1
1
Re-defining how mRNA degradation is coordinated with transcription and
2
translation in bacteria
3
4
5
Seunghyeon Kim
1
, Yu-Huan Wang
1
, Albur Hassan
1
, and Sangjin Kim
1,2*
6
7
1
Department of Physics, University of Illinois at Urbana-Champaign, Urbana, IL 61801,
8
USA
9
2
Center for Biophysics and Quantitative Biology, University of Illinois at Urbana–
10
Champaign, Urbana, IL 61801, USA
11
*Correspondence: [email protected]
12
13
14
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
2
Abstract
15
16
In eukaryotic cells, transcription, translation, and mRNA degradation occur in distinct
17
subcellular regions. How these mRNA processes are organized in bacteria, without
18
employing membrane-bound compartments, remains unclear. Here, we present
19
generalizable principles underlying coordination between these processes in bacteria. In
20
Escherichia coli, we found that co-transcriptional degradation is rare for mRNAs except
21
for those encoding inner membrane proteins, due to membrane localization of the main
22
ribonuclease, RNase E. We further found, by varying ribosome binding sequences, that
23
translation affects mRNA stability not because ribosomes protect mRNA from degradation,
24
but because low translation leads to premature transcription termination in the absence
25
of transcription-translation coupling. Extending our analyses to Bacillus subtilis and
26
Caulobacter crescentus, we established subcellular localization of RNase E (or its
27
homolog) and premature transcription termination in the absence of transcription-
28
translation coupling as key determinants that explain differences in transcriptional and
29
translational coupling to mRNA degradation across genes and species.
30
31
Keywords
32
mRNA degradation, RNase E, transcription-translation coupling, premature transcription
33
termination, transertion, co-transcriptional regulatcanion
34
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
3
Introduction
35
Unlike eukaryotic cells, bacterial cells do not have a nucleus, and the transfer of genetic
36
information from DNA to protein takes place within a common space, the cytoplasm,
37
permitting concurrence of translation and even mRNA degradation while mRNA is
38
transcribed
1,2
. How transcription, translation, and mRNA degradation are coordinated
39
during the life cycle of an mRNA, in the absence of membrane-bound compartments, is
40
a fundamental question that underpins gene expression regulation in bacterial cells.
41
Addressing this question will enable understanding of how protein expression levels are
42
regulated by the cell in response to different environments
3-5
and also inform the design
43
of synthetic gene expression systems, where precise manipulation of gene expression is
44
essential
6
.
45
Decades of research in a model bacterium, E. coli, has provided strong evidence that
46
transcription is coupled to translation
7-11
, that mRNA degradation can start during
47
transcription
12-15
, and that translation affects mRNA degradation
16,17
. However, whether
48
this picture is broadly applicable across all genes and different bacterial species remains
49
unclear. Identifying the molecular and sequence variables affecting coupling between
50
transcription, translation, and mRNA degradation will lead to a generalizable model for
51
understanding the regulation of bacterial gene expression. In this work, we investigated
52
when (during the life cycle of mRNA) and where (within a cell) mRNAs are degraded in
53
coordination with transcription and translation in bacterial cells and identified key factors
54
that contribute to commonalities and differences in co-transcriptional and post-
55
transcriptional control of gene expression in E. coli, B. subtilis, and C. crescentus.
56
The possibility of co-transcriptional mRNA degradation has been discussed since
57
early studies of long operons in E. coli. In lac and trp operons, mRNA sequences from
58
the promoter-proximal gene were shown to decay before the promoter distal genes were
59
transcribed
12-14
. A genome-wide measurement of mRNA lifetimes in E. coli compared
60
transcription elongation time and mRNA lifetime and suggested that many long genes
61
that exhibit transcription elongation times longer than mRNA lifetimes may experience co-
62
transcriptional mRNA degradation
15
. Co-transcriptional mRNA degradation can have a
63
significant impact by reducing the number of proteins made per transcript, which can be
64
beneficial when a quick stop in protein synthesis is needed to respond to changing cellular
65
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
4
needs. However, whether co-transcriptional degradation is indeed possible in E. coli
66
remains in question because the main ribonuclease controlling mRNA degradation,
67
RNase E, is found on the inner membrane of the cell, away from the nucleoid
18-21
. Co-
68
transcriptional mRNA degradation would therefore need to invoke the dynamic
69
relocalization of a gene locus to the membrane, which has been observed for certain
70
genes
22,23
. Furthermore, unlike in E. coli, RNase E is localized in the cytoplasm in C.
