GENE EXPRESSION

GSND 5113Q
LEARNING OBJECTIVES:
Understand how controlling access to translation initiation sites in bacteria governs extent of synthesis
Appreciate how RNA bacteriophage MS2 is controlled at the translational level
Recognize role of mRNA 3-dimensional structure in controlling access to translation starts
Be aware of effects of antisense RNA and translational coupling
Identify the roles of rare codons and codon bias in protein synthesis
Comprehend the principles underlying programmed translational frameshifts
Learn how translational repressors control expression, for example in the Str operon
Become acquainted with the principles underlying translational attenuation
Realize that mechanisms for translation activation also exist
Emanuel Goldman, Ph.D.
New Jersey Medical School
Rutgers Biomedical Health Sciences (RBHS)
Rutgers, The State University of New Jersey
Department of Microbiology, Biochemistry & Molecular Genetics
International Center for Public Health (ICPH)
225 Warren Street, room E450T
Newark, NJ 07103 USA
Telephone: 973-972-4367
Fax: 973-972-3644 or -8982
Email: [email protected]
 Regulation of Prokaryotic Initiation by RNA

• Initiation codon and methionine.

• Three-dimensional structure of the RNA can expose or bury the recognition site. RNA
bacteriophage MS2 regulation is a paradigm of this mechanism.

• Small ribosome recognition sequence in mRNA (Shine & Dalgarno) interacts with 16S
rRNA (covered in lecture 3).

• Varying sequences and 3-dimensional structures vary efficiencies of translation initiation.

• mRNA competition for ribosomes (phages T4 and T7); translational "enhancers" (T7).

• Antisense RNAs can block translation; quorum sensing: antisense RNAs control
expression.

• "Translational coupling:" initiation of downstream gene linked to synthesis of upstream

gene.
 Initiation codon and methionine
• AUG preferred (83% of E. coli genes)

• GUG acceptable (14% of E. coli genes), ~4 fold less efficient

• UUG acceptable (3% of E. coli genes), ~10 fold less efficient

• AUU is used for initiation of two genes in E. coli, infC and pcnB

• infC encodes initiation factor IF3, pcnB encodes poly(A) polymerase

• IF3 discriminates against non-canonical start codons, including AUU

• Thus, this is a form of autoregulation, since excess IF3 will inhibit translation initiation of its
own mRNA

• Excess IF3 also favors ribosome subunit dissociation, which slows translation, since 70S
ribosomes are necessary for translation

• Poly(A) polymerase is toxic when overproduced (normal role in bacteria is to facilitate

turnover of stable RNAs)

• Therefore, use of AUU for translation initiation of these two genes limits the extent of
translation of proteins that are harmful if made in excess

Sacerdot C, Chiaruttini C, Engst K, Graffe M, Milet M, Mathy N, Dondon J, Springer M. (1996) The role of the AUU initiation codon in
the negative feedback regulation of the gene for translation initiation factor IF3 in Escherichia coli. Mol Microbiol 21:331-46.

Binns N, Masters M.(2002) Expression of the Escherichia coli pcnB gene is translationally limited using an inefficient start codon: a
second chromosomal example of translation initiated at AUU. Mol Microbiol 44:1287-98.
 IF3 Autoregulation

Translation of IF3 is autogenously controlled

via a negative feedback loop acting at the
level of initiation codon detection.

(A) IF3 catalyzes continuous eviction of

tRNA from the ribosomal P-site. This
ejection mechanism is faster if the
initiator tRNA is noncognate for the
initiation codon being used. For this
reason, in the presence of IF3, initiation
on the IF3 mRNA is highly inefficient
due to the use of an AUU initiation
codon. In the absence of IF3, initiator
tRNA is not ejected, and IF3 translation
is begun.

(B) Additionally, IF3 selectively prevents

70S formation if a noncognate P-site
interaction is encountered. In the
absence of IF3, 70S formation is not
sensitive to initiator tRNA decoding
fidelity, thus initiation at the infC (IF3)
AUU codon is permitted.
Modified from Betney R, de Silva E, Krishnan J, Stansfield I. (2010)
Autoregulatory systems controlling translation factor expression:
thermostat-like control of translational accuracy. RNA 16:655-63.
 Coliphage MS2
• Two groups, closely related: i) MS2, f2, R17; ii) Qβ.
• Small, icosahedral particles (shape like a geodesic dome).
• Small, single-stranded RNA genome;
~3600 nt (MS2 group), ~4000 nt (Qβ).
• Particle consists of 180 copies of coat protein and
one copy of maturation (or "A") protein.
• Adsorption only to the F pilus; hence, male-specific phages;
coat is left behind.
• Extensive secondary/tertiary structure in the phage RNA;
considerable nuclease resistance.
• Burst sizes can go as high as 10,000 phage per infected cell!
• Approximately 50 min growth cycle.
 MS2 Genetics
• No recombination (RNA genome).
• Genes identified by conditional lethals, complementation,
and sequencing of RNA and protein.
• Three genes mapped for MS2 group
(overlapping 4th gene made in very small amounts)
| lysis protein |
5’ 3’
maturation ("A") protein coat protein (RNA) Replicase
(filled boxes represent intercistronic and/or untranslated regions)

