MIAMI UNIVERSITY
The Graduate School
Certificate for Approving the Dissertation
We hereby approve the Dissertation
of
Heather Joann Beck
Candidate for the Degree
DOCTOR OF PHILOSOPHY
______________________________________
Mitch Balish, Director
______________________________________
Eileen Bridge, Reader
______________________________________
DJ Ferguson, Reader
______________________________________
Xiao-Wen Cheng, Reader
______________________________________
Jack Vaughn, Graduate School Representative
ABSTRACT
Roles of Escherichia coli 5’-terminal AUG triplets in translation initiation and regulation
by
Heather J. Beck
The process of translation initiation involving the purine-rich Shine-Dalgarno sequence in
bacteria has been well studied. However, approximately one-half of the mRNAs across
prokaryotic genomes do not contain a Shine-Dalgarno sequence. Therefore, it is
important to understand the mechanism and ribosome recognition signals involved in the
Shine-Dalgarno-independent translation initiation, which is yet to be elucidated. To do
this, leaderless mRNAs lacking a 5’-UTR are utilized to examine the roles of 5’-terminal
AUGs in ribosome recognition. In this study, we identified AUGs at the 5’-terminus of
canonical Shine-Dalgarno-led mRNAs and examined their ability to act as initiation
codons for leaderless mRNA translation. The identified 5’-terminal AUGs were found to
specify open reading frames that were translated at varying levels of expression, with
some playing a role in downstream coding sequence regulation. Further investigation into
one 5’-terminal AUG-led canonical mRNA, ptrB, identified a novel regulation
mechanism that is distinct from the Shine-Dalgarno mechanism. This novel mechanism
is instead controlled by the 5’-terminal AUG and can regulate in a gene context
independent manner making it ideal for genetic engineering purposes. To further
understand how the 5’-AUG is recognized and bound by ribosomes, a cross-linking
method was developed to identify the initial interactions between leaderless mRNAs and
the ribosome. Overall, this study provides insight into 5’-terminal AUG-directed
translation and the role of 5’-AUGs in ribosomal recognition, binding and translational
regulation.
Roles of Escherichia coli 5’-terminal AUG triplets in translation initiation and regulation
A DISSERTATION
Presented to the Faculty of
Miami University in partial
fulfillment of the requirements
for the degree of
Doctor of Philosophy
Department of Microbiology
by
Heather J. Beck
The Graduate School
Miami University
Oxford, Ohio
2016
Dissertation Director: Mitch Balish
©
Heather J Beck
2016
TABLE OF CONTENTS
General Introduction ........................................................................................................... 1
Chapter 1 ........................................................................................................................... 14
Abstract ......................................................................................................................... 15
Introduction ................................................................................................................... 16
Materials and Methods .................................................................................................. 18
Bacterial strains ......................................................................................................... 18
Reagents..................................................................................................................... 18
Construction of lacZ fusions...................................................................................... 18
β-galactosidase assay ................................................................................................. 18
Preparation of in vitro synthesized transcripts .......................................................... 18
Ribosome isolation .................................................................................................... 19
Primer extension inhibition (toeprint) assay .............................................................. 19
Results ........................................................................................................................... 20
5’-uORFs support translation .................................................................................... 20
5’-uAUGs bind 70S ribosomes.................................................................................. 35
5’-uAUGs can influence downstream gene expression ............................................. 46
Discussion ..................................................................................................................... 50
Chapter 2 ........................................................................................................................... 53
Abstract ......................................................................................................................... 54
Introduction ................................................................................................................... 55
Materials and Methods .................................................................................................. 60
Bacterial strains and reagents .................................................................................... 60
Recombinant DNA procedures .................................................................................. 60
Construction of DNA molecules with randomized sequence regions .................. 60
Gene fusions.......................................................................................................... 61
Preparation of competent cells .................................................................................. 61
Transformations ......................................................................................................... 62
β-galactosidase assay ................................................................................................. 62
In vitro synthesis of RNA .......................................................................................... 62
Primer extension inhibition (toeprint) assay .............................................................. 62
iii
Results ........................................................................................................................... 64
ptrB 5’-uAUG as a ribosome binding signal ............................................................. 64
Predicted SD sequence has unexpected effects ......................................................... 64
Ribosome binding signals present within the ptrB 5’-UTR ...................................... 72
Spacing optimized between the 5’-uAUG and downstream RBS ............................. 78
Expression of the uORF negatively impacts downstream CDS expression .............. 81
Transplantability of regulation by the 5’-UTR of ptrB ............................................. 82
Initiation signals present within a bona fide CDS ..................................................... 87
Discussion ..................................................................................................................... 93
Appendix A ....................................................................................................................... 99
Concluding Remarks ....................................................................................................... 109
References ....................................................................................................................... 117
iv
LIST OF TABLES
Table 1-1. RegulonDB identified genes............................................................................ 21
Table 1-2. Selected genes examined that contain an AUG triplet at the 5’-terminus of its
transcribed canonical mRNA. ........................................................................................... 29
Table 1-3. Translation from the 5’-uORF as a percent of translation relative to its
downstream canonical CDS as determined by β-galactosidase assays performed in
triplicate. ........................................................................................................................... 36
Table 1-4. Expression levels from downstream CDS fusions to lacZ that contain an intact
5’-uAUG (CDS), a mutated 5’-uAUC (KO), or an intact 5’-AUG with a premature stop
codon (StartStop) (see Figure 1-1) including the standard deviation from triplicate
cultures. ............................................................................................................................. 41
Table 1-5. Sequences of mRNAs tested ........................................................................... 44
Table 2-1. DNA sequences of gene fragments, with and without mutations, from ptrB.. 65
Table 2-2. ptrB mutations expressed as a percentage of either the 5’-uAUG in-frame
leader fusion with lacZ (100%) (column 2) or the CDS fused with lacZ (100%) (column
3). ...................................................................................................................................... 82
v
LIST OF FIGURES
Figure I-1: Structural representation of the bacterial ribosome. ......................................... 2
Figure 1-1. Schematic of lacZ fusions constructed for each gene tested .......................... 31
Figure 1-2. 5’uAUGs support varying levels of translation compared to known leaderless
mRNA in E. coli. .............................................................................................................. 33
Figure 1-3. 5’-uAUGs bind 70S ribosomes following the proposed leaderless mRNA
initiation mechanism ......................................................................................................... 38
Figure 1-4. Loss of 30S subunit binding to CDS start codon as a result of the 5’-uAUG
mutation. ........................................................................................................................... 48
Figure 2-1. The ptrB gene sequence. ................................................................................ 58
Figure 2-2. Role of SD sequence in ptrB regulation ........................................................ 69
Figure 2-3. ptrB mRNA structure may play a secondary role in expression .................... 73
Figure 2-4. Scanning mutagenesis of ptrB 5’-UTR .......................................................... 75
Figure 2-5. 5’-UTR mutations and deletions to affect spacing between 5’-uAUG and
downstream RBS. ............................................................................................................. 79
Figure 2-6. Transplantability of regulation by the 5’UTR of ptrB. .................................. 85
Figure 2-7. ptrB 5’-UTR insufficient to stimulate expression of internal RNA fragments.
........................................................................................................................................... 88
Figure 2-8. Expression regulated by both upstream and downstream signals .................. 91
Figure A-1. Schematic of S1 domains and their amino acid positions in the S1 r-protein.
......................................................................................................................................... 101
Figure A-2. Cross-linking of 4S-U cI RNA to purified S1 protein ................................ 103
Figure A-3. Cross-linked positions on S1 ribbon structure. ........................................... 106
vi
DEDICATION
I would like to dedicate this dissertation to my late mentor, Dr. Gary R. Janssen.
Calling him my mentor does not do him justice. He was also my trusted colleague, my
source of constant encouragement and inspiration, my idol, but most importantly, my
beloved friend. I feel blessed to have learned how to be a researcher and a scientist from
him. He has made such an impact on me and will continue to influence my future. I will
always aspire to be the scientist, the teacher and the friend that he was. The man, the
beard, the legend. He was truly inspiring.
vii
ACKNOWLEDGEMENTS
First, I would like to thank my lab mates, Racheal Devine, Sarah Steimer, and
Jackie Giliberti for their continued guidance and assistance in the lab. To Racheal for
teaching me how to do most everything in lab, being patient with me and for being such a
dear friend. To Sarah, without whom I would have never made it through the constant
struggles we faced together in the worst year of my life. Thank you for the very bottom of
my heart for all the laughs and acting as my sounding board while writing even after you
graduated. To Jackie, for also teaching me so much my first year in lab and trusting me to
take over part of her project and for also being such a giant help in editing my
dissertation. I would also like to thank the wonderful undergrads I’ve worked with in the
lab, particularly Leah Carter and Brendan Radel. You have both become very close
friends of mine and I so enjoyed watching you grow in your confidence in the lab.
Next, I would like to thank my committee who all stepped in to help guide me in
the last year of my graduate career. To Dr. Bridge and Dr. Ferguson who graciously read
and edited my dissertation and for their help throughout the years in making my project
better. To Dr. Vaughn who has also helped in the development of my project throughout
the years. To Dr. Cheng who agreed to join my committee without much notice. Also to
Dr. Morgan-Kiss who was also a very supportive member of my committee, even if it
was for just a short time. Finally, to Dr. Balish who has become my new advisor. I know
how much work it was to edit my dissertation and I can never thank you enough for
agreeing to help me after Gary’s passing. Your guidance in this endeavor and throughout
the years has been enormously helpful.
I would also like to thank the rest of the microbiology department at Miami
University, both the faculty and the graduate students who have been supportive over the
years. I have made such wonderful life-long friends throughout the years, without whom,
I likely would have never made it this far. A special thank you to Ryan R., Bill P., Dan
Z., Liz F., Steve F., Jenna D., Ryann B., Steve, D., and Tzvia S. Also to our favorite
Canadian, Chris Sedlacek.
viii
I would like to thank my Janssen lab family. Attending the reunions over the
years has allowed me to meet so many previous graduate students who have been brought
together by Dr. Gary Janssen. They are some of the most kind, approachable and helpful
people I’ve ever met and they so easily welcomed me in as part of the family and have
been there for me, especially after Gary’s passing. In addition to the graduate students,
Kob, Leah and Ezra Janssen: thank you for being so kind and helpful to me.
Lastly, I would like to thank my family, whose unwavering support and guidance
have made me the person I am today. Thank you all.
ix
General Introduction
Leaderless mRNAs utilize a 5’-terminal AUG as the signal for translation initiation.
However, the exact mechanism for ribosome recognition and binding of these non-canonical
mRNAs is unknown. This work seeks to examine the 5’-terminal AUG as a ribosome
recognition signal to determine how leaderless mRNAs initially interact with the ribosome and if
the 5’-terminal AUG can be used as a signal for canonical Shine-Dalgarno (SD) - led mRNAs to
influence translation.
Translation is the process of protein synthesis via the ribosome, a macromolecular
ribonucleoprotein complex with enzymatic activity to catalyze peptide bond formation. The
ribosome has the ability to decode the template mRNA to synthesize a polypeptide product. Due
to the lack of a nuclear membrane in prokaryotes, translation is tightly coupled with transcription
in the cell, so much so that translation can occur before transcription is terminated. Translation
occurs in four steps: initiation, elongation, termination, and ribosome recycling. Translation
initiation is the rate-limiting step in protein synthesis, occurring on the order of seconds (Laursen
et al., 2005). During initiation, the ribosome complex consisting of a large and small subunit is
assembled on the mRNA to be translated. It then proceeds to elongation, a much faster process,
synthesizing polypeptides at 12 amino acids per second in Escherichia coli (Kennell and
Riezman, 1977). The mRNA is decoded within the ribosome complex three nucleotides at a
time with each triplet pairing with anticodons of its cognate tRNA until the ribosome reaches a
stop codon. The stop codon signals the termination step in which the ribosome complex
dissociates from the mRNA, the polypeptide is released, and the subunits are recycled.
The Ribosome
The bacterial ribosome consists of two subunits, 30S and 50S, which join together to
form a 70S ribosome that catalyzes protein synthesis. The small 30S ribosomal subunit contains
21 distinct proteins and one rRNA (16S) and has a mass of 0.8 MDa (Laursen et al., 2005). This
subunit is traditionally divided into an upper third, called the head, connected by the neck to the
body and platform with a protrusion called the spur (Figure I-1A) (Laursen et al., 2005). The 30S
subunit contains the decoding center, positioned at the upper part of the body and lower part of
the head, made entirely of RNA, which controls translational fidelity by monitoring base pairing
1
Figure I-1: Structural representation of the bacterial ribosome. Adapted from Simonetti et
al., 2009. (A) Diagram of 30S ribosomal subunit (brown) with view from above and the different
regions indicated. The mRNA binding platform is highlighted in pink/red and the position of the
anti-Shine-Dalgarno sequence (aSD) of the 16S rRNA is in cyan. (B) The 30S subunit with the
initial mRNA interaction: docking onto the platform and binding the aSD before being loaded
into the mRNA channel. Also included is a view of the ribbon structure of the 30S subunit preinitiation complex with an mRNA bound from a side vantage point. (C) Diagram of the large
50S ribosomal subunit outlining the three protuberances. (D) Depiction of the 50S subunit
(green) bound to the 30S subunit (brown) forming the 70S initiation complex with the mRNA
(purple) loaded into the mRNA channel and interacting with the initiator tRNA (dark red).
2
3
between each mRNA codon and its cognate tRNA (Green and Noller, 1997). Cognate tRNAs
have lower dissociation rates from the ribosome and, when bound, cause an induced-fit
conformational change of the ribosome (Ogle et al., 2001). This process allows for tightly
controlled decoding with an error rate of tRNA selection in E. coli from 10-3 to 10-4 (Laursen et
al., 2005). The 30S subunit also contains three tRNA binding sites, the acceptor (A), peptidyl
(P), and exit (E) sites. The functions of these sites will be discussed below.
The large 50S ribosomal subunit contains 34 distinct proteins and two rRNAs (5S and
23S) with a total mass of 1.5 MDa (Laursen et al., 2005). The 50S subunit is more compact than
the 30S subunit, consisting of a rounded base with three protuberances (L1 stalk, central
protuberance and L7/L12 stalk Figure I-1C) (Wilson and Nierhaus, 2003). The 50S subunit
contains the catalytic peptidyltransferase center (PTC) where formation of the peptide bond
occurs. The PTC consists primarily of RNA with no proteins within 18Å of the catalytic site
(Nissen et al., 2000). A tunnel starts at the PTC and the nascent polypeptide exits through this
tunnel, which is approximately 100 Å long with an average diameter of 15 Å (Nissen et al.,
2000).
The process of translation initiation
For polypeptide formation to occur, a charged cognate tRNA (with the exception of the
initiator tRNA which binds the P-site) must bind to the A-site (Laursen et al., 2005). The αamino group of the tRNA in the A-site attacks the carbonyl group of the P-site peptidyl group
tRNA to form a peptide bond (Laursen et al., 2005). The now deacylated tRNA in the P-site
moves to the E-site, which is specific for deacylated tRNAs, for ejection (Schmeing et al., 2003).
The A-site peptidyl-tRNA moves to the P-site and a new cognate aminoacyl-tRNA binds in the
A-site (Laursen et al., 2005). In the conventional pathway of initiation, a ternary complex is
formed from the 30S ribosomal subunit, the messenger RNA (mRNA), and the initiator tRNA.
30S ribosomal subunits bind the purine-rich SD sequence of the mRNA via complementary
pairing to the anti-Shine-Dalgarno (aSD) sequence of the 16S rRNA of the 30S subunit, thereby
forming a binary complex (Shine and Dalgarno, 1974; Yusupova et al., 2001; Benelli et al.,
2003). This initial interaction creates an unstable 30S pre-initiation complex which positions the
mRNA on the ribosomal platform (Marzi et al., 2007). In the pre-initiation complex, the 30S
subunit platform proteins S2, S7, S11, and S18 are near and may contact the mRNA,
4
suggesting a role for these proteins in docking the mRNA on the platform (Simonetti et al.,
2009). Once bound to the platform, the mRNA is then adjusted into the mRNA tunnel of the 30S
subunit. This may require unwinding the mRNA, a process for which several platform proteins
(S2, S7, S11, S18, S21) as well as S1 with its helicase properties may be responsible, to promote
movement into the tunnel and proper codon-anticodon pairing between the mRNA and the tRNA
(Simonetti et al., 2009). The initiator tRNA binds to form the ternary complex, which is
stabilized by the SD-aSD interaction and leads to proper placement of the start codon in the Psite of the ribosome (Ringquist et al., 1993; Wimberly et al., 2000). Final adjustments are then
made in the decoding center after which translation proceeds into elongation.
Three additional proteins, initiation factor 1 (IF1), initiation factor 2 (IF2), and initiation
factor 3 (IF3) are also involved in translation initiation. These factors exert their effects by
binding within the tRNA binding sites on the 30S subunit (Ramakrishnan, 2002). IF1 binds to
the A-site of the 30S ribosomal subunit to block tRNA binding to the A-site (Ramakrishnan,
2002), directing the initiator tRNA to the P-site (Laursen et al., 2005). IF1 also increases the
affinity of the 30S subunit for IF2 and IF3 and stimulates their activity (Gualerzi and Pon, 1990).
IF3 binds the E-site on the interface side of the platform and neck of the 30S subunit to directly
prevent subunit association, thereby also playing a role in ribosomal recycling (McCutcheon et
al., 1999; Ramakrishnan, 2002). IF1 and IF3 are ejected from the 30S subunit to allow for
association of the 50S subunit, which is stimulated by IF2 (Laursen et al., 2005). IF2 is thought
to bind over the A-site to interact with IF1 while also binding the aminoacyl end of initiator
tRNA in the P-site for proper positioning (Ramakrishnan, 2002). Once the initiator tRNA is
properly positioned in the P-site, the GTP bound to IF2 is hydrolyzed and IF2 is released
(Laursen et al., 2005). This process results in the ternary complex formation of mRNA, initiator
tRNA and 70S ribosome, with the initiator tRNA placed in the P-site codon-anticodon pairing
with the mRNA start codon.
Factors involved in the regulation of translation initiation
The variable element in translation initiation is the mRNA, which can differ in sequence
and structural components. Most mRNAs have a 5’-untranslated region (5’-UTR) upstream of
their initiation codon that contains ribosomal recognition features that influence the mRNA’s
translation efficiency. The start codon itself can also influence translation efficiency. Typically,
5
an AUG codon is most efficient for initiation and is the most prevalent start codon in E. coli
(90%) (Schneider et al., 1986). However, other codons also act as start codons in E. coli, with
GUG used in 8% of mRNAs, UUG in 1% of mRNAs (Schneider et al., 1986), and AUU used in
two mRNAs in E. coli (infC and pcnB) (Binns and Masters, 2002). The sequences flanking the
start codon comprise the ribosome binding site, which typically extends approximately 30
nucleotides (Laursen et al., 2005). This region can include the SD sequence, the spacer region
between the SD sequence and the start codon (7±2 nucleotides) and the start codon (Laursen et
al., 2005).
The ribosome can recognize the binding site through mRNA sequence and/or secondary
structural features. Multiple sequence features, both upstream and downstream of the start codon,
have been implicated in influencing ribosome binding and therefore translation efficiency. The
“cumulative specificity initiation mechanism” theory suggests that both upstream and
downstream sequence elements work cooperatively, yet independently to ensure no singular
interaction is essential, to select the translation initiation site (Nakamoto, 2011). The sequence
elements that influence expression and may therefore be part of this theory include, but are not
limited to, the SD sequence, CA repeats, AU richness, and enhancer elements. The SD sequence
directly interacts with the aSD sequence at the 3’end of the 16S rRNA (Shine and Dalgarno,
1974). This interaction is thought to contribute to ribosome binding and proper placement of the
start codon in the P-site of the ribosome (Shine and Dalgarno, 1974). AU-rich sequences
stimulate translation when located upstream or downstream of a start codon (Dreyfus, 1988;
Qing et al., 2003) through an interaction with ribosomal protein S1 (Komarova et al., 2005). It
has been hypothesized that the AU richness prevents the formation of secondary structure around
the start codon allowing for easier ribosomal entry and binding to the mRNA (Kaberdin and
Blasi, 2006). Downstream adenine-rich sequence motifs have been implicated as general
enhancers of E. coli translation (Brock et al., 2007). CA multimers introduced downstream of
the start codon increased expression independent of translation signals contained within the
untranslated leader region by increasing the ribosome binding strength (Martin-Farmer and
Janssen, 1999). The impact of these sequence features on translation demonstrate the importance
of the mRNA’s primary sequence in translation efficiency.
6
Apart from sequence features, mRNA secondary structure can play roles in translation
initiation. In general, ribosome binding requires local single-stranded structure for proper
interaction (Draper, 1987; Draper et al., 1998; Gualerzi and Pon, 1990). Therefore, secondary
structure can occlude regions of mRNA from ribosome binding and can inhibit translation.
Since RNAs are dynamic, they can be highly influenced by their environment and can alternate
between various secondary structures with different thermodynamic stabilities. Changes in
temperature can result in changes in RNA secondary structure, a property utilized by bacteria as
a mechanism of regulation (Klinkert and Narberhaus, 2009). Increase in temperature destabilizes
secondary structure that would otherwise occlude the translation initiation region, thereby
allowing for ribosome binding (Kaberdin and Blasi, 2006). Conversely, some mRNAs are
upregulated during a decrease in temperature due to structural changes that shift the site of a
hairpin loop and as a result eliminate the secondary structure at the translation initiation region
(Thieringer et al., 1998; Gualerzi et al., 2003; Kaberdin and Blasi, 2006).
In addition to secondary structure, trans-acting elements also affect translation. There are
many examples of proteins binding to translational operators to repress translation (Salavati and
Oliver, 1995; Kozak, 1999; Kaberdin and Blasi, 2006). This mechanism is especially effective in
operons in which downstream translation is coupled to and therefore dependent upon translation
of the preceding region. In this way, translation of multiple ORFs can be repressed by a single
binding event. Examples of this type of regulation include the α operon and the spc operon of E.
coli, both of which encode ribosomal proteins and are repressed by binding of a ribosomal
protein to the operon, thereby self-regulating expression (Singer and Nomura, 1985; Kaberdin
and Blasi, 2006). It has also been hypothesized that ribosome binding can indirectly increase
expression by protecting the mRNA from RNase cleavage by competing for binding sites
(Kaberdin and Blasi, 2006). Low-molecular weight effectors such as amino acids, coenzymes,
and vitamins also bind to structured noncoding segments of mRNA, known as riboswitches, to
control expression without the assistance of protein factors (Tucker and Breaker, 2005; Winkler
and Breaker, 2005). Riboswitches are able to recognize the effectors and fold into an alternative
structure to regulate the activity of the mRNA by opening or sequestering the RBS (Breaker,
2012).