71
crescentus
24-26
, raising a question about how mRNA degradation is differentially
72
controlled in C. crescentus cells in comparison to E. coli cells.
73
In contrast to the lack of clarity how mRNA degradation can be coupled to transcription
74
in bacterial cells, several studies have supported the coupling of mRNA degradation to
75
translation, such that mRNAs with a strong ribosome binding sequence have long
76
lifetimes
16,17,27-31
. This trend has been explained by the notion that ribosomes protect
77
mRNA from degradation. However, what aspect of ribosome activity—for example,
78
whether it is the rate of loading at the 5’ end of the mRNA or whether it is ribosomal
79
density across the mRNA—is responsible for the protective role remains unclear
16,17,32
.
80
Understanding the exact mechanism of translation that affects mRNA lifetime would help
81
make quantitative predictions for expression output for different genes.
82
In this work, we used lacZ as a model gene to study how transcription, translation,
83
and mRNA degradation are coordinated in bacterial cells. The lac operon in E. coli is a
84
paradigm of bacterial gene regulation, and our current understanding of transcription-
85
translation coupling in bacteria and dependency between translation and mRNA stability
86
has been established by seminal studies that used lacZ as a model gene
9,28,33,34
. Its
87
regulatory mechanisms are well characterized, allowing us to manipulate parameters for
88
lacZ gene expression and test hypotheses toward a generalizable model. For example,
89
we introduced the effect of transertion to lacZ to emulate what happens to genes encoding
90
inner membrane proteins
22
, we varied the 5’ untranslated region (5’-UTR) sequence of
91
lacZ to test variable translation efficiencies across the genome
3,35
, and we perturbed the
92
subcellular localization of RNase E to capture differences across bacterial species
36,37
.
93
From this approach, we identified spatial and genetic design principles that bacteria have
94
evolved to differentially regulate transcriptional and translational coupling to mRNA
95
degradation across various genes and species.
96
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
5
97
Results
98
lacZ mRNA is degraded post-transcriptionally, uncoupled from transcription
99
While the possibility of nascent mRNA degradation during transcription has been
100
discussed
12-15,27,38,39
, the actual rate of co-transcriptional mRNA degradation has never
101
been reported. We used lacZ gene under the lac promoter in E. coli as a model to
102
measure the rates of co-transcriptional and post-transcriptional mRNA degradation (k
d1
103
and k
d2
, respectively; Fig. 1A). Earlier studies discussed co-transcriptional degradation
104
of lacZ based on the observation that lacZ decays before the synthesis of lacY and lacA
105
in the original lac operon. However, this result can be explained by co-transcriptional
106
mRNA processing at the intergenic region between lacZ and lacY
40-42
, instead of real co-
107
transcriptional degradation of lacZ. Therefore, we deleted lacY and lacA genes from the
108
original lac operon in the chromosome of wild-type, MG1655 to yield a monocistronic lacZ
109
(strain SK98). The intrinsic terminator sequence after lacA follows the coding sequence
110
of lacZ to ensure the dissociation of RNA polymerase (RNAP) from DNA after finishing
111
the transcription of lacZ (Fig. 1B). To follow the degradation kinetics of lacZ mRNA after
112
the stoppage of transcription initiation, transcription of lacZ was induced with membrane-
113
permeable inducer isopropyl b-D-1-thiogalactopyranoside (IPTG) and re-repressed with
114
glucose 75 seconds (s) after addition of IPTG (Fig. 1B). Importantly, glucose was added
115
before the first RNAPs finished transcription, so that co-transcriptional mRNA degradation
116
can be observed. During this time course, a population of cells was acquired every 20-30
117
s, from which 5’ and 3’ lacZ mRNA levels were quantified by quantitative real-time PCR
118
(qRT-PCR) using probes denoted as Z5 and Z3, respectively. A key feature of this time
119
course experiment is the temporal separation of lacZ mRNA status between nascent and
120
released (Fig. 1B). Until the first RNAPs finish transcription of lacZ (T
3’
), all lacZ mRNAs
121
are expected to be nascent (time window i). After the last RNAPs finish transcription of
122
lacZ at t
3’
, all lacZ mRNAs are released (time window iii). In between, nascent and
123
released lacZ mRNAs co-exist (time window ii). Hence, we can measure the rates of co-
124
transcriptional and post-transcriptional mRNA degradation by fitting Z5 level changes with
125
an exponential decay function in the time windows i and iii, respectively.