Example of Complementation Test

Su− host Su+ host
wild type + +
coat amber − +
replicase amber − +
coat amber + replicase amber + +
(mixed infection)
 Structure and function of MS2 Genome
• Coat protein (~14 kD; 129/131 aa MS2/Qβ): structural component;
regulator of expression.
• Maturation ("A") protein (~40 kD; ~330 aa): structural component required for
infectivity in reassembled particles; RNA accessible to nuclease without it.
• Replicase (~62 kD; ~550 aa): viral RNA synthesis; regulator of expression. Uses
three additional host subunits: EFTu, EFTs, and S1 (from 30S ribosome).
Another host factor, HF-I (or HF-Q) needed for minus strand synthesis.
• Lysis protein (~8 kD; ~75 aa): has lysozyme activity; begins in +1 frame within
codon 114 of coat and continues into replicase coding region in +1 frame.
In Qβ, readthrough protein: UGA stop at end of coat gene read by tRNATrp
~5% efficiency; translation continues for ~150 codons (non-replicase frame).
• Intercistronic regions: help regulate translation; region between coat & replicase
translated in 4th (lysis [MS2] or ‘readthrough’ [Qβ]) gene.
• Extensive RNA 2o/3o structure needed for packaging, regulation & protection of genome.
• Phage RNA: + strand, translated to protein; also template for complementary (minus)
strand synthesis.
• Complementary strand: (negative) − strand; used only as template for + strand synthesis.
• RNA phage don't shut off the host; phage can be made to depend on host expression.
• Phage RNAs do compete effectively against host mRNA for ribosomes;
"mRNA competition."
 MS2 Regulated by Translational Control
• Disparate expression levels of the major genes: 1 maturation:20 coat:5 replicase proteins
• Control of ribosome access to translation start sites; different affinities for ribosomes
• Due to secondary structure of RNA; translational repressors;
also, "A" gene starts with GUG codon
• Coat translated first; translation past first 30-40 codons of coat exposes replicase start
• Evidence: i) amber at coat 6 – no replicase made; amber at coat 70 – replicase made
ii) Complementation of coat amber 6 by an A protein mutant shows that replicase can
copy in trans; i.e., effect of coat amber 6 is cis; thus, it's not that RNA isn't made (like
polarity) but that rbs (ribosome binding site) of replicase unavailable
iii) In vitro replication system: replicase can copy coat amber 6
• Disparate expression can be explained by RNA structure controlling access to rbs's
• Evidence: i) in vitro translation assay for 35S-fMet amino terminal peptides, identified by
peptide mapping, reproduces in vivo ratios: 75% coat, 20% replicase, 5% A protein
ii) Denature template (f2 RNA) structure by mild formaldehyde treatment, or elevated
temperature: find 1:1:1 ratio of the 3 genes in the in vitro translation initiation assay
• Low expression of lysis protein also due to RNA structure and because ribosomes are
already translating lysis start region in another reading frame (i.e., making coat)
• Another mechanism affects translation starts for replicase:
repression by coat protein dimer (covered later in this lecture)
• Parallel mechanism also affects translation starts for coat: repression by replicase
 Exposing or burying the translation start signals: phage MS2 RNA

coat start

replicase
start

The nucleotide sequence of the last part of the A protein gene, the first intercistronic region, the coat gene, the second inter-
cistronic region and the first part of the polymerase gene is given in the form of a secondary structure. The initiation codons
and termination signals are shown boxed. The arrows show the splitting points with RNase T1 and the number of feathers on
the arrow gives a semiquantitative measure of the susceptibility of the bond (arrows with four feathers correspond to links
which are always split). Some hairpins are strongly supported by experimental evidence, while other aspects are speculative.
 Schematic: RNA structure can expose or bury translation start site
coat protein gene
A. rbs AUG
5’ 30 40

ribosome GUA rbs

ribosome
blocked
3’
coat protein gene
rbs AUG
B. 5’ 30 40

ribosomes

3’
Schematic diagram how RNA structure governs relative translation of coat and replicase protein genes. Protein synthesis proceeds from the 5' to
3' direction (indicated) on the mRNA. In panel A, the ribosome binding site (rbs) and the start codon (AUG) of the coat protein are exposed and
accessible to ribosomes so that translation of the coat gene proceeds readily. However, the rbs and the AUG (read right to left) of the replicase
gene are buried in the RNA structure (selected hydrogen bonds depicted as short vertical lines between the upper and lower portions of the RNA
chain) and not accessible to ribosomes, so this gene is not translated unless the structure is opened up. In panel B, ribosomes translating the
upstream coat gene open the RNA structure when they proceed past approximately codons 30-40 (indicated above the upper portion of the RNA
chain), allowing the replicase gene rbs and AUG to be accessible to other ribosomes, which can now commence translating the replicase gene.
RNA Phage Systems
Historically, the first realization that protein synthesis could be controlled at the translational level came from studies of RNA bacteriophage
morphogenesis. The RNA phages that have been primarily studied infect male Escherichia coli (requiring the pilus for adsorption), and fall into
two main groups: the MS2 group, consisting of MS2 and the closely related phages f2 and R17; and Qbeta. Their genomes consist of a small,
single-stranded, highly structured RNA (~3600 nucleotides for the MS2 group, ~4000 nucleotides for Qbeta) which also serves as a messenger
RNA for the synthesis of phage proteins (i.e., these are '+' strand RNA viruses). The MS2 group of phages was instrumental in elucidating many
of the fundamental precepts of translation (such as the mechanism of translation initiation), in part because the isolated phage genome was an
excellent template for in vitro translation systems. Indeed, the first actual proof of the codon assignments in the genetic code occurred when MS2
coat protein, and the gene encoding this protein in the MS2 RNA genome, were independently sequenced in 1976, some 6 years after original
presentation of the genetic code table.

There were three phage-specific protein products identified during the early studies of the MS2 group (a minor fourth product, the 'lysis' protein,
was identified years later): the coat protein, the RNA-dependent RNA replicase protein, and the ‘maturation’ protein (also known as the 'A'
protein). These three proteins were synthesized in vastly disparate amounts during an infection, in a ratio of 1 maturation to 20 coat to 5 replicase
protein molecules. Because the RNA genome also served as template for protein synthesis, it was presumed that the differential expression of the
three phage proteins was due to translational control. A key to the elucidation of this control was the observation that the disparate ratios of
expression could be approximated in an in vitro translation system assaying for initiations of the respective proteins, using [35S]fMet-tRNA to tag
the N-termini, digesting the products with proteases, and separating and identifying the N-terminal peptides by high-voltage paper
electrophoresis. In such an assay, the ratio of initiations of coat:replicase:maturation was approximately 75:20:5. However, disruptions of the
phage RNA secondary/tertiary structure by mild pretreatment of the RNA with formaldehyde, or by conducting the in vitro translations at elevated
temperature, resulted in roughly equimolar ratios of initiations of the three phage proteins. Thus, the differential levels of expression, at least in
vitro, can be ascribed to the structure of the template RNA limiting ribosome access to translation start sites.

In the case of the maturation protein, two additional factors in vivo appear to play a role in minimizing its expression: (1) the sequence encoding
this protein begins with the GUG codon, now known to be less efficient for translation starts than the commonly used AUG codon; and (2) the
structure of the phage RNA so severely restricts ribosome access to the translation start that the maturation protein, encoded near the 5’ end of the
phage RNA, apparently is only (or primarily) synthesized from replicating (incomplete) RNA in the 'replicative intermediate' form.