7
Translation can also be regulated by binding of small noncoding RNAs termed antisense
RNAs. Antisense RNAs typically base-pair to complimentary sequences within the translation
initiation region of their target mRNAs to block ribosome binding and lead to degradation
(Gottesman, 2005). Trans-encoded antisense RNAs are less specific and can affect multiple
targets for widespread regulation but require the chaperone protein Hfq to exert their effects
(Wagner et al., 2002). Antisense RNAs can have positive or negative effects on the translation
initiation of their target mRNAs by either blocking ribosome binding or causing structural
rearrangements that make the ribosome binding site more accessible (Kaberdin and Blasi, 2006).
The expendable role of the Shine-Dalgarno sequence
The SD sequence can be important in both ribosomal recognition and proper spatial
mRNA binding. The mechanism for mRNA-ribosome binding via the SD-aSD interaction has
been well studied and is well understood (Laursen et al., 2005 and references therein). However,
the SD-aSD interaction is not the sole determinant of translational efficiency or selection of a
translational start site (Melancon et al., 1990; Calogero et al., 1988; Benelli et al., 2003).
Initiation complex formation can be achieved in the absence of a SD sequence but requires
ribosomal protein S1 (Tzareva et al., 1994). Often translation initiation regions are characterized
by signals other than the SD sequence, especially in translationally coupled open reading frames
that are more highly dependent on the AUG start codon for recognition of the ribosome binding
site (Andre et al., 2000). In multiple cases, deletion of the entire 5’-UTR leads to higher levels of
expression than mutation of only the SD sequence (Van Etten and Janssen, 1998; Condo et al.,
1999; Benelli et al., 2003), suggesting that mRNAs absent of a 5’-UTR are more efficiently
expressed that those with absent SD sequences. Although the SD site may be important in
proper positioning of the start codon in the P-site, it has been suggested that there is an initial
SD-independent binding interaction (de Smit and van Duin, 2003). This suggestion is based on
evidence that the rate constant for association of mRNA and the 30S subunit is independent of
SD-aSD interaction strength.
An increasing number of non-SD led genes (i.e., those with no recognizable SD sequence
within the untranslated leader) as well as genes lacking a 5’-UTR altogether, thereby constituting
leaderless mRNA, have been identified (Chang, 2006). A bioinformatic study of 162 prokaryotic
genomes concluded that the number of SD-led genes is equal to the number of non-SD led genes
8
(Chang, 2006). Although an SD sequence is not required for initiation, it is unknown how the
ribosome is recruited to non-SD led mRNAs. Especially in the case of leaderless mRNA, lacking
both a 5’-UTR and SD sequence, the ribosome binding signals and mechanism of start codon
positioning in the P-site have yet to be identified.
Leaderless mRNA
The study of leaderless mRNA has expanded our knowledge regarding the evolution of
translation. Leaderless mRNAs have been identified in every domain of life and are proposed to
be ancestral to all mRNAs (Grill et al., 2000). Translation factors from all three kingdoms share
a significant degree of homology, suggesting that current components might have been present at
the universal ancestor stage (Moll et al., 2002). Regardless of the origin of the leaderless
mRNA, it can be translated by the translational machinery from any of the three domains (Grill
et al., 2000). In contrast, leadered bacterial mRNAs are poorly translated in archaeal and
eukaryotic cell-free translational assays (Grill et al., 2000). Leaderless mRNA are more frequent
in Gram-positive bacteria than Gram-negative bacteria but have been identified in E. coli,
Caulobacter crescentus, and Thermus thermophilus and have been predicted to be
underrepresented due to the difficulties in identifying leaderless mRNAs bioinformatically (Moll
et al., 2002). Leaderless mRNAs have been identified in many Gram-positive genera including
Streptococcus, Lactococcus, Streptomyces, and Corynebacterium (Janssen, 1993) and appear to
be quite common in archaea (Benelli et al., 2003, 2009; Brenneis and Soppa, 2009). The trend in
archaea suggests that monocistronic transcripts and genes located at the 5’-proximal end of
operons are leaderless, but internal genes in operons are preceded by SD sequences (Tolstrup et
al., 2000; Slupska et al., 2001; Benelli et al., 2003). Distinct SD sequence-dependent and
leaderless mechanisms for ribosome/mRNA interactions have been demonstrated in archaeal
translation initiation (Benelli et al., 2003), supported by the prevalence of both SD sequence-led
and leaderless mRNAs in archaea.
In bacteria, leaderless mRNAs appear to undergo translation initiation through a novel
pathway initially binding a 70S ribosome (Balakin et al., 1992; Udagawa et al., 2004; Moll et al.,
2004). This binding is dependent on an AUG initiation codon and is stabilized by initiator tRNA
in E. coli (Van Etten and Janssen, 1998; O’Donnell and Janssen, 2002; Brock et al., 2008). This
binding is also positively influenced by IF2 through stabilization of fMet-tRNA binding, and
9
antagonized by IF3, which promotes dissociation of 70S ribosomes (Tedin et al., 1999; Grill et
al., 2000, 2001). However, ribosome binding to leaderless mRNAs (Balakin et al., 1992) and
progression from initiation into the elongation phase of translation occur in the absence of all IFs
(Udagawa et al., 2005).
Identification of a ribosomal particle deficient in up to 11 small subunit ribosomal
proteins (r-proteins) that retains the ability to exclusively translate leaderless mRNA has
provided further insight into the mechanism of leaderless mRNA translation and ribosomal
scaffolding (Kaberdina et al., 2009). This r-protein-deficient ribosome, termed 61S, forms in the
presence of the antibiotic kasugamycin (Ksg), which inhibits initiation complex formation by
binding the mRNA track at the P- and E-site codons, thereby indirectly dissociating the P-site
bound fMet-tRNAfMet from 30S subunits (Schulenzen et al., 2006; Schurwirth et al., 2006). After
treatment with Ksg, translation of leadered mRNAs is inhibited but leaderless mRNA translation
is still fully functional in vivo and in vitro (Kaberdina et al., 2009). The r-proteins absent in the
61S ribosome cluster in regions associated with the mRNA’s path through the ribosome,
including regions associated with the aSD sequence (Wilson, 2009). Therefore, these regions
might be dispensable for mRNAs lacking the SD sequence and 5’-UTR altogether. Although the
active centers of the ribosome are primarily constructed of rRNA, certain r-proteins are
necessary to support the rRNA structure (Wilson and Nierhaus, 2005). The case of the 61S
ribosome translating leaderless mRNAs highlights the existence of a minimal set of r-proteins
necessary for rRNA stabilization and active translation. The 61S ribosome also has evolutionary
implications, suggesting that an ancestral form of the small subunit with the proteins deficient
from the 61S may have been added later in history (Kaberdina et al., 2009). The 61S can also
provide insight into the mechanism of ribosome binding as well as mRNA: ribosome contacts
and mRNA movement through the ribosome.
The set of ribosomal contacts for canonical leadered mRNA has been established (La
Teana et al., 1995; Simonetti et al., 2009). Crosslink studies using E. coli 30S subunits have
shown that the mRNA’s +2 position contacts the 16S rRNA at the 1530 position with the
mRNA’s +3 position contacting the ribosomal platform proteins S18 and S21 (La Teana et al.,
1995). The crosslinks shift upon addition of IFs, with the 16S rRNA’s 1530 position instead
contacting the mRNA’s -3 position and a new crosslink forming between the mRNA’s +2 and -3
10
positions and the ribosomal head protein S7 (La Teana et al., 1995). These results suggest that
the ribosome shifts in the 5’ direction on the mRNA, situating the mRNA closer to the P-site in
the presence of IFs for canonical leadered mRNAs.
Examination of 5’-terminal AUGs
This study aims to provide insight into the mechanism of 5’-AUG recognition, ribosome
binding, and translation initiation of leaderless mRNA. Since the 5’-AUG has been implicated
as the signal for ribosome binding in leaderless mRNA, I hypothesize that an AUG triplet at the
5’-terminus of canonical SD-led mRNAs will also act as a signal for binding and as an initiation
codon for upstream open reading frame translation. I will then investigate the 5’-AUGs’
translational efficiency and regulatory effects. Lastly, to provide insight into the initial
interactions between the ribosome and the 5’-AUG, I will identify contact sites between
ribosomal protein S1 and leaderless mRNA through development of a novel method that utilizes
photochemical cross-linking and coupled mass spectrometry analysis.
The importance of a 5’-terminal AUG start codon has been demonstrated in leaderless
and “unleadered” (i.e., after removal of the leader sequence) mRNAs through the introduction of
non-cognate initiation codons, resulting in reduced ribosome binding and translational efficiency
(Van Etten and Janssen, 1998; O’Donnell and Janssen, 2002). However, the effect observed due
to the use of non-cognate initiation codons is specific to leaderless mRNAs and does not occur in
leadered mRNA (O’Donnell and Janssen, 2002). Addition of a 5’-terminal AUG to an internal
segment of lacZ RNA makes it competent to form ternary complexes with 70S ribosomes and
tRNA (Brock et al., 2008), suggesting that a 5’-terminal AUG is a sufficient signal for ribosomes
to identify a leaderless mRNA. Also, a naturally occurring leaderless mRNA with a deletion of
30 nucleotides from its 5’-terminus cannot bind ribosomes; however, addition of a 5’-terminal
AUG triplet to the truncated leaderless mRNA restores 70S ribosome binding and enables it to
compete successfully with naturally leaderless mRNA for ribosome binding in vitro (Brock et
al., 2008). These data illustrate that the presence of a 5’-terminal AUG is sufficient for an RNA
molecule to bind ribosomes and be translated as a leaderless mRNA and identified the 5’terminal AUG as a critical signal for recognition by a ribosome. In the present study, I examined
not only naturally occurring leaderless mRNAs but also 5’-terminal AUGs at the end of
11
canonical SD-led mRNAs, to further the understanding of leaderless mRNA prevalence and the
mechanism of translation initiation utilized by leaderless mRNA.
Because 5’-AUGs are signals for ribosomal recognition, 5’-AUGs at the end of canonical
SD-led mRNAs may also bind ribosomes. In this work I searched for these mRNAs in the E. coli
genome and sought to determine if the newly discovered 5’-terminal AUGs are efficiently
translated in a manner similar to leaderless mRNAs. I hypothesize that the 5’-terminal AUGs
present on canonical mRNAs are bound by 70S ribosomes and translated to varying degrees.
These 5’-terminal AUG-specified ORFs may exhibit a variety of functions, such as acting as
binding signals, exerting regulatory functions, or coding for small peptides. I further examined
one identified 5’-terminal AUG-led canonical mRNA, ptrB, to determine its ability to exert
regulatory functions on the downstream coding sequence. I hypothesize that the 5’-terminal
AUG, in addition to the predicted SD sequence, acts as a regulatory signal to control downstream
translation in a novel fashion.
This study also aimed to explore how leaderless mRNAs interact with the translational
machinery. Although contacts between the ribosome and mRNA are identified for canonical
mRNAs, it is unclear how leaderless mRNA are initially bound by ribosomes. The contacts
between specific ribosomal proteins and leaderless mRNA were investigated using 4-thiouridine
(4S-U) photochemical cross-linking. Previous cross-linking studies done with leaderless mRNA
showed association with several ribosomal proteins (S1, S3, S4, S7, and S10/S18) even in the
absence of initiator tRNA (Brock et al., 2008). To expand on this research, I cross-linked
purified ribosomal protein S1 to 4S-U-tagged, naturally occurring cI leaderless mRNA to identify
specific sites on the protein that contact the mRNA’s 5’-AUG. This study will help to expand the
knowledge on the alternative mechanism used by leaderless mRNAs for translation initiation.
Variations in conventional translation initiation mechanisms exist in both prokaryotes and
eukaryotes (Malys and McCarthy, 2010). It is important to understand these deviations from
traditional initiation and the signals used for ribosome recognition and binding because they
represent a significant proportion of all mRNAs (Chang et al., 2006). My research focuses on
how 5’-AUGs are recognized by ribosomes to initiate translation and the roles of 5’-AUGs in
regulation in E. coli. Due to the ancestral nature of leaderless mRNAs, it is likely that
understanding leaderless mRNA initiation mechanisms will provide insight into universal
12
principles of translation initiation in all organisms. Identification of ribosomal recognition and
binding signals could also be used in genetic engineering and designing expression vectors to
fine-tune translational expression levels.
13
Chapter 1
5’-terminal AUGs in Escherichia coli mRNAs with Shine-Dalgarno sequences:
identification and analysis of their roles in non-canonical translation initiation
Heather J. Beck1*, Ian M. C. Fleming1#a, Gary R. Janssen†
1
Department of Microbiology, Miami University, Oxford, Ohio, United States of America
#a
Guild BioSciences, Dublin, Ohio, United States of America
* Corresponding author
Email: [email protected] (HJB)
Published in PLoS One
14
Abstract
Analysis of the Escherichia coli transcriptome identified a unique subset of messenger
RNAs (mRNAs) that contain a conventional untranslated leader and Shine-Dalgarno (SD)
sequence upstream of the gene’s start codon while also containing an AUG triplet at the
mRNA’s 5’- terminus (5’-uAUG). Fusion of the coding sequence specified by the 5’-terminal
putative AUG start codon to a lacZ reporter gene, as well as primer extension inhibition assays,
reveal that the majority of the 5’-terminal upstream open reading frames (5’-uORFs) tested
support some level of lacZ translation, indicating that these mRNAs can function both as
leaderless and canonical SD-leadered mRNAs. Although some of the uORFs were expressed at
low levels, others were expressed at levels close to that of the respective downstream genes and
as high as the naturally leaderless cI mRNA of bacteriophage λ. These 5’- terminal uORFs
potentially encode peptides of varying lengths, but their functions, if any, are unknown. In an
effort to determine whether expression from the 5’-terminal uORFs impacts expression of the
immediately downstream cistron, we examined expression from the downstream coding
sequence after mutations were introduced that inhibit efficient 5’-uORF translation. These
mutations were found to affect expression from the downstream cistrons to varying degrees,
suggesting that some 5’-uORFs may play roles in downstream regulation. Since the 5’-uAUGs
found on these conventionally leadered mRNAs can function to bind ribosomes and initiate
translation, this indicates that canonical mRNAs containing 5’-uAUGs should be examined for
their potential to function also as leaderless mRNAs.
15
Introduction
Translation initiation is the rate-limiting step in protein synthesis and requires ribosomes
to recognize and bind messenger RNAs (mRNA). In the conventional pathway of initiation, a
ternary complex is formed between the 30S ribosomal subunit, the mRNA, and the initiator
tRNA with the aid of three initiation factors (for review see Laursen et al., 2005). Briefly, 30S
ribosomal subunits bind the Shine-Dalgarno (SD) sequence of the mRNA via complementary
pairing to the 16S rRNA anti-Shine-Dalgarno (aSD) sequence (Shine and Dalgarno, 1974;
Yusupova et al., 2001). The initiator tRNA binds to the complex, stabilized by the SD-aSD
interaction, and promotes the proper placement of the start codon in the P-site of the ribosome
with subsequent translation initiation (Ringquist et al., 1993; Wimberly et al., 2000). The
mechanism for mRNA-ribosome binding via the SD-aSD has been well studied and is thought to
be a prerequisite step for translation initiation (Laursen et al., 2005; Gualerzi and Pon, 2015 and
references therein). However, an increasing number of genes have been identified that lack SD
sequences or lack a 5’ untranslated region (5’-UTR) altogether (Chang et al., 2006). mRNAs
lacking a 5’-UTR are referred to as leaderless mRNAs and have been reported in all domains of
life (Janssen, 1993). Leaderless mRNA lack ribosome binding signals that would otherwise be
contained within the 5’-UTR, such as the SD sequence. The widespread occurrence of these noncanonical mRNAs suggests that they contain features allowing recognition by ribosomes from all
translation systems.
Leaderless mRNA appear to follow a novel pathway by which a 70S ribosome binds the
5’-terminal AUG to initiate translation (Udagawa et al., 2004; Moll et al., 2004). Binding
depends on an AUG initiation codon and is stabilized by initiator tRNA in Escherichia coli (Van
Etten and Janssen, 1998; O’Donnell and Janssen, 2002; Brock et al., 2008). Addition of a 5’terminal AUG to an internal segment of lacZ mRNA makes it competent to form ternary
complexes with 70S ribosomes and initiator tRNA (Brock et al., 2008), suggesting that a 5’terminal AUG might be a sufficient signal for ribosomes to identify a leaderless mRNA.
Furthermore, a thirty-nucleotide deletion from the 5’-terminus of the naturally occurring
leaderless cI mRNA results in the loss of the ability to bind ribosomes. However, addition of a
5’-terminal AUG triplet to the truncated cI leaderless mRNA restores 70S ribosome binding and
allows it to compete with the native leaderless mRNA for ribosome binding in vitro and form
translationally active complexes in vivo (Brock et al., 2008). Taken together, these results show
16
that the presence of a 5’-terminal AUG is sufficient for an RNA molecule to bind ribosomes and
be translated as a leaderless mRNA.
To further investigate the hypothesis that a 5’-terminal AUG is sufficient for translation
of an mRNA, we sought to identify AUG triplets that occur at the 5’-termini of canonical, SDled mRNAs. Such 5’-terminal AUGs would have the potential to bind ribosomes and allow for
translation of a second open reading frame (ORF), in addition to translation from the start codon
of the downstream SD-led coding sequence (CDS). Previous work has demonstrated the
abundance of small ORFs within 5’-UTRs in both eukaryotic and prokaryotic organisms. In
prokaryotic organisms, these small ORFs often encode small peptides (Hobbs et al., 2011) and
can regulate downstream ORF expression (Lovett and Rogers, 1996; Gaba et al., 2001). In this
study, we utilized the RegulonDB database (Gama-Castro et al., 2011) to conduct an in silico
search of E. coli for canonical, SD-led mRNAs that contain 5’-terminal AUG triplets. These 5’terminal upstream AUGs (5’-uAUGs) were assayed for their ability to act as initiation codons for
translation of the putative 5’-terminal upstream open reading frame (5’-uORF). The 5’-uORF
translational activity and ability to bind ribosomes was also analyzed, as well as their effect on
the translation efficiency of their respective downstream SD-led CDS. Our results suggest that a
number of canonical SD-led mRNAs contain 5’-uAUGs that bind 70S ribosomes and support
biologically relevant levels of translation. Our results also suggest that certain 5’-uAUGs and
their defined ORFs impact regulation of downstream expression.
17
Materials and Methods
Bacterial strains. E. coli DH5α (New England Biolabs [NEB]) was used as the host for all
plasmid DNA manipulations. E. coli RFS859 (F-, thr-1, araC859, leuB6, Dlac74, tsx-274, l-,
gyrA111, recA11, relA1, thi-1) (Schleif, 1972) was used as host for the expression and assay of
lacZ fusion mRNA constructs. E. coli K12 total genomic DNA was used as a template for PCR
amplifications to isolate the genes of interest used in this study.
Reagents. Radio-labeled [γ-32P] ATP (6000 Ci/mmol, 150 mCi/mL) was purchased from Perkin
Elmer. Restriction endonucleases, T4 DNA ligase, T4 polynucleotide kinase (PNK), and T7
RNA polymerase were purchased from New England Biolabs and used according to
manufacturer’s instructions. RNase-free DNase I (Roche), AMV reverse transcriptase (Life
Sciences), and Pfu DNA polymerase (Stratagene) were used according to the manufacturer’s
specifications. DNA oligonucleotides were purchased commercially (IDT). The lacZ-specific
oligonucleotide 5’-GTTTTCCCAGTCACGACGTTG-3’, which anneals to positions +99 to +78
of the lacZ coding sequence in lacZ fusions, was used in the primer extension inhibition
(toeprint) assays.
Construction of lacZ fusions. Codons 1-16 of each gene tested, including the putative upstream
open reading frame (uORF) expressed within 5’-UTR, were fused to the fifth codon of a lacZ
reporter gene and cloned into pUC18-derivative plasmids (Janssen and Bibb, 1993) containing
an ampicillin resistance marker. The constructs contained an upstream lac promoter (TATAAT).
The 5’-uORF for each gene tested was fused to the fifth codon of a lacZ reporter gene just
upstream of its in-frame stop codon.
β-galactosidase assay. β -galactosidase assays were performed as previously described (Miller,
1992).
Preparation of in vitro synthesized transcripts. The cloned plasmids were used as templates in
PCR amplifications utilizing a primer to incorporate the T7 RNA polymerase promoter sequence
(5’-TAATACGACTCACTATAG-3’). This produced DNA fragments containing a T7 promoter
sequence, allowing for in vitro transcription with T7 RNA polymerase and production of RNA
used in toeprint reactions. RNAs were synthesized and purified as described (Fredrick and
Noller, 2002). RNAs used in toeprint assays were synthesized by combining purified PCR
18
amplicons (constructs with lacZ fusions containing a T7 promoter) and T7 RNA polymerase in
1X buffer (40 mM Tris-HCl at pH 7.8, 25 mM MgCl2, 1 mM spermidine, 0.01% Triton X-100, 5
mM each NTP, and 30 mM dithiothreitol). Transcription reactions were incubated for
approximately 4 h at 37°C, and 40 mM ethylenediamineetetraacetic acid (EDTA) was added.
Samples were treated with DNase (Roche) for 15 min at 37°C. RNA was ethanol-precipitated
and suspended in RNA loading dye (50% formamide, 0.05% bromophenol blue, 0.05% xylene
cyanol). Samples were subjected to polyacrylamide gel electrophoresis (PAGE; 6% acrylamide,
7 M urea) and full-length products were excised using UV shadowing. Gel slices were incubated
overnight at room temperature in elution buffer [(300 mM NaOAc at pH 5.2, 0.1% sodium
dodecyl sulfate (SDS), 1 mM EDTA)] with gentle rocking. The supernatant was phenolextracted and ethanol-precipitated.
Ribosome isolation. Isolation of E. coli MRE600 70S ribosomes and 30S ribosomal subunits
was performed as previously described (Martin-Farmer and Janssen, 1999). The same batch of
ribosome preparations was used in each initial and duplicate primer extension inhibition assay.
For confirmation of unexpected binding signals, a second ribosomal batch preparation was used
and in each case the binding signals were reproduced (not shown).
Primer extension inhibition (toeprint) assay. DNA oligonucleotides were phosphorylated at
the 5’-terminus using [γ-32P] ATP (6000 Ci/mmol, 150 mCi/mL; Perkin Elmer) and T4 PNK in
1X kinase buffer for 30 min at 37°C and annealed to 3’-termini of RNA as previously described
(Fredrick and Noller, 2002). Annealed RNA was incubated with 30S subunits or 70S ribosomes
with or without tRNAfMet for 15 min at 37°C. Reactions were transferred to ice, and reverse
transcriptase was added to extend from the labelled oligonucleotide primer to produce cDNA.
The reactions were incubated for 15 min at 37°C and stopped by the addition of 0.3M NaOAc
and 100% ethanol and precipitated overnight at -80oC. Precipitated complexes were collected by
centrifugation and dissolved in loading dye (80% deionized formamide, 10 mM NaOH, 1 mM
EDTA, 0.5% bromophenol blue, and xylene cyanol), followed by heat treatment (95°C, 5 min)
and PAGE (6% acrylamide, 7 M urea) in 1X TBE. Gels were visualized via autoradiography. In
each case, toeprint assays were performed at least twice with reproducible results.
