largely unstructured region spanning the GTP
binding pocket (cf 11, 13). As mentioned, the two most sensitive sites for trypsin
cleavage of the native protein are just upstream of the gal-Lit promoted cleavage
(14) which also suggests that this region is exposed on the surface of the molecule.
This region may also play a regulatory role. The EF-Tus of many bacteria
including E. coli are known to be methylated in response to starvation conditions
(18,19). The one site of methylation in E. coli is the lysine at position 56 (10),
only four amino acids upstream of the cleavage site. It is of interest whether the
methylated form of EF-Tu is also cleaved.

Further testimony to the irnportance of the region of cleavage comes from
its evolutionary conservation. The EF-Tus of bacteria (called EF-la in
eukaryotes) are among the most highly conserved cellular constituents, and the site
of cleavage lies within one of the most highly conserved regions of all (20). In
fact, the two amino acids on one side of the cleavage site and the three on the other
side are probably the same in all organisms on earth (20, and Figure 5). This same
five arrrino acid sequence has also been conserved in the EF-Gs of all organisms
(called EF-2 in eukaryotes) even though the flanking sequences have diverged
(20). It will also be of interest to determine which elongation factors fi'om other
organisms can serve as substrates for the protease.

The only other phage exclusion to be understood in' any detail, the

exclusion of T4 by the prr element in a clinical isolate of E. coli, shows

49

remarkable similarity to the exclusion of T4 by e14. In both exclusions, a small
T4 encoded polypeptide seems to activate an endogenous enzyme to attack a
highly conserved cellular constituent. In the case of the prr exclusion, the host
lysine tRNA is cleaved (cf. 21). This system will even cleave the human lysine
tRNA (GLKaufmann, pers. coMun.) Because of the specificity of these phage
exclusions for highly conserved cellular components, they may be useful reagents
for studying cellular processes in many organisms.
ACKNOWLEDGMENTS

We thank David ‘L. Miller and Robert Bernlohr for their generous gifts of
anti EF-Tu antibodies and David L. .Miller for a clone of tqu. We also thank
Allen Nicholson and Richard Schwartz for their advice on the antibody
precipitation experiments; Nanette Guyer for help with strain construction; and
Joseph Leykarn of the Michigan State'University Macromolecular Facility for his
help with the N terminal polypeptide sequencing. This work was supported by a
grant fi'om the NSF to LS. and was submitted as part of the requirement for the
PhD. by Y.-T.Y.

50

References

l.

2
3.
4

.V'

10.

ll.
12.

l3.

14.

15.

Molineux, I. J. (1991) The New Biologist 3, 230-236.

Hill, C. W., Gray, J. A. & Brody, H. (1989) J. Bacteriol. 171, 4083-4084.
Kao C. & Snyder, L. (1989) J. Bacteriol. 170, 2056-2062.

Bergsland, K. J., Kao, C., Yu, Y.-T.N., Gulati, R. & Snyder, L. (1990) J.
Mol. Biol. 213, 477.494. '

Brody, H., Greener, A & Hill, C. (1985) J. Bacteriol. 161, 1112-1117.
Kao, C., Gumbs, E. & Snyder, L. (1987) J. Bacteriol. 169, 1232-1238.
Bourgaize, D. B. & Fournier, M. J. (1987) Nature (London) 325, 281-
284.

Bardwell, J.C.A., Regnier, P., Chen, S.-M., Nakamura, Y., Grunberg-
Manago, M. & Court, D. L. (1989) EMBOJ. 8, 3401-3407.

Laemmli, UK. (1970) Nature (London) 227, 680-685.

Arai, K., Clark, B.F.C., Duffy, L., Jones, M. D., Kaziro, Y., Laursen, R. A.,
L'Italien, J., Miller, 0. L., Nagarkatti, S., Kanamura, S., Nielson, K. M.,
Petersen, T. E., Takahashi, K., & Wade, M. (1980) Proc. Natl. Acad Sci.
USA 77, 1326-1330.

Jurnak, F. (1985) Science 230, 32-36.

Gulewicz, K., Faulhammer, H. G. & Sprinzl, M. (1981) Eur. J. Biochem.
121, 155-162.

Riis, B., Suresh, I., Rattan, S., Clark, B.F.C. & W. C. Merrick, (1990)
Trends Biochemical Science 13, 420-424.

Whittinghofer, A., Frank, R. & Leberrnan, R. (1980) Eur. J. Biochem.

- 108, 423-431.

Gulewicz, D., Faulhammer, H. G. & Sprinzl, M. (1981) Eur. J. Biochem.
121, 155-162.

l6.

17.

18.

19.
20._

21.

51

Morikawa, K., LaCour, T.F.M., Nyborg, J., Rasmussen, K. M., Miller, D.
L. & Clark, B.F.C. (1978) J. Mol. Biol. 125, 325-338.

Arai, K., Ota, Y., Arai, N., Nakamura, S., Henneke, C., Oshirna, T. &
Kaziro, Y. (1978) Eur. J. Biochem. 92, 509-519.

Ferro-Luzzi Ames G. & Niakido, K. (1979) J. Biol. Chem. 254, 9947-
9950.

Young, C. C. & Bemlohr, R. W. (1991) J. Bacterial. 173, 3096-3100.
Ibawe,N., Kruna, K.-I., Hasegawa, M., Osawa, S. & Migata, T. (1989)
Proc. Natl. Acad Sci. USA 86, 9355-9359.

Levitz, R., Chapman, D., Amitsur, M., Green, R, Snyder, L. & Kaufinann,
G. (1990) EMBOJ. 9,1383-1389.

52

Figure Captions

Figure 1. The "imitation infected cell" used in these experiments. The E.
coli W3110 10ch recA cells have been cured of e14 so lack the normal
chromosomal source of Lit protein. They contain two compatible plasmids
as described in Methods. One plasmid expresses the lit gene of e14
constitutively from its own promoter; while the other plasmid has the gal
region of T4 gene 23 cloned so that its transcription is induced by IPTG and

it will be translated in the gene 23 frame from the lacZ ribosome initiation

site.

53

 

ro¥-@>mo<o

  

LNavmoao

54

Figure 2. A defect in translation by supematants from induced cells
correlates with the absence of a 43 kD polypeptide. Panel A: Incorporation
of 35S methionine into acid insoluble material with different mixtures of
supematants. (-°-) S30 uninduced + S150 uninduced; (-A-) $30 uninduced
+ $150 induced; (-O-) S30 uninduced; (-I-) S30 induced. Panel B: SDS
polyacrylamide gel electrophoresis of the proteins in the supematants of
Panel A. Lane 1: S30 supernatant extract of induced cells. Lane 22 S30
supernatant of uninduced cells. Lane 3: 8150 supernatant of induced cells.
Lane 4: $150 supernatant of uninduced cells. The 43 kD polypeptide that
is missing from the supematants of the inhibited extract as well as the
polypeptide which replaCes it are indicated by the arrows. As discussed in
Results, the other difference between the two extracts at about 75 kDa is
not reproducible and is probably due to a protein that sporadically

sediments in the preparation of the supematants.

55

HZ Sawo aw>2mr>joz

0C3

 

Emoo

Lace 1

LEGO i

LOCO l

moo

moo

 

 

 

 

 

A00

 

 

 

 

 

 

|\)
O
O
T

Panel A

A5 mo
.33 2:6

56

Panel B

 

57

Figure 3. Antibody precipitation of EF-Tu in inhibited and uninhibited
cells. The cells were those in figure 1 except they were derived from E.
coli JM101 and the lit gene was overexpressed from e14 in the chromosome
because of an up-promoter mutation rather than from pACRV9. Shown is
an autoradiograrn of the dried gel prepared as in figure 2. Lanes 1-3: Cells
without Lit protein or pUC84PZl but containing the plasmid, pTA9, that
overproduces EF-Tu. Lanes 4-7: Induction of pUC84PZ1 in cells with no
Lit protein. Lanes 8-11: Induction of pUC84PZ1 in cells with Lit protein:
The antibody precipitations in lanes 1, 5 and 9 had three times as much
antiserum as those in lanes 2, 6 and 10. Lanes 3, 7, 112 one preparation of
nonimmune serum. Lanes 4, 8: a different preparation of nonimmune

serum. The nonimmune serums were added at the higher concentration.

58

um l.

mwl

zeoaaom an.“

F

 

59

Figure 4. Multiplication of T4 in cells with an excess of EF-Tu. The cells
were the same as those in figure 3, but some of the cells had an excess of
EF-Tu because they contained the plasmid pTA9 with the tqu gene. Panel
A: Commassie blue stained SDS-PAGE gels of proteins from cells infected
by T4. Lanes 1-3: Cells with Lit protein and normal amounts of EF-Tu.
Lanes 4-6: Cells with no Lit protein and normal amounts of EF-Tu. Lanes
7-9: Cells with Lit protein and excess EF-Tu. Lanes 10-12: Cells with no
Lit protein and excess EF-Tu. Times after infection are 0', 20', 40' for each
set. Panel B: One step growth experiment of T4 in cells fi'om Panel A. (-
0-) no Lit protein, normal amounts of EF-Tu; (-V-). no Lit protein,
excess EF-Tu; .(-O-) Lit protein, excess EF-Tu; (-O-) Lit protein, normal
EF-Tu.

Panel A

H 3
31.3: a
a. . ..
.3. .. T i: ...I

.3... . .1 3...;

Fl! '2,

I...

.1.

a

. 3. .33 3.1% 4.13:

3

 

m. .,:.r m. .3» .lede
33 33.... 3...... 3. a... 3.:
r. i .L I s... . . .

All

Panel B

phage no./ per cell

60 i I f I
LitO LitO(EFTu)
50 — V v _
l
I
30 3 Lit+++ _.
(EFTu)
,0
20 » / / 1
1o 3 . -
’ O
/ V/ Lit+++
0‘ H43 H
O 15 3O 45 60

61

ONE STEP GROWTH EXPERIMENT

 

 

 

 

 

 

TIME (min)

62

Figure 5. A comparison of the sequences of amino acids in EF-Tu (BF-1a
in eukaryotes) from the region of cleavage. The arrow shows the site of
cleavage. The identical amino acids on either side of the cleavage site are
highlighted. The asterisk shows the lysine that is methylated in E. coli in
response to starvation conditions. The sequences were obtained from Gene

Probe.

63

 

 

«
50:3u-5m...->_m->_.m.0_<-—_?H__...=n->m=-._,_:i 5 m. 833 Er?

---O_=um->3m -0333-..» 8-0_<-=?._._:..:?>m=-=m 3 l: 233.83.12.23 mun?
---Q_=o~->3m ->_m->..n-n_<-__ou—._:r__?mon -3513 -- m. 32.5.33... 73382525... ES...
59.33-3533 -Q.=->..m-0_<-__?._._:...=?>mc-=o d 5. m. 8.335.... mm-..“

---Q_=ao->3m -0.:->..m-n_<-=?._._.7:?>mc-:a .3 l: :55: ea :28 ES...