126
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
6
If the co-transcriptional degradation of lacZ mRNA takes place (k
d1
>0), Z5 level will
127
decrease in the time window i, as demonstrated by a mathematical model (Fig. 1C and
128
S1A). However, our data shows that Z5 level stays constant in the time window i,
129
suggesting that k
d1
is close to zero (Fig. 1D). From biological replicates, we obtained k
d1
130
= 0.042 ± 0.0598 min
-1
and k
d2
= 0.43 ± 0.067 min
-1
(Fig. S1B). Essentially, the mean
131
lifetime of nascent lacZ mRNA (1/k
d1
) is 24 min, much longer than transcription elongation
132
time (~3.5 min), suggesting that lacZ mRNA is unlikely to experience degradation during
133
transcription elongation.
134
135
Membrane localization of RNase E accounts for uncoupling of transcription and
136
degradation of lacZ mRNA
137
Among various ribonucleases in E. coli, the endoribonuclease RNase E has been
138
considered the main enzyme to initiate mRNA degradation
21,36,43-45
, including lacZ
139
mRNA
27,46,47
. To confirm that the observed k
d1
and k
d2
of lacZ mRNA are controlled by
140
RNase E, we repeated the experiment in a strain carrying a temperature-sensitive RNase
141
E allele (rne3071, strain SK519), in which RNase E can be inactivated by a 10-min shift
142
to 43.5°C
48
. We performed IPTG induction of lacZ transcription at 43.5°C after 10 min of
143
the temperature shift. Because transcription elongation is faster at this high temperature
144
(T
3’
= 100 s), glucose was added at 50 s after IPTG induction, so that we can still capture
145
the time window i to measure k
d1
. When RNase E was inactivated, k
d1
and k
d2
were about
146
7 times smaller than those measured in wild-type RNase E at 43.5°C and about 2-3 times
147
smaller than those measured at 30°C (Fig. S2A), confirming that RNase E controls k
d1
148
and k
d2
of lacZ mRNA. In E. coli cells, RNase E is associated with the inner membrane
149
via the membrane targeting sequence (MTS)
20
. Therefore, the lack of co-transcriptional
150
degradation of lacZ mRNA (very low k
d1
) could be accounted for by the membrane
151
localization of RNase E, away from the nucleoid (or transcription site).
152
E. coli cells are viable even when the MTS sequence of RNase E is removed and
153
RNase E is localized to the cytoplasm
20,49
(RNase E ΔMTS, Fig. 2A). When RNase E is
154
in the cytoplasm, instead of anchored to the membrane, it can interact with nascent and
155
released mRNAs more frequently and likely affect k
d1
and k
d2
of lacZ mRNA. To check
156
this possibility, we measured k
d1
and k
d2
of lacZ mRNA in the strain expressing RNase E
157
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
7
ΔMTS. We found that cytoplasmic RNase E increases both k
d1
and k
d2
; especially, k
d1
158
increases about 7 fold, to 0.31 ± 0.084 min
-1
, in comparison to the wild-type RNase E
159
strain (Fig. 2B).
160
We tested another cytoplasmic RNase E mutant, RNase E (1-529) (Fig. 2A). This
161
mutant lacks the MTS as well as the C-terminal domain, which provides binding sites for
162
other RNA degradosome components
50
(Fig. S2B). To interpret the effect of RNase E
163
localization in the absence of the C-terminal domain, we compared k
d1
and k
d2
of lacZ
164
mRNA from cells expressing the RNase E (1-529) mutant with those from cells expressing
165
a RNase E (1-592) mutant, which also lacks the C-terminal domain but is localized to the
166
membrane via MTS. We found that k
d1
and k
d2
of lacZ mRNA are higher in the cytoplasmic
167
RNase E (1-529) than in the membrane-bound RNase E (1-592) (Fig. 2B), supporting
168
that mRNA degradation is faster when RNase E is localized in the cytoplasm. We note
169
that the absence of C-terminal domain in RNase E (1-592) results in lower k
d1
and k
d2
of
170
lacZ mRNA in comparison to those in the wild-type RNase E, suggesting the importance
171
of having the C-terminal domain for the catalytic activity of RNase E happening at the N-
172
terminal domain
51,52
(See Fig. S2C for additional data). Altogether, our results show that
173
the membrane localization of RNase E slows down the degradation of lacZ mRNA,
174
especially during transcription, giving rise to the uncoupling of transcription and mRNA
175
degradation.