19
Results
5’-uORFs support translation
Bioinformatic analysis of the E. coli RegulonDB transcriptome database (Gama-Castro et
al., 2011) of all promoter types identified several canonical mRNAs with an untranslated leader
and SD-led open reading frame (ORF) that also contained an AUG triplet within three
nucleotides of the mRNA’s 5’-terminus (i.e., 5’-AUG, NAUG, NNAUG, NNNAUG) (Krishnan
et al., 2010). Of the 3,456 E. coli transcripts in RegulonDB, 115 transcripts have an AUG at the
experimentally demonstrated or predicted 5’-terminus and 287 transcripts have an AUG within
three nucleotides of their 5’-terminus (Table 1-1). In addition to undergoing translation as
canonical SD-led mRNAs from an internal start codon, we predicted that these mRNAs might
also be translated as leaderless mRNAs from AUG triplets located at, or near, their 5’-termini,
thereby categorizing them as bicistronic mRNA. We selected a subset of thirteen genes for
further study, chosen on the basis of characteristics that include the predicted gene’s function,
length of the putative peptide encoded from the 5’-uORF, the distance of the 5’-uAUG triplet
from the 5’-terminus, and the position of the uORF’s stop codon relative to the downstream
ORF’s start codon (Table 1-2). We chose genes in which the 5’-uAUG was out of frame with the
downstream start codon to select bicistronic mRNAs rather than genes in which translation of the
5’-uORF could produce longer isoforms of the canonical gene product.
The translational activity of each of the 5’-uORFs was assessed by an in-frame fusion to
a lacZ reporter gene and transcribed from the lac promoter (Figure 1-1A). β-galactosidase assays
(Miller, 1992) were performed to measure translation from the 5’-uORFs. Activity measured
from the products of lacZ fusions to naturally leaderless E. coli bacteriophage λ cI (Ptashne et
al., 1976) and transposable element Tn1741 tetR (Klock and Hillen, 1986) mRNAs were used as
controls for comparison of expression levels.
Varying levels of expression were seen from the 5’-uORFs, with many falling within the
expression range measured for the tetR and cI leaderless mRNAs (Figure 1-2). cmk, glpF, and
luxS 5’-uORFs were expressed at levels 10-fold lower than tetR, whereas uvrY and pncB 5’uORFs were expressed at levels only 2-fold lower than tetR (Figure 1-2, inset). Conversely, ptrB,
rhaB and xap 5’-uORFs were expressed at levels very similar to tetR (Figure 2, inset). Several
20
Table 1-1. RegulonDB identified genes. Complete list of genes identified by in silico analysis
of the RegulonDB E. coli transcriptome as having a Shine-Dalgarno sequence within their 5’UTR as well as an AUG triplet within three nucleotides of the 5’-terminal (i.e., 5’-AUG, NAUG,
NNAUG, and NNNAUG). The nucleotides defined as the 5’-uAUG for each gene are
underlined. The table includes the gene name, the first 21 nucleotides at the 5’-terminal end of
the mRNA, its start position on the E. coli chromosome, strand of location, accession number
(ECK#), and the sigma factor associated with its transcription.
21
22
23
24
25
26
27
28
Table 1-2. Selected genes examined that contain an AUG triplet at the 5’-terminus of its
transcribed canonical mRNA. An overlap refers to a uORF stop codon that is within the
downstream coding sequence, albeit out of frame.
29
30
Figure 1-1. Schematic of lacZ fusions constructed for each gene tested. lacZ fusions to RNA
encoding the 5’-uORF (A) or to the coding sequence (CDS) of the downstream gene (B-D). The
5’-uAUG is either maintained as AUG (CDS), mutated to AUC (KO) or contains a premature
stop codon (StartStop). The arrow indicates the in-frame AUG start codon defining the ORF
providing expression of the lacZ reporter gene.
31
32
Figure 1-2. 5’uAUGs support varying levels of translation compared to known leaderless
mRNA in E. coli. Translation from select 5’-uORFs fused to lacZ (see Figure 1-1A) using βgalactosidase assays performed in triplicate. Translation is compared to β-galactosidase activity
measured from cI-lacZ (= 100%) fusions. A subset of 5’-uORFs displaying lower levels of
translation are shown as compared to β-galactosidase activity measured from cI-lacZ (=100%)
fusions (inset).
33
34
5’-uORFs were actually expressed at levels 2-fold to 10-fold higher than tetR (e.g., mngR, fucP,
rcnR and pnp) (Figure 1-2), albeit still significantly lower than cI (Figure 1-2). The most highly
expressed 5’-uORF tested was iscR, which was expressed at a level 2-fold greater than cI (Figure
1-2). These results show that the 5’-uAUGs can act as an initiation codon to support
translational expression, thereby identifying previously unexplored ORFs.
lacZ fusions were also constructed to assess the translational activity of the canonical SDled downstream CDS (Figure 1-1B). In each instance, the 5’-uORF and the downstream SD-led
ORF are in different reading frames or the uORF’s stop codon is upstream of the annotated ORF
and could therefore be compared individually via LacZ fusions. This is important to note because
in some instances the two cistrons overlap but will still produce two different gene products
(Table 1-2). Only one gene tested, rcnR, supported equal levels of expression from both ORFs
(Table 1-3). However, three genes, pnp, iscR, and mngR, had 5’-uORFs that were translated at
levels ranging from 10-50% of downstream CDS expression (Table 1-3). The majority of 5’uORFs tested were not expressed at comparable levels (Table 1-3). There is large variation in the
levels of translation from these 5’-uAUGs and the basis for these differences is still unclear.
These results demonstrate that an AUG codon at the 5’-terminus of an ORF in the untranslated
region of a canonical, SD-led mRNA has the potential to function as a start codon and support
significant translational activity. However, the presence of a 5’-uAUG is not always sufficient
for translation. Comparable expression to known leaderless mRNAs, as well as to their
downstream cognate CDS in some cases, suggests the potential importance of 5’-uAUGs.
5’-uAUGs bind 70S ribosomes
Primer extension inhibition (toeprint) assays were performed to assess ribosome binding
patterns to the 5’-uAUG start codons and the internal SD-led start codons. In vitro transcribed
mRNAs, corresponding to 5’-uORFs exhibiting varying expression levels (fucP, iscR, rcnR, and
ptrB) (Figure 1-2), were tested by toeprint assays to analyze the inherent affinity of ribosome
complexes for the 5’-uAUG. Assays were performed in the presence of initiator tRNA and either
30S ribosomal subunits or intact 70S ribosomes. As expected, 70S ribosomes bound the 5’uAUGs, whereas 30S subunits bound the internal CDS start codon for each gene tested (Figure
1-3). This ribosome binding pattern suggests that these mRNAs interact with ribosomes both as
35
Table 1-3. Translation from the 5’-uORF as a percent of translation relative to its
downstream canonical CDS as determined by β-galactosidase assays performed in
triplicate.
36
37
Figure 1-3. 5’-uAUGs bind 70S ribosomes following the proposed leaderless mRNA
initiation mechanism. Primer extension inhibition (toeprint) reactions contain mRNA (A) fucP,
(B) iscR (C) rcnR (D) ptrB, as well as initiator tRNA, and 30S subunits or 70S ribosomes as
indicated by + or - symbols; reaction products are separated on denaturing polyacrylamide gels
and visualized by autoradiography. Predicted position of toeprint signals (+15) to the 5’-uAUG
(open arrow) and downstream AUG (closed arrow) are indicated. Additional positions displaying
ribosome dependent bands are also indicated. Figure is representative of experiments performed
in biological triplicate.
38
39
canonical leadered mRNAs, with 30S subunits binding to the CDS start codon, and as leaderless
mRNAs with 70S ribosomes binding to the 5’-uAUGs. This further supports the notion that the
5’-uAUGs are functioning as initiation codons for uORF translation.
In the case of fucP, a 70S toeprint signal was observed at the 5’-uAUG which displays
weak intensity compared to the downstream 30S subunit signal. The relationship between the
two signals correlates with the fucP expression data (Table 1-3). The presence of a 70S toeprint
signal at the 5’-uAUG confirms the ability of the fucP 5’-uORF to bind ribosomes (Figure 1-3A)
and be expressed as a leaderless mRNA (Figure 1-2). The fucP downstream CDS was shown to
be relatively highly expressed (Table 1-4), which agrees with the strong internal toeprint signals.
Interestingly, 70S ribosomes also appeared to bind fucP’s internal AUG in vitro. The internal
70S ribosome binding phenomenon was also seen in the case of iscR mRNA.
30S subunits bound to the SD-led CDS start codon of iscR mRNA and 70S ribosomes
bound to the 5’-uAUG (Figure 1-3B), as expected. 70S ribosomes binding to iscR mRNA’s 5’uAUG (Figure 1-3B) corresponds with a high level of iscR 5’-uORF expression compared to
leaderless cI mRNA (Figure 1-2). In addition, 70S ribosome binding to the internal SD-led start
codon of iscR was observed, as with fucP above. The toeprint signal strength of 70S ribosome
binding of the SD-led start codon was nearly as strong as that of the 30S subunit binding to the
internal start codon. There have been other reports of internal 70S ribosome binding in toeprint
assays, but the 70S ribosome binding is typically weaker than 30S subunit binding at the internal
position (Hartz et al., 1989; Udagawa et al., 2004). It is interesting that internal 70S ribosomal
binding was seen for both fucP and iscR, although not for rcnR or ptrB (Figure 1-3), suggesting
that, for reasons that are still unclear, certain mRNAs have features that allow 70S internal
binding in vitro. It is possible that 70S internal binding is always present, but less stable for
certain mRNAs although further investigation must be completed to examine this line of inquiry.
The observation that 70S ribosomes did not bind the CDS of all tested transcripts (Figure 1-3)
demonstrates the absence of contamination of 30S subunits and absence of ribosomal splitting
occurring within the 70S ribosomal preparations. Assays reproduced with different 70S batch
preparations and different mRNA preparations consistently gave the same results (data not
shown).
40
Table 1-4. Expression levels from downstream CDS fusions to lacZ that contain an intact
5’-uAUG (CDS), a mutated 5’-uAUC (KO), or an intact 5’-AUG with a premature stop
codon (StartStop) (see Figure 1-1) including the standard deviation from triplicate cultures.
NA=not available
41
42
As expected, 30S subunits bound to the SD-led CDS start codon of rcnR mRNA and 70S
ribosomes bound to the 5’-uAUG (Figure 1-3C). The toeprint signal strength for 70S ribosomes
bound to the 5’-uAUG of rcnR was equal to or greater than that of 30S subunits binding to the
internal SD-led AUG. This indicates similar toeprint signal strength, and may reflect the
observation that translation from lacZ fusions to the rcnR 5’-uORF and the downstream SD-led
CDS were equivalent (Table 1-3). The 70S ribosomes binding at the 5’-uAUG, with or without
initiator tRNA, displayed multiple toeprint signals at the expected +16 position, but also at the
+5 position and the +25 position (Figure 1-3C lanes 4 and 6). These signals are ribosomedependent, indicating that they are not the product of secondary structure. Ribosomes appear to
stably bind the 5’-terminus without the AUG start codon positioned in the P-site, demonstrating
that tRNA codon-anticodon pairing is unnecessary, supporting tRNA-independent binding.
Although this signal pattern has been reproduced in subsequent toeprint assays (data not shown),
further experimentation is necessary to explore if this ribosome binding pattern reflects the
process of ribosomal loading on 5’-terminal AUGs.
Aside from the predicted 30S subunit toeprint binding signal binding to the downstream
rcnR CDS start codon, there was another tRNA-dependent signal at position +102 corresponding
to 30S subunit binding to an AUG at position +88 within the 5’-UTR (Figure 1-3C lane 3, Table
1-5). This potential start codon is out of frame with the rcnR CDS start codon and specifies an
uORF that would produce a 22-amino acid long putative peptide before encountering a stop
codon. These results indicate that there are additional ORFs within the rcnR mRNA that may
represent additional peptide production.
Similarly to the other mRNAs tested, 30S subunits bound to the internal SD-led start
codon of ptrB mRNA, and 70S ribosomes bound to the 5’-uAUG. In the case of ptrB, however,
the 70S toeprint signal intensity at the 5’-uAUG was not a reliable predictor of expression level.
ptrB’s 5’-uAUG showed very strong 70S binding as well as tRNA-independent 70S binding
(Figure 1-3D), but expression from the 5’-uORF was low compared to either its downstream
CDS (Table 1-3) or known leaderless mRNA (Figure 1-2). In this case, the ribosome binding
data does not correlate to translational activity seen in β-galactosidase assays which may indicate
that select 5’-uAUGs are bound by ribosomes for purposes other than translation initiation.
43
Table 1-5. Sequences of mRNAs tested. List of mRNAs tested in this study and their
sequences including their 5’UTRs (lower case) and the first 15 codons of the coding sequences
(upper case). The 5’-uAUGs and their in-frame stop codons are upper case and bold. The
underlined sequences correspond to the additional rcnR putative uORF identified using toeprint
assays.
44
45
5’-uAUGs can influence downstream gene expression
The 70S toeprint signal intensity of ptrB led us to suspect that 5’-uAUGs function not
only in peptide production but also as regulatory features. Genes in the same transcriptional unit,
such as in an operon, typically have related functions, and in some cases the uORFs can affect
expression of downstream cistrons. One example of this type of regulation is referred to as
translational coupling (Sarabhai and Brenner, 1967) and is typically necessary due to
sequestration of the downstream ribosome binding regions by secondary structure (Rex et al.,
1994) and is most efficient when the upstream stop codon is in close proximity to the
downstream start codon (Cone and Steege, 1985). To assess effects on downstream CDS
expression that might result from disruption of 5’-uORF translation, lacZ was fused to the SDled CDS in the presence of a 5’-uAUG knockout mutation (i.e., uAUGAUC) (Figure 1-1C).
Mutation of the 5’-terminal start codon would be expected to prevent ribosomal binding as well
as prevent translation of the 5’-uORF. Mutation of a subset of genes’ 5’-uAUGs to AUCs
affected expression of the SD-led downstream CDS to varying degrees (Table 1-4). Many
remained minimally affected, whereas some increased in expression (e.g., cmk, iscR, luxS and
mngR) and others decreased in expression (e.g., fucP, glpF, ptrB, and rcnR) (Table 1-4).
To determine whether the change in downstream CDS expression was due to loss of
translation of the 5’-uORF or loss of ribosome binding to the mRNAs’ 5’-terminus, we made a
separate start-stop mutation to introduce a premature stop codon two codons after the 5’-uAUG
start codon (i.e., AUGxxxUGA) in a subset of genes (Figure 1-1D). The 5’-uAUGs may
stabilize the mRNA due to their inherent ability to bind ribosomes thereby protecting the mRNA
from degradation by RNases (Komarova et al., 2005). This protection may contribute to the
positive effects the 5’-uAUG has on downstream expression, independent of 5’-uORF
translation. In some cases, the start-stop mutation produced results similar to those of the
corresponding 5’-uAUG knockout mutations, emphasizing the importance of translation of the
5’-uORF (Table 1-4). The luxS and mngR mRNAs both had increased expression levels in the
knockout mutant and decreased expression levels in the start-stop mutant (Table 1-4),
highlighting the negative effect of the 5’-uAUG codon on CDS expression.
Of the thirteen different genes assayed containing 5’-uAUG5’-AUC mutations, the
ptrB gene was particularly affected by the 5’-uAUG mutations. As a result of the 5’-uAUG 5’46
AUC mutation, expression from the downstream ptrB CDS was drastically reduced by 93%
(Table 1-4). Toeprint assays revealed that 30S subunit binding to the downstream start codon in
the presence of the 5’-AUC mutation was nearly eliminated (Figure 1-4), correlating with the
expression data (Table 1-4). However, in the presence of the start-stop mutation, expression was
only reduced by 59% compared to the wild-type level (Table 1-4) and reappearance of a 30S
subunit toeprint was observed (Figure 1-4). These data show that ptrB’s 5’-uAUG influences
30S subunit binding to the downstream start codon and concomitantly influences expression.
This suggests there is a potential regulatory function of ptrB’s 5’-uAUG on downstream
expression that is not dependent on uORF translation.
47
Figure 1-4. Loss of 30S subunit binding to CDS start codon as a result of the 5’-uAUG
mutation. Primer extension inhibition (toeprint) reactions contain mRNA, initiator tRNA, and
30S subunits or 70S ribosomes as indicated by + or - symbols; reaction products are separated on
denaturing polyacrylamide gels and visualized by autoradiography. Predicted position of toeprint
signals (+15) to the 5’-AUG (open arrow) and downstream AUG (closed arrow) are indicated.
ptrB sequencing ladder is included on the left. Figure is representative of experiments performed
in biological triplicate.
48
49
Discussion
The results presented here indicate that AUG triplets at the 5’-termini of SD-led mRNAs
can function as start codons and have the ability to bind ribosomes and be translated as leaderless
mRNAs. Using β-galactosidase and toeprint assays, we have shown that many 5’-uORFs were
translated at biologically relevant levels and bound by 70S ribosomes, with some appearing to
play a role in expression of the downstream CDS.
A number of the previously uncharacterized 5’-uORFs supported levels of expression
comparable to known leaderless mRNAs or their respective downstream CDS. rcnR is one
example of an mRNA this study has identified as an efficiently translated bicistronic mRNA.
Stable ribosome binding at the 5’-terminus supports high expression levels observed from rcnR’s
5’-uORF. Since rcnR’s 5’-uORF and CDS expression were similar (Table 1-3), this suggests
nearly equivalent amounts of polypeptide product are being made for each cistron. This is
surprising because leaderless mRNA is typically thought to be less efficiently translated than
leadered mRNA (Moll et al., 2002). Since the rcnR 5’-uAUGAUC mutation reduced
downstream CDS expression (Table 1-4), this 5’-uAUG may also play a role in downstream
expression, possibly through translational coupling. Therefore, rcnR is an example of a 5’-uAUG
that specifies a highly expressed uORF whose expression is linked to downstream expression.
Additional regulatory features, possibly independent of the putative translated peptide,
may be present within the 5’-uORF that could influence expression efficiency or function of the
downstream CDS. This form of regulation is widespread and has been seen in both prokaryotic
(Dincbas et al., 1999; Delgardo-Olivares et al., 2006; Sharma et al., 2012) and eukaryotic
systems (Morris and Geballe, 2000 and references therein; Meijer and Thomas, 2002;
Medenbach et al., 2011). Mutation of the 5’-uAUG to AUC impacted downstream expression in
iscR, mngR, ptrB, and rcnR mRNA (Table 1-4). In each case, this suggests that disruption of
ribosome binding and/or 5’-uORF translation has an effect on the downstream CDS. Conversely,
in some mRNAs (i.e., rhaB, uvrY, and xap), no change was seen in CDS expression as a result of
the 5’-uAUG mutation (Table 1-4) indicating that the CDS is expressed independently from any
5’-uORF ribosome binding or translation. This reinforces the idea that 5’-uAUGs have diverse
functions, and may function differently in different contexts.
50
The ptrB CDS showed a dramatic decrease in expression, as well as loss of internal 30S
subunit binding, in the presence of the 5’-uAUG knockout (Table 1-4, Figure 1-4). The
dependency on the 5’-uAUG, but not translation of the putative peptide, for ribosome binding
and expression of the ptrB CDS, suggests regulatory roles. Since the ptrB 5’-UTR is only 26
nucleotides long (Table 1-5), the regulatory effects may be related to an overlap in ribosomal
occupancy causing a steric hindrance and prohibiting both 70S ribosomes and 30S subunits from
being stably bound at the same time. We propose that the 5’-uAUG is the major signal for
ribosome recruitment and binding to the mRNA for CDS expression, although the mechanism
remains unclear. In one possible scenario, rather than translating the 5’-uORF, this region may
act as a standby site (de Smit and van Duin, 2003) for ribosome loading onto the mRNA, taking
advantage of inherent binding strength to 5’-AUGs. Once the ribosome is bound to the 5’terminus, it could then more efficiently access and bind the downstream CDS translation
initiation region. This pre-loading may result in more stable ribosome binding and more efficient
internal CDS translation. Further investigation into this potential model may elucidate a novel
mechanism of initiation or regulation of translation via a 5’-uAUG.
While the majority of the 5’-uAUGs we tested function in either 5’-uORF translational
expression or regulation, some 5’-uAUGs showed no obvious role in these activities. The 5’uORFs of cmk, glpF, and luxS mRNA were translated at levels much lower than tetR (Figure 12), and do not appear to have a substantial effect on downstream CDS expression (Table 1-4). It
is possible that these 5’-uAUGs may have formed by chance and have not been evolutionarily
selected for a regulatory role. Alternatively, it is possible that there is an unknown function for
these 5’-uAUGs that we have not considered or tested.
Overall, the 5’-uAUGs we tested may possess a variety of potential functions, such as
providing protection to stabilize the mRNA transcript via ribosome binding, producing peptides,
and contributing to regulation of the downstream CDS. The ribosome binding studies revealed
the translation potential for unannotated uORFs, which may be more widespread than previously
thought. Ribosome binding studies similar to the work we have performed could also reveal
ORFs that are not detected by visual inspection or bioinformatic analysis. This study provides
insight into pitfalls of our current methods of identifying translation initiation sites, and imply
that in silico analyses may be biased by imposing size limitations and overlooking uORFs when
51
analyzing genomes. The 5’-uORFs may represent an additional subtype of cistron that should
continue to be considered when annotating genomes because this study shows their potential to
be functional at biologically relevant levels.
52
Chapter 2
Novel translation initiation regulation mechanism in Escherichia coli ptrB directed by 5’terminal AUG and downstream coding sequence elements
Heather Beck, Gary Janssen
53
Abstract
Translation of Escherichia coli ptrB requires an AUG triplet at the 5’-terminus of its messenger
RNA (mRNA). The 5’-terminal AUG (5’-uAUG) acts as a ribosome recognition signal to attract
ribosomes to the ptrB mRNA. A mutation in the 5’-uAUG of ptrB reduced downstream ptrB
expression by more than 90%, even in the presence of the predicted ptrB Shine-Dalgarno (SD)
sequence. Strengthening of the ptrB SD sequence relieved the necessity for the 5'-uAUG,
suggesting that the regulation imposed by the 5'-uAUG is distinct from canonical SD sequencemediated regulation. Additional sequences within the ptrB 5’-untranslated region (5’-UTR) were
identified that contribute to ribosome binding and expression of the downstream ptrB coding
sequence (CDS). Replacement of 5’-UTRs from other mRNAs with the ptrB 5’-UTR sequence
showed a similar dependence on the 5’-uAUG for downstream expression, suggesting that the
regulatory features contained within the ptrB 5’-UTR are sufficient to control the expression of
other CDS. Mutational analysis also identified two specific adenine nucleotides within the ptrB
CDS at position +6 and +9 that contribute to ptrB expression. The primary sequence features
identified in ptrB mRNA reveal a novel form of translation regulation driven by the 5’-uAUG.