 

 

CHAPTER 3

ACTIVATION OF THE E14 ENCODED LIT PROTEASE

BY A 25 AMINO ACID OLIGOPEPTIDE FROM
THE T4 GOL SITE

64

65

Abstract

Cleavage of translation elongation factor EF-Tu has been shown to occur
in an exclusion phenomenon of E. coli phage T4 caused by a cryptic prophage e14
encoded protein, pLit(5, 16). The cleavage results from the interaction between
Lit protein and a short sequence, go]. The gol sequence has a minimum length of
75 base pairs within T4 gene 23. The cleavage of EF-Tu apparently inhibits all
cellular and phage translation. To determine the roles of factors that might be
involved in the proteolytic reaction, in vitro assays were performed. Here we
report that the Escherichia coli translation elongation factor EF-Tu can be cleaved
after mixing two cell free extracts, one from cells containing a Lit recombinant
plasmid, another containing a go] fusion plasmid. A computer search for the
function of the Lit protein revealed a possible zinc-dependent metalloprotease
motif at sequences close to the middle of the protein. RNase treatment of the go]
extract prior to mixing did not abolish proteolytic activity, suggesting that the Go]
peptide rather than gol RNA is required for cleavage of EF-Tu. Also adding in
vitro synthesized gol RNA did not activate the protease. In contrast, adding the
synthesized Gol polypeptide did activate the cleavage. Our current hypothesis for
the proteolytic mechanism is that the Lit protein functions as a protease and the
Go] peptide works as an activator for gpLit's catalytic activity.

Introduction

Bacteriophage T4, one of the largest viruses, is composed of a circular
double strand DNA genome close to 160 kb long with the capacity of 200 genes. It
propagates in a majority of E. coli laboratory strains, but hosts sometimes acquire
the ability to exclude infecting phages by functions exerted from resident plasmids

or prophage. The exclusion mechanisms usually provoke cellular dysfunction that

66

lead to a cessation of all macromolecular synthesis, terminating phage
development, thus preventing phage progeny from spreading into the bacterial
population. Exclusion of phage T4 rII mutants and other related phage by the rex
genes of lambda, and exclusions of T7 by the pif of F plasmid are classic examples
and documented (1, 2). But the detailed mechanisms are still not fully understood.

The cryptic prophage e14, which is located at 25 minutes of E. coli K strain
genome, also excludes phage T4 at late stage of phage infection, particularly when
one of its gene products Lit is overproduced, either by an up-promoter mutation or
expression from a high copy number plasmid (4). The Lit polypeptide has a
molecular weight of 30 kD and is believed to be an inner membrane protein by
virtue of its copurification with a cellular inner-membrane fraction (3). It causes a
severe blockage of translation late in T4 infection, owing to its interaction with a
short region, gol, about 75 bp long and 300 bp downstream from the T4 gene 23
ribosome binding site (6). The Gene 23 encodes the phage T4 major head protein
and is only expressed during the late stage of development.

We recently reported that the translation deficiency in the phage exclusion
system , caused by the interaction of gpLit with the inducible gol activity (RNA or
peptide), is at least partly due to cleavage of translation elongation factor Tu (5:
Yu and Snyder submitted). Partial amino acid sequencing data from the N
terminus of the major 37 kD cleaved fragment revealed the location of the
cleavage site to be in between glycine 59 and isoleucine 60, which is an unusual
sequence for protein hydrolysis. There are a few interesting features around the
cleavage site context. First, a computer database search indicated that the
surrounding 5 amino acids including the target site is the second longest stretch of
identity, beside the one within the GTP binding motif ,in all the prokaryotic EF-
Tus and their eucaryotic counterparts, EF- las (GCG database). Secondly, one of

the two trypsin sensitive sites in the native form of EF-Tu is located just one

67

residue ahead of the Lit-gol mediated cleavage site (13). Furthermore, EF-Tu of E.
coli is known to be methylated under starvation conditions and the only lysine
residue found to be methylated is located at position 56 only 4 amino acids away
from the gol-Lit mediated cleavage site (14,15). The intriguing interconnection
between cleavage and methylation is also discussed in this report.

The evidence obtained from genetic and molecular experiments indicated
that the gol-Lit interaction somehow triggers a specific proteolytic reaction to
attack the essential protein EF-Tus. The consequence of the cleavage is blockage
of translation; thus abortion of the phage infection. Subjects remaining to be
resolved include how does activation occurs; what the roles of Lit and gol are; and
which protease is involved. In order to approach those questions, in vitro assays
were undertaken to characterize the proteolytic reaction. In this paper we
demonstrate the cleavage of EF-Tu in vitro and propose potential roles for the

Lit protein and the gol sequence in the catalytic reaction.

Materials and Methods

Bacteria, phages, and plasmid constructs: Table 1 summarizes the
relevant features and references of the bacterial strains and plasmids used in our
experiments. Crossing a kan resistance gene into e14 was carried out by the
means of restricting ColEl plasmid in a polA(ts) strain as described(16). The kan
resistance cassette obtained from Strategene Inc. was cloned into the BamHl site
of the pAG plasmid (obtained from Dr. Charles Hill) carrying an e14 fragment
(17). The recombinant clone cannot replicate in a polA(ts) strain at 420C,
therefore hosts can only acquire the kan resistant trait through homologous
recombination at the e14 region. Kan resistance was then transduced by P1 phage
from the kan1'-lit+ polA donor to JM101 lit+ and J M101 lit(con) respectively by
the method of Miller (18). To make an isogenic lit0 strain, a JM1011it(con) kanr

68

strain 609 was cured of e14 by UV irradiation, selecting for kans and testing for
the loss of the lit(con) phenotype by cross-streaking with wild type T4. A lacZ
fusion plasmid, pUC84PZl, translates the gol region in the frame of T4 gene 23,
and exhibits the gol phenotype which cannot transform lit(con) cells due to
blockage of cellular translation. pTTQ 18lit contains the entire lit gene oriented
clockwise downstream of the tac promoter at the polylinker site of pTTQ18. The
Lit protein was made from its own promoter and the ribosome binding site.
pBlue75gol transcribes the 75 bp gol sequence from the T7 promoter of
pBluescript ( the vector is obtained from a hybrid clone given by Dr. J. Dogdson).
pETl lb-S transcribes and translates gol-lacZ fusion RNA from the T7 promoter of
pETll (N ovagen).

Media and culture conditions: E. coli strains were usually grown in LB
media (1% tryptone, 1% NaCl, 0.5% yeast extract). In protein labeling
experiments, cells were grown in M9 media with 0.4% glucose plus 20 amino
acids except L-methionine. For making cell free-extracts, cells were grown in
tryptone broth (1% tryptone, 1% NaCl, and 50 ug/ml vitamin B1) supplemented
with 0.4% glucose. When desired, antibodies were used in the following
concentrations: arnpicillin, 50 ug/ml; kanamycin, 50 ug/ml; IPTG (BMB)was
added to 1 mM to induce the lacZ promoter.

Plasmid purification: All plasmids used in experiments were purified by
Qkit obtained from Qiagen. The restriction enzymes used to construct recombinant
clones were obtained from BMB.

In vitro cleavage assay: S30 cell extracts were made according to the
procedure described in the previous paper(5). Substrate proteins were prepared
from crude extract of S35-methionine labeled cells. Cells cured of e14 (litO) were
transformed with a tqu clone, pTA9, to express increased levels of Tu protein

from a inducible promoter. The transformants were grown in M9+Bl media to

69

OD=O.4, then induced with IPTG to 1mm and labeled with radioactive S35-
methionine for 10 min. Cells were washed lx with cold saline and resuspended in
resuspension buffer. The labeled crude proteins were then extracted by fi'eezing
and thawing 5x. Assays were performed at 300 water bath by mixing equal
volume of substrate proteins and S30 from inhibited and uninhibited cells. Samples
were taken at various time points and analyzed in 12.5% SDS Page. The results
were visualized by autoradiography with Kodak X-ray film.

In vitro complementation assay: To prepare cell free extract, the isogenic
strains JM1011it0 and JM1011it(con) harboring pUC84PZ1 and pTTQl8lit
respectively, were grown in 50 ml of tryptone broth, supplemented with ampicillin
at 37°C overnight. The overnight cultures were then diluted into 250 ml of fresh
media and grown at 300 until an 0D625mu of 0.4, and IPTG was added to 1 mM
to induce gol activity from the lacZ promoter. 30 minutes afier induction, the cells
were centrifuged 5K for 10 minutes and washed once with cold saline solution and
resuspended in 3 ml of resuspension buffer ( lOmM Tris-acetate, pH 8.2, 60 mM
KC], 14 mM MgoAc, to which 1 mM DTT was added fresh). The cells were then
sonicated for three cycles and centrifuged 10K for 10 minutes to remove cell
debris. The supernatant was then aliquoted and quickly frozen in acetone and dry
ice before storing in a 40°C freezer. The in vitro complementation assays were
done by mixing the two cell extracts, one made from cells overproducing Lit
protein, the other made from cells carrying a gol fusion clone. Each cell extract
was used separately as a control. All the in vitro cell extract assays were
performed in a 30°C water bath, and appropriate amounts of samples were
withdrawn at different time points, and the reaction was stopped by an addition of
equal volumes of 2x loading buffer. The samples were boiled for 3 minutes and
applied to 12.5% polyacrylamide SDS PAGE and electrophoresed 16 hours at 80V
(19). Chemically synthesized Gol polypeptides fiom Chiron inc. are dissolved in

70

distilled water to make up a stock solution of 2uM. 30 ul of "Lit" cell extract is
mixed with a various amount of 601 polypeptide and the mixture is incubated for
30 min. and processed as described above.

Computer Aide analysis: The functional domain of gpLit was searched by
the “ Motif ” program of the Wisconsin GCG protein analysis package. The
homology search of the cleavage site in EF-Tus and EF- la was done according the

database of the Gene probe program.

Results

Cleavage of EF-Tu in an inhibited S30 cell extract: Blockage of cellular
translation by the gol-lit initiated exclusion system requires two major factors: one
is production of the Lit protein, and the other is go] region expression in the gene
23 frame. The use of a recombinant plasmid with the gol sequence expressed
under control of the lacZ promoter enabled us to investigate the biological cellular
defect without dealing with the complexity of phage infection. In a previous paper
we have presented evidence that a S30 cell fraction made from cells harboring
both lit and go] clones in two compatible vectors failed to incorporate radioactive-
labeled arrrino acid in in vitro translation experiments. The translational defect of
the inhibited extract (IPTG-induced) was correlated with cleavage of the
predominant protein EF-Tu, which is responsible for bringing aa-tRNAs to
ribosomes during polypeptide synthesis.

This observation led to a way of testing the specific protease function in the
830 cell fractions. In order to distinguish added EF-Tu from the already existing
proteins in the preparation, and to differentiate the relative levels of proteolysis in
between the inhibited and uninhibited extracts, the substrate proteins were
radioactive-labeled in cells containing a tqu plasmid, which expresses an excess

of EF-Tu. An equal amount of the labeled proteins were added to the inhibited and

71

uninhibited S30 extracts respectively. In the autoradiogram of Fig] (compare
Lane 1 and 5) , we saw only about one half of as much intact EF-Tus in the
inhibited cell extract compared to the uninhibited one, but no difl‘erence in the
amount of other proteins. Thus, the EF-Tu protease was solely in the induced
(inhibited) extract. We could not detect the cleavage fragment. Perhaps it is
unstable under these conditions.