176
177
Proximity of nascent mRNAs to the membrane alone does not affect their
178
degradation rates
179
Since slow co-transcriptional mRNA degradation is likely due to the spatial separation
180
between membrane-localized RNase E and nascent mRNAs, we considered a scenario
181
where nascent mRNAs are positioned close to the membrane. When mRNAs coding for
182
a transmembrane protein are transcribed, co-transcriptional translation may be
183
accompanied by membrane insertion of the nascent protein, a process known as
184
transertion
53-55
. A previous study showed that expression of lacY (encoding the lactose
185
permease localized in the inner membrane) brings the lacY locus and nearby DNA region
186
(~90 kb) close to the membrane
23
. This suggests that even a gene encoding a
187
cytoplasmic protein (such as lacZ) can be localized close to the membrane if it is adjacent
188
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
8
to an actively transcribed lacY gene locus on the chromosome. Therefore, we inserted a
189
constitutively expressed lacY gene downstream of lacZ (strain SK435; Fig. 3A) to test if
190
the transertion of lacY can bring lacZ closer to the inner membrane and increase k
d1
. As
191
a control, we made a strain where lacY is replaced with aadA, a gene encoding a
192
cytoplasmic protein, spectinomycin adenylyltransferase, that does not undergo
193
transertion (strain SK390).
194
To test the effect of transertion on the localization of nascent lacZ mRNA, we
195
performed fluorescence in situ hybridization (FISH) using Cy3B-labeled probes binding
196
to the first 1-kilobase region of lacZ mRNA (Z5
FISH
; Fig. 3A)
56
. Transcription of lacZ was
197
induced with IPTG and re-repressed with glucose as in the qRT-PCR experiment (Fig.
198
1B). Cells were sampled every 1 min interval and fixed immediately. The Z5
FISH
signal
199
appeared as diffraction-limited foci (Fig. 3A and S3A). Their centroid coordinates along
200
the short and long axes of the cell were normalized to cell width and length, respectively,
201
and combined into a 2D histogram (Fig. 3B-3C). Notably, until T
3’
(or the end of the time
202
window i, t = 210 s), most of the Z5
FISH
signals are expected to be from nascent mRNAs
203
tethered to gene loci (Fig. 1B). Hence, the location of Z5
FISH
at t = 1, 2, and 3 min after
204
induction allows us to examine the subcellular localization of the nascent mRNAs (and
205
their gene loci) exclusively. We observed that already at t = 1 min, Z5
FISH
in SK435 (lacZ
206
followed by constitutively transcribed lacY) were localized off the center long axis, while
207
those in SK390 (lacZ followed by constitutively transcribed aadA) were close to the center
208
long axis of the cell (Fig. 3B-3C). As time progresses to t = 2 and 3 min, Z5
FISH
in both
209
strains localized away from the center long axis, likely due to the lacZ gene locus moving
210
to the periphery of the nucleoid upon induction—an effect previously observed in the lacZ
211
locus in E. coli
57
. In all three time points, Z5
FISH
in SK435 were localized closer to the
212
membrane than those in SK390 (Fig. 3B-3C), suggesting that the transertion of lacY,
213
which does not occur with aadA, results in the neighboring lacZ gene transcription taking
214
place close to the inner membrane.
215
Next, we measured k
d1
and k
d2
of lacZ mRNA in SK435 and SK390 by qRT-PCR to
216
check if the proximity to the membrane allows the nascent and released lacZ mRNAs to
217
be degraded faster. We found that k
d1
and k
d2
of lacZ mRNAs were invariable between
218
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
9
the two strains and almost identical to the original lacZ-only strain without lacY or aadA
219
(Fig. 3D).