Due to the abundance of AUG triplets at the 5’-termini of E. coli mRNAs and the ability of the
ptrB 5’-UTR regulation to function independent of gene context, the regulatory effects of 5’uAUGs on downstream CDS may be widespread throughout the E. coli genome.
54
Introduction
Protein synthesis, in which the ribosome translates a messenger RNA (mRNA) sequence
to form polypeptides, is a four-step process, including initiation, elongation, termination, and
ribosome recycling. Translation initiation is energy-dependent and the most highly regulated
phase of translation (Laursen et al., 2005). Both the mRNA’s sequence and secondary structure
can affect how it interacts with the translational machinery and therefore influence its translation
efficiency. The translational machinery interacts with the mRNA at its ribosome binding site
(RBS), along with initiator tRNA, to form a ternary complex that is equipped for polypeptide
production. The RBS of a canonical mRNA contains an initiation codon, a purine-rich ShineDalgarno (SD) sequence complementary to the anti-Shine-Dalgarno (aSD) sequence of the 16S
rRNA (Shine and Dalgarno, 1974), and an appropriately sized spacer region between the two
aforementioned elements (Ringquist et al., 1992; McCarthy and Brimacombe, 1994). The SD
sequence and its spacing upstream of the initiation codon are important for ribosome binding and
proper placement of the initiation codon in the ribosomal P-site to initiate translation (Laursen et
al., 2005). These primary structural elements, as well as potential upstream and downstream
enhancer sequences can contribute to translation efficiency. The cooperative effect of multiple
nucleotides at specified regions is referred to as the “cumulative specificity initiation
mechanism,” and the multiple elements can collectively contribute to ribosome recognition,
strength of ribosome binding, and reading frame selection (Nakamoto, 2011).
The mRNA’s secondary structure can also impact the rate of translation. The ribosome
requires a local single stranded region for efficient binding (Draper, 1987; Draper et al., 1998;
Gualerzi and Pon, 1990). Secondary structure that occludes the RBS can cause the region to be
unrecognizable to the ribosome and negatively impact translation (de Smit and van Duin, 1990;
de Smit and van Duin, 1994; Gu et al., 2010). Nevertheless, secondary structure can act to
regulate translation and allow for certain mRNAs to be translated only under specific conditions.
For example, cold or heat stress can cause thermodynamic changes in mRNA structure that could
either open or occlude the RBS and therefore result in a change in initiation efficiency (Storz,
1999; Narberhaus et al., 2006; Klinkert and Narberhaus, 2009).
There are also examples of cis-acting elements, such as upstream open reading frames
(uORFs), which can influence the expression of downstream open reading frames (ORFs) on the
55
same transcript. In prokaryotes, uORFs exert their effects primarily through translational
coupling (Sarabhai and Brenner, 1967; Min Jou et al., 1972; Andre, 2000). These effects occur
after the ribosome completes translation of the uORF and, rather than dissociating, remains
bound to the mRNA. The ribosome then repositions to the downstream initiation region using a
scanning-like movement (Adhin and van Duin, 1990). This process is beneficial because it
allows polycistronic mRNAs to be controlled from a single point (Spanjaard and van Duin,
1989) and may eliminate inhibitory RNA structures that would otherwise disrupt initiation of the
downstream ORF (Rex et al., 1994; Takyar et al., 2005). uORFs can also encode nascent
peptides that cause ribosomal stalling by binding to regions of the peptidyltransferase center,
blocking the ribosome exit tunnel or through attenuation (Lovett and Rogers, 1996; Tenson and
Ehrenberg, 2002). This stalling can then impact downstream translation by blocking the RBS,
causing conformational changes in secondary structure, or increasing degradation of the mRNA
(Nakatogawa and Ito, 2002; Cruz-Vera et al., 2011).
We recently reported another example of the influences of uORFs on downstream
expression (Beck et al., 2016). In that study, we identified novel examples of uORFs contained
within the 5’-untranslated region (5’-UTR) of SD sequence-led mRNAs in Escherichia coli. The
newly discovered uORFs were all positioned at the 5’-terminus, thereby classifying them as
leaderless mRNAs. Leaderless mRNAs lack a 5’-UTR and are thought to initiate translation via a
mechanism distinct from leadered mRNAs, in which an intact 70S ribosome binds to the 5’terminus of the mRNA (Balakin et al., 1992; Udagawa et al., 2004; Moll et al., 2004). The
leaderless uORFs tested bound 70S ribosomes and were translated to varying degrees (Beck et
al., 2016). A subset of the uORFs also contributed to downstream ORF expression levels, either
through uORF translation or through their respective 5’-terminal AUG’s ability to act as a
ribosome binding signal. Downstream expression of ptrB, an excellent example, was dependent
upon the 5’-terminal AUG (5’-uAUG) even though the 5’-terminal uORF was not efficiently
translated (Beck et al., 2016).
The necessity of the ptrB 5’-uAUG and its role in ptrB coding sequence (CDS) regulation
is the focus of this study. The annotated ptrB CDS produces a 686-amino acid protease II
predicted to be a member of the prolyl endopeptidase family (Kanatani et al., 1991), but little is
currently known about its regulation. The ptrB mRNA has a 26-nucleotide untranslated leader
56
region, containing a predicted SD sequence located nine nucleotides upstream of the ptrB
initiation codon with imperfect complementarity to the aSD sequence (Figure 2-1). The 5’uAUG defined uORF overlaps with the downstream CDS reading frame, such that the uORF
stop codon is at the +35 position (with the ‘A’ of the CDS start codon at +1). Therefore, if
translated, the uORF would produce a 21-amino acid peptide.
This study reinforces the dependency of ptrB translation on the 5’-uAUG based on its
ability to act as a ribosome-binding signal rather than through translation of the uORF. Ribosome
binding and expression assays revealed a regulatory process distinct from SD sequence-mediated
regulation, suggesting a novel mechanism for translation regulation driven by a 5’-uAUG. We
propose two potential regulatory models that could play a role in the translation initiation of
ptrB. These findings are significant because the ptrB 5’-UTR regulatory regions can act
autonomously regardless of gene context, suggesting that this form of regulation may be
widespread in E. coli.
57
Figure 2-1. The ptrB gene sequence. The 5’-ATG is bolded at the 5’-terminus. The 5’-UTR is
indicated by the lowercase letters. The predicted SD sequence is underlined. The CDS is
indicated by the uppercase letters. The 5’-uAUG defined ORF’s stop codon is within the coding
sequence and also bolded (TGA). The SalI site used to construct the in-frame lacZ fusion is
indicated by the dashed underline.
58
59
Materials and Methods
Bacterial strains and reagents. E. coli DH5α (New England Biolabs [NEB]) was used as the
host for all plasmid DNA manipulations. E. coli RFS859 (F-, thr-1, araC859, leuB6, Dlac74,
tsx-274, l-, gyrA111, recA11, relA1, thi-1) (Schleif, 1972) was used as the host for the expression
and assay of lacZ fusion mRNA constructs. Chromosomal DNA from E. coli K12 was used as a
template for initial PCR amplifications of the ptrB gene. Two plasmids were used in the study,
pM1108 (for the incorporation of a mutated lac promoter) and pA906 (to allow for the lacZ gene
fusion) (Janssen and Bibb, 1993).
Radiolabeled [γ-32P] ATP (6000 Ci/mmol, 150 mCi/mL) was purchased from Perkin Elmer.
Restriction endonucleases, T4 DNA ligase, T4 polynucleotide kinase (PNK), and T7 RNA
polymerase were purchased from NEB and used according to the manufacturer’s instructions.
RNase-free DNase I (Roche), AMV reverse transcriptase (Life Sciences), and Pfu DNA
polymerase (Stratagene) were used according to the manufacturers’ specifications. DNA
oligonucleotides were purchased (IDT).
Recombinant DNA procedures. PCR amplifications were performed using E. coli K12
chromosomal DNA or purified plasmids containing the gene of interest as templates. Specific
oligonucleotides (IDT) were added to incorporate specific mutations (Table 2-1), in addition to
the template, 5X Pfu buffer (commercially bought with subsequent addition of 1 mM dNTPs),
and Pfu DNA polymerase, to complete PCR amplifications. After amplification, PCR products
were electrophoresed on 1.3% low-melting point agarose gels for DNA purification. DNA bands
were then excised from the gels and purified using a phenol: chloroform: isoamyl alcohol
(25:24:1) mixture and isopropanol precipitation. Purified DNA products were then digested with
restriction endonucleases and ligated into plasmid pA906, containing an ampicillin resistance
marker and the mutated lac promoter (AATAAT), to fuse the regions of interest in-frame to the
fifth codon of the lacZ reporter gene.
i. Construction of DNA molecules with randomized sequence regions
Oligonucleotides with regions of randomized nucleotides were purchased (IDT). The
randomized regions in each primer were three nucleotides long and designed to allow for the
60
possibility of each nucleotide (A, T, G, or C) at each position within a single primer mixture,
except in an instance that would result in the introduction of a termination codon. The
randomized region extended from the +4 to +16 position (with the ‘A’ of the 5’-uAUG at
position +1). Mutants were cloned as previously described and screened via lacZ blue-white
screening on Luria-Bertani plates with the addition of 40 µg/mL 5-bromo-4-chloro-3-indolylbeta-D-galactopyranoside (X-gal). Colonies with detectable color differences compared to the
wild-type ptrB were selected and sequenced on an Applied Biosystems 3730 DNA analyzer to
identify the sequence of the randomized region.
ii. Gene fusions
Fusions of regions from two separate genes (Table 2-1) were constructed using trimolecular ligations. The primers at the amplicons’ junction were phosphorylated with T4 PNK,
T4 PNK buffer (2 M Tris-HCl, pH 7.6, 1 M MgCl2, 1 M DTT) and ATP (25 mM) for 30 minutes
at 37oC. The reactions were then heat-inactivated for 20 minutes at 65oC. The regions within the
genes used for the fusions were amplified using PCR with the phosphorylated primers, purified
as previously described, and digested to allow for incorporation into pM1108. The two
amplicons and the plasmid were then ligated using T7 ligase and 10X ligation buffer (NEB) and
incubated at 16oC overnight. The ligation was then used as a template for a subsequent PCR
amplification using a 5’ primer specific for the plasmid and a 3’ primer specific to the
downstream amplicon to ensure proper ligation. The resulting amplicon was then purified and
subcloned in pA906 as previously described for constructing lacZ fusions.
Preparation of competent cells. This is a modified version of the previously described Hanahan
procedure (Hanahan, 1983). An overnight stationary phase culture of E. coli DH5α or RFS859
cells was diluted 1:50 with L broth (10 g/L Difco Bacto tryptone, 5 g/L Difco Bacto yeast
extract, 10 g/L NaCl, 1 g/L glucose) supplemented with 20 mM MgCl2 and incubated at 37oC,
with shaking at 200 rpm, for two hours. Cultures were then incubated on ice for 10 minutes and
centrifuged at 2795 × g for 5 minutes. The supernatant was discarded and the cell pellet was
resuspended in 20 mL of ice-cold 10 mM NaCl. After resuspension, the culture was again
centrifuged at 2795 × g for 5 minutes. The supernatant was discarded and the cell pellet was
resuspended with 20 mL of 30 mM CaCl2 and incubated on ice for 1 hour. The resuspension was
61
again centrifuged at 2795 x g for 5 minutes and the pellet was resuspended in 3 mL of 30 mM
CaCl2 and stored at 4oC for up to two weeks.
Transformations. Transformations were performed by adding ligation mix to chemically
competent E. coli cells and incubated on ice for 30 minutes. The mixture was then heat shocked
at 37oC for 2 minutes then incubated on ice for 2 minutes. The transformation mixture was then
plated onto L agar (L broth with agar, 15 g/L) containing 200 µg/mL ampicillin for selection and
40 µg/mL X-gal for screening and incubated at 37oC for approximately 24 hours.
β-galactosidase assay. β -galactosidase assays were performed as previously described (Miller,
1992).
In vitro synthesis of RNA. The cloned plasmids were used as templates in PCR amplifications
utilizing a primer to incorporate the T7 RNA polymerase promoter sequence (5’TAATACGACTCACTATAG-3’). This produced DNA fragments containing a T7 promoter
sequence, allowing for in vitro transcription with T7 RNA polymerase and production of RNA
used for toeprint reactions. RNAs were synthesized and purified as described (Fredrick and
Noller, 2002). RNAs used in toeprint assays were synthesized by combining purified PCR
amplicons (constructs with lacZ fusions containing a T7 promoter) and T7 RNAP in 1X buffer
(40 mM Tris, pH 7.8, 25 mM MgCl2, 1 mM spermidine, 0.01% Triton X-100, 5 mM each NTP,
and 30 mM DTT). Transcription reactions were incubated for approximately 4 h at 37°C, and 40
mM EDTA was added. Samples were treated with DNase (Roche) for 15 min at 37°C. RNA was
ethanol-precipitated and suspended in RNA loading dye (50% formamide, 0.05% bromophenol
blue, 0.05% xylene cyanol). Samples were subjected to PAGE (6% acrylamide, 7 M urea) and
full-length products were excised using UV shadowing. Gel slices were incubated overnight at
room temperature in elution buffer (300 mM NaOAc at pH 5.2, 0.1% SDS, 1 mM EDTA) with
gentle rocking. The supernatant was phenol-extracted and ethanol-precipitated.
Primer extension inhibition (toeprint) assay. DNA oligonucleotides were phosphorylated at
the 5’-terminus using [γ-32P] ATP (6000 Ci/mmol, 150 mCi/mL; Perkin Elmer) and T4 PNK in
1X kinase buffer for 30 min at 37°C and annealed to 3’-termini of RNA as previously described
(Martin-Farmer and Janssen, 1999). Annealed RNA was incubated with 30S subunits or 70S
ribosomes with or without tRNAfMet for 15 min at 37°C. Reactions were transferred to ice, and
62
reverse transcriptase was added. The reactions were incubated for 15 min at 37°C and stopped by
the addition of 0.3M NaOAc and 100% ethanol and precipitated overnight at -80oC. Precipitated
complexes were collected by centrifugation and dissolved in loading dye (80% deionized
formamide, 10 mM NaOH, 1 mM EDTA, and 0.5% bromophenol blue and xylene cyanol),
followed by heat treatment (95°C, 5 min) and PAGE (6% acrylamide, 7 M urea) in 1X TBE.
Gels were visualized via autoradiography.
63
Results
ptrB 5’-uAUG as a ribosome binding signal
Our recent study showed that the ptrB 5’-uAUG is necessary for downstream CDS
expression, although the 5’-uORF is expressed at a low level (Beck et al., 2016). lacZ fusions
after five, ten, and fifteen codons of the uORF also resulted in low levels of expression (data not
shown) suggesting that no region of the uORF is efficiently translated. Because the 5’-uORF is
not efficiently translated, we hypothesize that the 5’-uAUG’s effect on CDS expression is due to
the use of the 5’-uAUG as a signal for ribosome binding, rather than through translation of the
uORF. However, it is possible that the two ORFs may be translationally coupled (Andre et al.,
2000) because there are examples of coupling that do not exhibit a linear relationship of a gene
pair with fixed stoichiometry (Rex et al., 1994; Yu et al., 2001), so we chose to first investigate
this possibility.
To rule out the possibility of translational coupling in ptrB, the stop codon of the uORF
was mutated, allowing for ribosomal read-through and therefore disrupting potential coupling.
Mutation of the uORF stop codon resulted in no discernible change in CDS expression (data not
shown), suggesting that this regulation is not a result of translational coupling. Additional stop
codons in-frame with the 5’-uAUG were also introduced upstream of the natural uORF stop
codon (Table 2-1, pAptrB.prestop1, pAptrB.prestop2, pAptrB.prestop3) to examine the
influences of ribosomal progression on the transcript. The additional stop codons also had no
effect on CDS expression (data not shown) supporting the conclusion that translational coupling
does not contribute to ptrB regulation. This also reinforces the notion that it is the 5’-uAUG itself
that is necessary for downstream CDS expression, rather than uORF translation.
Predicted SD sequence has unexpected effects
Expression of the downstream CDS was drastically reduced in the presence of the 5’uAUG mutation despite the fact that the predicted SD sequence was still intact (Beck et al.,
2016). These data suggest that the SD sequence is not the sole ribosome binding signal and its
presence cannot overcome the negative impact of the 5’-uAUG mutation. To address the SD
sequence’s role in ptrB CDS expression, the SD sequence was mutated to its complement
64
Table 2-1. DNA sequences of gene fragments, with and without mutations, from ptrB. The
predicted SD sequence is underlined, the mutated nucleotides are bolded and the stop codons are
italicized. The pA series of plasmid constructs contained the ptrB coding sequence, and the pAB
series contained the CDSs of different genes.
65
Plasmid
Sequence
Variable 5’UTR
Variable coding sequence
pAptrB.WT
ATGTTTCAACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.5’KO
ATCTTTCAACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.prestop1
ATGTTTTAACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.prestop2
ATGTTTCAACCAGAATGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.prestop3
ATGTTTCAACCAGAAAGAACAATAAC
ATG CTA CTA AAA GCC GCC
pAptrB.SDmut
ATGTTTCAACCACTTTCAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.5’KO.SDmut
ATCTTTCAACCACTTTCAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.SDstr
ATGTTTCAACCAGGAGGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.5’KO.SDstr
ATCTTTCAACCAGGAGGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.structmut
ATGAAACAACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.structmutcomp
ATGAAACAACCAGAAACTTCAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.scan1
ATGAAAGTACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.scan2
ATGTTAGTTGCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.scan3
ATGTTTCATGGTCAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.scan4
ATGTTTCAACGTCCCAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.scan5
ATGTTTCAACCACCCCCAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.scan6
ATGTTTCAACCAGAACCTTGTATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.scan7
ATGTTTCAACCAGAAAGAACTTATTC
ATG CTA CCA AAA GCC GCC
pAptrB.5ntdel
ATGTTCAACCAGAAAGAACAA
ATG CTA CCA AAA GCC GCC
pAptrB.9ntdel
ATGCAACGAAAGAACAA
ATG CTA CCA AAA GCC GCC
pAptrB.SDdel
ATGTTTCAACCAAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.9ntadd
ATGTTTCAACCATTTCAACCAGAAAGAAC
AATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.21ntadd
ATGTTTCAACCACTTTCAACAATAACATT
TCAACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
66
pAptrB.33ntadd
ATGTTTCAACCACTTTTTCAACCACTTTC
AACAATAACATTTCAACCAGAAAGAACA
ATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.6GCA
ATGTTGCAACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.6TCG
ATGTTTCGACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.6GCG
ATGTTGCGACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.7CGC
ATGTTTCGCCCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pAptrB.7AGA
ATGTTTAGACCAGAAAGAACAATAAC
ATG CTA CCA AAA GCC GCC
pABptrB-pnp
ATGTTTCAACCAGAAAGAACAATAAC
TTG CTT AAT CCG ATC GTT
pABptrB-pncB
ATGTTTCAACCAGAAAGAACAATAAC
ATG ACA CAA TTC GCT TCT
pABptrB5’KO-pnp
ATCTTTCAACCAGAAAGAACAATAAC
TTG CTT AAT CCG ATC GTT
pABptrB5’KO-pncB
ATCTTTCAACCAGAAAGAACAATAAC
ATG ACA CAA TTC GCT TCT
pABptrB-pncBdblmut
ATGTTTCAACCAGAAAGAACAATAAC
ATG ACG CAG TTC GCT TCT
pABptrB5’KO-pncBdblmut
ATCTTTCAACCAGAAAGAACAATAAC
ATG ACG CAG TTC GCT TCT
pAptrBCDSdblmut
ATGTTTCAACCAGAAAGAACAATAAC
ATG CTG CCG AAA GCC GCC
pAptrB5’KO.CDSdblmut
ATCTTTCAACCAGAAAGAACAATAAC
ATG CTG CCG AAA GCC GCC
pABptrB-tna.IN
ATGTTTCAACCAGAAAGAACAATAAC
ATG GTG GCG TTC TCT AAC
pABptrB-aroL.IN
ATGTTTCAACCAGAAAGAACAATAAC
ATG ACG GTC GCG GAG ATC
pABptrB5’KO-tna.IN
ATCTTTCAACCAGAAAGAACAATAAC
ATG GTG GCG TTC TCT AAC
pABptrB5’KO-aroL.IN
ATCTTTCAACCAGAAAGAACAATAAC
ATG ACG GTC GCG GAG ATC
67
(5’-CTTTC-3’) to disrupt its complementarity with the 16S rRNA aSD sequence (Table 2-1,
pAptrB.SDmut). This change significantly reduced CDS expression by approximately 60%
(Figure 2-2A). However, an SD sequence mutation in E. coli would typically result in complete
disruption of expression because it is the sole ribosome binding signal (de Boer et al., 1983; Hui
and de Boer, 1987; Van Etten and Janssen, 1998), which does not appear to be the case for ptrB.
Also, in the ptrB SD sequence mutant (pAptrB.SDmut, Table 2-1), a 30S subunit binding signal
at the internal start codon is present in the primer extension inhibition assay (toeprint assay)
(Figure 2-2B, lane 15), supporting the levels of expression observed (Figure 2-2A). These data
suggest that ribosomes can still initiate translation, albeit at reduced efficiency, in the absence of
the predicted ptrB SD sequence.
Next, the SD sequence was strengthened such that its sequence had absolute
complementarity to the 16S rRNA aSD sequence (i.e. 5’-GAAAG-3’ 5’-GGAGG-3’,
pAptrB.SDstr, Table 2-1). The mutation increased the downstream expression levels by 2.5-fold
(Figure 2-2A). To understand the relationship between the 5’-uAUG and the SD sequence, the
5’-uAUG was mutated to 5’-AUC in concert with the strengthened SD sequence
(pAptrB.5’KO.SDstr, Table 2-1). An insignificant change in ptrB CDS expression was observed
as a result of the 5’-uAUG mutation in the presence of the strengthened SD sequence (Figure 22A), demonstrating that the necessity for the 5’-uAUG to regulate ptrB CDS expression was
relieved by strengthening the SD sequence. The loss of 5’-uAUG dependence was supported by
the toeprint assay, which displayed a loss of 70S binding to the 5’-terminus due to the 5’-uAUG
mutation, whereas the strong internal 30S binding signal was maintained (Figure 2-2C).
Therefore, the SD sequence appears to be acting as a part of a mechanism completely separate
from 5’-uAUG regulation.