Of course, one can argue that the dirnunition of EF—Tu might have resulted
from the cessation of cellular translation instead of from a gol-Lit activated
proteolytic frmction. To eliminate this possibility, a similar assay was done with a
chloramphenicol-treated cell extract, in which the translation process was severely
blocked. The result showed no reduction in the level of EF-Tu (shown in Fig 2).
Therefore it is unlikely that EF-Tu cleavage is the consequence of translational
deficiency, otherwise we should expect to see a similar effect when translation is
blocked in other ways, such as the chloramphenicol experiment.

Characterization of the EF-Tu protease: The use of this assay system
enabled us to characterize this enzyme in a few general aspects. First, we found
that catalytic activity of the protease was impaired in an alkaline condition but not
affected in an acidic one. At pH=10, there was no obvious difference in the
amount of EF-Tu substrate proteins after normalizing it with the amount of protein
loaded. (Fig 3A), whereas in the buffer of pH=4.8 the reduction of EF-Tu
occurred normally in the inhibited S30(Fig 3B). Secondly, in an attempt to
examine the possibility that the enzymatic function activated by gol-Lit interaction
is via a phosphorylation mechanism, we treated the S30 extracts with phosphatase
to counteract a possible kinase, prior to mixing with the labeled substrate. The
preliminary result showed that the treatment did not abolish the proteolytic
reactions, implying the unlikeness of phosphorylation involvement ( Fig 3C ). It

should be noted that, unlike the in vivo assays, all of the experiments presented

72

above only allowed us to detect the protease activity by the fact of decreased
amount of intact EF-Tu (43kD) but not by the appearance of the cleaved fragment
(37kD), which we usually can detect in vivo. This could be explained by its
instability in the incubated cell extracts as mentioned above.

Cleavage of EF—Tu' can be generated in an in vitro complementation
assay: To further characterize the EF-Tu proteolytic event and learn more about
the roles of go] and Lit , we devised a simple in vitro complementation system.
Instead of using purified components, the in vitro assay was performed by mixing
two cell-free extracts; one provides Lit protein from the pTTQlSlit transformed
Lit(con) cells and the other provides the gol function from the isogenic Lit0 strain
carrying a pUC84PZl clone expressing the gol region . The results shown in figure
4 indicated that the EF-Tu did get cleaved by mixing both extracts (Lane2), but
not by either one alone(Lane4 and Lane6). Interesting enough, in this
complementation experiment we not only observe the reduction of intact EF-Tu
but also can detect the cleaved fragment at the position of 37kD. So far we have
no explanation for this difference.

An independent assay showed that the reaction happened very rapidly. We
saw the presence of the cleaved fragment about 2 nrin. after mixing both extracts
. The result supported the idea that EF-Tu cleavage is the cause of translation
dysfunction instead of vice versa. From Commassie Blue stained gels we detected
no other proteins being cleaved than EF-Tu (data not shown). However, we cannot
rule out the possibility that some minor proteins are subject to the same protease.
The specificity of EF-Tu was tested on its substrate eukaryotic counterpart EF-la
(about 53 kD) using the same in vitro mixing assay plus purified EF -1a from
rabbit reticulocytes (obtained from Dr. William C. Merrick). The EF-lor failed to
be cleaved irnpling the specificity of EF-Tu protease, even though these two share

73

extensive homology in the target site microenvironment. Alternatively, some
modification of EF-lot may prevent it from being cleaved.

Which is the protease? Lit protein or gol?: A key challenge is to identify
the EF-Tu protease. Overproduction of either the gpLit or gol sequence (RNA
or/and peptide)alone did not cause cleavage of EF-Tu in cells. The proteolytic
reaction depends on the presence of the two components, go] and pLit. Therefore
it is reasonable to propose that one of the key factors actually exerts enzymatic
frmction, while the other activates it somehow.

A computer searCh for a functional motif of Lit protein revealed that pLit
possesses a common pattern of primary structure (shown in Fig 5) which is also
shared by a family of neutral zinc-dependent metallopeptidases composed of at
least 22 members. This suggested that Lit is the EF-Tu protease. However, to
convince ourselves that the protein itself is indeed directly involved as an
enzyme, the purification of the Lit protein and evaluation of its catalytic role in
vitro is necessary.

Gol polypeptide is required for the proteolytic reaction: If Lit does play a
protease role, it can not fimction unless a cofactor is provided from extracts made
of the isogenic litO strain with a ”gol" clone in it. From analysis of several fi’ame
shift and nonsense mutations in that region, it is clear that, to exhibit its maximum
effect, translation of the region in the same frame as in gene 23 is crucial. These
results suggest that the go] peptide sequence is the active component.
Nevertheless, we cannot rule out the possibility that the action of translation itself
instead of the product causes the effect of go]. For example, the translating
ribosome could release activity by opening up the secondary structure of gol RNA.

The likelihood of RNA vs. polypeptide activity was tested by a RNase-
treating the go] extract. The rational behind this experiment is that if the gol RNA

molecule is responsible for the activation, we might expect that removal of RNAs

74

from the extract by RNase digestion would prevent the activation of the protease.
Fig 6 ( Lane8 ) showed that pre-RNase incubation of the ”gol” cell extract,
thereby eliminating RNA species prior to mixing with the "Lit" cell extract, still
retained the proteolytic activity. The reduction of RN As resulting from RNase
treatment in "gol" extract was significant compared to the proteolytic activity (data
not shown). Further experiments showed that addition of full length gol RNA
made from a T7 system, to the "Lit” extract, failed to exhibit the catalytic activity
(data not shown). These two lines of evidence argued against the RNA model but
favored the idea that polypeptide made from the go] region could be the key
activator. Recently, in a attempt to confirm the role of Gol polypeptide, we added
the chemically synthesized Gol oligopeptide (29 amino acid long made from
Chiron Inc.) to the "Lit" extracts to investigate its function in EF-Tu proteolysis. In
Fig 7, Lane 3-6 is shown an SDS PAGE gel of the proteins after adding various
amount of Go] polypeptide to the "Lit " extracts. More than 95% of the EF-Tu was
cleaved even in the presence of the lowest concentration ( Lane 3, about 0.025uM)
of the 601 peptides. The proteolytic reaction also generated the 37 kD fragment
with the same mobility as the cleavage fragment in control experiment in Lane 1,
obtained by mixing the "Lit" and "gol" cell extracts. This dramatic effect is“
detected neither in "Lit " extracts alone (Lane 2) nor in the presence of 1% non-
ionic detergent NP4O (Lane 7). The result clearly indicates that Go] polypeptide is
the activator for the function of EF-Tu protease.

Characterization of gol mutations: The minimal nucleotide length of gol
for its activity encodes a 25 aa long polypeptide. The sequence does not have any
functional motifs, nor does it contain highly charged residues. In fact, the molecule
is quite hydrophobic, with 9 nonpolar groups out of 25 . Fig 7 shows the wild type
gol peptide sequence and several mutant sequences with a single base substitution

in nucleotide sequences that abolished the go] phenotype, by the criterion of

75

whether a plasmid construct containing the sequence was capable of transforming
Lit(con) cells . Based on whether their mutations can be rescued in phages to
overcome Litcon phenotype we can classify them into two major classes, those
that can not be crossed back into phage T4 to cause the go] mutant phenotype and
those that can. The first class includes six independently isolated mutations in four
different positions where changes are apparently lethal to the phage. For example,
NTGl and MTD 8 both created a stop codon in gene23 that would be lethal
because the gene 23 product is required for T4 maturation. The rest of the group
may either generate an unacceptable amino acid or the substitution can not inhibit
its wild type effect in the context of a full length ng3 molecule. The class H
mutations overcome the go] effect and allow the recombinant phage to grow in
lit(con) cells; presumably the amino acids are not essential for gene23 function.
We assume that the T4 phages carrying this type of non-lethal mutation prevent
the cleavage of EF-Tu. So far we have only tested two of them, gol6B and
MTD16; both change a methionine into threonine and isoleucine respectively and

fail to cleave EF-Tu (data not shown).

Discussion

In this paper, we demonstrated cleavage of EF-Tu in vitro and
complementation of the protease activity with the mixture of the “ Lit “ and the
“gol” cell extracts. We also presented evidence that the Go] polypeptide activates
the proteolytic function of Lit protein to cleave EF-Tu. The mode for activating the
protease will not be fully understood until experiments with purified components
are done. However, the activity itself holds equal interest as a highly specific
action on translation elongation factor Tu, which has at least 70% homology in
protein sequences with its functional counterparts in all prokaryotic and some

lower eukaryotic species.

76

EF-Tu, a single polypeptide chain of molecular weight 43 kD, is the most
abundant of cellular proteins, about 5% of the total (7). When bound to GTP, EF-
Tu is in the "on" state to carry amino acid-tRNAs to the ribosome A site during
protein synthesis and is released from the ribosome-RNA complex following
GTP hydrolysis. Its abundance makes it the most prominent component of E. coli.
In addition to elongating polypeptides, EF-Tu has other functions, including
serving as one of the four subunits in RNA phage QB replicase (8). A regulatory
role in rRNA synthesis has also been proposed but the specific mechanism has not
yet been characterized (9,10).

As the best known ubiquitous guanosine nucleotide binding protein with
counterparts in both prokaryotes and eucaryotes, EF-Tu has attracted much
attention fi'om researchers, particularly for the investigation of function-structure
relationships. Its 393 a peptide chain is folded into three distinct domains. The N
terminal GTP binding domain (residues 1-200) has structural features common to a
family of ras oncogene p21 proteins. The comparison of primary amino acid
sequences, homology and tertiary structural similarities between EF-Tu and H-ras
p21 has been well described (11,12).

The cleavage site (residues 59-60), together with the preferential uypsin
cleavage site (residues 58-59) and methylation site (residue 56) reflect that in the
three dimensional structure of EF—Tu the cleaved region is probably exposed on
the surface. This prediction is also supported by a computer aided simulation of
EF-Tu protein structure, which indicates this region has a high water accessibility.