220
We wondered if nascent lacZ mRNAs in SK435 were not close enough to the
221
membrane to facilitate their degradation. To bring nascent lacZ mRNAs even closer to
222
the membrane, we constructed a translational fusion of lacZ with the first two
223
transmembrane segments of lacY (lacY2), so that lacZ is directly linked to the transertion
224
element (Fig. 3E). We also fused the venus gene at the 3’ end of lacZ sequence to verify
225
the membrane localization of LacZ proteins by fluorescence imaging (strain SK575; Fig.
226
S3B). FISH imaging of 5’ lacZ mRNA expressed from the lacY2-lacZ-venus fusion at t =
227
1, 2, 3 min after induction showed that as soon as two transmembrane segments (lacY2
228
of mRNA length 222 nt) are transcribed, the lacZ sequence is strongly enriched near the
229
membrane in comparison to the original lacZ strain without the lacY2 fusion (Fig. 3F-3G
230
and more information in Fig. S3C). This result suggests that transertion takes place
231
immediately after induction and brings nascent transcripts to the membrane. Also, the
232
direct translational fusion of lacY2 element to lacZ placed the nascent lacZ mRNAs closer
233
to the membrane in comparison to the previous lacZ-lacY context where the transertion
234
effect came from the neighboring lacY gene locus (strain SK435; Fig. 3B).
235
While nascent lacZ mRNAs were closer to the membrane, their k
d1
was not larger than
236
that of the original lacZ-only strain (strain SK98; Fig. 3H). This result further supports the
237
notion that proximity to the inner membrane (where RNase E is localized) is not sufficient
238
to increase the rate of co-transcriptional lacZ mRNA degradation.
239
In the lacY2-lacZ-venus fusion strain, we can also measure the degradation kinetics
240
of the lacY2 region of the transcripts using a set of qRT-PCR primers amplifying that
241
region. Remarkably, we found that lacY2 exhibits fast co-transcriptional mRNA
242
degradation with k
d1
= 0.34 ± 0.041 min
-1
(Fig. 3I and S3D). This likely represents the
243
characteristics of the original lacY transcript, as we measured a similar rate of co-
244
transcriptional mRNA degradation from the full-length lacY gene (Fig. S3E). This finding
245
indicates that coupling between transcription and mRNA degradation is possible for
246
transcripts encoding inner-membrane proteins. Considering that the proximity of nascent
247
mRNAs to the membrane alone does not affect the co-transcriptional degradation rate of
248
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
10
lacZ mRNA, the fast co-transcriptional degradation observed with lacY mRNA is likely
249
due to additional factors other than its proximity to the membrane (see Discussion).
250
251
Another cytoplasmic protein-coding gene, araB, exhibits high k
d1
due to its RBS
252
sequence
253
To determine if the result we observed for lacZ is generalizable to other genes in E. coli,
254
we examined k
d1
and k
d2
of another gene encoding a cytoplasmic protein, araB, which is
255
under the control of the arabinose-inducible promoter (P
ara
) of the araBAD operon on the
256
chromosome of E. coli strain MG1655 (Fig. 4A). We deleted araA and araD genes to
257
make araB a monocistronic gene (P
ara
-araB; strain SK472). Transcription from the P
ara
258
promoter was induced with arabinose and re-repressed by glucose 50 s afterward. In
259
contrast to very slow co-transcriptional degradation observed in lacZ mRNA, 5’ araB
260
mRNAs were degraded before 3’ araB mRNAs were transcribed (time window i indicated
261
as a blue box in Fig. 4B), resulting in k
d1
= 0.55 ± 0.134 min
-1
(Fig. 4C). This is quite
262
striking because in E. coli, co-transcriptional degradation does not seem to occur for
263
genes encoding cytoplasmic proteins due to the membrane localization of RNase E (Fig.
264
2B).
265
We found that this high k
d1
is not due to any aspects of the araB sequence, because
266
replacing araB’s coding region with that of lacZ (P
ara
-lacZ; SK477) resulted in similarly
267
high k
d1
of 0.56 ± 0.146 min
-1
of lacZ mRNA, in contrast to the low k
d1
observed at the
268
native lac locus (SK98; Fig. 4C). The high k
d1
of lacZ mRNA produced from P
ara
was not
269
due to the chromosomal position either, because bringing the lacI-lacZ region from the
270
native lac locus to the ara locus (P
lac
-lacZ; SK499) did not change the original (low) k
d1
of
271
lacZ mRNA (SK98; Fig. 4C). The high k
d1
likely originates from what P
ara
-araB (SK472)
272
and P
ara
-lacZ (SK477) have in common: the sequence in 5’-UTR (Fig. 4A), which is
273
different from 5’-UTR of native lacZ (SK98 and SK499). Therefore, the high k
d1
of P
ara
274
may originate from a certain feature of the 5’-UTR sequence.