Because the 5’-uAUG and SD sequence appear to be acting through distinct regulatory
mechanisms, we tested whether ptrB expression would be completely abolished by disrupting
both. In the presence of both the 5’-uAUGAUC mutation and the SD sequence mutated to its
complement (pAptrB.5’KOSDmut, Table 2-1), ptrB CDS expression was reduced by 45%
(Figure 2-2A). Taken together, ptrB CDS expression is higher as a result of the mutations in
68
Figure 2-2. Role of SD sequence in ptrB regulation. (A) Expression from the ptrB CDS fused
to lacZ in the presence of the various SD sequence mutations (see Table 2-1), shown as a
percentage of ptrB wild-type (WT) CDS (pAptrB.WT) fused to lacZ (100%). * refers to a
statistical significance compared to pAptrB.WT where p<0.001. (B) Primer extension inhibition
(toeprint) reactions contain mRNA, 30S subunits or 70S ribosomes, and initiator tRNA as
indicated by + or - symbols; reaction products are separated on denaturing polyacrylamide gels
and visualized by autoradiography. Predicted positions of toeprint signals (+15) to the 5’ AUG
(open arrow) and downstream AUG (closed arrow) are indicated. (C) Toeprint assay as
previously described containing strengthened SD sequence mutations (see Table 2-1). Reaction
components are indicated.
69
70
71
tandem than it is in the presence of the mutations separately (Figure 2-2A). This finding was
supported by the ribosome binding data as well, exhibiting a stronger band corresponding to 30S
subunit binding to the CDS start codon in the presence of the tandem mutations, compared to
each single mutation (Figure 2-2B, lane 21 compared to lanes 9 and 15).
Using secondary structure prediction software (Zuker, 2003), both singular mutations
(pAptrB.5’KO and pAptrB.SDmut, Table 2-1) were predicted maintain the wild-type structure,
including a hairpin loop within the 5’-UTR (Figure 2-3A). However, the double mutation
(pAptrB.5’KO.SDmut, Table 2-1) caused a loss of the secondary structure, resulting in an open
conformation with no steric hindrance to prevent ribosome binding (Figure 2-3B). The complete
loss of secondary structure in the double mutant could allow for SD sequence-independent
initiation, because the absence of secondary structure is necessary and sufficient to initiate SD
sequence-independent translation (Scharff et al., 2013).
To examine the importance of secondary structure in ptrB regulation, a construct was
created to mutate one side of the aforementioned hairpin loop in the 5’-UTR to disrupt the
predicted secondary structure (pAptrB.structmut, Table 2-1). Secondly, a compensatory mutant
was constructed to reconstitute the secondary structure by mutating the second side of the hairpin
loop (pAptrB.structmutcomp, Table 2-1). If structure were the primary means of ptrB regulation,
the compensatory mutant would be expected to restore wild-type expression levels. The mutant
that disrupted the secondary structure, resulting in an open conformation, caused a 1.5-fold
increase in ptrB CDS expression (Figure 2-3C), following the trend suggested by the previous
expression data (Figure 2-2A). In contrast, the compensatory mutant caused a significant
reduction in CDS expression of over 80% (Figure 2-3C). The same hairpin structure is present in
both the wild-type mRNA and the pAptrB.structmutcomp mRNA suggesting that the reduction
in expression seem in the pAptrB.structmutcomp mRNA is a result of the changes in the primary
sequences rather than secondary structure. Therefore, although secondary structure may play a
role in regulation, the sequence features are more important in influencing CDS expression.
Ribosome binding signals present within the ptrB 5’-UTR
The occurrence of ptrB CDS expression in the absence of both the 5’-uAUG and the SD
sequence suggests there may be other sequence features within the 5’-UTR that act as ribosome
72
Figure 2-3. ptrB mRNA structure may play a secondary role in expression. (A) Predicted
secondary structure of ptrB WT mRNA using computational modeling (Zuker et al., 2003). The
predicted SD sequence is underlined and the CDS initiation codon is boxed. (B) Predicted
secondary structure of ptrB mRNA with 5’-uAUG mutated to AUC and SD sequence mutated to
its complement (pAptrB.5’KO.SDmut, Table 2-1) using computational modeling (Zuker et al.,
2003). The mutated SD sequence is underlined and the CDS initiation codon is boxed (C)
Expression from the ptrB CDS fused to lacZ and in the presence of structural mutations (see
Table 2-1) shown as a percentage of ptrB WT CDS (pAptrB.WT) fused to lacZ (100%). * refers
to a statistical significance compared to pAptrB.WT where p<0.05.
73
74
Figure 2-4. Scanning mutagenesis of ptrB 5’-UTR. (A) Schematic of regions within the 5’UTR mutated to its complementary DNA sequence, corresponding to lines 1-7 beneath the ptrB
sequence. (B) Expression from the ptrB CDS fused to lacZ in the presence of the various 5’-UTR
mutations (1-7), shown as a percentage of ptrB WT CDS (pAptrB.WT) fused to lacZ (100%). *
refers to a statistical significance compared to pAptrB.WT where p<0.05. (C) Toeprint assay
comparing pAptrB.WT mRNA to pAptrB.scan2 and pAptrB.scan6 mRNA performed as
previously described (see Figure 2-2). Reaction components are indicated. Predicted positions of
toeprint signals (+15) to the 5’-uAUG (open arrow) and downstream AUG (closed arrow) are
indicated.
75
76
77
binding signals. To test whether there are other regions of importance within the 5’-UTR,
scanning mutagenesis was conducted, spanning the length of the ptrB 5’-UTR with overlapping
mutations (Figure 2-4A). The mutations changed the wild-type sequence to its complement,
except in the case of the SD region, in which the nucleotides were all mutated to cytosines. The
majority of the mutations significantly disrupted CDS expression, some more drastically than
others (Figure 2-4B). Mutations of the regions closer to the 5’-terminus caused a more severe
decrease in CDS expression, corresponding with the region of ribosomal coverage when bound
to mRNA, overlapping to approximately position +19 (Huttenhofer and Noller, 1994). A
toeprint assay was also conducted to analyze the ribosome binding pattern in two of the scanning
mutant constructs (pAptrB.scan2 and pAptrB.scan6, Table 2-1). In both mutants, the 70S binding
to the 5’-terminus and the internal 30S binding was greatly reduced compared to the ptrB wildtype even in the presence of an intact 5’-uAUG (Figure 2-4C). These data suggest that there are
signals within the 5’-UTR that may help to stabilize the ribosome once bound to the mRNA’s 5’terminus.
Spacing optimized between the 5’-uAUG and downstream RBS
The spacing between primary sequence elements in the mRNA also contributes to
efficient expression. The importance of spacing is seen with regards to the SD sequence, which
must be spaced an appropriate distance from the initiation codon to place the initiation codon in
the P-site when the SD sequence is properly aligned with the 16S rRNA aSD sequence
(Ringquist et al., 1992). To determine if the spacing between the 5’-uAUG and the CDS
initiation codon is similarly important, five and nine nucleotides within the 5’-UTR, shown to be
nonessential to CDS expression, were deleted to shorten the 5’-UTR (pAptrB.5ntdel,
pAptrB.9ntdel, Table 2-1) and therefore reduce the distance between the 5’-uAUG and the
downstream RBS. The predicted SD sequence, also five nucleotides in length, was deleted
(pAptrB.SDdel, Table 2-1) to determine if a deletion of the SD sequence would produce results
similar to the complementary mutation, pAptrB.SDmut (Figure 2-2). Deletion of five nucleotides
or deletion of the SD sequence each significantly reduced CDS expression by 98%, whereas
deletion of nine nucleotides abolished CDS expression completely (Figure 2-5A). This suggests
that the 5’-UTR cannot be shortened without a significant reduction in downstream expression
78
Figure 2-5. 5’-UTR mutations and deletions to affect spacing between 5’-uAUG and
downstream RBS. (A). Expression from the ptrB CDS fusion to lacZ in the presence of the
deletion or insertion mutants (see Table 2-1) to either reduce or increase the length of the 5’UTR expressed as a relative percent of ptrB wild-type CDS expression (100%). * above the
error bars refer to a statistical significance compared to pAptrB.WT where p<0.05 and ** where
p<0.001. A horizontal line and ** above two bars denotes to a statistical comparison between
those two constructs where p<0.001. (B) Toeprint assay containing mRNA corresponding to
either the pAptrB.WT or pAptrB.21ntadd constructs performed as previously described (see
Figure 2-2). Reaction components are indicated. Double plus sign (++) refers to component
which was added first for pre-binding. Predicted positions of toeprint signals (+15) to the 5’uAUG of pAptrB.WT (open arrow), the 5’-uAUG of pAptrB.21add (shaded arrow) and the
downstream AUG (closed arrow) are indicated.
79
80
and the positioning of the sequence elements relative to one another may be influential for CDS
expression.
Due to the short nature of the ptrB 5’-UTR, there is not enough space to allow for
simultaneous binding of both a 70S ribosome at the 5’-terminus and a 30S subunit at the internal
RBS. Therefore, the 5’-UTR was lengthened, which would presumably allow for dual ribosome
binding. Nine, twenty-one and thirty-three nucleotides of intervening sequence were then added
to the 5’-UTR, immediately following the 5’-uAUG (pAptrB.9ntadd, pAptrB.21ntadd,
pAptrB.33ntadd, Table 2-1). The addition of nine nucleotides had no impact on downstream
expression however, as the region between the 5’-uAUG and the CDS initiation codon was
extended, CDS expression was reduced in a linear fashion (Figure 2-5A).
To understand the relationship between ribosome binding now possible on the elongated
ptrB 5’-UTR, a toeprint assay was performed with the ptrB mRNA containing the 21-nucleotide
insertion (pAptrB.21ntadd). 30S subunits and 70S ribosomes were added separately and
simultaneously to compare the differences seen as a result of ribosomal interactions (Figure 25B). When added separately (Figure 2-5B, lanes 7-12), no significant difference was observed in
the ribosomal binding pattern when compared to ptrB wild-type mRNA (Figure 2-5B, lanes 1-6).
However, when added simultaneously (Figure 2-5B, lane 13), internal binding was observed
with no discernable binding at the 5’-terminus. The 30S subunits were then pre-bound to the
mRNA for 15 minutes with subsequent addition of the 70S ribosomes and vice versa. When the
30S subunits were pre-bound (Figure 2-5B, lanes 14-16), binding was observed at both the
internal RBS and the 5’-terminus. Binding at both RBSs was also seen when the 70S ribosomes
were pre-bound (Figure 2-5B, lanes 17-19), however 70S ribosome pre-binding resulted in a
more pronounced band at the 5’-terminus when compared to the internal RBS. Overall, the
toeprint assay shows that there are differences that occur depending on the order of ribosomal
addition to the system, suggesting that the 70S ribosome and 30S subunits may be interacting or
competing for binding regions, which could contribute to CDS expression levels.
Expression of the uORF negatively impacts downstream CDS expression
81
The ptrB uORF is not efficiently translated despite strong 70S ribosome binding to the
5’-uAUG (Figure 2-2), leading to the prediction that expression from the 5’-uAUG is repressed
to allow for CDS translation. Therefore, we introduced randomized sequence mutations across
the 5’-UTR to test whether the ptrB 5’-uORF translational repression could be relieved. A
tandem mutation at positions +6 and +8 (with the A of the 5’-uAUG at +1) from TCA to GCG
was identified (pAptrB.6GCG, Table 2-1) that caused a 16-fold increase in 5’-uORF expression
(Table 2-2). Conversely, in the presence of the GCG mutation (pAptrB.6GCG), expression from
the CDS was reduced to 30% of wild-type expression (Table 2-2). Individual mutations were
also made at each position (pAptrB.6GCA and pAptrB.6TCG, Table 2-1) but neither produced as
drastic results as seen by the mutations together (Table 2-2). Although relieving the apparent
repression of the uORF translation reduced CDS expression, other sequence mutations within the
5’-UTR that did not allow for efficient uORF translation also reduced CDS expression (Table 22). Therefore, inhibition of uORF may indirectly influence downstream regulation, while specific
sequence features within the 5’-UTR have significant effects on CDS translational efficiency.
The mutation at position +8 is predicted to cause a change in amino acid sequence in the
event of uORF translation, with glutamine (CAA) being changed to an arginine (CGA) at the
third position. Because the CGA codon is rare in E. coli, we tested whether the amino acid
sequence influenced expression levels, by mutating the third codon of the uORF to either CGC
(pAptrB.7CGC, Table 2-1), a non-rare arginine codon, or AGA (pAptrB.7AGA, Table 2-1),
another rare arginine codon. The CGC mutation led to a 22% reduction in uORF expression and
the AGA mutation resulted in a nearly 4-fold increase in uORF expression (Table 2-2).
Conversely, the CGC mutation resulted in an 83% reduction in CDS expression, whereas the
AGA mutation resulted in a 25% reduction in CDS expression (Table 2-2). These data indicate
that the changes in both uORF expression and downstream CDS expression are not a result of the
amino acid changes within the uORF or tRNA availability, but rather the nucleotide sequence
changes, highlighting important sequence features within the 5’-UTR.
Heterologous regulation by the 5’-UTR of ptrB
There appear to be multiple regions of importance contained within the ptrB 5’-UTR that
influence downstream CDS expression. To determine if the ptrB 5’-UTR could similarly
influence other genes’ CDSs, gene fusions were constructed linking the ptrB 5’-UTR to two
82
Table 2-2. ptrB mutations expressed as a percentage of either the 5’-uAUG in-frame leader
fusion with lacZ (100%) (column 2) or the CDS fused with lacZ (100%) (column 3).
83
Construct
Relative percentage 5-uAUG
Leader frame
Relative percentage CDS
frame
pAptrB.6GCA
107 ± 18
50 ± 3
pAptrB.6TCG
176 ± 40
40 ± 4
pAptrB.6GCG
1640 ± 30
30 ± 10
pAptrB.7CGC
80 ± 9
17 ± 5
pAptrB.7AGA
379 ± 32
74 ± 13
84
Figure 2-6. Heterologous regulation by the 5’UTR of ptrB. (A) LacZ activity expressed from
ptrB 5’UTR (AUG or AUG  AUC KO) fusions to pnp or pncB coding sequence fused to lacZ
(Table 2-1). * refers to a statistical significance where p<0.001. (B) Toeprint reactions (as
described in Figure 2) with mRNA containing the ptrB 5’UTR (5’-uAUG or AUG  AUC KO)
fused to the pnp coding sequence. Reaction components are indicated. Predicted positions of
toeprint signals (+15) to the 5’-uAUG (open arrow) and downstream CDS AUG (closed arrow)
are indicated.
85
86
other genes’ ORFs, pncB and pnp (pABptrB-pnp and pABptrB-pncB, Table 2-1). To examine
the 5’-uAUG regulatory effect on the other genes, the ptrB 5’-uAUG was again mutated to AUC
to disrupt ribosome binding and 5’-uORF translation. The change caused the expression of both
pncB and pnp CDSs to be significantly reduced by more than 90% (Figure 2-6A), following the
trend that was seen previously for the ptrB CDS. The ribosome binding pattern of the ptrB-pnp
fusion supported the expression data, with the 5’-uAUGAUC mutation resulting in the loss of
binding signals at both the 5’-terminus and the internal start codon (Figure 2-6B). Therefore, the
ptrB 5’-UTR regulatory signals can exert their effects on other E. coli genes when transferred
upstream of their CDS.
Initiation signals present within a bona fide CDS
Because the ptrB 5’-UTR conferred positive regulation for downstream ORFs, we sought
to determine if the addition of the ptrB 5’-UTR would allow for translation initiation from an
internal fragment of RNA. Presumably, an internal fragment would be devoid of any translation
initiation signals, but would still have the ability to be translated due to its coding potential.
Internal fragments of two E. coli genes, tna and aroL, were used. An internal in-frame
methionine, 381 nucleotides and 126 nucleotides from the native tna and aroL initiation codons,
respectively, was selected to act as an initiation codon. Each truncated ORF was then fused to the
ptrB 5’-UTR (pABptrB-tna.IN and pABptrB-aroL.IN, Table 2-1). This fusion would allow the
ptrB 5’-UTR to regulate expression of the downstream ORF to produce a truncated version of
the Tna or AroL proteins, respectively.
The ptrB 5’-UTR was not sufficient for efficient expression of the internal tna or aroL
ORFs (Figure 2-7). In the case of the aroL fusion, expression was just above levels of detection
Figure 2-7). The tna fusion produced low levels of expression but interestingly, when the ptrB
5’-uAUG was mutated to AUC, it caused a drastic reduction in expression of approximately
90%, similar to that seen with the bona fide coding sequences of ptrB, pncB and pnp (Figure 22A, Figure 2-6A, and Figure 2-7). It would appear that the lack of translation initiation signals
present on the internal fragments of RNA prevent efficient translation. Taken together, there are
signals within translation initiation regions of bona fide coding sequences that are important for
translation, and the ptrB 5’-UTR is not able to confer efficient expression independently and
must instead work in concert with signals within the CDS to identify a translation initiation
87
Figure 2-7. ptrB 5’-UTR insufficient to stimulate expression of internal RNA fragments.
Expression levels of internal RNA fragments (aroL.IN/tna.IN) fused to lacZ with or without ptrB
5’-uAUG mutation (see Table 2-1). All values compared to ptrB CDS WT (pAptrB.WT) fused to
lacZ (100%). * refers to a statistical significance compared to pAptrB.WT where p<0.001. A
horizontal line above two bars denotes a statistical significance between those two constructs
where p<0.001.
88
89
region. However, the 5’-uAUG still acts as an important signal for even low levels of translation,
supporting its regulatory role.
To further examine the role of CDS regions, we utilized pncB, which relies on two
nucleotides within its CDS for expression (Brock et al., 2007). pncB has a weak SD sequence,
does not possess a 5’-uAUG, and is not influenced by its native 5’-UTR for expression. Instead,
two nucleotides, both adenines, at positions +6 and +9 of the CDS, appear to influence
translation. When these two adenines are mutated to guanines, pncB expression is completely
abolished (Brock et al., 2007). A construct was made replacing the pncB 5’-UTR with the ptrB
5’-UTR (pABptrB-pncB, Table 1) to determine if the positive effects of the ptrB 5’-UTR could
overcome the negative impact of the pncB CDS mutations. In the presence of the ptrB 5’-UTR,
the CDS mutations (pABptrB-pncBdblmut, Table 2-1) caused pncB expression to be
significantly reduced by approximately 80% (Figure 2-8A), rather than completely abolished,
suggesting that the signals within the ptrB 5’-UTR partially rescue the negative impact of the
pncB CDS mutations on expression. The ptrB 5’-uAUG to AUC mutation (pABptrB5’KOpncB, Table 2-1) resulted in an even more dramatic reduction in pncB CDS expression compared
to the pncB CDS mutations (pABptrB-pncBdblmut, Table 2-1) in the presence of the ptrB 5’UTR. In this case, expression was reduced by approximately 88% (Figure 2-8A), which follows
the trends previously seen with ptrB 5’-uAUG mutation (Figure 2-2A). These data indicate that,
in this background, the 5’-uAUG has more of an impact on pncB CDS expression than the
influential regions within the pncB CDS itself. Finally, when the ptrB 5’-uAUG and the two
nucleotides within the pncB CDS were mutated in tandem (pABptrB5’KO-pncBdblmut, Table 21), pncB CDS expression was again abolished (Figure 2-8A), supporting the notion that the 5’uAUG and the two nucleotides within the CDS are the signals that control pncB CDS expression
in this context.
Interestingly, similar to the pncB CDS, the ptrB CDS also has adenines at position +6 and
+9 (relative to the A of the CDS start codon at +1). When both adenines were mutated to
guanines in ptrB (pAptrB.CDSdblmut, Table 2-1), CDS expression was also reduced by
approximately 70% (Figure 2-8B). In addition, when the 5’-uAUG and the two nucleotides
within the ptrB CDS are mutated in tandem (pAptrB5’KO.CDSdblmut, Table 2-1), expression
90
Figure 2-8. Expression regulated by both upstream and downstream signals. (A) Expression
values represented as a percent of the ptrB 5’-UTR fused to the pncB CDS with the pncB CDS
in-frame with lacZ (100%). Comparison of constructs with mutated upstream and downstream
signals (Table 1). (B) Expression values represented as a percent of that of the ptrB WT CDS inframe with lacZ (100%). Comparison of constructs with same mutated upstream and downstream
signals as in (A) (also see Table 1). * refers to a statistical significance compared to pABptrBpncB or pAptrB.WT respectively where p<0.001.
91
A
B
92
was abolished (Figure 2-8B). These results are concordant with the expression data derived from
similar mutations in the ptrB-pncB construct (Figure 2-8A). Taken together, these data suggest
that specific upstream and downstream sequence elements are responsible for CDS expression in
this novel regulation mechanism to compensate for the weak SD sequence.
93
Discussion
Our results suggest that translation initiation of E. coli ptrB mRNA is controlled via its
5’-terminal AUG, which represents a novel form of regulation that can act as an autonomous
regulatory element independent of gene context. We present two models that might explain the
mechanism of ptrB regulation. The models differ in the stability of the initial interaction between
the mRNA and the ribosome. In the first model, the 5’-uAUG acts as a recognition signal to
attract ribosomes through transient, unstable binding at the 5’-terminus. The transient binding
increases the local concentration of ribosomes to allow for more efficient binding to the
downstream ORF RBS resulting in ternary complex formation and increased CDS initiation. In
the second model, the 5’-uAUG acts as both a recognition signal and stable binding platform.
Ribosomes can stably bind the 5’-terminus of the mRNA through their inherent ability to bind
5’-AUGs, i.e., leaderless mRNAs. Additional sequences in the ptrB 5’-UTR then help stabilize
the ribosome at the 5’-terminus to prevent the ribosome from dissociating. The ribosome can
then transition down the mRNA until it reaches the CDS initiation codon. In each model, the
CDS also plays a role, likely supporting the proper positioning of the ribosome on the CDS
initiation codon, thereby defining the reading frame. In both models, the ptrB mRNA contains
upstream and downstream binding signals to compensate for its weak SD sequence which
function cooperatively in a novel regulatory mechanism. The models only differ as to whether
the ribosome dissociates from the 5’-terminus of the mRNA or remains associated to the mRNA.
Both models agree with the previously described “cumulative specificity initiation mechanism”
of Nakamoto (2011), in which multiple preferred upstream and downstream bases at specific
positions work cumulatively, but independent of each other, to regulate translation initiation,
thereby making no singular interaction completely essential.
In both models presented, the 5’-uAUG is required, but as a binding signal rather than
acting as an initiation codon. We have shown the inefficiency of uORF translation to the extent
in which translation appears to be repressed to prevent elongation and instead directs ribosomes
to the downstream RBS. This repression allows ptrB to take advantage of a sequence, the 5’terminal AUG, that strongly binds ribosomes without the added energetic consequence of uORF
translation. It is still unclear exactly how translation is repressed but poor coding potential
appears to be involved. Although the uORF is not efficiently translated, the 5’-uAUG still plays
a major role in ribosome binding to the ptrB mRNA to influence expression of the downstream
94
CDS. It is evident that the regulation is acting via the 5’-uAUG, rather than another trans-acting
element, because the effect of the 5’-uAUG mutation is also seen in the closed system of the
toeprint assay. Therefore, we can conclude that the 5’-uAUG is interacting with either the
initiator tRNA or the ribosomes rather than a trans-acting element such as a RNA-binding
protein or small RNA. To corroborate this hypothesis, we also measured ptrB expression levels
in an Hfq deletion strain (Baba et al., 2006) and saw no change in expression (data not shown),
greatly reducing the possibility of the involvement of a small RNA interaction.