The lysis protein of the MS2 group is made in even lower amounts than the maturation protein, which is why it was not discovered until years
later. The mode of this control is also severe restriction of access to the translation start site (though there were some earlier reports in the
literature incorrectly proposing another model). This protein is encoded in an overlapping section of the genome, starting in another reading frame
in the vicinity of the 114th codon of the 129 codon coat protein gene, continuing into the replicase-encoded section of the RNA in a different
reading frame, for a total of 75 codons in length. It has been demonstrated that the lysis protein translation start is perfectly functional when
placed in another context; therefore, the low levels of expression are due to restricted ribosome access, possibly exacerbated by the high traffic of
ribosomes traversing the site in another reading frame during translation of the predominant coat protein gene.
continued on next page
RNA phage systems, continued

Translational polarity
Several observations in vivo buttressed the model of RNA secondary structure controlling ribosome access as the explanation for differential
expression of RNA phage proteins. An amber mutation (UAG stop codon) in the coat protein gene at codon 6 was found to cause the lack of
expression of the downstream RNA replicase protein as well. By contrast, an amber mutation in the coat protein gene at codon 70 showed full
expression of the RNA replicase. This phenomenon was termed 'translational polarity' in analogy to polarity in expression that had been found in
mutants of the lac operon and some other systems. However, in classic (transcriptional) polarity, the absence of translation (due to the premature
stop codon) led to premature transcription termination (due to coupling of RNA and protein synthesis); thus no messenger RNA (mRNA) was
present for the downstream gene(s). That was not the case with the RNA phage, where the mRNA template for the downstream gene was present
in the (coat protein) codon 6 amber mutant.

Further, translational polarity was found to be cis-acting. Complementation of the coat protein mutant by an A protein mutant in a mixed infection
of a nonpermissive host led to replication of both mutant genomes (there is no recombination with RNA phage), indicating that the replicase
supplied in trans by the A protein mutant could copy the genome of the coat protein mutant.

The simplest explanation for translational polarity again is the model of secondary/tertiary structure of the RNA restricting access to translation
starts. Models of the phage RNA secondary structure suggest that the replicase translation start is buried in the structure and inaccessible to
ribosomes, while the coat translation start is readily accessible. As ribosomes translate through the coat gene, they open up the secondary structure
of the RNA so that the replicase start now becomes available. If ribosomes can only get to codon 6 of the coat, the structure has not been opened
up, hence no replicase can be made. However, if ribosomes get to position 70, now the structure has been opened and replicase translation can
commence. Inspection of putative phage RNA secondary structures suggested that ribosomes only need to move past codons 30–40 to open the
structure sufficiently so that the replicase protein can be synthesized. In fact, it has been shown that synthesis past codon 32 of the coat is
sufficient to allow replicase translation.

Translational operators
One additional feature of RNA phage translational control was the finding of translational repression by both the coat on replicase translation, and
the replicase on coat translation. It is worth noting here that the RNA phage presented the first known example of translational repressors, and that
the targets for binding by the repressors are sequences of RNA which have been termed 'translational operators' in analogy to the 'operator' targets
in DNA for binding of transcriptional repressors (such as the lac repressor binding the lac operator). The interaction between the target operator
RNA sequence and a dimer of the coat protein (in the early literature, it was thought that the coat protein repressed as a hexamer) continues to be
studied of interest today, and has found use in genetic screens in the “three-hybrid system” to detect RNA-protein interactions, and in
biotechnology applications. [discussed later in the lecture]

excerpt from Goldman E (2008) Translation control by RNA. Encyclopedia of Life Sciences, John Wiley & Sons, UK; http://www.els.net
 RNA structure controls sigma-S translation start
Primary mRNA transcript: 63 nt
Sopr-GUAUUCCACCGAGGAUGGGCACUAGGG CCCAAAUGCCCAAAGGGGAAC-5’

mRNA folds into

secondary structure:
AUU A
5’
63 nt
A A A A C
CCC A UGCCCA GG GGAA
GGG U ACGGGU A G CC CCUU AUG -rpoS
A C A
G A
G
UA U UU

ribosome access limited

Model of RNA secondary structure upstream of translation start for rpoS (encoding sigma-S) in Salmonella typhimurium.
nt = nucleotides. The ribosome binding site (AGGAG) and the AUG start codon are boxed. The folded structure makes
ribosome access to the start site difficult, and this ordinarily requires the help of either of two small antisense RNAs,
DsrA and RprA, which bind upstream of the translation start region in a complex with the RNA chaperone Hfq. In the
absence of Hfq, little sigma-S is produced. Unlike other examples in which antisense RNAs inhibit expression, DsrA and
RprA RNAs (in conjunction with Hfq protein) stimulate sigmaS expression, by altering the inhibitory structure of the
mRNA to allow ribosome access.The arrows to/from nucleotides in the figure denote mutations that lead to increased
production of sigma-S in the absence of Hfq, presumably because these mutations disrupt and open up the RNA
secondary structure, thereby allowing ribosomes greater access to the translation start signals. The model is a
modification from that of L. Brown and T. Elliott. 1997. J. Bacteriol. 179: 656-662.
 RNA secondary structure control of translation starts in phage T4 and other systems
Example of prokaryotic RNA secondary structure control. Schematic diagram depicts the principle of how alternate forms of messenger RNA
(mRNA) can lead to either inhibition or expression. A longer form of the mRNA, starting from promoter P1, folds into a structure such that the
ribosome-binding site (rbs) and the start codon (AUG) are occluded in a stem–loop, and inaccessible to ribosomes. A shorter form of the mRNA,
starting from promoter P2, does not form the inhibitory structure, therefore the rbs and AUG are available for ribosomes to commence translation.