Our results suggested that EF-Tu may exist in two forms and one is subject
to proteolysis. From Figures 1, 4 and 6, the low mobility of remaining EF-Tus
might be due to modification. One of the known modifications in EF-Tu is
methylation and the only known methylation site is Ly556 which is four amino

acids away from the cleavage site. Our preliminary methylation results also

77

indicated that the H3 methylated EF—Tu did run slower than the S35 labeled EF-Tu
( data not shown ). Following current thought about the influence of methylation
in EF-Tu proteolysis, it is possible that the addition of methyl groups on residue
56 lysine may reduce the accessibility of incoming proteases to the target site, thus
conferring resistance to cleavage (the influence of methylation does affect the rate
of trypsin degradation in EF-Tu) (21). The best way to prove this is to isolate the
remaining EF-Tus from the gel and use mass-spectrometry to examine the
methylation status of the protein. This hypothesis, if true, can also help account for
the failure of the protease to cleave EF- 1a, since it is highly methylated when
purified from rabbit reticulocytes. But one thing that needs to be reconciled is the
dilemma based on the recent result that the majority of EF—Tu did get cleaved (Fig
7) by adding the purified 601 into the same "Lit" extracts in vitro. It is possible
that an excess of Gol oligopeptides can overcome the hindrance caused by
methylation However, more experiments need to be done to elucidate this point.
The observation that some amount of intact EF-Tu was still present in the in
vitro assays are consistent with what we knew from previous studies that EF-Tu
can somehow retain resistance to the go] -induced protease when cells were
grown in poor conditions, such as in minimal M9+Bl labeling media. Apparently,
there is a correlation between cell physiological conditions and the resistance of
EF-Tu to protease. This discrepancy between them could be explained by the EF-
Tu methylation which was discussed above. Several researchers have observed
that the level of EF-Tu methylation increased in cells grown in media depleted of
carbon , nitrogen or phosphorus source, and cells of different developmental stages
such as in mycelia form of the fungus Mucous racemosus but not in the spores.
The vitro extract assay which exhibits EF-Tu protease activity may enable us to
characterize the EF-Tu of cells in different growth environments and

developmental stages by the criterion of its sensitivity to cleavage.

78

The EF-Tu proteolysis by the potential protease gpLit and its gol peptide
cofactor not only reveals the molecular mechanism of phage exclusion but also
provides an excellent system 'to exploit the foundations of protein-protein
interactions for an oligopeptide activated catalytic reaction. A short peptide
activating an enzyme is not a novelty. Mange] and Anderson (22) have reported
that a proteinase activity of hmnan Adenovirus (AD2) requires two cofactors. One
of them was found to be a 11 aa long oligopeptide with high negatively charged
density. The sequence was present at the carboxyl terminal of pIV, the 250 a
precursor protein to a virion component. The details of the activation mechanism
were not discussed . It is possible that these two systems may work in a similar
fashion.

Our level of understanding of the phage T4 exclusion phenomenon by a
cryptic DNA element E14 encoded Lit protein is already sufficient to draw a clear
and reasonable connection between the host-parasite gene interaction and the
occurrence of cellular dysfunction and tie into the observable abortive infection
phenotype. What remains to be answered are many questions of fundamental
importance: Is any specific secondary structure involved? When does EF-Tu get
cleaved, since it usually associated with other translation factors (GTP, GDP, aa-
tRNAs, EF-Ts and ribosomal components)? What is the significance of EF-Tu
methylation when cells encounter stressful conditions? Does the cleaved form of
EF—Tu retain any functions? Is it possible that the 37kD cleaved product plays a
similar role to methylated EF-Tu, since phage infection is a dreadful circumstance
to cells? We are not sure that every question will have defined answers, but they

definitely give scope for imagination.

79

References

1.Duckworth, D. H., J. Glenn, and D. J. Mccoquodale. 1981. Inhibition of
bacteriphage replication by extrachromosomal genetic elememt. Microbiol.
Review 45:52-71

2.Molineux, I. J ., 1991. Host-parasite interactions: recent development in the
genetics of abortive infections. The New Biologist 3:230

3.Kao, C.and L. Snyder. 1988. The lit gene product which blocks bacteriophage
T4 late gene expression is a membrame protein encoded by a cryptic DNA
elememt,e14. J.of Bactriol. 170:2056-2062

4.Kao, C., E. Gumbs and L. Snyder. 1987. Cloning and characterization of the
Echerichia coli lit gene,which blocks bacteriophage T4 late gene expression. J
of Bacteroiol. 169: 1232- 1238.

5.Yu, Y.-T. N. and L. Snyder. 1993. A Phage exclussion duo cleavage of
translation elongatio factOr Tu. Submmited.

6.Bergaland K. J., C. Kao, Y.-T. N.Yu, R. Gulati and L. Snyder. 1990. A site in ,
the T4 bacteriophage major head protein gene that can promote the inhibition of
all translation in Escherichia coli. J .Mol.Biol. 213:477-496.

7 Jacobson, G. R., and J. P. Rosenburch. 1976. Abundance and membrame
association of elongation factor Tu in E. coli. Nature (London)361:23-26.

8.Blumenthal, T. 1979. RNA replication : Function and structure of Qb-replicase.
Ann.Rev.Biochem. 48:525-548.

9.Travers, A. A., R I Karnen and R. F. Schleif. 1970. Factor necessary for
ribosomal RNA synthesis. Nature(London) 228:748-751.

10.Haseltine, W. 1972. Nature(London) 235:329-333.

11.Jurnak, F. 1985. Structure of the GDP domain of EF—Tu and Location of the
amino acids homologous to ras oncogene protein. Science 230:32—36.

12.Valencia, A.., M. Kjeldgaard, E. F. Pai,and C. Sander. 1991. GTPase domains
of ras p21 oncogene protein and elongation factor TuzAnalysis of three-
dimensional structures,sequence families,and function sites. Proc.Natl.Acad.Sci.
USA 88:5443-5447. ‘

8O

13.Arai, K.. et al. 1980. Primary structure of elongation factor Tu fi'om
Escherichia coli. Proc.Natl.Acad.Sci. USA 77: 1326-1330.

14.Young, C. C, J. D. Alvarez, and R. W. Bemlohr. 1990. Nutrition-dependent
methylation of a membrame-associated protein of Escherichia coli. J .of
bacteriol. 172:5147-5153.

15. Young, C. C. and R. W. Bemlohr. 1991. Elongation factor Tu is methylated in
response to Nutrient deprivation in Escherichia coli. J. Bacteriol. 173:3096-
3100.

16.Greener, A., and C. W. Hill. 1980. Identification of a noval genetic elememt in
Escherichia coli K-12. J .of Bacteriol. 144:312-321.

17.Brody, H., A. Greener, and C. W. Hill. 1985. Excision and reintegration of the
Escherichia coli chromosomal elememt e 14. J .of Bacteriol. 161:1112-1117.

l8.Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor
Press, Cold Spring Harbor,N.Y.

19.Laemrnli, U. K. 1970. Cleavage of structure proteins during assembly of the
head of bacterophage T4. Nature (London) 227:680—685.

20.Bardwell, J. C. A., etal. 1989. autoregulation of RNAase III operon by MRNA
processing. The EMBO 813401-3407.

21.Toledo, H., and C. A. Jerez. 1989. Methylation of elongation factor EF-Tu
affects the rate of trypsin degradation and tRNA-dependent GTp hydrolysis.
Febs. Lett. 252:37-41.

22.Mangel, W. F., W. J. Macgratth, D.L.Toledo, and C.W.Anderson. 1992. Viral
DNA and a viral peptide can act as cofactor of adenovirus virion proteinase
activity. Nature(London)

23.Sherman, M., and P. S. Sypherd. 1989. Role of Lysine methylation in the
activities of elongation factor 1a. Arch.Biochem.Biophys 275:371-378.

81

Table

Table l: The characteristics and references of bacterial strains, plasmid constructs

and phage mutants used in this acticle.

82

 

 

E

kaie
«3
age
m~§

55% .. BE.

5%

25:3.
3.3.2..

cgfimnl—z

23:83....

EB.“

dun:— .3

A 93. 93. 9.. 2.1.”. 5+. 55...... 36>? TEN Z;
:52 :7... 5... Ragga." 9.3.2.... 2. a: 9.5:— v.2...
hiss =55... . 33333533.. 3.. E...

:53 :8. 2.3.. can: 5. :< mafia—2...:

.§=% .Mbeafl 9:3. ». Emu 3...: .3 ”2.» 19:32.3...
.53.... Pen—2m 3.3.5.2..

Ge 5. 5.2.... cut”... 3...... n..— nnnme= =6... 32...; .e _
9:23.. r. 3...... EnN =33...

Pu E. Mae—2. ea 2.83 z. «.2... 23.2— .= 13.0;
.532. :4... EeN $2.353...

Mae—2-52:: me. wanna: aw ccnmfiuN. $.52. .55
nee—2 .53.: = 2:3: 9.. wiggle—Km. mug—gash a...
”2) :5... .3 13:522..

mega.» 5..."...2: 9.2.8.5..“ me. Raw... 5...... cCQXfiN—
223.. m: 9.5:— mmz en emu—42w. 23.2%... .e 5...... me—
.e22.e. EaN 93...... x2?

wee-.9335; .2333 5.5:: «35.8 5:52.... 5.3%
2.3.... R?! r. :28... £3.52.

«winging... 2... Sam
2.? ten—n
3...» $5....
2.... :5...»

13.32... 5...

wanna—5.... ”:2...
=..... $5....

:1.» $5....

=5. 3......»

Q_.=.:_.=aaa.fi e...— mamas»?

 

 

83

Figure Captions

Fig l: Cleavage of EF-Tu in the inhibited $30 extract. The S35-labeled protein
substrate were added to an equal volume of inhibited S30 (Lanes 1, L 3, 4) and
uninhibited S30 (Lanes 5, 6, 7, 8) and incubated at 300 water bath. Samples were
taken at different time points, 10‘ (Lanes 1 and 5), 30' (Lanes 2 and 6), 50' (Lanes
3 and 7), and 70' (Lanes 4 and 8). The samples were applied to 12.5% SDS Page,

and the autoradiogram was visualized with X-ray film.

 

08$ PGIIQNUI

ocs panqmurun

 

 

What
numb»: ‘-

I
ma! W F. "3'.

"1213

T
:lnlléll 61
Hal”! 11
tilllill ll
{MINI ll
nut”! ll
‘f‘llifl‘ll fl
illillill 9|
Hill 11 ‘l

"1'53

85

Fig2: Chloramphenicol-treated cell extract did not cleave EF-Tu. Cells harboring
both pTTQlit and pUC84PZ1 (strain #582) were grew to OD=O.4 and treated with
chlorrarnphenicol to 300ug/ml to block translation. The cell extract was made by
the method of sonication and used in the in vitro assay (see material and method).
Lane 3 and 4 showed the autoradiogram of substrate proteins taken from the
mixture at 20 (Lane 3) and 40 min. (Lane 4). Controls were run side by side in
Lanes 1 and .2 with cell extract made from un-induced (uninhibited) cells by the

same procedure.

86

; : mztl mar
~  ”WIFE?!
. rs; " 1!! “'{H‘3
é 53' ' ii
11?“ IN:
1

 

 

"133

87

Fig 3: Effect of buffer pH value and phosphatase on the EF-Tu proteolytic
activity. Panel A showed autoradiogram of substrate proteins in pH=8.2 (Lanes 1
and 2) and in pH=10 (Lanes 3 and 4) at different time points 0 (Lanes 1 and 3) and
30 min. (Lanes 2 and 4). Panel B compared the effect in pH=8.2 (Lanes 1 and 2)
and in pH=4.8 (Lanes 3 and 4). Panel C demonstrated the substrate proteins in
phosphatase-treated cell extracts (Lanes 3 and 4) and untreated control (Lanes 1
and 2). Phosphatase was added to the IPTG-induced #582 cell extract in vitro to
300u/ml. The control was done with the un-induced #582 in the absence of
phosphatase.