If the 5’-uAUG acts as the sole recognition signal rather than the SD sequence, the SD
sequence could be mutated without loss of expression. However, mutation of the SD sequence
significantly decreased CDS expression, indicating that the SD sequence may be used as a signal
to properly position the start codon in the ribosomal P-site. Thus, the ribosome can bind to the
mRNA but cannot be efficiently directed to the proper start codon in the absence of an intact SD
sequence, negatively impacting expression. The SD sequence appears to be too weak to
effectively act as both the binding and positioning signal, causing the 5’-uAUG to adopt one of
these roles. However, when the SD sequence is strengthened, it can again take on both roles,
making the 5’-uAUG obsolete (Figure 2-2).
The 5’-uAUG might be necessary to compensate for secondary structure within the 5’UTR that negatively impacts ribosome binding, because a local single stranded region is needed
for efficient ribosome binding (Draper, 1987; Gualerzi and Pon, 1990; de Smit and van Duin,
1990; Draper et al., 1998). Single-stranded regions upstream of the RBS have been shown to act
as “standby sites” to allow for temporary ribosome binding until the inhibitory secondary
structure opens (de Smit and van Duin, 1994; Studer and Joseph, 2006; Unoson and Wagner,
2007). It has been suggested that the ribosome can then transition downstream in a scanning-like
movement to reach the start codon (Petersen et al., 1978; Adhin and van Duin, 1990). However,
it appears that, in the case of ptrB, the 5’-UTR must be of a certain length to accommodate
ribosome binding and downward progression (Figure 2-5A). Examples of such energyindependent ribosome movement have also been demonstrated during translational reinitiation
(Schmidt et al., 1987; Berkhout et al., 1987; Adhin and van Duin, 1990; Osterman et al., 2013;
Shell et al., 2015, Yamamoto et al., 2015), translation of the MazF-regulon in E. coli (Sauert et
al., 2016), and in chloroplasts (Hirose and Sugiura, 2004). Therefore, the 5’-uAUG may act as a
95
standby-like site for initial ribosomal loading, after which the ribosome can then slide/scan down
the mRNA to the RBS when the structure relaxes momentarily. However, increasing the
distance that the ribosome must travel leads to a decrease in CDS initiation efficiency (Figure 25A). The same effect occurs in translation reinitiation in which longer intercistronic gaps lead to
reduced downstream initiation in operons due to ribosome drop-off (Cone and Steege, 1985).
Although it is unclear whether the 30S subunit or 70S ribosome is responsible for ptrB
regulation, our data raise the possibility that the 30S subunit may be initially loaded onto the 5’terminus and occupy the upstream region of the mRNA. When 30S subunits are pre-bound or
added simultaneously with 70S ribosomes, they appear to block the ribosome binding of the 70S
ribosome at the 5’-terminus (Figure 2-5B). However, the 30S subunit binding is known to be
unstable and therefore unlikely to block the action of the reverse transcriptase (Brock et al.,
2008), which can explain why we are unable to capture the 30S subunit binding at the 5’terminus in the toeprint assay (Figure 2-5B). This idea also explains why there is reduced 30S
subunit binding internally when the 70S ribosome is pre-bound (Figure 2-5B), because the 70S
ribosome may be blocking the 30S subunit entry site at the 5’-terminus. Although 70S
ribosomes are thought to primarily bind at the 5’-terminus, Brock et al. (2008) demonstrated that
the ribosomal proteins implicated in 5’-terminus binding are all 30S subunit proteins, suggesting
that the 30S subunit could also bind the 5’-terminus, and may therefore be implicated in ptrB
regulation.
It is also possible, due to the short nature of the ptrB 5’-UTR, that binding of the
ribosome at the 5’-terminus can resolve secondary structure at the RBS. It is unclear how the 70S
ribosome is loaded on to the 5’-terminus of leaderless mRNA; however, it is thought that the 5’end is pulled through the channel between the 30S and 50S subunits (Moll et al., 2002). The 5’AUG then finds its way through the tunnel until it is positioned in the P-site through interactions
with the initiator tRNA (Moll et al., 2004). The tunnel is only, on average, 15 Å in diameter
(Nissen et al., 2000; Takyar et al., 2005), so it cannot accommodate secondary structure.
Therefore, the structure must be eliminated as the mRNA is fed through the tunnel such that the
region of mRNA protected by the ribosome, from the 5’-terminus to approximately position +19
(Huttenhofer et al., 1994), becomes single-stranded. Because the ptrB 5’-UTR is only 26
96
nucleotides long, the loaded ribosome may cover the majority of the 5’-UTR, thereby
eliminating secondary structure.
Regardless of the model, this 5’-uAUG regulation mechanism has the ability not only to
influence ptrB CDS expression, but also to regulate other E. coli CDS. However, the inefficient
use of internal RNA fragments (Figure 2-7) suggests that signals within a genuine CDS must
nonetheless be present to direct the ribosome downstream for start site selection. Mutational
studies indicate that adenines at certain positions play a large part in the signal important for
CDS expression. Statistical analyses have shown an abundance of CA and A-rich downstream
sequences (Scherer et al., 1980; Rudd et al., 1992) with a large bias in codons at positions +2,
+3, and +4 (Sato et al., 2001) supporting this notion. Adenine-rich regions enhance expression
through increasing the rate and amount of ternary complex formation during initiation (Brock et
al., 2007). It is thought that the increased efficiency is due to ribosomal protein S1’s high affinity
for polypyrimidines (Draper and von Hippel, 1978; Subramanian, 1983) and S1’s ability to bind
different base motifs (Tzareva et al., 1994). Similarly, our results demonstrate a decrease in
expression of both pncB and ptrB as a result of a decrease in adenine richness (Figure 2-8). A
similar effect was also seen when analyzing the uORF expression levels via mutational analysis.
The mutations that introduce adenines increase uORF expression by relieving the apparent uORF
repression (Table 2-2), presumably by improving the codon usage and potential ribosome
interactions.
Overall, these data suggest a novel mechanism of translational regulation, functioning in
a distinct, yet cooperative mechanism from the canonical SD sequence mechanism, and instead
highlight the 5’-uAUG. Our previous study demonstrated the abundance of 5’-uAUGs within E.
coli transcripts (Beck et al., 2016); therefore, it is possible that this mechanism of regulation is
widespread in E. coli and may contribute to regulation of mRNA with weak or absent SD
sequences. Understanding mechanisms of non-SD sequence translation is essential because
approximately one half of prokaryotic mRNAs do not contain a SD sequence (Chang et al.,
2006). There are numerous examples of translation events occurring in the absence of an SD
sequence (Boni et al., 2001; Nakamoto, 2006 and references therein); leaderless mRNAs
efficiently initiate translation without a 5’-UTR altogether. There is also evidence that suggests
that, in certain instances, the SD sequence is important only if the initiation codon is not an AUG
97
or the RBS is masked by secondary structure (Munson et al., 1984; de Smit and van Duin, 1994;
Fargo et al., 1998), and the aforementioned S1-mediated initiation acts independently of the SD
sequence (Boni et al., 2001; Osterman et al., 2013). Clearly, there are more diverse mechanisms
for translation initiation than the conventional SD sequence-regulated mechanism. Others agree
that alternatives to the SD sequence-dependent mechanism are much more prevalent in bacteria
than anticipated (Hering et al., 2009; Malys and McCarthy, 2011). It is important to understand
the different mechanisms of translational regulation because this information can be applied to
synthetic biology and genetic engineering to finely tune protein production. Other studies
exploring optimization of recombinant gene expression have discovered upstream and
downstream sequence elements that work cooperatively to influence expression (Berg et al.,
2012). However, the elements identified work in a gene-dependent manner (Berg et al., 2012).
The sequence elements identified in this study, however, appear to function in a geneindependent manner, suggesting that they may be more easily adapted to control gene
expression.
98
Appendix A
Leaderless mRNAs initiate translation via a mechanism of ribosome binding distinct
from canonical leadered initiation. Rather than the multi-step process of a 30S subunit binding
followed by a 50S subunit binding to form the 70S initiation complex, the intact 70S ribosome
binds directly to the 5’-terminus of leaderless mRNA (Balakin et al., 1992; Udagawa et al.,
2004; Moll et al., 2004). While the process of leadered mRNA recognition and binding by the
30S subunit has been elucidated, it is currently unknown how leaderless mRNAs undergo these
initial steps involving the 70S ribosome. In Escherichia coli, the ribosomal proteins (r-proteins)
that appear to be involved in leaderless mRNA loading were discovered using 4-thiouridine (4SU) cross-linking studies with intact ribosomal subunits in either the presence or absence of
initiator tRNA and initiation factors (IFs) (Brock et al., 2008). The r-proteins found to cross-link
leaderless mRNA were all part of the 30S subunit and include S1, S3, S4, S7, and S10/S18.
Since the r-proteins involved in the initial mRNA: ribosome interaction have been
identified, it is now of interest to determine the specific cross-link positions on each purified rprotein (Giliberti, dissertation) to map the path of the leaderless mRNA when loaded into the 70S
ribosome. The objective of this study is to develop a method to identify specific cross-linked
positions between a RNA molecule and a protein that can be utilized to determine how leaderless
mRNAs interact with the 70S ribosome.
Similar to previous reports (Brock et al., 2008), a 4S-U at the +2 position (the U of the 5’AUG) of the leaderless mRNA cI from bacteriophage λ was used as a substrate for
photochemical cross-linking. 4S-U is used as a cross-linker due to its ability to covalently crosslink to nucleotides or amino acids, preferentially lysines, that are within 4 Å after UV excitation
(Favre, 1990). Although the initiator tRNA and IFs have been implicated in contributing to
efficient translation initiation of leaderless mRNAs, the cross-links to the ribosomal subunits
were unchanged in the absence of the initiator tRNA and IFs (Brock et al., 2008). This suggests
that their influence on translation efficiency occurs after the initial ribosome interaction and were
therefore not considered in this study.
For initial method development, r-protein S1 from E. coli was used due to its size, ease of
purification and abundance of previously identified cross-links (Brock et al., 2008). S1 also
99
contains four known RNA binding domains (R1-R4) located at the C-terminus and two domains
at the N-terminus involved in ribosome binding (Figure A-1) (Subramanian, 1983; McGinness
and Sauer, 2004). It is thought that domains R1-R3 are involved in mRNA binding whereas R4 is
involved in the autogenous regulation of S1 expression (McGinness and Sauer, 2004).
For cross-linking, 1 nmol of purified His-tagged S1 (Giliberti, dissertation) was incubated
with 2 nmol of a 20-nucleotide RNA oligomer. The oligonucleotide sequence was 5’-A [4S-U]
GAGCACAAAAAAGAAACC-3’, corresponding to the first 20 nucleotides of the bacteriophage
 cI leaderless mRNA. The cross-linking was performed for 20 minutes at 37oC in 5X sample
buffer (300 mM NH4Cl, 50 mM Tris-acetate pH 7.6, 50 mM Mg(OAc)2, 3 mM βmercaptoethanol). The reaction mixture was then distributed in 20-µL spots onto a Petri dish
resting on ice to reduce temperature changes as a result of UV light exposure. The pre-warmed
hand held UV light was then placed approximately 0.5 inches above the samples and the
reactions were exposed for 15 minutes to UV light (366 nm). Reactions were then pooled and
purified to remove non-cross-linked RNA from the S1 protein using metal ion affinity
chromatography (PrepEase Histidine-tagged Protein Purification Midi Kit, Affymetrix). The
column was first equilibrated with wash buffer (10 mM Tris, 0.3 M NaCl) prior to sample
loading. After the sample was loaded onto the column, it was then washed three times with the
wash buffer and eluted three times using the elution buffer (10 mM Tris, 0.3 M NaCl, 250 mM
imidazole). Fractions were collected and analyzed by SDS-PAGE and Coomassie staining
(Figure A-2). The gel indicates that the process of cross-linking is not 100% efficient because a
band representing uncross-linked S1 remains. However, the shifted band represents purified S1
protein that has been cross-linked with the cI RNA oligonucleotide (Figure A-2). Fractions
containing S1 were pooled for further analysis. After purification, the samples were trypsindigested, treated with nuclease P1 and subjected to inductively coupled plasma mass
spectrometry (ICP-MS) and electrospray ionization mass spectrometry (ESI-MS), both coupled
with on-line high-performance liquid chromatography (HPLC) for identification of cross-linked
amino acid residues (Giliberti, dissertation).
100
Figure A-1. Schematic of S1 domains and their amino acid positions in the S1 r-protein.
The gray boxes are the two N-terminal domains and the black boxes are the four C-terminal
RNA-binding domains (R1-R4). Adapted from McGinness and Sauer, 2004.
101
102
Figure A-2. Cross-linking of 4S-U cI RNA to purified S1 protein. Purified His-tagged S1
(lane 2) was cross-linked to 4S-U cI RNA oligonucleotide and bound to PrepEase His-tag
purification column. Lane 3 represents the fraction eluted off the column using imidazole. Lane 4
represents the fraction collected after washing the column prior to the elution step. Lane 1
contains the molecular weight marker and sizes of each band are indicated.
103
104
Several cross-links were detected reproducibly in multiple biological replicates and at
least five technical replicates suggesting these are structured interactions between the RNA
fragment and r-protein S1. These cross-links were identified at lysine residues K196 (domain
R1), K260 (domain R1), K272 (domain R1/R2), K450 (domain R3/R4), K457 (domain R4) and
K464 (domain R4) (Figure A-1, Figure A-3). Spatially, the cross-links appear to lie in two
distinct regions and within those regions, the cross-links all lie on the same face of the protein
(Figure A-3). The cross-links all also lie within mRNA binding domains, with no cross-links
occurring within N-terminal domains known to play a role in association with the 30S ribosomal
subunit. Unfortunately, high resolution X-ray crystal structures of the ribosome are intentionally
depleted of S1 due to its high flexibility (Schuwirth et al., 2005). Therefore, it is difficult to
relate, with any certainty, our identified cross-links to the ribosome crystal structure.
S1 was expected to allow for mRNA: protein cross-links due to its multiple RNA binding
domains (Subramanian, 1983). The mRNA binding domains R1-R4 in the C-terminus have been
predicted to extend at least 10 nm from the 30S subunit into the cytoplasmic space to interact
with mRNA (Walleczek, et al., 1990). A flexible hinge region has been identified that could
allow the C-terminus to scan the area surrounding the ribosome for mRNA and then contract to
position the mRNA close to the mRNA track (Byrgazov et al., 2015) supporting the prediction
that S1 is the first r-protein to interact with leadered mRNA (Komarova et al., 2002). S1 also
interacts with polyuracil regions of the 5’-untranslated region to facilitate 30S binding to
mRNAs with weak Shine-Dalgarno sequences (Boni et al., 2001; Simonetti et al., 2009). S1 is
highly involved in leadered mRNA binding (Suryanarayana and Subramanian, 1983; Boni et al.,
1990), and thought to be required for most, if not all, mRNA translation (Sørensen et al., 1998).
Although it has been suggested that leaderless mRNA initiation can occur in the absence of S1 in
vitro (Tedin et al., 1997; Krishnan, 2010), these results suggest that S1 might play an influential
role in ribosome binding to cI leaderless mRNA due to the reproducibility and non-random
distribution of cross-links.
These results suggest that a method has been successfully developed to identify crosslinks between mRNA and r-proteins, generating reproducible results with biological relevance.
Further use of this method to determine the interactions between leaderless mRNA and r-proteins
105
Figure A-3. Cross-linked positions on S1 ribbon structure. Crystal structure of ribosomal
protein S1 in green with cI 20mer oligonucleotide highlighted in red depicting potential crosslinked sites shown from two different views.
106
107
is important because the r-proteins have been implicated to play a role in progression of the start
codon to the P-site. Cross-linking results produced by Brock et al. (2008) indicated that the
presence of IFs and tRNA does not change the cross-link pattern between the ribosomal subunits
and the leaderless mRNA. This result differs from previous leadered mRNA cross-linking
studies performed by La Teana et al. (1995), suggesting that rather than IFs and tRNA causing
the progression of the start codon to the P-site, the r-proteins may instead be playing this role in
leaderless mRNA, reinforcing the differences between the two mechanisms of initiation. It is
now of interest utilize this method to determine the cross-link positions for the other previously
identified cross-linked r-proteins (S3, S4, S7, and S10/S18) since they may play larger roles in
leaderless mRNA binding and progression to the P-site. Using this newly developed method,
researchers can determine the numerous roles the r-proteins may be playing in translation
initiation through their interactions with leaderless mRNAs.
108
Concluding Remarks
The process of translation initiation in bacteria is typically simplified to include a single
factor, the SD sequence, which governs efficiency and is used to define open reading frames.
However, advancements in bioinformatics and transcriptomics have uncovered novel types of
RNAs and regulation mechanisms revealing the abundance of non-canonical mRNAs in
prokaryotic organisms. Through in silico analysis, this study reinforces the abundance of noncanonical mRNAs in Escherichia coli by identifying previously unannotated open reading
frames (ORFs) at the 5’-terminus of canonical Shine-Dalgarno (SD) sequence-led mRNAs that
are translated at biologically relevant levels. In these messages, an AUG triplet at the 5’-terminus
(5’-uAUG) acts as an initiation codon and is bound by 70S ribosomes, similar to the 70S
initiation mechanism used by leaderless mRNA (Balakin et al., 1992; Udagawa et al., 2004;
Moll et al., 2004). A subset of 5’-uAUGs identified were also implicated in playing a role in
regulation of the downstream coding sequence (CDS). In one mRNA, ptrB, we were able to
define a novel regulation mechanism dependent upon the 5’-uAUG as a signal itself, rather than
its ability to act as an initiation codon, that works in concert with downstream sequence features
to govern expression of the CDS. Overall, this study supports the idea that leaderless mRNAs are
more abundant than originally expected.
Since leaderless mRNAs are relatively prevalent, it is important to understand how they
are recognized by the ribosome in the absence of conventional binding signals. The translation
initiation mechanism used by leaderless mRNAs is distinct from the mechanism used by
canonical leadered mRNAs but the process of mRNA loading into the 70S ribosome is yet to be
elucidated. This study provided further insight into a leaderless mRNA’s initial interactions with
the ribosome through the use of photochemical cross-linking. We contributed to the development
of a method that makes use of 4-thiouridine (4S-U) cross-linking to identify specific contact sites
between a protein and an RNA molecule, specifically ribosomal protein S1 and the naturally
occurring leaderless mRNA cI from bacteriophage λ. This method will allow researchers to map
the initial interaction between leaderless mRNAs and the ribosome, providing insight into the
long-sought after mechanism of leaderless mRNA loading into the 70S ribosome.
Efforts to correct the underrepresentation of leaderless mRNAs and small peptides
109
Bioinformatic analysis performed in this study identified 287 transcripts in E. coli with an
AUG triplet within three nucleotides of the 5’-terminus. Although only thirteen transcripts have
been tested, the 5’-uORFs were expressed at varying levels and all the 5’-uAUGs appear to
perform diverse functions, indicating that extensive testing of all transcripts identified could
uncover novel functions and peptides. This study highlights the current underrepresentation of
leaderless mRNAs in E. coli, as only three leaderless mRNAs were previously described
(Ptashne et al., 1976; Christie and Calendar, 1985; Klock and Hillen, 1986). If leaderless
mRNAs are underestimated in one of the most thoroughly studied model organisms used in
molecular genetics, the occurrence of leaderless mRNAs is likely to be miscalculated throughout
the prokaryotes. Similar in silico analyses performed in prokaryotes with published
transcriptomes are likely to uncover a large number of 5’-uAUGs on canonical mRNAs that can
function to initiate translation of an uORF. Although leaderless mRNAs in E. coli must initiate
with an AUG triplet, bacteria of other genera contain leaderless mRNAs that can also initiate
with a GUG triplet (Cortes et al., 2013; Schrader et al., 2014), expanding the potential number of
transcripts analyzed in other bacteria and revealing a vast number of putative peptides encoded
by uORFs.
Although the number of small peptides is likely to be underestimated because they are
harder to identify through both experimental and bioinformatic methods, the small peptides that
have been identified participate in a wide array of cellular processes in both bacteria and archaea
(Wang et al., 2008; Hobbs et al., 2011). Since small peptides require less energy to be translated
and folded, organisms often use them to signal environmental changes and to globally regulate
relevant pathways (Seligmann, 2003; Wang et al., 2008), suggesting there may be a selective
advantage to maintaining small peptides in the cell. There are many examples of peptides as
small as three amino acids in length with a variety functions such as detoxification of
electrophiles (Ferguson, 1999), penicillin biosynthesis (Martin et al., 2010) and protease
inhibition (Tashiro et al., 1987), demonstrating the ability of very small peptides to have specific
functions. All of the uORFs in this study encode peptides that exceed three amino acids in
length, suggesting that they might also exhibit relevant functions and should be analyzed further.
Whereas some small peptides function in a variety of cellular processes, others are used
to regulate those processes. Some uORFs as small as six codons in length are used for
110
translation of nascent peptides that influence the functionality of the ribosome on which they are
synthesized by inducing ribosomal stalling (Tenson and Ehrenberg, 2002; Cruz-Vera et al.,
2011). Nascent peptides are able to interact with the ribosome in variety of ways, including
binding within and blocking the ribosome exit tunnel, acting within the peptidyltransferase
center to inhibit its catalytic activity, or altering 23S rRNA conformation (Lovett and Rogers,
1996; Tanner et al., 2009; Yan et al., 2010). Although some uORFs produce small regulatory
peptides, others as small as two codons in length are used as regulators to increase the rate of
tRNA starvation, thereby inhibiting further protein synthesis (Dincbas et al., 1999; DelgadoOlivares et al., 2006). Yet another uORF function was identified in this study, in which the ptrB
5’-uAUG acts to regulate downstream translation from a SD sequence-led initiation codon,
rather than initiating translation to produce a peptide. The 5’-uORFs identified through the
bioinformatic analysis in this study could function similarly to any of the aforementioned
uORFs, or could show novel types of regulation. Further experimentation should be performed to
define possible functions of these 5’-uORFs.
Mechanisms to regulate protein production
Translation is highly regulated in a variety of ways because it is an energetically
expensive process. Since initiation is the rate-liming step of translation, many systems are in
place to specifically govern initiation to prevent an unnecessary use of energy. The abundance of
non-SD-led mRNAs has led to the belief that there are a variety of different mechanisms utilized
by prokaryotes to regulate initiation, including many alternatives to the canonical mechanism
(Malys and McCarthy, 2011). Therefore, it is not surprising that another regulation mechanism
has been identified to control expression of ptrB mRNA. However, more work must be
completed to fully understand the energetics and ribosomal movement in the mechanism.