275
5’-UTR of an mRNA contains ribosome binding site (RBS), including Shine-Dalgarno
276
(SD) sequence, which governs translation initiation and protein expression level
58-60
.
277
Henceforth, we refer to the SD sequence and its surrounding sequence as the RBS. RBS
278
sequences are known to affect the energetics of ribosome binding and translation
279
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
11
initiation, such that one can quantitatively predict the RBS strength, or protein expression
280
outcome from the sequence
61-63
. However, weakening RBS strength by changing its
281
sequence has also been known to destabilize the mRNA
27-31
, thus reducing the overall
282
protein expression by reducing both translation initiation rate and mRNA lifetime.
283
Conforming to this expectation, lacZ transcripts with the native RBS of araB (P
ara
-lacZ
284
in SK477) produced 30-fold lower LacZ protein expression than P
lac
-lacZ at the same
285
chromosome location (Fig. 4D). This result came from measuring LacZ protein
286
expression by Miller assay. To corroborate this finding, we replaced the RBS sequence
287
in P
ara
-lacZ (SK477) with a strong RBS sequence designed using an RBS calculator
63
288
(SK613 in Fig. 4A). The synthetic RBS sequence yielded increased LacZ protein
289
expression, higher than that from native lacZ RBS (SK499; Fig. 4D). Also, lacZ mRNA
290
with this strong synthetic RBS sequence exhibited low k
d1
as observed in the native lacZ
291
RBS (SK98 or SK499; Fig. 4C). These results support that the hypothesis that the weak
292
RBS sequence in P
ara
-araB is responsible for the high k
d1
of araB mRNA.
293
294
RBS strength affects lacZ mRNA localization due to premature transcription
295
termination
296
We have shown that P
lac
and P
ara
, two inducible promoters widely used in gene
297
expression studies
64-66
, have vastly different RBS strengths. Indeed, RBS sequences and
298
their expected strengths vary widely among genes in the E. coli genome
3,35
. While
299
mRNAs with a weaker RBS are expected to have shorter lifetime
16,17
, how RBS
300
sequences affect co-transcriptional and post-transcriptional mRNA degradation rates has
301
not been studied. To address this question, we compared the original lacZ strain (with the
302
native RBS; SK98) with a weak RBS mutant, which was created by changing five bases
303
in the original SD sequence (Fig. 5A and S4A for LacZ protein expression). We found
304
that mutating the RBS sequence increases k
d1
by 15 fold to 0.65 ± 0.171 min
-1
without
305
affecting k
d2
(Fig. 5B). We confirmed that the high k
d1
is largely controlled by RNase E
306
because the temperature-sensitive RNase E allele (rne3071) showed much lower k
d1
for
307
this weak RBS at the non-permissive temperature in comparison to the wild-type RNase
308
E at the same temperature (Fig. S4B-S4C). This brings us to the next question: How does
309
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
12
membrane-bound RNase E carry out co-transcriptional degradation of mRNAs with a
310
weak RBS?
311
To test the possibility that nascent mRNAs are localized differently depending on the
312
RBS strength, we visualized 5’ lacZ mRNAs by FISH (Z5
FISH
). We reasoned that nascent
313
mRNAs with a weak RBS sequence would be difficult to detect by FISH because they are
314
quickly degraded (Fig. 5B). Therefore, we performed FISH in strains harboring the
315
rne3071 allele to inactivate RNase E (strain SK519 and SK591 for the native and weak
316
RBS sequences, respectively). At the non-permissive temperature, lacZ expression was
317
induced with IPTG and re-repressed with glucose at 50 s after the induction. qRT-PCR
318
analysis of RNA samples from this time-course experiment showed that Z3, probing the
319
3’ end of the mRNA, appears above the basal level at t = 100 s after induction (Fig. S4C),
320
indicating that before t = 100 s, all 5’ lacZ mRNAs would be nascent and visualized as
321
diffraction-limited foci originating from the gene loci that they are tethered to. Surprisingly,
322
the 2D histogram of the relative positions of Z5
FISH
in this time window showed different
323
mRNA localization patterns between native RBS and weak RBS strains (Fig. 5C-5D).