 

88

1234

. minim: "
~ I’M]
- [11:212:
- mama:

 

89

Fig 4: Coomassie Blue stained SDS PAGE of in vitro cell extracts assay. Samples
were taken at different time points, 0' (Lanes 1, 3, 5) and 20' min. (Lanes 2, 4, 6,)
during incubation. Lanes 1 and 2: shows the proteins of the mixed cell extracts
containing "Lit" and "gol". Lanes 3 and 4: "gol" cell extract only. Lanes 5 and 6:
"Lit” cell extract only. Arrows show the position of EF-Tu and EF-Tu cleaved
fragment (EF-Tu*). The presence of cleavage product is indicated by the star (‘)
along the lanes of the gel. The molecular weight markers were shown on the left

site of the gel.

cztmllv

:tm Iv

 

91

Fig 5: The Zinc-binding protease motif of Lit protein. The majority of zinc-
dependent metallopeptidases share a common pattern of primary sequence
highlighted in the above. The sequences of Lit protein detected by the pattern lie
from position 162-171, ILHHEISHVV. From the study of thermolysin, that
belongs to this family, two of the zinc ligands are histidines that are very close
together in the sequences. The glutamic acid following the histidine dimer acts as a
nucleophile and promote the attack of a H20 molecule. A signature pattern
including the two histidines and the glutamic acid is sufficient to detect the

proteases of this family.

2:

mx
1h

1
0!

ill
h:

53

m:

92

fiOTIFS from: lit:rn.pep

Xismatches: 0

January 26. '993

.aa
’J
bl
{J

Littrn.?ep Check: .06? Length: 297 2 REFORflAI of: Littrn.?ep check: 4067

 

 

Zinc_Protease (S.T.A.2.?)x2§E(L.I.V.3.?.Y.9)-(D.E.H.R.K.P)HX(L.I.V.3.F.Y.V.Q)
(I)xl2:HttI)-(D.t.a.R.K.P)Hx(V)
157: q§;au riuurrsavv
ma;;.-a.‘..‘.saauss“2-2;;;---_..--x;wh:m*mm

J—L‘L-

m- an . a
I““““-ii‘KIXAOIIXiN ‘‘‘‘‘‘‘‘‘‘‘

* Neutral zinc metallopeptidases. tine-binding r

egion signature *

.....
K‘I‘Iw “““““““““““ n

a

.......
Ki. KKKKKK “I KKKKKKK

 

Thermostable thermolysins (EC 3.4.24.4).
proteases from various species of Bacillus.
Hammalian neutral endopeptidase (EC 3.4.24.11) (NE?) (enkephalinase).
Hammalian extracellular matrix metalloproteinases (EC 3.4.24.7) (known as
natrixins) i3]: HEP-l (interstitial collagenase). any-2 (72 Rd gelatinase),
HEP-9 (92 Rd gelatinase). SHE-7 (pump-l). HEP-8 (neutrophil collagenase).
HHP-3 (stromelysin-l), axe-10 (stromelysin-Z). and.8HP-ll (stromelysin-B).

iotensin-converting entyme (EC 3.4.15.1) (dipeptidyl carboxypeptidase I)
'(ACE) the enzyme responsible

for hydrolyting angiotensin I to angiotensin

II. There are two forms of ACE: a testis-specific isozyme and a somatic
isocyme which has two active centers {4}.

Extracellular zinc-metalloprotease from Serratia.

Secreted proteases B and C from Erwinia chrysantheni.

Extracellular elastase from Pseudomonas aeruginosa (gene lass).
'Extracellular proteinase proA from Legionella pneumophila [S].

Immune inhibitor A from 3acillus thuringiensis (gene ina) [6]. Ina degrades
two classes of insect antibacterial proteins. attacias and cecropins.

Cell surface protease (surface glycoprotein 5963) from various species of
Leishmania.
Extracellular neutral metalloprotease from Streptomyces cacaoi.

Astacin. a crayfish endoprotease.

Heprin A. a mouse kidney brush border netalloendopeptidase [7].
PABA-peptide hydrolase. a human intestine membrane protease [7].

Bone morphogenic protein 1 (BHP-l). a protein which induces cartilage and
bone formation and'which expresses metalloendopeptidase activity [71 and a
Drosophila homolog. the dorsal-ventral patterning colloid protein.
Hemorrhagic metalloproteinases from snakes venom [8}. ‘
Sea urchin hatching enzyme (envelysin) (EC 3.4.24.12) [9]. A protease that

allows the embryo to digest the procective envelope derived from the egg
extracellular matrix.

Hammalian aminopeptidase N (EC 3.6.11.2).
Escherichia coli aminopeptidase N (gene pop").

and related thermolabile neutral

BP-l/6C3 antigen [10], an early B-lingggg_cells phggphgrylated cell surface

glycoprotein which is most probably a zinc metallopeptidase.
role in regulating growth and differentiatio
Kali blood group glycoprotein [ll]
erythrocytes. The Kell protein is v

It may play a
n of early B-lineage cells.

. a major antigenic protein of human
ery probably a zinc endopeptidase.

Leukotriene a-A hydrolase (EC 3.3.2.6). is the enzyme responsible for the
hydrolysis of an epoxide moiety of LIA-6 to form LIB-4. It has been shown
[12] that this enzyme binds tine and is capable of peptidase activity.

from

LQHPL

93

Fig 6: Coomassie Blue stained SDS PAGE of RNase-treated cell extracts. Samples
were taken at different time points, 0' (Lanes 1, 3, 5, 7,) and 25' (Lanes 2, 4, 6, 8,).
Lanes 1 and 2 : positive control of EF-Tu cleavage by mixing ”Lit” and "gol".
Lanes 3 and 4: negative control of EF—Tu cleavage by incubating ”Lit" alone.
Lanes 5 and 6 :Pre-incubate "gol" 30' prior to mixing with ”Lit”, serving as a
parallel control for Lanes 7 and 8. Lanes 7 and 8: "gol" was pre-irrcubated with
RNase (0. lu\ ul) for 30' prior to mixing with "Lit". The activity of EF—Tu protease
is reflected by the presence of EF-Tu“ shown by the star (*) signs along lanes of
the gel.

94

3h. .6 I

«Nm :0 I
MNm Eu I

Susanna: :30

o r:

o no.

a no. 30.2.2.an
22:: 32mm:

 

95

Fig 7: Activation of the EF-Tu protease by the Col polypeptide. Lane 1 shows
cleavage of EF-Tu by mixing the "Lit" and "gol" cell extracts serving as a positive
control. Lane 2 contains the "Lit" extracts alone. Lanes 3-6 indicate cleavage of
EF—Tu and generation of the cleavage fragment (indicated by * signs) by adding
various amounts of purified Go] peptide: 0.5x (Lane 3), 1x (Lane 4), 2x (Lane 5)
and 5x (Lane 6), respectively. Lane 7 contains the "Lit” and ”gol" extracts and 1%
NP40. IX of the Go] petide is equal to 0.05uM . The position of the intact EF-Tu
(43 kD) is indicated by arrows.

49.5 kD —
EF-Tu <————

EF-Tu*<————

32.5kD —

27.5 kD —

clJt
cgol
Gol

96

 

——> EF-Tu*

97

Fig 8: The amino acid sequences of wild type gol and its mutants: All the mutants
were selected by the criterion of their plasmid constructs capable of transforming
Lit (con) cells (indicated by "+" sign). The peptide sequence (represented by single
letter) is deduced from the nucleic acid sequences of the Pqu- Ball fragment of
go] region and mutations are identified by DNA sequencing data. MTD8 and
NTGl single base substitution generated a stop codon at residues C and Q,
respectively. The dashed line represents the same amino acid residue. Class I
includes mutations that can't be crossed back into the phage to give the gol mutant
phenotype. Class II contains mutations that can be crossed back into phage and

allow phage grow in Lit(con) cells.

98

3.. no—

23m

2&0.

2.2.3..

Zfifia

ZHD—c

E;

5;

mafia

><gfiz<ww>nw2mt~>wcunnu<ofizz

 

 

 

 

 

 

 

 

 

 

 

:H in”- use

- .............. miti:
....... - R.---......
- - -- ............. Z -
....... a, ......... t... ........ . ........
- - m ............ ii

.............. ~
- Ar...

 

 

an
we
21
o
and
Q
I
llll
nd

 

Ogmm

=

=

=

5233

.25

CHAPTER 4
MOLECULAR BASIS FOR THE LOCAL INHIBITION

CAUSED BY COL-LIT INTERACTION AND THE
ROLE OF THE T4 PHAGE GOL SITE

99

100

Abstract

Transcriptional termination in the T4 gol site, blocking the expression of
downstream genes, occurs in the presence of a wild-type, low amount of the Lit
protein. This gol-Lit mediated local inhibition requires translation of the go] region
for its full activity. However, the gol region, which contains a nut-like
antitermination sequence, enhances the level of transcription in the absence of Lit,
suggesting that the function of the gol sequence maybe exert an antipausing eflect
and increase the transcription of gene23. Here, we also propose a model to

correlate the function and structure of the gol sequences.

Introduction

The cryptic prophage e14 (1) is integrated within the icd gene (2) located at
25' of the Escherichia coli K chromosome. The cryptic prophage generates a
protein , Lit, (3) that can inhibit bacteriophage T4 late gene expression at the level
of tanslation, particularly when it is overproduced by an up-promoter mutation or
made from a multiple-copy plasmid (4, 5). The inhibition is due to the interaction
of Lit with a short sequence , gol, within the phage major capsid protein Gene23
(6). Transformation of Lit-overproducing cells with a plasmid , expressing the gol
region in the Gene23 fiame, can generate a global translational defect similar to
the phage exclusion. No other T4 sequences are required .

The minimal length of gol region was determined to be a 75m fragment
located about 300m downstream of the Gene23 ribosome binding site (4). We
have shown in the previous paper (7, Yu and Snyder submitted) that the gol-Lit
interaction somehow caused cleavage of translational elongation factor Tu (27),
thereby blocking cellular translation. In addition to this dramatic global effect,
Bersgland et al (4) also observed another effect of the gol-Lit interaction.

Transformants of a gol-LacZ construct that block translation in a Lit-

lOl

overproducing strain (Litcon) are normally blue on X-gal plates in the absence of
Lit protein, whereas in the presence of a wild-type low amount of Lit protein, the
transformants are white, indicating no functional LacZ is made. We referred to
this as local inhibition since only genes downstream of the go] region and
expressed from the same promoter are affected. The previous study (Bergsland et
al 1990) suggested that the local inhibition acts at the level of transcription and
only requires wild-type sequences of gol RNA.

In this report, we focus on reevaluating the molecular mechanism of this
local effect and illustrate the unexpected discovery that go] region has a significant
enhancement effect for downstream gene expression in the absence of Lit protein.
Discussions of the possible linkage between the apparent transcriptional
termination of the local inhibition and the putative antitermination effects in the

same region are also included here.