Regulation of expression of bacteriophage T4 lysozyme was puzzling in initial studies of T4 morphogenesis, because it was clearly a protein
expressed late in infection, but evidently was transcribed at early times after infection as well. The explanation turned out to be that the early
transcript began from a promoter some distance upstream from the lysozyme-encoding reading frame, and this transcript folded into a stable RNA
secondary structure which obscured the translation initiation region for the lysozyme gene (blocking ribosome access). However, late in infection,
following DNA replication of the phage genome, two late promoters closer to the lysozyme sequence were activated; these late transcripts did not
contain the inhibitory sequences which fold to obscure the lysozyme start site in the early transcripts, therefore the protein could be translated.
Phage T4 also has at least one other example of this type of control, in the regulation of the soc gene, which like lysozyme, is transcribed both
early and late. However, the early transcript starts from an upstream promoter and folds to obscure the translation start, while the late transcript
starts further downstream and allows ribosome access to the start site. In a curious variation on the theme, phage T4 gene 25 has been reported to
contain a secondary structure in the leader that enhances translation of the protein, by sequestering potentially competing Shine–Dalgarno
ribosome binding sites in the structure, thereby facilitating ribosome access to a productive Shine–Dalgarno for initiation of synthesis.

excerpt from Goldman E (2008) Translation control by RNA. Encyclopedia of Life Sciences, John Wiley & Sons, UK; http://www.els.net
mRNA competition
As a corollary to the principle of RNA structure controlling ribosome access to translation starts, different mRNAs display
different degrees of access to their initiation regions. This has led to the concept of 'mRNA competition' between different
transcripts for available ribosomes. Obviously, under conditions when ribosomes are in vast excess, this would not be of
great importance, however, whenever there are conditions of limiting ribosomes, the mRNAs with the best translation starts
will out-compete poorer mRNAs to get their proteins made. This mode of control is known to be responsible for the switch
from early to late protein synthesis during phage T7 infection, when early mRNAs remain in the infected cell during the
late stages of infection but are not translated, because the late mRNAs out-compete the early ones for ribosomes. It has also
been proposed that this kind of control is partly involved in the switch from early to late phage T4 gene expression, and
also may be responsible for high levels of RNA phage coat protein expression during infection, especially since the RNA
phage do not shut off host transcription.

Translational Enhancers
In bacteria, the quality of the ribosome-binding sequence (Shine–Dalgarno) is a primary determinant of the translational
efficiency of an mRNA. However, it was also observed on occasion that other primary sequences in translation initiation
regions appeared to enhance the ability of an mRNA to be translated, including for example the three nucleotides
immediately preceding the start codon, or an upstream uridine-rich sequence. The concept of a 'translational enhancer' was
formally established with the identification of a sequence upstream from the phage T7 gene 10 translation start, termed
'epsilon', which conferred significant increases in the translatability of messages in which it was placed. Epsilon is
complementary to nucleotides 458–466 of the 16S rRNA, and enhances translation even when placed downstream of the
start codon. Increasing the complementarity of epsilon from 9 to 16 nucleotides in 16S rRNA caused this element to even
be capable of replacing the Shine–Dalgarno, with 80% efficiency of translation. However, other studies have not supported
complementarity to 16S rRNA as the basis for translational stimulation by epsilon, and the mechanism remains unresolved.
Enhancement of translation by a 'downstream box' (i.e., a sequence just downstream of the start codon) has also been
reported for several genes, including aminoacyl-tRNA synthetases, phage T7 gene 0.3, and other genes. Complementarity
to nucleotides 1469–1483 of 16S rRNA was originally proposed as the basis for this activity, however, tests of this model
by altering the corresponding sequence in 16S rRNA failed to support complementarity as the explanation.
Eukaryotic translational enhancers have also been reported in both the 5' and the 3' ends of a plant viral (Tobacco
Mosaic Virus) message. The 5' element, termed "omega", functions as a binding site for a heat shock protein which
stimulates translation and confers translatability on uncapped mRNA.
excerpts from Goldman E (2008) Translation control by RNA. Encyclopedia of Life Sciences, John Wiley & Sons, UK; http://www.els.net
 Antisense RNA can block translation
ribosome
Quorum sensing
Many bacterial species sense the density of their
rbs AUG coding region populations by a process called "quorum
sensing", in which "autoinducer" (pheromones,
A. 3’
5’ e.g. N-acteyl-homoserine-lactones) signal
molecules are secreted by cells into the medium.
ribosome When cell densities are high, the concentrations
blocked
of the autoinducers become high enough to
change patterns of gene expression, leading the
bacterial populations to exert new properties
such as virulence or biofilm formation. This
rbs AUG coding region process, which has been studied extensively in
antisense RNA 5’ 3’ Vibrio species, turns out to be regulated in part
B. 5’ 3’ bound by antisense RNAs.

At low cell densities (i.e., in the absence of high concentrations of autoinducers), small RNAs complementary to the translation initiation
region of a 'master' transcriptional activator inhibit synthesis of this activator, preventing the patterns of gene expression seen at high cell
densities. As was seen previously for osmolarity and SigmaS regulation, the RNA chaperone Hfq is needed for the antisense RNAs to exert their
control. Synthesis of the antisense RNAs is governed by signal transduction pathways that respond to the levels of autoinducers via
phosphorylation/dephosphorylation.
Interestingly, the antisense RNAs not only directly block ribosome accesss to the mRNA for this activator protein, but they also destabilize
the mRNA encoding the activator. Thus, the antisense RNAs involved in Vibrio quorum sensing are acting in a dual fashion, both directly
inhibiting translation initiation and facilitating degradation of mRNA. The degradative function is reminiscent of the mechanism by which
antisense RNA functions in eukaryotic cells, "RNA interference" (RNAi), discussed previously.
An example of translational activation by an antisense RNA in an alternate pathway of quorum sensing in Vibrio cholerae has also been
found. In a similar fashion to the DsrA/RprA activation of translation of SigmaS mRNA described above, this V. cholerae antisense RNA
activator disrupts a translation-inhibitory structure in the mRNA for a putative diguanylate cyclase encoded by that mRNA, allowing ribosome
access. See Ng & Bassler (2009) Annu Rev Genet 43:197-222 for a full discussion of quorum sensing.

excerpt from Goldman E (2008) Translation control by RNA. Encyclopedia of Life Sciences, John Wiley & Sons, UK; http://www.els.net
 Translational Coupling
example: tryptophan biosynthetic operon of E. coli

P O leader trp E trp D trp C trp B trp A

DNA

P = promoter Full-length
O = operator trp mRNA
Attenuated trp mRNA
Anthranilate Phosphoribosyl Phosphoribosyl Tryptophan Tryptophan
synthetase anthranilate anthranilate synthetase β synthetase α
transferase isomerase

trpE-Thr-Phe-end trpB-Glu-Ile-end
ACU UUC UGAUG GCU GAA AUC UGAUG GAA
Met-Ala-trpD Met-Glu-trpA

Oppenheim DS, Yanofsky C. (1980) Translational coupling during expression of the tryptophan operon of
Escherichia coli. Genetics 95: 785-795.
 Regulation of Prokaryotic Elongation by RNA
• Limitation of aa-tRNA leads to synthesis of pleiotropic regulator ppGpp (stringent response) (lecture 9)
(ribosomes stall at codons for which there are no charged tRNAs to decode them)

• Codon bias and expression levels: highly expressed proteins avoid rare codons;
tRNA levels correlate to codon demand, e.g., ribosomal proteins.