324
While Z5
FISH
signals from the native RBS were localized at a specific location with a high
325
probability (red bins in the histogram) as seen earlier in WT RNase E (Fig. 3G), Z5
FISH
326
signals from the weak RBS were localized at random places throughout the cytoplasm,
327
such that the dense region (red color) did not show up in the histogram (Fig. 5D).
328
Additionally, before t = 100 s, the weak RBS strain contained a higher number of Z5
FISH
329
spots per cell that have weaker fluorescence intensity in comparison to those in the native
330
RBS strain (Fig. S4D-S4E). For example, at t = 60 s, there are up to two lacZ gene loci
331
per cell (Fig. S4F), but the weak RBS strain had 16% of cells with 3 or more Z5
FISH
spots
332
per cell, in contrast to 4% observed in the native RBS strain (Fig. 5E). These results are
333
consistent with a scenario, in which 5’ lacZ mRNAs with the weak RBS become physically
334
separated from gene loci even when all of them are expected to be tethered to the gene
335
loci and form only one or two diffraction-limited fluorescence spots per cell (Fig. S4F).
336
The spatial dispersion of mRNAs with the weak RBS in the time window i is
337
reminiscent of premature transcription termination, previously shown to follow
338
transcription-translation uncoupling due to nonsense mutation, antibiotic treatment, and
339
amino acid starvation
33,67-69
. To check the possibility of premature RNAP termination in
340
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
13
our weak RBS construct, we examined Z5 and Z3 levels at steady state after induction.
341
In the native RBS strain (SK98), Z5 and Z3 levels were equal at the steady state (Fig.
342
5F). Considering that Z5 and Z3 have equal lifetimes (Fig. S1B), the equal steady state
343
level means that 100% of RNAPs that passed the Z5 probe region reached the Z3 probe
344
region at the end of the gene, i.e., 0% premature transcription termination. However, in
345
the weak RBS lacZ strain (SK421), we observed that the steady state level of Z3 is about
346
half of that of Z5, indicating significant premature transcription termination (Fig. 5G). We
347
confirmed that this premature transcription termination is controlled by the rho factor
348
because treatment with bicyclomycin (BCM) rescued the Z5 and Z3 difference, bringing
349
the steady-state Z5 and Z3 levels to equal in the weak RBS strain (Fig. 5H).
350
These results are consistent with the hypothesis that transcription-translation coupling
351
requires a strong RBS, which allows the loading of the first ribosome to the RBS as soon
352
as the RBS sequence is transcribed by an RNAP. In the case of a weak RBS, in which
353
the first ribosome loading event is delayed, an RNAP might not be coupled with a
354
ribosome and experience premature termination by the rho factor
68,70
(Fig. 5I). Then, the
355
prematurely released (short) transcripts may diffuse to the membrane and get degraded
356
by RNase E on the inner membrane.
357
358
Translation affects mRNA degradation via premature transcription termination, not
359
via ribosome protection
360
A notable lesson from the weak RBS strain is that there are prematurely released
361
transcripts in the time window i, in which we measured k
d1
assuming all transcripts are
362
nascent. Hence, the high k
d1
observed in weak RBS strains (including the ones observed
363
with araB’s RBS in Fig. 4C) likely includes the degradation of prematurely released
364
mRNAs and is not a true rate of co-transcriptional mRNA degradation. To address this
365
problem, we modeled k
d1
as a weighted average of real co-transcriptional degradation
366
rate of nascent mRNAs (k
d1*
) and post-release degradation rate of prematurely terminated
367
mRNAs (k
dPT
):
368
(1)
where PT is the probability of premature termination during transcription.