Materials and Methods

Baterial strains and media: The strains used in this work JmlOl,
JM1011it0, JM1011it+ and litcon were described in the precious paper (8, Yu and
Snyder in preparation). W31 lOIq (#539) is a strong laclq recA E.coli K strain (9).
Unless otherwise mentioned, Transformants for the IPTG induction were grown in
Tryptone Broth (1% tryptone, 1%NaCl, 50ug/ml vitamin B1) supplemented with
the appropriated antibiotics. Supplements for media were at the following
concentration: ampicillin (Amp), 50ug/mi; chloramphenicol(Cm), 25ug/ml;
tetraclycin,12.5ug/ml; IPTG, lmM; X-gal, 50ug/ml. All the experiments were
performed in a 30°C water bath or incubator.

DNA manipulations and plasmid constructs: All the endonucleases used
for plasmid cloning were obtained from BMB and used with the buffer and
conditions recommended by BMB. DNA fragments of interest were isolated from

102

agarose gels by the method of squeezing and thawing followed by several cycles
of phenol extraction followed by ethanol precipitation. The DNAs were
resuspended in H20. The vector M13mp18 and mp19 were obtained from BMB
(10). The plasmid pSP64 was a gift of Dr. Richard Schwartz (1 l). The pTA9 that
contains an EF-Tu structural gene under control of lac, T7 and its own promoters
was a generous gift of Dr. David Miller (12). The construction of the plasmids
used in this article are described as the following. The pUC84PZ1(4) has a 159 bp
Pstl- 1 (HindIII linker) fragment of the go] region cloned into pUC84, that has a 4
bp filled in BamI-II site. The pUCgol-cat contains a pvuII-Xhol 75 bp fiagment of
the go] region from pUClZPZXR (4) at the Smal and Sall sites and a CAT gene at
the HindIII site of pUC8 vector. The pUCUAAgol-cat is like pUCgol-cat, in that it
contains the same gol region and CAT segments but has a 4bp filled in EcoRI site
that generates a UAA ochre stop codon upstream of the gol region. All three
plasmids described above translate the gol region in gene23's flame and CAT in
the Cat ribosome binding site (RBS) from the same transcript made from lac
promoter. The pUC 12-cat served as a control without the go] fragment. The
M13PZl contains the Pl] fragment from pUC84PZl cloned into EcoRl and
HindIII sites of M13mp18 vector. With the same scheme, MTD16 that has a
single-base substitution in the gol region was cloned into M13mp18 to make
M13MTD16(8). Both the M13 constructs transcribed the hybrid genes from a lac
promoter and confer a-complementation when the phages plate on the E. coli
male strains JmlOl litO- The pSP64PZ1-JD has a EcoRI-HindIII PZl fragment
cloned at the opposite orientation at the HindIII and EcoRl sites and a 110 bp
chicken histone gene segment at the Pstl site of pSP64. The pACYCtqu(R) and
pACYCtqu(W) were made by cloning the Pqu segment from the plasmid pTA9
into the Smal site of the Kan resistance cassette in a pACYC derivative that

contains the Karir cassette (obtained from Pharmacia) in the EcoRI site of

103

pACYC184. The R and W stand for clockwise and counterclockwise transcription
of tqu relative to the Kant gene's promoter, respectively.

Plating with M13-gal hybrid phages:M13PZl, M13MTD16 along with
the vector M13mp18 were introduced into the male E. coli strains via transfection
of competent cells (13). Transfectants were plated on LB plates with 3ml of top
agar( 1% tryptone, 1%NaCl, 0.5% yeast extracts, 0.7% agar). A single plague was
isolated by toothpick and incubated in 5 ml pre-diluted culture. The hybrid M13
particles were partially purified 6-8 hours post- infection by reserving the
supernatant of centrifugation at 5K for 5min. The M13 RF forms left in the cell
pellets were isolated and tested for inserts. The appropriate E .coli male strains
(lito, lit+, and litcon) grown to OD=0.5. 3 ml of top agar supplemented with
IPTG,X-gal and 0.5m] culture of indicator strains were plated on LB plates and
spotted with various dilution of the M13-go] hybrid phages, and the plates were
incubated at 340-350C. .

RNA protection with S] nuclease: The RNA probe that can hybridize the
P21 segment was labeled uniformly by the method of in vitro transcription (14).,
using a Sall-linearized template of pSP64PZ1-JD and SP6 RNA polymerase.
Cellular RNA samples were extracted by a modification of the method of Young
and Furano (15). A 10ml culture of the pUC84PZ1 transformants was induced by
IPTG for 20 min. and lysed in corex tubes containing 2.5 ml lysis buffer (1M
NaCl, 1%SDS and 50mM EDTA pH 8 ) and incubated in a boiling water bath. The
lysate was cooled down to room temperature and phenol- extracted twice and
ethanol precipitated before dissolving the pellets in 0.3 ml H20. The cellular RNA
samples were hybridized overnight with an excess of the labeled probe at 550C in
30 ul hybridization buffer. (70% formamide, lmM EDTA, 0.5M NaCl, and 20mM
Tris-HCl pH8). The hybrids were chilled on ice and digested for 1 hour at 370C
with S1 nuclease(BMB) in 300ul S1 buffer (7% forrnamide, 0.25M NaCl, 30mM

104

Na-acetate pH 4.5 and lmMZn-sulfate) (16). The ethanol-precipitated pellets were
resuspended in sample buffer and applied to 10% polyacrylamide, 7M urea gels.
The gels were then dried and exposed to a Kodak X-ray film.

CAT ELISA assay: Transformants containing various plasmids were
innoculated into 5 ml fresh tryptone media supplemented with ampicillin and
grown until OD=O.4. Aliquots of cultures were taken and washed once with 0.25M
Tris-HCl pH 7.5 , then lysed by freezing and thawing. The cell extracts were then
quickly frozen using a dry ice / ethanol bath and stored at -700C. Protein
concentrations of each sample were determined by the method of Dradford (17).
The amount of CAT in each cell extract was measured by a CAT ELISA assay kit
(BMB) following the procedures recommended by the supplier. The final CAT

concentration was normalized to the total cellular protein .

Results

The use of M13-gal constructs to detect local inhibition: Instead of
analyzing the gol-induced inhibition with the gol-LacZ fusion clone of the pUC
vector, it is more convenient to monitor the gol-Lit initiated effects using the go]
site cloned in an M13 vector. If the entire gol region, the EcoRI-HindIII fragment
of pUC84PZ1, is cloned into the polylinker site of filamentous phage vector,
M13mp18, the fusion construct will transcribe and translate the go] region fused to
lacZ gene as in the pUC system. This strategy had at least two advantages over the
previous assay: It allowed us to screen a large quantity of samples for the desired
gol mutations, and that may help to identify nucleotides essential for the global
and the local effects or both; Additionally, we can directly sequence mutations of
interest without extra cloning procedures.

M13 phage particles containing the proper construct were prepared and

spotted on various indicator strains (Table l). Plates were supplemented with

105

IPTG and X-gal to test (it-complementation of B-galactosidase. The M13-go]
hybrid phage generated blue plaques in cells cured of e14 (no Lit produced, Lit°),
whereas in the wild-type cells (Lit+) the plaques were completely white, and no
plaques were formed in the Lit-overproducing strain (Litcon). The M13mp18
vector served as a control producing blue plaques in all of the tester strains
regardless of the presence or absence of Lit protein. The result agreed with the
expectation of the local inhibition exhibited in Lit+ and the global effect in Litcon.
M13 phages have a poor plating efficiency at 300, at which the gol-Lit mediated
inhibitions were usually detected . However, this plating assay works well at
temperatures of 340-350.

We isolated several gol mutants and localized their changes within the gol
region. One of the mutants, MTD16, generated by a MutD strain , was cloned into
M13 and used in this plating assay as a control to show the depression of the local
and global inhibition by a single- base mutation. It produced equally blue plaques
in all of the indicators. The single-base mutation of MTD16 changes residue
methionine into isoleucine, which is shown in Fig8 of Chapter 3.

The local inhibition occurs at the level of transcription: It is difficult to
interpret the cause of blockage of gene expression in a translational fusion, such as
pUC84PZl, in our previous study. Inhibition of the LacZ expression could be at
the level of transcription or translation. In order to further differentiate these
possibilities, a construct pUC75gol-cat containing the CAT ( chloramphenicol
acetyltransferase) gene at the HindIII site downstream of go] was made. This
plasmid transcribes through the go] region into the CAT gene from an inducible
lacZ promoter, and translates CAT protein independently from its own ribosome
binding site. If the blockage caused by gol-Lit interaction occurs at the level of

transcription, one would expect that no complete messages would be made through

106

the reporter CAT gene, and consequently cells would be sensitive to the
antibiotic, chloramphenicol (Cm).

Transformants of pUC75gol-cat did fail to grow on Cm+IPTG plates in the
presence of a wild-type amount of Lit, but made colonies comparable to the vector
control, pUC-cat, on plates supplemented with ampicillin (Table 2). This result is
significant for two reasons. First, the failure of plating on Cm+IPTG is indeed due
to the local blockage of the CAT gene expression downstream of the gol region,
not the global effect , otherwise one would see a similar phenotype on the
ampicillin plates. Secondly, according to the rationale described above, the data
suggested that the local inhibition is at the level of transcription.

This conclusion received further support from the result of a RNA
protection assay. I used a uniformly labeled probe, in which the first 180 bp is
complementary to the gol transcript of the EcoRI-HindIII fragment in pUC84PZl
and the last 110 bp is chicken histone gene specific, to detect the length and the
amount of protected transcripts from total RNA extracts of the appropriate
transformants. If transcription is not affected by the function of Lit, one should
expect a similar pattern of RNA protection regardless of the presence of Lit. The
data in Figl showed that 20' after IPTG addition to induce transcription of
pUC84PZ1, the amount of message protected by the anti-go] probe decreased
proportionally to the increase of Lit protein in the corresponding transformants
(Lane 2, 3, 4). The length of the protected RNA was virtually the same among all
3 types of host (Lit°, Lit+, Litcon), except that a low concentration of S1 nuclease
revealed a smear of shorter species protected in Litcon cells (Lane 8) as well as in
Lit+ (Lane 9) with a weaker signal. Those shorter bands were not found in the
control (Lane 5), that contained the labeled probe and a excess of tRNAs. This

result suggested that the appearance of the shorter species was due to protection

107

of the truncated transcripts in Lit-containing cells rather than to the incomplete
digestion of the un-hybridized probes.

Translation of the upstream gol region is crucial for the transcriptional
blockage: The transcriptional termination in the go] region could be due to a
polarity effect of blocking translating ribosomes upstream. To examine whether
the putative transcriptional termination in the gol region was due to block of
translation possibly by the gol-Lit interaction, thereby exposing a termination
signal for RNA polymerase, an ochre stop-codon was generated 5' approximal to
the gol region. The logic behind this is that the UAA construct should arrest
ribosome at the EcoRI site and cause a polarity effect in transcription, if a p-
dependent terminator exists in the go] region or beyond. As the data in Table 2
show, pUCUAAgol-cat transformed Lit+ cells were chloramphenicol resistant,
indicating the CAT enzyme was made. Apparently, the entire gol RNA doesn't act
like a terminator in the course of transcription, otherwise the transformants
wouldn't grow on Cm+IPTG plates.