• Programmed translational frameshifts; example: RF2 has in-frame UGA at 26th codon,
must frameshift +1 to make rest of protein; internal Shine & Dalgarno facilitates; autoregulatory.

• Other "recoding" events: rare codon stimulated frameshifts;

stem-loops and pseudoknots impeding ribosome progress, as in selenocysteine incorporation (lecture 2);
readthrough of termination codons (especially UGA);
"simultaneous slip" –1 frameshifts (X XXY YYN motif);
ribosome "hops" (T4 gene 60).

• "Back translocation" on ribosome by LepA protein after defective translocation restores ribosome to a
pre-translocation state, allowing EF-G a second chance to translocate the tRNAs correctly (lecture 3)

• tmRNA for proper termination of incomplete proteins coded by truncated or defective mRNAs (lecture 3)

• mRNA stability:
Stem-loop at 3' end (via RNaseIII processing, or intrinsic [rho-independent] termination) stabilizes.
Stem-loop at 5' end stabilizes, protects from endonuclease RNase E.
Poly A tails destabilize in bacteria (stabilize in eukaryotes).

5’ 3’
 Codon Bias can affect translational efficiency

• Although there are many synonymous codons in the genetic code, usage of these
synonyms is nonrandom, and varies widely from organism to organism; this is
called 'codon bias'. Concentrations of cognate tRNAs generally correlate to the
frequencies of codon usage, i.e., to codon demand.

• Codon choices affect translation, with common codons in a given organism

believed to increase the efficiency of translation. On the other hand, multiple rare
codons have been shown to slow translation in E. coli and yeast, frequently
observed when attempts are made to express heterologous cloned genes. However,
there are also some instances where this occurs naturally. Increasing the
concentrations of cognate tRNAs generally reverses inhibition by rare codons.

• Codon “de-optimization” has been shown to attenuate Poliovirus virulence.

• Especially notable is the translation of the gene for silk fibroin protein, which is
largely composed of codons for glycine, alanine and serine. In order to
accommodate expression of this highly specialized message, the tRNA
composition in differentiated worm silk glands consists primarily of cognate
tRNAs to codons for these three amino acids.
 Codon usage in E. coli
total codons: 323059, total proteins: 968
U C A G
UUU Phe 6089 UCU Ser 3382 UAU Tyr 4875 UGU Cys 1552 U
UUC Phe 5838 UCC Ser 3048 UAC Tyr 4295 UGC Cys 1961 C
U UUA Leu 3397 UCA Ser 2105 UAA Stp 644 UGA Stp 259 A
UUG Leu 3659 UCG Ser 2549 UAG Stp 65 UGG Trp 4169 G
CUU Leu 3204 CCU Pro 2123 CAU His 3666 CGU Arg 7981 U
CUC Leu 3135 CCC Pro 1354 CAC His 3471 CGC Arg 6946 C
C CUA Leu 960 CCA Pro 2623 CAA Gln 4223 CGA Arg 989 A
CUG Leu 17477 CCG Pro 7724 CAG Gln 9589 CGG Arg 1491 G
AUU Ile 8809 ACU Thr 3499 AAU Asn 5265 AGU Ser 2380 U
AUC Ile 8712 ACC Thr 7873 AAC Asn 7868 AGC Ser 4831 C
A AUA Ile 1273 ACA Thr 2108 AAA Lys 12104 AGA Arg 692 A
AUG Met 8507 ACG Thr 4050 AAG Lys 3857 AGG Arg 428 G
GUU Val 6716 GCU Ala 5770 GAU Asp 10382 GGU Gly 9202 U
GUC Val 4552 GCC Ala 7490 GAC Asp 7116 GGC Gly 9824 C
G GUA Val 3896 GCA Ala 6760 GAA Glu 14134 GGA Gly 2246 A
GUG Val 7972 GCG Ala 10641 GAG Glu 6149 GGG Gly 3110 G

DNA sequences analyzed were from GenBank version 63 (15 March 1990).
Boldface indicates the 7 lowest usage sense codons in E coli.

Zhang S, Zubay G, Goldman E. (1991) Low-usage codons in Escherichia coli, yeast, fruit fly and primates. Gene 105:61-72.
 Programmed translational frameshifts in E. coli

Summary of the DNA and peptide sequence of the region of RF-2 containing the in-frame termination codon. Amino
acid sequence confirms the presence of the amino acids Tyr, Leu, Asp, Tyr, and Asp on either side of the termination codon.
The 3 thymidine nucleotides at which tRNALeu at codon 25 rephases, and the internal Shine-Dalgarno sequence at codons 22-
23, are underlined.

The dnaX frameshift cassette. The boxed AAAAAAG sequence is the frameshift site. UGA (boxed) is the stop codon in
the –1 frame. The stem-loop structure 3' of the shift site enhances frameshifting. SD-region, internal Shine-Dalgarno
sequence; spacer 1, nucleotides between the SD region and the frameshift site. Numbering begins at the first nucleotide of
the SD region.
 Programmed translational frameshifts

+1 frameshifts move the reading frame to the right by one nucleotide. This requires a vacant
A site which can occur by different mechanisms, for example, a stop codon in the A site, or a
rare codon that has little if any cognate tRNA to read it. Extrapolating from the rare codon
condition, any starved codon in the A site (due to lack of charged tRNA to read the codon)
can also promote a +1 shift. A stimulator or facilitator is generally needed to amplify the
frameshift to the level of significance. In bacteria, an upstream translated Shine-Dalgarno
close (5 nt) to the shift site stimulates a +1 shift. The tRNA that read the last codon before
the shift does not rephase. Instead, a tRNA reading the next triplet in the +1 frame resets the
frame. In effect, a base in the mRNA is skipped to enter the new frame. In +1 shifts, tRNA
anticodons are not required to read triplets in both the 0 frame and the +1 frame.