Although ribosomal movement along a transcript when not actively translating is energyindependent (Schmidt et al., 1987; Yamamoto et al., 2015), the energy consumption in ptrB
regulation should be analyzed as well. The conformation of the ribosome and its initial binding
site must also be elucidated. Also, because the regulation operates in a gene context-independent
manner, this regulation may be more widespread. Therefore, it is of interest to identify other
transcripts that utilize the same method of regulation, potentially even in other genera.
111
The regulation of translation initiation employed by ptrB, though unique, has parallels
with that of E. coli tisAB mRNA, in which the uORF, tisA, is not translated but is simply used as
a platform for ribosome binding due to the absence of secondary structure (Darfeuille et al.,
2007). In the case of tisAB, upstream ribosome loading onto the standby site is required for
downstream initiation due to a highly structured SD sequence-led translation initiation region
(Darfeuille et al., 2007). The uORF in this case, although not at the 5’-terminus, is essential due
to inhibitory secondary structure, in contrast to the regulation of ptrB, which has a weak SD
sequence and instead must take advantage of the inherent binding strength of the ribosome for
5’-AUGs to allow for ribosomal binding and CDS expression. Therefore, the regulation observed
in translation initiation of ptrB may be a result of evolutionary selection to possess a 5’-terminal
AUG to be used as a ribosome binding platform to enhance translation that would be inefficient
if it were reliant upon canonical ribosome binding signals.
Leaderless mRNAs provide insight into the evolution of translation
Leaderless mRNA translation initiation has been hypothesized to reflect the mechanism
used by ancestral mRNAs because, regardless of origin, an mRNA with a 5’-terminal start codon
can be translated using the translational machinery from any of the three domains (Grill et al.,
2000). This evidence suggests that the inherent ability to translate leaderless mRNAs was present
prior to domain separation and has been retained throughout evolution. Genomic studies have
shown that more slowly-evolving bacteria and those located near the root of the 16S rRNA
phylogenetic tree have more leaderless mRNAs, and phylogenetically related species have a
similar abundance of leaderless mRNAs (Zheng et al., 2011). It is thought that distinct
mechanisms of initiation evolved due to the presence of the nuclear envelope in eukaryotes,
which decouples transcription and translation and requires added protection from RNases
through the addition of the 5’-7-methylguanylate cap and 3’-polyA tail (Malys and McCarthy,
2011). Although leaderless mRNAs are present in every domain of life (Janssen, 1993), the
reduced abundance of leaderless mRNAs in prokaryotes has also been attributed to an increased
number of operons (Zheng et al., 2011). Organizing genes into operons ensures that only the
proximal cistron can be leaderless, whereas the distal cistrons must have upstream sequence
features to recruit the ribosome. These sequence features are necessary for ribosomal retention
and proper alignment of the ribosome on downstream initiation codons (Cone and Steege, 1985).
112
This trend is also seen in archaea, in which proximal genes in operons are leaderless and the
distal genes are SD sequence-led (Tolstrup et al., 2000). This operon organization appears to be
the case for several of the genes identified in this study because they are the first cistron in their
operon, i.e., fucP, glpF, iscR, and uvrY, and have been shown in this study to also contain a
proximal leaderless 5’-uORF.
Although the mechanism of leaderless mRNA initiation reflects an ancestral initiation
mechanism, the translational machinery still possesses an inherent functionality to translate
leaderless mRNAs. The number of leaderless mRNAs present in prokaryotes suggests there may
be an advantage to retaining the ability to translate mRNAs lacking a 5’-UTR. During
translation, the mRNA must pass through a closed channel of the 30S ribosomal subunit, so an
initiation codon at the 5’-terminus allows for leaderless mRNAs to be threaded directly into the
decoding channel of the 70S ribosome (Benelli and Londei, 2009). Therefore, there are structural
advantages to lacking a 5’-UTR. These advantages manifest under distinct conditions such as
cold stress, carbon starvation and amino acid depletion (Uchida et al., 1970; Ruscetti and
Jacobson, 1972; Jones and Inouye, 1996; Vesper et al., 2011). Under these conditions, the
concentration of 70S ribosomes increases, which may influence their binding efficiency to
leaderless mRNAs (Vesper et al. 2011). This response to different conditions would allow 70S
binding to be globally regulated and certain mRNAs, such as leaderless mRNAs, to be translated
only under those specific conditions.
Under specific stress conditions, changes occur not only in the concentration of 70S
ribosomes but also in their composition. The composition of the 70S ribosome is altered under
kasugamycin stress, resulting in the loss of multiple ribosomal proteins from the 30S subunit,
giving rise to the formation of the 61S particle (Kaberdina et al., 2009). This reduced ribosomal
particle can still efficiently translate leaderless mRNAs, but not leadered mRNAs, suggesting
that the 61S particle reflects an ancient form of the ribosome that was present before domain
separation (Kaberdina et al., 2009). The proteins absent from the 61S particle were likely added
later in evolution to facilitate leadered mRNA translation. It would be of interest to determine if
the expression levels of the newly identified 5’-uORFs vary under stress conditions that alter the
concentration and composition of 70S ribosomes.
Applying newly discovered regulatory sequences to genetic engineering
113
A deeper understanding of microbial genetics has allowed scientists to begin the
enterprise of creating synthetic life (Hutchison et al., 2016). However, this study has
demonstrated that there are extensive areas of the process of translation, a fundamental
component of the central dogma, that are still unclear. These includes the different levels of
initiation regulation, and the abundance and mechanism of initiation of non-canonical mRNAs.
Because translation is such a tightly regulated and important process in the microbial life cycle, it
is important to understand how to control translational efficiency so we can effectively
manipulate microorganisms in the laboratory to benefit scientific research, as well as exploiting
them as protein production centers with maximum efficiency to provide limitless supplies of indemand proteins. The goal is to identify regulatory elements that can be applied to any variety of
CDSs to produce the expected concentrations of each protein.
This study has identified novel regulation mechanisms that can be utilized to help control
translation initiation, and therefore help control protein production. The presence of an uORF
has been shown in genetic engineering studies to be beneficial for translation of a downstream
cistron due to the positive effects of translational coupling (Mutalik et al., 2013). Coupling was
also shown to overcome the effects of a weak SD sequence, similar to the results seen in this
study. The presence of an uORF was beneficial when the same regulatory element was used to
control the translation of different CDSs, by reducing the variability of expression that naturally
arises when comparing expression of different CDSs due to the differences in coding potential
(Mutalik et al., 2013). Nevertheless, there was still variability in expression, demonstrating that
the regulatory elements tested function in a CDS-dependent manner. This variability could cause
problems in complex synthetic genetic systems because the expression is not as tightly controlled
to maintain the balance needed to sustain cellular metabolism and growth. However, this study
successfully identified regulatory sequence features that function in a CDS-independent manner.
The ptrB regulation mechanism influenced expression of multiple CDSs in E. coli similarly,
suggesting that we can utilize the ptrB regulatory elements to control expression of diverse
CDSs. The use of regulatory signals that function independently of gene context with low
variability has vast implications for its use in designing synthetic transcripts. The use of ptrB
regulatory sequences may be one step towards the goal of designing controllable regulatory
elements.
114
Using forward engineering, others could incorporate the sequence features identified in
this study to synthetically design a ribosome binding site to drive translation at a specified rate
(Salis et al., 2009). Designing synthetic translation initiation regions would allow for fine-tuning
of protein production, especially through the discovery of additional regulatory signals.
Therefore, rather than simply controlling translation through the strength of the SD sequence, a
transcript could also contain a variety of important upstream and downstream signals to result in
even more efficient expression. This tightly regulated control will be necessary as construction of
larger and more complicated genetic systems progresses.
Leaderless mRNAs advancing antibiotic research
The process of translation and the ribosome are both targets of numerous antibiotics
(Lambert, 2012). Further understanding of the process of translation and how it is regulated may
help us to improve current antibiotics or develop new antibiotics. There are antibiotics that are
ineffective in repressing leaderless mRNA translation, specifically kasugamycin and pactamycin
(Chin et al., 1993). Leaderless mRNAs can be translated under kasugamycin inhibition because
the binding region between the mRNA and the antibiotic is reduced due to the absence of the 5’UTR, and the presence of the 50S subunit stabilizes the binding of tRNA such that kasugamycin
is unable to cause tRNA dissociation to elicit its effect (Schuwirth et al., 2006). Pactamycin is
thought to specifically disrupt the SD-anti-SD interaction to block protein production (Brodersen
et al., 2000); however, because leaderless mRNAs are not reliant upon this interaction they are
successfully translated in the presence of this drug.
Although this information may currently be of little concern, certain bacteria undergo a
stress response that induces sequence specific endonuclease cleavage of mRNA 5’-UTRs
resulting in leaderless mRNAs (Vesper et al., 2011; Sauert et al., 2016). If a significant portion
of bacteria could acquire a similar response system through horizontal gene transfer, or already
possess a homologous system, the organisms may respond to antibiotic stress by generating
antibiotic-resistant leaderless mRNAs. This study has demonstrated the potential abundance of
previously unannotated leaderless mRNAs in E. coli, suggesting that there are likely many
leaderless mRNAs throughout prokaryotes, any of which might have resistance to certain
antibiotics. Therefore, it is important to not only understand the details of translation to generate
new antibiotics but to also understand how leaderless mRNAs are translated, in case bacteria can
115
utilize them as a countermeasure to antibiotic stress. The methods developed in this study can be
utilized to provide this invaluable information in understanding leaderless mRNA interactions
with the ribosome.
Conclusions
Overall, the information gained in this study has illustrated the need for more
comprehensive genomic and transcriptomic studies to identify potential uORFs. Although
advancements in high-throughput DNA sequencing have expanded our knowledge at a genomic
level, there are biases in the annotation of those genomes, with small ORFs and uORFs being
drastically underrepresented. Performing laboratory experiments is necessary to properly and
meticulously identify the transcriptome and proteome of various organisms.
Unfortunately, complete annotation of the transcriptomes will not provide us with
detailed mechanisms of translational regulation. We have worked to identify more sequences
involved in controlling protein production; however, we hypothesize that there are still many that
are unknown. It appears that each mRNA may possess different combinations of regulatory
sequences. This idea is termed the “cumulative specificity initiation mechanism,” in which the
initiation site is selected by the cumulative and cooperative actions of preferred nucleotides at
certain positions, both upstream and downstream relative to the initiation site (Nakamoto, 2011).
These independent interactions allow the ribosome to have broad specificity and ensure that no
singular interaction is absolutely essential (Nakamoto, 2011). Therefore, researchers should
continue to try to identify as many important regulatory sequences as possible to further our
understanding of a cellular process essential to every domain of life. Because there appears to be
an advantage to maintaining a 5’-uAUG and producing small peptides, we should continue our
research in both of these areas to utilize the millions of years of natural selection performed by
bacteria. This information can then be applied to antibiotic manipulation, genetic engineering
and –omics annotation.
116
References
Adhin, M.R. and van Duin, J. 1990. Scanning model for translational reinitiation in eubacteria. J. Mol.
Biol. 213: 811-818.
Andre, A., Puca, A., Sansone, F., Brandi, A., Antico, G., Calogero, R.A. 2000. Reinitiation of protein
synthesis in Escherichia coli can be induced by mRNA cis-elements unrelated to canonical
translation initiation signals. FEBS Letters 468:73-78.
Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okurmura, Y., Baba, M., Datsenko, K.A., Tomita, M.,
Wanner, B.L., Mori, H. 2006. Construction of Escherichia coli K-12 in-frame, single-gene
knockout mutants: the Keio collection. Mol. Syst. Biol. 2:1-11.
Balakin, A.G. Skripkin, E.A., Shatsky, I.N., Bogdanov, A.A. 1992. Unusual ribosome binding properties
of mRNA encoding bacteriophage λ repressor. Nuc. Acids Res. 20:563-571.
Benelli, D., Maone, E., Londei, P. 2003. Two different mechanisms for ribosome/mRNA interaction in
archaeal translation initiation. Mol. Microbiol. 50:635-643.
Benelli, D., Londei, P. 2009. Begin at the beginning: evolution of translation initiation. Res Microbiol.
160:493-501.
Berkhout B., Schmidt, B.F., van Strien, A., van Boom, J., van Westenen, J., van Duin, J. 1987. Lysis
gene of bacteriophage MS2 is activated by translation termination at the overlapping coat gene.
J. Mol. Biol. 195: 517-524.
Binns, N., and 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. 44:1287-1298.
Boni, I.V., Isaeva, D.M., Musychenko, M.L., Tzareva, N.V. 1990. Ribosome-messenger recognition:
mRNA target sites for ribosomal protein S1. Nuc. Acids Res. 19:155-162.
Boni, I.V., Artamonova, V.S., Tzareva, N.V., Dreyfus, M. 2001. Non-canonical mechanism for
translational control in Bacteria: synthesis of ribosomal protein S1. EMBO J. 20: 4222-4232.
Breaker, R.R. 2012. Riboswitches and the RNA world. Cold Spring Harb Perspect Biol. 4:a003566.
117
Brenneis, M., Soppa, J. 2009. Regulation of translation in haloarchaea: 5’- and 3’-UTRs are essential
and have to functionally interact in vivo. PLOS One 4:e4484.
Brock, J.E., Paz, R.L., Cottle, P., Janssen, G.R. 2007. Naturally occurring adenines within mRNA
coding sequences affect ribosome binding and expression in Escherichia coli. J. Bacteriol.
189:501-510.
Brock, J.E., Pourshahian S., Giliberti, J., Limbach, P.A., Janssen, G.R. 2008. Ribosomes bind leaderless
mRNA in Escherichia coli through recognition of their 5’-terminal AUG. RNA. 14:2159-2169.
Brodersen, D.E., Clemons, W.M., Carter, A.P., Morgan-Warren, R.J., Wimberly, B.T., Ramakrishnan,
V. 2000. The structural basis for the action of antibiotics tetracycline, pactamycin, and
hygromycin B on the 30S ribosomal subunit. Cell. 103: 1143-1154.
Byrgazov, K., Manoharadas, S., Kaberdina, A.C., Vesper, O., Moll, I. 2012. Direct interaction of the Nterminal domain of ribosomal protein S1 with protein S2 in Escherichia coli. PLoS One 7(3):
e32702. doi:10.1371/journal.pone.0032702.
Byrgazov, K., Grishkovkaya, I., Arenz, S., Coudevylle, N., Temmel, H., Wilson, D.N., DjinovicCarugo, K., Moll, I. 2015. Structural basis for the interaction of protein S1 with the Escherichia
coli ribosome. Nuc. Acids Res. 43: 661-673.
Calogero, R.A., Pon, C.L., Canonaco, M.A., and Gualerzi, C.O. 1988. Selection of the mRNA
translation initiation region by Escherichia coli ribosomes. Proc. Natl. Acad. Sci. USA. 85:64276431.
Chang, B., Halgamuge, S., Tang, S. 2006. Analysis of SD sequences in completed microbial genomes:
Non-SD-led genes are as common as SD-led genes. Gene. 373:90-99.
Chin, K., Shean, C.S., Gottesman, M.E. 1993. Resistance of λ cI translation to antibiotics that inhibit
translation initiation. J. Bacteriol. 175:7471-7473.
Christie, G.E., Calendar, R. 1985. Bacteriophage P2 late promoters II. Comparison of the four late
promoter sequences. J Mol Biol. 181: 373-382.
118
Condo, I., Ciammaruconi, A., Benelli, D, Ruggero, D. Londei, P. 1999. Cis-acting signals controlling
translational initiation in the thermophilic archaeon Sulfolobus solfataricus. Mol. Microbiol.
34:377-384.
Cone, K.C., and Steege, D.A. 1985. Functional analysis of lac repressor restart sites in translational
initiation and reinitiation. J. Mol. Biol. 186: 733-742.
Cortes, T., Shubert, O.T., Rose, G., Arnvig, K.B., Comas, I., Aebersold, R., Young, D.B. 2013.
Genome-wide mapping of transcriptional start site defines an extensive leaderless transcriptome
in Mycobacterium tuberculosis. Cell Reports. 5: 1121-1131.
Cruz-Vera, L.R., Sachs, M.S., Squires, C.L., Yanofsky, C. 2011. Nascent polypeptide sequences that
influence ribosome function. Curr. Opin. In Microbiol. 14: 160-166.
Darfeuille, F., Unoson, C., Vogel, J., Wagner, E.G.H. 2007. An antisense RNA inhibits translation by
competing with standby ribosomes. Mol. Cell. 26: 381-392.
de Boer, H.A., Hui, A., Comstock, L.J., Wong, E., Vasser, M. 1983. Portable Shine-Dalgarno regions: a
system for a systematic study of defined alterations of nucleotide sequences within E.coli
ribosome binding sites. DNA. 2:231-235.
Delgardo-Olivares, L., Zamora-Romo, E., Guarneros, G., Hernandez-Sanchez, J. 2006. Codon-specific
and general inhibition of protein synthesis by the tRNA-sequestering minigenes. Biochimie. 88:
793-800.
de Smit, M.H., van Duin, J. 1990. Secondary structure of the ribosome binding site determines
translational efficiency: a quantitative analysis. Proc. Natl. Acad. Sci. USA. 87: 7668-7672.
de Smit, M.H., van Duin, J. 1994. Translational initiation on structured messengers. Another role for the
Shine-Dalgarno interaction. J. Mol. Biol. 235:173-184.
de Smit, M.H. and van Duin, J. 2003. Translational standby sites: How ribosomes may deal with the
rapid folding kinetics of mRNA. J. Mol. Biol. 331:737-743.
Dincbas, V., Heurgue-Hamard, V., Buckingham, R.H., Karimi, R., Ehrenberg, M. 1999. Shutdown in
protein synthesis due to the expression of min-genes in bacteria. J Mol Biol. 291: 745-759.
119
Draper, D.E., von Hippel, P.H. 1978. Nucleic acid binding properties of Escherichia coli ribosomal
protein S1. J. Mol. Biol. 122: 321-359.
Draper, D.E. 1987. Translational regulation of ribosomal proteins in Escherichia coli. In Translational
Regulation of Gene Expression. Ilan, J. (ed.) New York: Plenum Press, pp.1-26.
Draper, D.E., Gluick, T.C., Schlax, P.J. 1998. Pseduoknots, RNA folding and translational regulation. In
RNA Structure and Function. Simons R.W., and Grubery-Manago, M. (eds). Cold Spring
Harbor, New York: Cold Spring Harbor Laboratory Press, pp. 415-436.
Dreyfus, M. 1988. What constitutes the signal for initiation of protein synthesis on Escherichia coli
mRNAs? J. Mol. Biol. 204:79-94.
Fargo, D.C., Zhang, M., Gillham, N.W., Boynton, J.E. 1998. Shine-Dalgarno-like sequences are not
required for translation of chloroplast mRNAs in Chlamydomonas reinhardtii chloroplasts or in
Escherichia coli. Mol. Gen. Genet. 257:271-282.
Favre, A. 1990. Photochemistry and the nucleic acids. In Bioorganic photochemistry (ed. H. Morrison),
pp. 379–425. John Wiley &Sons, New York.
Ferguson, G.P. 1999. Protective mechanisms against toxic electrophiles in Escherichia coli. Trends in
Microbiology. 7:242-247.
Fredrick, K., Noller, H.F. 2002. Accurate translocation of mRNA by the ribosome requires a peptidyl
group or its analog on the tRNA moving into the 30S P site. Mol. Cell. 5: 1125-1131.
Gaba, A., Wang, Z., Krishnamoorthy, T., Hinnebusch, A., Sachs, M. 2001. Physical evidence for
distinct mechanisms of translational control by upstream open reading frames. EMBO. 20: 64536463.
Gama-Castro, S., Salgado, H., Peralta-Gil, M., Santos-Zavaleta, A., Muniz-Rascado, L., Solano-Lira, H.
et al. 2011. RegulonDB version 7.0: transcription regulation of Escherichia coli K-12 integrated
within genetic sensory response unit (Gensor Units). Nuc. Acids Res. 39: D98-D105.
120
Gottesman, S. 2005. Micros for microbes: non-coding regulatory RNAs in bacteria. Trends Genet.
21:399-404.
Green, R., and Noller, H.F. 1997. Ribosomes and translation. Annu. Rev. Biochem. 66:679-716.
Grill, S., Gualerzi, C.O., Londei, P., Blasi, U. 2000. Selective stimulation of translation of leaderless
mRNA by initiation factor 2: evolutionary implications for translation. The EMBO Journal. 19:
4101-4110.
Grill, S., Moll, I., Hasenohrl, D., Gualerzi, C.O., Blasi, U. 2001. Modulation of ribosomal recruitment to
5’-terminal start codons by translation initiation factors IF2 and IF3. FEBS Letters 495:167-171.
Gu, W., Zhou, T., Wilke, C.O. 2010. A universal trend of reduced mRNA stability near the translationinitiation site in prokaryotes and eukaryotes. PLoS Comput Biol. 6:e1000664.
Gualerzi, C.O., Pon, C.L. 1990. Initiation of mRNA translation in prokaryotes. Biochemistry 29:58815889.
Gualerzi, C.O., Giuliodori, A. M., Pon, C.L. 2003. Transcriptional and post-transcriptional control of
cold-shock genes. J. Mol. Biol. 331:527-539.
Gualerzi, C.O., and Pon, C.L. 2015. Initiation of mRNA translation in bacteria: structural and dynamic
aspects. Cell Mol. Life Sci. 72: 4341-4367.
Hanahan, D. 1983. Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 4:557-580.
Hartz, D., McPheeters, D.S., Gold, L. 1989. Selection of the initiator tRNA by Escherichia coli
initiation factors. Genes Dev. 3: 1899-1912.
Hering, O., Brenneis, M., Beer, J., Suess, B., Soppa, J. 2009. A novel mechanism for translation
initiation operates in haloarchaea. Mol. Microbiol. 71: 1451-1463.
Hirose, T., Sugiura, M. 2004. Multiple elements required for translation of plastid atpB mRNA lacking
the Shine-Dalgarno sequence. Nuc. Acids Res. 32: 3503-3510.
Hobbs, E.C., Fontaine, F., Yin, X., Storz, G. 2011. An expanding universe of small proteins. Curr. Opin.
in Microbiol. 14: 167-173.
121
Hui, A., de Boer, H.A., 1987. Specialized ribosome system: Preferential translation of a single mRNA
species by a subpopulation of mutated ribosomes in Escherichia coli. Proc. Natl. Acad. Sci.
USA. 84:4762-4766.
Hutchison III, C.A., Chuang, R., Noskov, V.N, Assad-Garcia, N., Deerinck, T.J., Ellisman, M.H., Gill,
J., Kannan, K., Karas, B.J., Ma, L., Pelletier, J.F., Qi, Z., Richter, R.A., Strychalski, E.A., Sun,
L., Suzuki, Y., Tsvetanova, B., Wise, K.S., Smith, H.O., Glass, J.I., Merryman, C., Gibson, D.G.,
Venter, J.C. 2016. Design and synthesis of a minimal bacterial genome. Science. 351: DOI:
10.1126/science.aad6253.
Huttenhofer, A., Noller, H.F. 1994. Footprinting mRNA-ribosome complexes with chemical probes. The
EMBO Journal. 13:3892-3901.