In addition, the Lit protein didn't cause the same local inhibition in the stop-
codon construct as in the translatable equivalent plasmid pUCgol-cat, illustrating
that translation of the gol region is necessary for the function of Lit to block
message synthesis to the downstream CAT gene.

To further quantitatively analyze the CAT reporter gene expression, we
used a CAT ELISA assay to measure the actual amount of CAT products in the
transformants of the various CAT constructs (Fig 3). There was no distinct
difference between CAT production of the vector pUC-cat in the absence and
presence of Lit protein, whereas, in the presence of Lit, pUCgol-cat produced only
0.3% of CAT that was made in Lit0 cells and 2% of that found in absence of the
go] region (pUC-cat). Generally speaking, the results obtained from the ELISA

assay were in agreement with the plating assay.

108

Gal sequence stimulates gene expression in the absence of Lit protein:
Two additional observations emerged from the data of the LitO transformants of
pUCgol-cat and the UAA construct. In the absence of Lit protein , pUCgol-cat
made almost 5 fold more CAT protein than the vector control (pUC-cat). The
elevating effect was apparently abolished when translation of the go] region was
blocked by a up-stream stop codon of the pUCUAAgol-cat. In fact, CAT
production of the UAA exhibited a significant decrease compared to that of the
vector control. Possible explanations for the decrease are that the context
surrounding the cat ribosome binding site could interfere with the untranslated gol
structure, or that RNA polymerase, which has to travel further, could be partially
terminated in the go! region.

We hypothesize that the enhanced CAT production resultes from an
increased rate of transcription by the presence of the upstream gol region. Of
course one could argue that the increase might be due to the function of an
additional promoter in the go] construct instead of a stimulating effect of go] in
transcription. However, if most of the transcription is coming from the lacZ
promoter, then, in the absence of IPTG, we should expect less Cat to be
synthesized. To improve the experiment, we have transformed a strong lac Iq
strain W3110, cured of e14, with the same set of plasmids to monitor the amount
of Cat before and after IPTG induction (Fig 3). The addition of IPTG to induce the
lacZ promoter did cause a significant increase in CAT synthesis of all of the CAT
constructs , implying that the increased transcription utilized the Lac promoter
(comparison of Cat in Fig 3, -IPTG vs. +IPTG). The low amount of CAT produced
in pUCUAAgol-cat, in which the gol sequence is virtually the same as in the
pUCgol-cat, also argued against the possibility that this region contains an

additional promoter.

109

Based on the rationale described above, it is likely that the 8-fold increase
of CAT in the pUCgol-cat transformants before addition of IPTG resulted from an
enhancement of the basal transcription by the presence of the translatable gol
region. This result also explains the observation that LitO transformants of
pUCgol-cat grew much better on Cm + glucose plates than the cells containing the
control alone, obviously the enhanced production of CAT even fiom the basal
transcription of the uninduced Lac promoter supported cell growth on the
chloramphenicol plate. The level of CAT found in transformants of the UAA
construct was consistently about 6-7 times lower than that of control both before
and after IPTG. The RNA structure of the entire gol sequence, that may influence
the efficiency of ribosome binding on the neighbor cat gene RBS, could account
for the reduction of CAT protein, even through it did not exhibit the phenotype of
a transcriptional terminator on the plate assay.

Given the evidence presented above, we concluded that transcription of the
Cat gene comes from a single promoter, plac, and that the increase of CAT
produced in transformants of pUCgol-cat in the absence of Lit is due to a
transcriptional stimulation by the gol sequence. The stimulatory effect also
requires translation of the go] region. Interesting enough, translation of the go]
region also is an important factor to determine the transcription termination in the
local inhibition caused by Lit protein.

Is Cleavage of EF—Tu responsible for the local inhibition ?: When the
majority of cellular EF-Tu gets cleaved, a global inhibition of translation occurs
(7). We wondered whether cleavage of EF-Tu also takes place in wild-type hosts
(Lit+), that only inhibits the transcription of genes downstream of the gol site.
From the comparison of the total protein profile of the isogenic Lit strains (Fig 4)
that contained the gol construct, pUC84PZl, we found that ,40' after IPTG

induction, no more than 5% of intact EF-Tu was left in Litcon (Lane 1, 2, 3) but

110

half of EF-Tu remained intact in the Lit+ cells even after a long period of
induction (Lane 5, 6, 7, 8). The data showed that some EF-Tu did get cleaved in
the presence of a wild-type amount of Lit protein. Apparently the degree of EF-Tu
cleavage is proportional to the Mount of Lit produced in the hosts.

From the previous paper(7), we knew that cleavage of EF-Tu is at least a
contributing factor for phage exclusion and we have demonstrated that the global
defect can be relieved to some extent, in terms of supporting phage growth, by an
overproducing tqu clone that encodes EF-Tu. Presumably, the phage can multiply
because not all of cellular EF-Tu got cleaved after T4 infection of the EF-Tu-
overproducing Litcon host. Therefore, we applied the same philosophy in the
situation of local inhibition. If .a wild-type host can maintain the same amount of
EF-Tu after induction of go] activity as before by the aide of a tqu clone , it
would be interesting to see whether the uncleaved EF—Tu can suppress the local
inhibition.

We approached this by transforming the Lit+ host with two compatible
plasmids, one containing the EF-Tu structural gene in a pACYC 184 derivative and
the other carrying the gol region in pUC84PZl. Overproduction of EF-Tu
somehow made the cells sick and the degree of sickness was varied from strain to
strain. However we found that a clone carrying the tqu gene in the opposite
direction of the interrupted kanamycine resistance gene promoter can be tolerated
by the hosts and allowed us to pursue the question. If an excess of EF—Tu can
prevent the occurrence of the local inhibition in Lit+ when the gol-activity is
induced in pUC84PZ1 , then we should expect to see the colonies as blue as in a
LitO cell that does not have a local inhibition for functional LacZ synthesis. What
we found was that the Lit+ transformants were white on X-gal plates just like the
parental strain without a tqu (Table 3). The data demonstrated that the remaining

EF-Tu, although more than the normal amount , did not overcome the local

111

inhibition. Instead ,it suggested that the cleaved form of EF-Tu, which is the only
known difference between. a Lit+(tqu) and a LitO host, determines whether the
local inhibition takes place in the gol region. If this idea turns out to be true, it
also could explain why the incomplete recovery of phage growth occurred in the
Litcon hosts containing a tqu clone, since the presence of the cleaved form may
cause the local inhibition in Gene23 to block the synthesis of T4 major head
protein.This theory also accounts for the observation that wild-type T4 grew at a
much lower rate in Lit+ hosts than in Lit0 according to their performance in one

step growth experiments (Data not shown ).

Discussion

In this article, we uncovered the molecular basis of the gol-Lit initiated
local inhibition. The inhibition is thought to occur at the level of transcription and
somehow prevents RNA polymerase reading through to reporter genes located
downstream. The transcriptional termination depends on translation of the gol
region. Interestingly, in the absence of Lit protein, we discovered that production
of the downstream gene product is highly stimulated by the presence of the gol
sequence provided it is translated. We also demonstrated that ,unlike global
inhibition , local inhibition occurs even if not all cellular EF-Tu gets cleaved.
Overproduction of EF-Tu does not overcome the defect.

To understand the mechanism of the gol-Lit initiated effects, the putative
secondary structure of go] RNA sequences could be informative. The entire RNA
encoded by the minimal gol fragment ,PvuII-Al, can form a long stem-loop
structure with a stability of AG=-35.l Kcal (Fig 5). This stable structure could be
generated in the transcript When ribosomes are arrested upstream. Additionally,
the go] RNA possesses an A antitermination nut-like site composed of a typical

boxA sequence and a boxB hairpin structure. The nut-like sequences located in the

112

middle of the second part of the long stem are expected to be exposed by a
coupling translating ribosome (Fig 5). With the evidence presented in this paper,
we postulate that in the absence of Lit protein, the transcription rate of the gol-
CAT fusion unit could be enhanced by the unique nut-like site in the go] RNA
segment, that can be exposed with movement of translating ribosomes. In other
words, the nut-like site may frmction as a transcriptional enhancer or an
antiterminator signal that could be converted into a terminator by the function of
Lit. '

Of course, one could argue that the increase in CAT production and in the
protection of full-length gol message in the absence of Lit may result from an
enhanced stability of the transcripts instead of from a stimulation of transcription
by the presence of gol in the construct. If so, it would be hard to imagine how the
function of Lit destabilizes the gol-containing transcripts. Based on our current
knowledge about gol- and Lit, the possibility of RNA stability seems unlikely.
However, experiments to measure of stability of RNA degradation and protein
production have to be done in the future.

The most intriguing phenomenon of the gol-Lit interaction may be the
interdependence of the global and local inhibition. All the mutations isolated thus
far for their ability to overcome the global effect always suppress the local
inhibition. Our preliminary result also showed that mutations in gol that failed to
activate Lit protease to cleave EF-Tu retained the stimulating effect in the absence
of Lit. A simple explanation for this phenomenon is that the cleaved EF-Tu
molecules or some other proteolysis product resulting from the gol-Lit activated
proteolysis perhaps recognized the putative antitermination site of gol and
converted it into a termination signal, thereby causing the local inhibition. If it is

due to the cleaved form of EF—Tu, it may lead to a reasonable speculation,

113

concerning the general nature of cellular antitermination systems. The intact EF-
Tu might play an important role in the function of antitermination apparatuses.
The studies of transcription antitermination of E. coli phage A (19, 20) and
the highly conserved cellular rRNA operons (20, 21, 22) have given insights into
the constitution of the antitermination complex and allowed some specific
functions to be assigned to the known components. Unlike the A nut site whose
function requires a particular transcription antiterminator pN, there is no known
cellular analogy of N protein involved in the antitermination utility of cellular
rRNA operons that also carry a version of boxA and boxB substructure(21). The
hint of linkage between EF-Tu and the rRNA antitermination came from
characterization of EF-Tu in functions other than its traditional role in translation,
including the study of its participation in phage QB RNA replication (23, 24) and
that of being a positive regulator for rRNA synthesis(25, 26). In fact, our
observation of the local inhibition that might be due to the consequence of EF-Tu
cleavage provides encouraging evidence for this provocative concept. Thus , it is
plausible that the cleaved EF-Tu product not only aborts its antitermination effect
in transcription of go] but also may compete off the remaining intact antiterminator
for it has a stronger affinity for the nut-like site. If a physical binding of the nut-
like site does exist, then the binding would block transcription through the go]
region. Although we can not detect reduction of 16s and 23s RNA syntheses in
cells (pUC84PZl in Lit+)of an overnight plate, we did notice that the colonies
became sick after a prolonged period of incubation. The delayed sickness of Lit+
transformants may be attributed to the inhibition of rRNA transcription by the
presence of the EF-Tu cleaved product.
In summary, based on the logic discussed above, the model of EF-Tu's
involvement in the gol-Lit induced local inhibition and the cellular antitermination

is extremely attractive but lacks solid support. Experiments to directly analyze the

114

effect of EF-Tu cleavage on antitermination of rRNA operons are emergent.
Continuing investigation of the enhancing effect of gol on transcription may also

shed some light on general mechanism of transcription antitermination.