-1 frameshifts move the reading frame to the left by one nucleotide. This generally requires
simultaneous movement of the tRNAs in the P and A sites to the left by one nucleotide,
generating a new reading frame codon in the A site. For the two tRNAs to move, the
nucleotide sequence has to be configured so that both tRNA anticodons can still acceptably
pair with the triplets in the -1 frame. Thus, the nucleotide sequence generally is in the form
X XXY YYN so that tRNAs reading XXY YYN in the zero frame can pair with XXX YYY
in the -1 frame. X and Y are often the same nucleotide, meaning that a run of six of the same
nucleotide has the possibility of being a -1 frameshift site. One or more stimulators or
facilitators is generally needed to amplify the frameshift to the level of significance. In
bacteria, a downstream stem-loop fairly close to the shift site is required. At least one
bacterial -1 shift additionally uses an upstream (11 nt) translated Shine-Dalgarno as a
stimulator. In eukaryotes, a downstream stem-loop and/or pseudoknot (stem loop where the
loop is also base paired) is required to facilitate a -1 frameshift.
 Regulation of Prokaryotic Initiation by Proteins

• Translational repressors which bind to the messenger RNA and block ribosome binding.

• Some translational repressors are also autoregulatory (i.e., block their own translation).

• Translational repressors are also important in eukaryotic systems. For example, the gene
involved in Fragile X Mental Retardation, FMR, encodes a pleiotropic translational repressor
expressed in neural tissue. The failure to express this gene is responsible the disease.

• "Molecular mimicry" between RNA binding sites for r-proteins on rRNA and on mRNAs

• "Translational attenuation:" translation of upstream leader peptide controls downstream start for
some antibiotic resistance genes (e.g., chloramphenicol); effects of nascent peptides.

• Translational activators: com product of phage Mu binds to mom mRNA to expose start site.
Activation of mitochondrial cytochrome oxidase translation.
 Translational repression in RNA phage gene regulation

Collision course averted in RNA phage growth. (a) Synthesis of phage minus-strand RNA requires the
template to be copied in the opposite direction from which ribosomes move on the same RNA during
translation. To avoid collision, translational repression makes the RNA translationally dormant, so that
replication can proceed unimpeded. (b) Synthesis of phage RNA plus strands by replicase proceeds in the
same direction as ribosome movement. See next page for a full discussion.
Model of "Traffic Control" to Explain MS2 Repression Patterns
• Replication of the RNA requires synthesis in the opposite direction as the
ribosomes move in translating the same template; i.e., a collision course
between replicase and ribosomes.
• Early times: i) first + strand translated to coat protein.
ii) RNA secondary structure opened up to allow replicase synthesis.
iii) Replicase will translationally repress further coat initiations;
ribosomes in progress will run off; no more replicase will be made, either,
due to loss of access of rbs by RNA structure.
iv) Now, replicase replicates the + strand into − strand,
then back to new + strands. Each new + strand repeats same events.
• Late times: i) enough coat protein accumulates to translationally repress rbs
for replicase.
ii) − strands continue to serve as templates for + strands, which are translated
as they are made.
iii) No collision this time, since replicase and ribosomes are moving in same
direction.
iv) When − strand copied, structure is called the "replicative intermediate."
This form of phage RNA may be the only effective template for A protein
translation, because of RNA structure.
RNA Phage Gene Regulation
RNA phage systems were the first in which regulation at the level of translation was shown to occur.

In addition to the early analysis demonstrating that RNA structure governed ribosome access to translation start sites, it was also observed that the RNA phage
coat protein could bind to phage RNA and repress the translation of the RNA phage replicase gene. The target in phage MS2 RNA was identified as a 21-
nucleotide sequence, which can form a hairpin encompassing the Shine–Dalgarno ribosome binding site and the AUG start codon for translation of the replicase
protein. It was thought for a long time that the phage coat protein bound as a hexamer, but it was later shown that this repression involved a coat protein dimer.
The binding of the coat protein dimer to its target translational operator in the RNA prevents ribosome access to the translation start signal for the replicase
gene, thus repressing translation of the replicase.

In addition to the phage coat protein repressing translation of the replicase gene, the phage replicase protein, in turn, similarly represses translation of the coat
protein gene. This involves binding of the replicase just upstream of the coat protein translation start. This rather elaborate dance of mutual repressions by the
two major phage gene products is actually a highly sophisticated switching mechanism so that the phage will be able to successfully replicate its genome.

Model of 'traffic control' to explain repression patterns

The single-stranded plus-strand RNA of RNA phage serves a dual function of mRNA and genome, and thus is subject to translation by ribosomes as well as
replication by the phage replicase. However, replication of the phage RNA complementary (minus)-strand, the first step in replication, requires RNA synthesis
in the opposite direction from which the ribosomes move when translating the same template; that is, replicase and ribosomes would be on a collision course
unless some regulatory event intervened. This is because, like all nucleic acid synthesis, the minus-strand RNA is synthesized from the 5' to 3' direction (in the
product, which means the template is being read from the 3' end), while ribosomes also move over mRNA in the 5' to 3' direction, which is the opposite
direction to replicase movement during minus-strand synthesis.

At early times after infection, the first plus strand is translated to coat protein. During this process, the RNA secondary structure opens up to allow replicase
synthesis. As replicase molecules accumulate, they will translationally repress further coat initiations; ribosomes in progress will run off. Because coat protein
synthesis is stopped, no more replicase will be translated, either, due to folding of RNA structure which prevents ribosome access to the replicase translation
start site. Thus, the consequence of translational repression of coat protein synthesis is to create a translationally dormant phage RNA, which is now a suitable
template for replication. Replicase will now copy the plus strand into a minus strand, then use the minus strand as a template to make new plus strands. Each
new plus strand repeats the same events, synthesizing coat and then replicase, then shutting down translation through translational repression by the replicase,
followed by synthesis of complementary and then phage RNAs.

At late times after infection, enough coat protein has accumulated to translationally repress the translation start sites for replicase. Thus, coat protein can
continue to be synthesized without stimulating continued synthesis of replicase. Enough replicase has also accumulated so that minus strands continue to serve
as templates for synthesis of plus strands, which are translated as they are made. There is no collision this time, because replicase and ribosomes are moving in
same direction when the replicase is synthesizing plus strands.