Islam, S., Shaw, R., Frankel, G., Pallen, M.J., Busby, S.J.W. 2012. Translation of a minigene in the
5’leader sequence of the enterohaemorrhagic Escherichia coli LEE1 transcription unit affects
expression of neighboring downstream gene. Biochem J. 441: 247-253.
Janssen, G.R. Eubacterial, archaebacterial, and eukaryotic genes that encode leaderless mRNA. In
Industrial Microorganisms: Basic and Applied Molecular Genetics. 1993. Baltz, R.H., Hegeman,
G.D. and Skartrud, P.L. (eds). Washington, DC: American Society for Microbiology Press. Pp.
59-67.
Janssen, G.R., Bibb, M.J. 1993. Derivatives of pUC18 that have BglII sites flanking a modified multiple
cloning site and that retain the ability to identify recombinant clones by visual screening of
Escherichia coli colonies. Gene. 124: 133-134.
Jones, P.G. and Inouye, M. RbfA, a 30S ribosomal binding factor, is a cold-shock protein whose
absence triggers the cold-shock response. Mol. Microbiol. 1996; 21: 1207-1218.
Kaberdin, V.R., and Blasi, U. 2006. Translation initiation and the fate of bacterial mRNAs. FEMS
Microbiol. Rev. 30:967-979.
Kaczanowska, M., and Ryden-Aulin, M. 2007. Ribosome biogenesis and the translation process in
Escherichia coli. Microbiol. and Mol. Biol. Rev. 71:477-494.
122
Kanatani, A., Masuda, T., Shimoda, T., Misoka, F., Sheng Lin, X., Yoshimoto, T., Tsuru, D. 1991.
Protease II from Escherichia coli: Sequencing and expression of the enzyme gene and
characterization of the expressed enzyme. J. Biochem. 110: 315-320.
Kennell, D. and Riezman, H. 1977. Transcription and translation initiation frequencies of the
Escherichia coli lac operon. J. Mol. Biol. 114-1-21.
Klinkert B, Narberhaus F. 2009. Microbial thermosensors. Cell. Mol. Life Sci. 66:2661-2676.
Klock, G., and Hillen, W. 1986. Expression, purification and operator binding of the transposon
Tn1721-encoded Tet repressor. J. Mol. Biol. 189: 633-641.
Komarova, A.V., Tchufistova, L.S., Supina, E.V., Boni, I.V. 2002. Protein S1 counteracts the inhibitory
effect of the extended Shine-Dalgarno sequence on translation. RNA. 8:1137-1147.
Komarova, A.V., Tchufistova, L.S., Dreyfus, M., Boni, I.V. 2005. AU-rich sequences within 5’
untranslated leaders enhance translation and stabilize mRNA in Escherichia coli. J Bacteriol.
187:1344-1349.
Kozak M. 1999. Initiation of translation in prokaryotes and eukaryotes. Gene. 234: 187-208.
Krishnan, K.M., Van Etten, III W.J., Janssen, G.R. 2010. Proximity of the start codon to a leaderless
mRNA is a strong positive determinant of ribosome binding and expression in Escherichia coli. J
Bacteriol. 192: 6482-6485.
Lambert, T. 2012. Antibiotics that affect the ribosome. Rev. Sci. Tech. 31:57-64.
La Teana, A., Gualerzi, C.O., Brimacombe, R. 1995. From stand-by to decoding site. Adjustment of the
mRNA on the 30S ribosomal subunit under the influence of the initiation factors. RNA. 1:772782.
Laursen, B., Sorensen, H., Mortensen, K., Sperlink-Petersen, H. 2005. Initiation of protein synthesis in
bacteria. Micro. and Mol. Biol. Rev. 6:101-123.
Lovett, P.S. and Rogers, E.J. 1996. Ribosome regulation by the nascent peptide. Microbiol. Rev. 60:
366-385.
123
Malys, N., McCarthy, J.E.G. 2011. Translation initiation: variations in the mechanism can be
anticipated. Cell. Mol. Life Sci. 68: 991-1003.
Martin, J.F., Ullan, R.V., Garcia-Estrada, C. 2010. Regulation and compartmentalization of β-lactam
biosynthesis. Microbial Biotech. 3:285-299.
Martin-Farmer, J., and Janssen, G.R. 1999. A downstream CA repeat sequence increases translation
from leadered and unleadered mRNA in Escherichia coli. Mol. Micro. 31:1025-1038.
Marzi, S., Myansnikov, A.G., Sergavov, A., Ehresmann, C., Romby, P., Yuspupov, M., Klaholz, B.P.
2007. Structured mRNAs regulate translation initiation by binding to the platform of the
ribosome. Cell. 130:1019-1031.
McCarthy, J.E., Brimacombe, R. 1994. Prokaryotic translation: the interactive pathway leading to
initiation. Trend Genet. 10:402-407.
McCutcheon, J.P., Agrawal, R.K., Rhilips, S.M., Grassucci, R.A., Gerchman, S.E., Clemons, W.M., Jr.,
Ramakrishnan, V., and Frank, J. 1999. Location of the translation initiation factor IF3 on the
small ribosomal subunit. Proc. Natl. Acad. Sci. USA. 96:4301-4306.
McGinness, K.E., Sauer, R.T. 2004. Ribosomal protein S1 binds mRNA and tmRNA similarly but plays
distinct roles in translation of these molecules. Proc. Natl. Acad. Sci. 101:13454-13459.
Medenbach, J., Seiler, M., Hentze, M.W. 2011. Translational control via protein-regulated upstream
open reading frames. Cell. 145: 902-913.
Meijer, H., and Thomas, A.M. 2002. Control of eukaryotic protein synthesis by upstream open reading
frames in the 5’-untranslated region of an mRNA. Biochem J. 367: 1-11.
Melancon, P., Leclerc, D., Destroismasisons, N., Brakier-Gingras L. 1990. The anti Shine-Dalgarno
region in Escherichia coli 16S ribosomal RNA is not essential for the correct selection of
translational starts. Biochem. 29:3402-3407.
Miller, J. A short course in bacterial genetics. In A laboratory manual for Escherichia coli and related
bacteria. Vol 1, pp. 72-74. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 1992.
124
Min Jou, W., Haegeman, G., Ysebaert M., Fiers, W. 1972. Nucleotide sequence of gene coding for the
bacteriophage MS2 coat protein. Nature. 237:82-88.
Moll, I., Grill, S., Gualerzi, C.O., Blasi, U. 2002. Leaderless mRNAs in bacteria: surprises in ribosomal
recruitment and translational control. Mol. Microbiol. 43: 239-246.
Moll, I., Hirokawa, G., Kiel, M.C., Kaji, A., Blasi,U. 2004. Translation initiation with 70S ribosomes:
an alternative pathways for leaderless mRNAs. Nuc. Acids Res. 32:3354-3363.
Morris, D.R. and Geballe, A.P. 2000. Upstream open reading frames as regulators of mRNA translation.
Mol. and Cell Biol. 20: 8635-8642.
Munson, L.M., Stormo, G.D., Niece, R.L., Reznikoff, W.S. 1984. lacZ translation initiation mutations. J.
Mol. Biol. 177: 663-683.
Mutalik, V.K., Guimaraes, J.C., Cambray, G., Lam, C., Christoffersen, M.J., Mai, Q., Tran, A.B., Paull,
M., Keasling, J.D., Arkin, A.P., Endy, D. 2013. Precise and reliable gene expression via standard
transcription and translation initiation elements. Nature Methods. 10: 354-360.
Nakamoto, T. 2006. A unified view of initiation of protein synthesis. Biochem. and Biophys. Res.
Comm. 341: 675-678.
Nakamoto, T. 2011. Mechanisms of the initiation of protein synthesis: in reading frame binding of
ribosomes to mRNA. Mol. Biol. Rep. 38: 847-855.
Nakatogawa, H., Ito, K. 2002. The ribosomal exit tunnel functions as a discriminating gate. Cell 108:
629-636.
Narberhaus, F., Waldminghaus, T., Chowdhury, S. 2006. RNA thermometers. FEMS Microbiol Rev. 30:
3-16.
Nissen, P., Hansen, J., Ban, N., Moore, P.B., Steitz, T.A. 2000. The structural basis of ribosome activity
in peptide body synthesis. Science. 289:920-930.
O’Donnell, SM., and Janssen, GR. 2002. Leaderless mRNAs bind 70S ribosomes more strongly than
30S ribosomal subunits in Escherichia coli. J. Bacteriol. 184:6730-6733.
125
Ogle, J.M., Brodersen, D.E., Clemons, W.M., Tarry, M.J., Carter, A.P., Ramakrishnan V. 2001.
Recognition of cognate transfer RNA by the 30S ribosomal subunit. Science. 292:897-902.
Osterman, I.A., Evfratov, S.A., Sergiev, P.V., Donstova, O.A. 2013. Comparison of mRNA features
affecting translation initiation and reinitiation. Nuc. Acids Res. 41: 474-486.
Petersen, H.U., Joseph, E., Ullmann, A., Danchin, A. 1978. Formylation of initiator tRNA methionine in
prokaryotic protein synthesis: in vivo polarity in lactose operon expression. J. Bacteriol. 135:
453-459.
Perego, M. 1997. A peptide export-import control circuit modulating bacterial development regulates
protein phosphatases of the phosphorelay. Proc. Natl. Acad. Sci. USA. 94:8612-8617.
Ptashne, M., Backman, K., Humayun, M.Z., Jeffrey, A., Maurer, R., Meyer, B., et al. 1976.
Autoregulation and function of a repressor in bacteriophage lambda. Science. 194: 156-161.
Qing, G., Xia, B., Inouye M. 2003. Enhancement of translation initiation by A/T –rich sequences
downstream of the initiation codon in Escherichia coli. J. Mol. Microbiol. Biotechnol. 6:133144.
Ramakrishnan, V. 2002. Ribosome structure and the mechanism of translation. Cell. 108:557-572.
Rex, G., Surin, B., Besse, G., Schneppe, B., McCarthy, J.E.G. 1994. The mechanism of translational
coupling in Escherichia coli. J. Biol. Chem. 269: 18118-18127.
Ringquist, S., Shinedling, S., Barrick, D., Green, L., Binkley, J., Stormo, G.D., Gold, L. 1992.
Translation initiation in Escherichia coli: sequence within the ribosome-binding site. Mol.
Microbiol. 6:1219-1229.
Ringquist, S., MacDonald, M., Gibson, T., Gold, L. 1993. Nature of the ribosomal mRNA track:
Analysis of ribosome-binding sites containing different sequences and secondary structures.
Biochem. 32: 10254-10262.
Ruscetti, F.W. and Jacobson, L.A. 1972. Accumulation of 70S monoribosomes in Escherichia coli after
energy source shift-down. J. Bacteriol. 111: 142-151.
126
Rudd, K.E., Schneider, T.D. 1992. Complication of E.coli ribosome binding site, p 17.19-17.46. In J.
Miller (ed.), A short course in bacterial genetics. Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, NY.
Salavati R, Oliver D. 1995. Competition between ribosome and SecA binding promotes Escherichia coli
secA translational regulation. RNA. 1:745-753.
Salis, H.M., Mirsky, E.A., Voigt, C.A. 2009. Automated design of synthetic ribosome binding sites to
control protein expression. Nature Biotechnology. 27: 946-950.
Sarabhai, A., Brenner, G. 1967. A mutant which reinitiates the polypeptide chain after chain
termination. J. Mol. Biol. 27: 145-162.
Sato, T., Terabe, M., Watanabe, H., Gojobori, T., Hori-Takemoto, C., Mirua, K. 2001. Codon and base
biases after the initiation codon of the open reading frames in Escherichia coli genome and their
influence on the translational efficiency. .J Biochem. 129: 851-860.
Sauert, M., Wolfinger, M.T., Versper, O., Muller, C., Byrgazov, K., Moll, I. 2016. The MazF-regulon: a
toolbox for the post-transcriptional stress response in Escherichia coli. Nuc. Acids Res. 10:1-16.
Scharff, L.B., Childs, L., Walther, D., Bock, R. 2013. Local absence of secondary structure permits
translation of mRNAs that lack ribosome-binding sites. PLoS Genetics 7: 1-11.
Scherer, G.F.E., Walkinshaw, M.D., Arnott, S., Morr, D.J. 1980. The ribosome-binding sites recognized
by E.coli ribosomes have regions with signal character in both the leader and protein coding
segments. Nucleic Acids Res. 8: 3895-3907.
Schleif, R. 1972. Fine-structure deletion map of the Escherichia coli L-arabinose operon. Proc. Natl.
Acad. Sci. USA. 11: 3479-3484.
Schluenzen F., Toakemoto, C., Wilson, D.N., Kaminishi, T., Harms, J.M., Hanawa-Suetsugu, K.,
Szaflarski, W., Kawazoe, M., Shirouzu, M., Nierhaus, K.H. 2006. The antiobiotic kasugamycin
mimics mRNA nucleotides to destabilize tRNA binding and inhibit canonical translation
initiation. Nat. Struct. Mol. Biol. 13:871-878.
127
Schmeing, T.M., Moore, P.B., Steitz, T.A. 2003. Structures of deacylated tRNA mimics bound to the E
site of the large ribosomal subunit. RNA. 9:1345-1352.
Schmidt, B.F., Berkhout, B., Overbeek, G.P., van Strien, A., van Duin, J. 1987. Determination of the
RNA secondary structure that regulates lysis gene expression in the bacteriophage MS2. J Mol.
Biol. 195: 505-516.
Schneider, T.D., Stormo, G.D., Gold, L., Ehrenfeucht, A. 1986. Information content of binding sites on
nucleotide sequences. J. Mol. Biol. 188:415-431.
Schrader, J.M., Zhou, B., Li, G., Lasker, K., Childers, W.S., Williams, B., Long, T., Crosson, S.,
McAdams, H.H., Weissman, J.S., Shapiro, L. 2014. The coding and noncoding architecture of
the Caulobacter crescentus genome. PLOS Genetics. 10(7): e1004463.
doi:10.1371/journal.pgen.1004463.
Schuwirth, B.S., Borovinskaya, M.A., Hau, C.W., Zhang, W., Vila-Sanjurjo, A., Holton, J.M., Cate,
J.H.D. 2005. Structure of the 70S ribosome at 3.5 Å resolution. Science. 310:827-834.
Schuwirth, B.S., Day, J.M., Hau, C.W., Janssen, G.R., Dahlberg, A.E., Cate, J.H., Vila-Sanjurjo, A.
2006. Structural analysis of kasugamycin inhibition of translation. Nat. Struct. Mol. Biol.
13:879-886.
Seligmann, H. 2003. Cost-minimization of amino acid usage. J. Mol. Evol. 56: 151-161.
Sharma, P., Yan, F., Doronina, V.A., Escuin-Ordinas, H., Ryan, M.D., Brown, J.D. 2012. 2A peptides
provide distinct solutions to driving stop-carry on translation recoding. Nuc. Acids Res. 40:
3143-3151.
Shell, S.S., Wang, J., Lapierre, P., Mir, M., Chase, M.R., Pyle, M.M., Gawande, R., Ahmad, R.,
Sarracino, D.A., Ioerger, T.R., Fortune, S.M., Derbyshire, K.M., Wade, J.T., Gray, T.A. 2015.
Leaderless transcripts and small proteins are common features of the mycobacterial translational
landscape. PLoS Genet. 11(11): e1005641. doi:10.1371/journal.pgen.1005641
Shine, J., and Dalgarno, L. 1974. The 3’-terminal sequence of Escherichia coli 16S ribosomal RNA:
Complementarity to nonsense triplets and ribosome binding sites. Proc. Natl. Acad. Sci. USA.
71:1342-1346.
128
Simonetti, A., Marzi, S., Jenner, L., Mayansnikov, A., Romby, P., Yusupova, G., Klaholz, B.P.,
Yusupov, M. 2009. A structural view of translation initiation in bacteria. Cell Mol. Life Sci.
66:423-436.
Singer, P. and Nomura, M. 1985. Stability of ribosomal protein mRNA and translational feedback
regulation in Escherichia coli. Mol. Gen. Genet. 199:543-546.
Slupska, M.M., King, A.G., Fitz-Gibbon, S., Besemer, J., Borodovsky, M., Miller, J.H. 2001. Leaderless
transcripts of the crenarchaeal hyperthermophile Pyrobaculum aerophilum. J. Mol. Biol.
309:347-360.
Sørensen, M.A., Fricke, J., Pedersen, S. 1998. Ribosomal protein S1 is required for translation of most,
if not all, natural mRNAs in Escherichia coli in vivo. J. Mol. Biol. 280: 561-569.
Spanjaard, R.A., and van Duin, J. 1989. Translational reinitiation in the presence and absence of a Shine
and Dalgarno sequence. Nuc. Acids Res. 17: 5501-5507.
Storz, G. 1999. An RNA thermometer. Genes Dev. 13:633-636.
Studer, S.M., Joseph, S. 2006. Unfolding of mRNA secondary structure by the bacterial translation
initiation complex. Mol. Cell. 22:105-115.
Subramanian, A.R. 1983. Structure and functions of ribosomal protein S1. Prog. Nucleic Acid Res. Mol.
Biol. 28:101-142.
Suryanarayana, T., Subramanian, A.R. 1983. An essential function of ribosomal protein S1 in messenger
ribonucleic acid translation. Biochemistry. 22:2715-2719.
Takyar, S., Hickerson, R.P., Noller, H.F. 2005. mRNA helicase activity of the ribosome. Cell 120:49-58.
Tanner, D.R., Cariello, D.A., Woolstenhulme, C.J., Broadbent, M.A., Buskirk, A.R. 2009. Genetic
identification of nascent peptides that induce ribosome stalling. J. Biol. Chem. 284:34809-34818.
Tashiro, M., Klenk, H.D., Rott, R. 1987. Inhibitory effect of a protease inhibitor, leupeptin, on the
development of influenza pneumonia, mediated by concomitant bacteria. J. Gen. Virol. 68:20392041.
129
Tedin K., Resch A., Blasi U. 1997. Requirements for ribosomal protein S1 for translation initiation of
mRNAs with and without a 5’ leader sequence. Mol. Microbiol. 25: 189-199.
Tedin, K., Moll, I., Grill, S., Resch, A., Grashopf, A., Gualerzi, C.O., Blasi, U. 1999. Translation
initiation factor 3 antagonizes authentic start codon selection on leaderless mRNAs. Mol.
Microbiol. 31:67-77.
Tenson, T., Ehrenberg, M. 2002. Regulatory nascent peptides in the ribosomal tunnel. Cell. 108:591594.
Thieringer H.A., Jones, P.G., Inouye, M. 1998. Cold shock and adaptation. Bioessays. 20:49-57.
Tolstrup, N., Sensen, C.W., Garrett, R.A., Clausen, I.G. 2000. Two different and highly organized
mechanisms of translation initiation in the archaeon Sulfolobus solfataricus. Extremophiles.
4:175-179.
Tucker, B.J. and Breaker, R.R. 2005. Riboswitches as versatile gene control elements. Curr. Opin.
Struct. Biol. 15:342-348.
Tzareva, N.V., Makhno, V.I., Boni, I.V. 1994. Ribosome-messenger recognition in the absence of the
Shine-Dalgarno interactions. FEBS Letters 337:189-194.
Uchida, T., Abe, M., Matsuo, K., Yoneda, M. 1970. Amounts of free 70S ribosomes and ribosomal
subunits found in Escherichia coli at various temperatures. Biochem. Biophys. Res. Commun.
41: 1048-1054.
Udagawa, T., Shimizu, Y., Ueda, T. 2004. Evidence for the translation initiation of leaderless mRNAs
by the intact 70S ribosome without its dissociation into subunits in Eubacteria. J. Biol. Chem.
279:8539-8546.
Unoson, C., Wagner, E.G.H. 2007. Dealing the stable structures at ribosome binding sites: bacterial
translation and ribosome standby. RNA Biol. 4:113-117.
Van Etten, W.J. and Janssen, G.R. 1998. An AUG initiation codon, not codon-anticodon
complementarity, is required for the translation of unleadered mRNA in Escherichia coli. Mol.
Microbiol. 27:987-1001.
130
Vesper, O., Amitai, S., Belitsky, M., Byrgazov, K., Kaberdina, A.C., Engelberg-Kulka, H., Moll, I.
2011. Selective translation of leaderless mRNAs by specialized ribosomes generated by MazF in
Escherichia coli. Cell. 147: 147-157.
Wagner, E.G., Altuvia, S. Romby, P. 2002. Antisense RNAs in bacteria and their genetic elements. Adv.
Genet. 46:361-398.
Walleczek, J., Albrecht-Ehrlich, R., Stoffler, G., Stoffler-Meilicke, M. 1990. Three-dimensional
localization of the NH2- and carboxyl-terminal domain of ribosomal protein S1 on the surface of
the 30S subunit from Escherichia coli. J. Biol. Chem. 265: 11339-11344.
Wang, F., Xiao, J., Pan, L., Yang, M., Zhang, G., Jin, S., Yu, J. 2008. A systematic survey of miniproteins in bacteria and archaea. PLOS One. 3: e4027. doi: 10.1371/journal.pone.0004027.
Wilson, D.N., and Nierhaus, K.H. 2003. The ribosome through the looking glass. Angew. Chem. Int. Ed.
Engl. 42:3464-3486.
Wilson, D.N., Nierhaus, K.H. 2005. Ribosomal proteins in the spotlight. Crit. Rev. Biochem, Mol. Biol.
40: 243-267.
Wilson, DN. 2009. Less is more for leaderless mRNA translation. Mol. Cell. 33:141-142.
Wimberly, B., Broderson, D., Clemons, W., Morgan-Warren, R., Carter, A., Vonrhein, C., Hartsch, T.,
Ramakrishnan, V. 2000. Structure of the 30S ribosomal subunit. Nature. 407:327-339.
Yamamoto, H., Wittek, D., Gupta, R., Qin, B., Ueda, T., Krause, R., Yamamoto, K., Albrecht, R., Pech,
M., Nierhaus, K.H. 2016. 70S-scanning initiation is a novel and frequent initiation mode of
ribosomal translation in bacteria. Proc. Natl. Acad. Sci. USA. 113:E1180-1189.
Yan, F., Doronina, V.A., Sharma, P., Brown, J.D. 2010. Orchestrating ribosomal activity from the
inside: effects of the nascent chain on the peptidyltransferase centre. Biochem. Soc. Trans. 38:
1576-1580.
Yusupova, G., Yusupov, M., Cate, J., Noller H. 2001. The path of the messenger RNA through the
ribosome. Cell. 106:233-241.
131
Winkler, W.C. and Breaker, R.R. 2005. Regulation of bacterial gene expression by riboswitches. Annu.
Rev. Microbiol. 59:487-517.
Zheng, X., Hu, G., She, Z., Zhu, H. 2011. Leaderless genes in bacteria: clue to evolution of translation
initiation mechanisms in prokaryotes. BMC Genomics. 12: 361-374.
Zuker, M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nuc. Acids
Res. 31: 4306-3415.
132