115

References

1.Greener, A., and C. W. Hill. 1980. Identification of a novel genetic element in
Escherichia coli K—12. J .of Bacteriol. 144:312-321.

2.Hill, C. W., J. A. Gray and H. Brody. 1989. Use of the isocitrate dehydrogenase
structural gene for attachment of e14 in E. coli K-12. J. Bacteriol. 17124083-84.

3a.Kao, C. and L. Snyder. 1988. The lit gene product which blocks bacteriophage

T4 late gene expression is a membrane protein encoded by a cryptic DNA
elememt, e14. J .of Bactriol. 17022056-2062

3b.Kao,C., E.Gumbs and L.Snyder. 1987. Cloning and characterization of the
Escherichia coli lit gene, which blocks bacteriophage T4 late gene expression. J
of Bacteroiol. 169: 1232-1238.

4.Bergaland K. J ., C. Kao, Y.-T.N. Yu, R. Gulati and L.Snyder. 1990. A site in the
T4 bacteriophage major head protein gene that can promote the inhibition of all
translation in Escherichia coli. J .Mol.Biol. 213:477-496.

5.Champness, W. and L. Snyder. 1984. Bacteriophage T4 gol site: sequence
analysis and effects of the site on plasmid transformation. J. Viol. 50:555-562.

6.Champness, W. and L. Snyder. 1982. The gol site : a cis-acting bacteriophage T4
regulatory region that can affect expression of all the T4 late genes. J. Mol. Biol.
155:395-407.

7.Yu,Y.-T.N. and L.Snyder. 1993. A Phage exclusion duo cleavage of translation
elongation factor Tu. Submitted.

8.Yu, Y.-T. N. and L. Snyder. 1993. Cleavage of E. coli elongation factor Tu in
vitro. In preparation.

9.Panayotatos, N. 1984. Nucleic Acids Res. 12:2641-48.
l0.Roberts, R J. 1987. Nucleic Acids Res. 15, Supplement, rl89-217.
11.Melton, D. A., et al. 1984. Efficient in vitro synthesis of biologically active

RNA and RNA hybridization probes from plasmids containing a bacteriophage
SP6 promoter. Nucleic Acids Res. 12;7035.

116

12.Hwang, Y.-W., A. Sanchez and D.L. Miller. 1989. Mutagenesis of bacterial
elongation factor Tu at lysine 136:A conserved amino acid in GTP regulatory
protein. J. Biol. Chem. 264:8304-09.

14.Maniatis et a1. 1989. Molecular cloning: A laboratory manual second edition.
Cold Spring Harbor Press.

15.Young, F. S. and A. V. Furano. '1981. Regulation of the synthesis of E. coli
elongation factor Tu. Cell 24:695-06.

l6.Sarrnientos, P., et a1. 1983. Differential stringent control of the tandem E. coli
ribosomal RNA promoters from the mnA operon expressed in vivo in multicopy
plasmids. Cell 32:1337—46.

17.Bradford, M. 1976. Anal. Biochem. 72:248.

19.Ward, D. F. and M. E. Gottesman. 1982. Suppression of transcription
termination by phage lambda. Science 216:946-51.

20.Roberts, J. W. 1993. RNA and protein elements of E. coli and lambda
transcription antitermination complexes. Cell 72:653-655.

21.Suzanne,C. et al. 1984. Antitermination of E. coli rRNA transcription is caused

by a control region segment containing lambda nut-like sequences. Cell 38:851-
60.

22.Blumentha1, T. 1979. RNA replication : Function and structure of QB-replicase.
Ann.Rev.Biochem. 48:525-548.

23.Biebricher, C. K. and R. Luce. 1992. In vitro recombination and terminal
elongation of RNA by Qb replicase. EMBO 11:5129—35.

24.Travers, A. A., R. I. Karnen and R. F. Schleif. 1970. Factor necessary for
ribosomal RNA synthesis. Nature(London) 228:748-751.

25.Haseltine, W. 1972. In vitro transcription of Escherichia coli ribosomal RNA
genes.Nature(London) 235:329-333.

26.Arai, K., et a1. 1980. Primary structure of elongation factor Tu from
Escherichia coli. Proc.Natl.Acad.Sci. USA 77:1326-1330.

27.Vieira, J. and J. Messing. 1982. Gene 19:259.

117

Table

Table 1: Characterization of M13-go] hybrid phages on Lit-containing cell. The
hybrids M13PZl, M13MTD16 prepared as in Materials and Methods harbor a
wild-type and a single-base mutant sequence of go] ,respectivelyx M13PZ1
produces white plagues on Lit+ cell and no plagues on Lit-overproducing strain

(Litcon).

118

 

 

 

Ujmmm/mqmi L73 on :8 L2: 2 _:+ ,=<: on :80:
_<:w_uN4 69$ mEm 23cm <<3=m 28cm so 285

Emma! a mic 23cm 2.5 23% 2.5 Emcee
2.333 m 2% 23.5 mfm 33cm 95 Emacm

 

 

 

 

 

119

Table-2: A transcriptional gol-Cat fusion clone does not confer chloramphenicol
resistance in the presence of a wild-type low amount of Lit protein. The host is
JM1011it+ as in Materials and Methods. The results were recorded after overnight

culture in a 300 incubator.

120

 

 

 

 

 

EN...“ 20%...._. ewamwfi- .8 Es

>n+ Q + + + L. + + + + +
.03 + =u._.m II + + + + +
28:30. +++ +++ +++

 

 

5,390.53” mace. nmq 3.. 02033338.” mmcm no“ 3_. 54.06033. Q” obaxc

QEOOmm

 

121

Table 3 : Overproduction of EF-Tu does not overcome the local inhibition. The
color of each transformant is analyzed on X-gal plates supplemented with
appropriate antibiotics for ON at 300. Overproduction of EF-Tu is obtained by

introducing a multi-copy plasmid containing tqu.

122

 

@819

 

 

3835 50 _:+ :8 Q52 5+ Q53
occaspfl mac Ego _ mac Ego
ocosm 2% mac Eco w mac

 

 

 

 

 

 

123

Figure Captions

Fig l: Plasmid constructs of pUC8PZl and p64PZl-JD. pUC8PZl shown above
contains a 180 bp Pst l-HindIII fragment of gol fused in frame with lacZ and
transcribes fiom lacZ promoter. p64PZl-JD shown at bottom contains a go]
sequence cloned in the opposite orientation (log) followed by a 110bp chicken
Histone specific Pstl fragment (JD-H). The riboprobe made from p64PZl-JD in a
in vitro transcription system by sp6 polymerase is indicated by a heavy line with a
arrow. The putative protected region of hybridized gol RNA is indicated by the
two vertical dashed lines. The thin lines beneath pUC8PZl represent various
length of transcripts from lacZ promoter. Restriction sites are simplified by single
letters: H, HindIII; S, Sall; P, Pstl.

 

 

C11" lZd179d

l

 

 

.
1
.
l

 

60'
r 9de r

H'Cll‘

aqordoqu

> --___._

 

 

 

 

 

 

124

'5
~
\
\
s.
s
\
\
x
\
\
\
\

\

\

 

IIIIIIIIIJfIIIII’I

2021

125

Fig 2: Autoradiogram of the probe protected transcripts. RNA samples made from
pUC8PZl-containing LitO (Lane 4, 10), Lit+(Lane 3, 9) and Litcon(Lanes 2, 8)
were hybridized with the riboprobes of which full length are shown by the first
arrow on the left side (Lanes 1, 11). The mixture were then digested with various
concentration of S1: 400u/sample (Lanes 8, 9, 10a), 1200u/sample (Lanes 2, 3, 4,
5, 6 ), and 3200u/samp1e(Lane 10b). The riboprobe that has been incubated ON
alone is shown in Lane 7( without S1 digestion) and Lane 8(with S1). Result of S1
digestion of the tRNA-riboprobe mixture is shown in Lane 5. The amount of tRNA
is approximately 10 times the concentration of the RNA sample used. Full-length
protection of the 180 bp gol transcripts is indicated by the second arrow fi'om the
top on the left-handed side.

126

 

 

Io. 2‘. ‘

 

127

Fig 3: Inhibition of downstream gene transcription by the gol-Lit interaction. The
plasmid constructs shown on the left-handed side are used to transform Lit+ and
Lit0 cells. The downstream Cat enzymes synthesized are determined by the CAT
ELISA assays and normalized to total cellular proteins. -Lit and +Lit indicate that
the samples are prepared from the Lit0 transformants and Lit+ transformants,

respectively.

128

room. 3:830: on 823383 mono #358230: 3
m2 ._.= 5639.0:

 

 

 

 

 

 

 

.. r: + E

EboN. .

ESE-ca _v 1.. T mmbom obmm
me. On.

armoN .
3:982, $1.1 T obi 93A
09..an me. On. . _

trnoN . . .
ncodmémfi r T.1 whom #bmw

 

on: .

0m" mEm> >mmm< ” 0:0 :3: u one 350033 on om" m3~<3o boa 3.9833 on" 206.3.

129

Fig 4: Stimulation of gene transcription by the gol site in the absence of the Lit
protein. The plasmids shown on the left-handed side are used to transform Lit0
cells and the cell extracts made from these transformants ,before ITPG (-IPTG)

and afier IPTG (+IPTG) are used to determined the Cat enzyme concentrations.

130

9.30.0203 0.. 003230.003 @030 #0300330: 9.
.30 02 00000300 .3 .30 0000300 0.. 0....

 

 

 

 

 

 

.. N .. =u.._.Q +=u._.®.

n 00

Echoes. hlflnr a... T 0.»: 8.0%.
2.00N

0co.mc>>

@900. T: 0.6mm 9mm.—

oo. 00. .

urmoN .

0:05-00. r T.1 .9de .9de

00.

00. mt?» >000.) 030 :3: n 030 3030903 0. 00. 03530 00.. 3.08903 0. 03.0.3.

131

Fig 5: EF-Tus get cleaved in the presence of a wild-type, low amount of the Lit
protein. pUC84PZl is used to transform Litcon (Lanes 1-4), Lit+ (Lanes 5-9) and
Lit°(Lanes 10-12). The protein contents of the cell samples, made in the absence
of IPTG for 60' (Lanes 4, 9), and the presence of IPTG for various time period: 0 '
(Lanes 1, 5, 10); 20' (Lanes 2, 6,‘ 11); 40' (Lanes 3, 7, 12 and 60' (Lane 8). are
visualized in a coomassei-blue stained SDS PAGE. The position of EF-Tu is

indicated by arrows.

132

 

a Op :3.

"iiiiiiiiiiii