In addition to the simple elegance of this switching system, the interaction of coat protein dimer with its target translational operator has received widespread
interest as a model system for RNA–protein interactions. Further, investigators have devised applications for this interaction in biotechnology, including ‘three-
hybrid’ systems to generally detect RNA–protein interactions.
excerpt from Goldman E (2007) Translation control by proteins. Encyclopedia of Life Sciences, John Wiley & Sons, UK; http://www.els.net
Three Hybrid System to detect RNA binding proteins based on MS2

(A) Overview of the three-hybrid strategy to detect RNA–protein interactions. A hybrid protein containing a DNA-binding domain (e.g.,
LexA) with RNA-binding domain 1 (e.g., MS2 coat protein) localizes to the promoter of an appropriate reporter gene. A second hybrid
protein containing a transcriptional activation domain (e.g., from Gal4) with RNA-binding domain 2 (e.g., IRP1) will activate transcription
of the reporter gene when in close proximity to the gene’s upstream regulatory sequences. A hybrid RNA containing sites recognized by the
two RNA-binding proteins links the two hybrid proteins to one another, and the tripartite complex results in detectable expression of the
reporter gene. For simplicity, the following details of the typical experimental design are not depicted in A. Multiple LexA binding sites are
present in the promoter. Both MS2 coat protein and LexA bind to their target sites as dimers. Thus multiple copies of the hybrid RNA, each
one of which contains two potential binding sites for MS2, may be bound simultaneously to the promoter. (B) Schematic structure of the
hybrid RNAs. The hybrid RNAs are expressed from an RNase P promoter, using RNA polymerase III. They retain the 59 stem–loop
structure(s) and 39 end of RNase P RNA, but are missing its internal portion. To facilitate cooperative binding to MS2 coat protein, the coat
protein binding site is present in two copies, each of which contains a point mutation (indicated by a black dot) that enhances this RNA–
protein interaction. In the example shown, the stem–loop of the IRE has been inserted into the hybrid RNA vector, generating a plasmid
encoding IRE–MS2 hybrid RNA. Thin line represents RNaseP RNA sequences; bold lines represent IRE or MS2 binding site sequences;
black dot represent point mutation in the loop of the MS2binding site. From SenGupta et al. (1996) Proc Natl Acad Sci USA 93:8496-501.
 Translational repression in T4 phage gene 32 regulation

A pseudoknot in the mRNA at the 5’ end is the nucleation point for binding to the mRNA of gene 32
protein monomers, depicted as filled half-circles. The horizontal hatchmarks in the pseudoknot represent
hydrogen bonds in the two stems of this structure. Binding of gene 32 protein is cooperative and proceeds
downstream until the ribosome binding site (rbs) and the start codon (AUG) are coated with gene 32
protein, thereby blocking access of ribosomes to the translation start of the gene 32 coding region. See
next page for a full discussion.
 Phage T4 gene 32 autoregulation
T4 gene 32 encodes a single-stranded nucleic acid-binding protein involved in T4 deoxyribonucleic acid (DNA)
replication, recombination and repair. Mutants in gene 32 were found to overproduce the gene 32 protein. The
mutations were in the coding sequence for the protein, thus the mutant product was somehow responsible for
overproduction of itself. Even fragments resulting from premature stop codon mutations were overproduced in
infected cells. Measurements of mRNA from gene 32 indicated that this phenomenon was posttranscriptional. Thus,
it was suggested that the wild-type gene 32 protein was autoregulatory at the level of translation, i.e. it somehow
blocked translation of mRNA coding for itself. The gene 32 mutants, or premature termination fragments of the
protein, were unable to block this translation, hence the product was overproduced.

Because the gene 32 protein seemed to bind nonspecifically to unstructured nucleic acid, the source of specificity
for this autoregulation was not obvious. The binding of gene 32 protein to its mRNA was shown to block the
translation start region, fulfilling a requirement for a translational repressor. The clue to the source of specificity
was that this binding was cooperative with upstream binding of other molecules of gene 32 protein, and appeared to
originate from a nucleation site considerably upstream of the initiation site on the mRNA. This nucleation site was
proposed, and later shown, to be an RNA pseudoknot at the 5' end of gene 32 mRNA; it is this pseudoknot that
confers the specificity of gene 32 autoregulation. (A pseudoknot, in its simplest description, is a stem–loop in
which the loop is also involved in base pairing with another stretch of RNA.) The figure on the previous page
illustrates autoregulation by gene 32 protein. The evolutionary rationale of this autoregulation acting at the
translational level makes sense when considering the function of gene 32 protein, which is to bind single-stranded
DNA during phage replication. When the protein is needed for this function, it will not be available in sufficient
quantities to translationally repress itself. Only when its role in phage replication is fulfilled is excess protein
available to repress its own production, using the same property of binding to single strands, in this case mRNA.

excerpt from Goldman E (2007) Translation control by proteins. Encyclopedia of Life Sciences, John Wiley & Sons, UK; http://www.els.net
 Translational repression in the r-protein operons

Regulation of gene expression in vivo and in vitro, by the identified repressor r-protein. Genes are represented by
the protein product, and the direction of transcription from the promoter (P) of each operon is indicated by the
arrow. It should be noted that in vivo, the primary rif (formerly called the β operon) and α-operon transcripts are the
result of read-through from the upstream L11 and spc operons, respectively. Regulatory r-proteins are indicated by
the boxes, and the effects of the boxed proteins on the in vitro or in vivo synthesis of proteins in the same operon are
indicated. Results obtained in vivo are those from experiments using hybrid plasmids to achieve overproduction of
the regulatory protein. Symbols: +, specific inhibition of synthesis; -, no significant effect on synthesis; +/-, weak
inhibition of synthesis: (+), inhibition presumed to occur in vivo; ND, not determined. It has not been established
how L14, L24, or S12 is regulated, but the results of (unpublished) gene dosage experiments suggest that L14 and
L24 each regulates the translation of its own mRNA, respectively.
 Translational Repression, Coupling & mRNA degradation

P rpsL rpsG fusA tufA

DNA
rbs AUG rbs AUG rbs AUG
AUG mRNA

S12 S7 EF-G EF-Tu proteins

translationally
rbs AUG rbs AUG rbs AUG AUG repressed mRNA
no ribosome
mRNA ribosomes access
degraded blocked

ribosomes
The ‘streptomycin operon’ is comprised of four genes, rpsL and rpsG, encoding ribosomal proteins S12 and S7, and
fusA and tufA, encoding elongation factors EF-G and EF-Tu, respectively. Mutations in protein S12 confer resistance to
streptomycin. S7 acts as a translational repressor of this operon, binding to target sites near the translation starts for S7
(i.e., it’s autoregulatory) and EF-G synthesis, blocking ribosome access. Ordinarily, EF-G synthesis is required for
EF-Tu to be made, because these genes are translationally coupled. Thus, blocking EF-G synthesis also blocks EF-Tu.
Binding of the S7 repressor also destabilizes the mRNA upstream of the binding site (encoding S12), thus preventing
S12 synthesis as well. P = promoter; rbs = ribosome binding site; AUG = translation start codon.
 Example of molecular mimicry

(a) 16S rRNA (b) S7 mRNA

Model of secondary structure of Escherichia coli r-protein S7 binding sites on 16S rRNA (a)
and on mRNA from the operon containing the gene encoding S7 (b). Boxed sequences indicate
homology. The structure in (a) was taken from work of Noller and coworkers. In (b), the S12
coding region ends at 501 and the S7 coding region begins at 604.
Figure 2 in Nomura et al. (1980) Proc Natl Acad Sci USA 77: 7084–7088.
 Chloramphenicol resistance: “translational attenuation”
No translation The mRNA has a leader peptide coding region
upstream of the gene coding for chloramphenicol
acetyl transferase (CAT). In the absence of
chloramphenicol (CMP), a ribosome translates the
leader peptide to the stop codon (UGA), and the bulk
of this ribosome obscures the rbs and start codon for
the CAT gene, which is therefore not translated.
No CMP
In the presence of sub-lethal doses of CMP, the
ribosome translating the leader peptide stalls during
the translation of that peptide, because the primary
sequence of the amino acids in the leader peptide
potentiate the ribosome that has translated this peptide
to be more sensitive to CMP than it normally would
be. The peptide itself has been shown to inhibit
Sub-lethal CMP peptidyl transferase. This stalled ribosome has not
progressed far enough to obscure the rbs and start
codon for the CAT gene (in fact, it even improves the
exposure of the rbs and start codon for the CAT gene),
therefore another ribosome will start the translation of
the CAT gene, and the protein will be made.

The sequences shown here are for the catA86 gene

from a gram+ Bacillus species. Similar results have
been obtained for the cmlA gene of (gram-) E. coli.

Model for induction of cat-86. A ribosome stalled at leader codon 6 destabilizes the stem-loop structure, freeing RBS-C,
allowing translation initiation of the cat-86 coding sequence. Based on Lovett & Rogers (1996) Microbiol Rev 60:366.
 Translation Activation Systems: Phage Mu

ORF: open reading frame

Figure 13 in Hattman (1999) Pharmacol Ther 84:367-388

Lysogenic bacteriophage Mu is perhaps best known as a mobile, transposable element. Part of Mu's survival strategy
relies on modification of DNA to make the viral genome resistant to many host restriction endonucleases. The enzyme that
accomplishes this modification is encoded by the mom gene, which is the second gene in an operon containing an upstream,
overlapping gene called com (control of mom) The com gene ("com ORF" in the figure) encodes a small protein of 62
residues that is necessary for mom gene ("mom ORF" in the figure) expression.
Analysis of the sequence and structure of the com–mom mRNA revealed that the translation start region of the mom
reading frame was obscured in a complex stem–loop, and thus unavailable to ribosomes. This is similar to other examples of
RNA structure regulating access to translation start sites. What is unique about this system is that the Com protein binds to
mom mRNA in the vicinity of the translation-inhibiting stem–loop, thereby changing the structure of the mRNA, such that
the inhibitory stem–loop is disrupted and the mom gene translation start signals are now accessible to ribosomes.
In addition to this translational switch, expression of the mom gene is also under transcriptional controls, including
requirement for a phage-encoded transcriptional activator protein ("C" in the figure), and activity of a host methylase enzyme
("Dam" in the figure) to antagonize binding of a host repressor protein ("OxyR" in the figure). Apparently, the phage exerts
all these layers of control on mom expression because overproduction of the Mom modification enzyme is lethal to the host
cell. The requirement for translational activation by Com protein is thus the last regulatory check before the phage commits
the cell to production of Mom protein, a process that has clearly been designed to be under very tight regulation.
excerpt from Goldman E (2007) Translation control by proteins. Encyclopedia of Life Sciences, John Wiley & Sons, UK; http://www.els.net
 Translation Activation Systems: Mitochondrial gene regulation
Only eight major proteins are encoded and synthesized by Saccharomyces
cerevisiae mitochondria, of which seven are involved in energy transduction.
ribosome
These hydrophobic proteins reside in the inner mitochondrial membrane.
Investigation of the synthesis and localization of these proteins elucidated a
pattern of translation activation involving proteins encoded in the nucleus, with
membrane
targets in the 5' untranslated leaders (5' UTLs) of the mRNAs encoding these
yeast mitochondrial proteins.

The nuclear-encoded activator proteins are transported and bound to the

COX1 mitochondrial inner membrane, thereby tethering the mRNAs encoding the
mRNA
mitochondrial membrane proteins. This facilitates targeting the mitochondrial
proteins to their proper destination, the inner membrane, as they are synthesized.

Among the mitochondrial proteins studied in this regard are three subunits of
Cox1 protein cytochrome c oxidase, encoded by the COX 1, COX2 and COX3 genes.
modified from Fig. 6 in Perez-Martinez et al.
Translation of COX1 mRNA requires two nuclear-encoded translational activator
(2009) Mol Cell Biol 20:4371-4380 proteins, PET309 and Mss51 (see figure); translation of COX2 mRNA requires a
single nuclear-encoded translational activator protein, PET111, while translation
of COX3 mRNA uses three different activators, PET54, PET122 and PET494.

The target on COX3 mRNA has been studied by deletion analysis, and appears to be contained within a 151-base
region in the middle of the 613-base 5' UTL. None of the 5' UTLs contain Shine–Dalgarno ribosome-binding sites to
facilitate ribosomes finding the AUG translation starts; thus, other features of the 5' UTLs apparently compensate and
direct translation initiation.
text modified from Goldman E (2007) Translation control by proteins. Encyclopedia of Life Sciences, John Wiley & Sons, UK; http://www.els.net