369
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
14
If premature transcription termination leads to high k
d1
in a weak RBS strain, we expect
370
to see a good correlation between k
d1
and PT. To test this prediction, we used nine strains
371
harboring lacZ with varying RBS sequences at the ara and lac loci (see Table S4). We
372
measured k
d1
and k
d2
of lacZ mRNAs from re-repression (with glucose addition; Table S5)
373
and calculated PT from steady-state levels of Z5 and Z3 after induction (without glucose
374
addition). Because the ratio between stead-state levels of Z5 and Z3 (e.g. Fig. 5F-5H) is
375
related to PT as well as k
dPT
(Fig. S5A), we performed iterative fitting of k
d1
and estimated
376
PT using equation (1) and obtained best PT value for each strain and k
dPT
common among
377
nine strains (Fig. 5J; see method details). The optimal fitting of equation (1) gave k
d1*
of
378
0.025 ± 0.0372 min
-1
and k
dPT
of 0.80 ± 0.0587 min
-1
(R
2
= 0.93). We note that k
d1*
value
379
is very similar to k
d1
of strong RBS cases (where premature termination is 0%; e.g. SK98),
380
and k
dPT
value is larger than most of k
d2
we have observed for transcripts released after
381
transcription is completed. Possibly, prematurely released transcripts are degraded faster
382
because they diffuse faster than longer, full-length mRNAs and/or because they lack
383
certain features at the 3’ end that full-length mRNAs have, such as a stem-loop structure,
384
making them more easily degraded, by 3’-to-5’ exonuclease, PNPase.
385
One of the models explaining the RBS effect on mRNA lifetime is based on the notion
386
that ribosomes protect mRNA from the attack of RNase E
16,17
. According to this model,
387
transcripts with a weak RBS sequence, or those showing high probability of premature
388
transcription termination (PT), would undergo fast degradation because there are fewer
389
ribosomes on the mRNAs. To test this model across different RBS sequences, we
390
examined k
d2
, the decay rate of Z5 after t
3’
(last RNAP passes the end of lacZ gene) in
391
time window iii. k
d2
is largely determined by the degradation rate of full-length transcripts
392
that have the 3’ sequence and not affected by prematurely released transcripts, which
393
are degraded rather quickly (k
dPT
) and minimally contribute to the Z5 signal in this time
394
window. Nine strains of varying RBS sequences showed that k
d2
is independent of PT
395
(Fig. 5K) with very little correlation (P = -0.078). This result is in contrast to what would
396
be expected from the ribosome protection model, which would expect higher k
d2
in
397
transcripts with weaker RBS, or higher probability of premature transcription termination,
398
because the transcripts carry fewer ribosomes on average. Therefore, our results suggest
399
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
15
that translation affects mRNA lifetime mainly by affecting the percentage of prematurely
400
released transcripts (Fig. 5J).
401
402
Premature transcription termination and subcellular localization of RNase E (or its
403
homolog) affect the degradation of lacZ mRNA in other bacterial species
404
Our results so far imply that in E. coli, the fate of mRNA is determined by the RBS
405
sequence because of its effect on transcription-translation coupling. Next, we examined
406
if this conclusion can be generalized to other bacterial species. For example, in B. subtilis,
407
RNAP was shown to translocate faster than the ribosome during expression of lacZ,
408
preventing the ribosome from coupling to RNAPs
34,71
.We tested if the transcription-
409
translation uncoupling in B. subtilis results in premature transcription termination and
410
potentially a high k
d1
. First, we repeated the experiment done by previous papers
411
measuring the transcription and translation times of lacZ by qRT-PCR and Miller assay in
412
B. subtilis (strain GLB503; Fig. 6A), respectively. The transcription time was acquired
413
from the initial increase of Z3 signal from the baseline after induction with IPTG, indicating
414
the moment first RNAPs reach the end of the gene. The translation time was acquired
415
from the initial increase of LacZ protein levels from the baseline after induction, indicating
416
the moment first ribosomes reach the end of lacZ mRNA. In a slow growth condition
34
,
417
we observed that the translation time was 2.6 ± 0.054 min, much longer than the
418
transcription time of 1.3 ± 0.56 min (Fig. S6A-S6C). The steady-state levels of Z5 and Z3
419
were similar (Fig. 6B), implying that premature transcription termination does not take
420
place even if transcription and translation are uncoupled in B. subtilis.
421
Next, we measured k
d1
and k
d2
of lacZ mRNAs by re-repressing transcription by
422
adding rifampicin, a drug that stops transcription initiation
72-74
, at t = 30 s after induction
423
(Fig. 6C). We obtained k
d1
of 0.025 ± 0.0036 min
-1
and k
d2
of 0.14 ± 0.026 min
-1
(Fig.
424
S6D). Since premature transcription termination is not observed (Fig. 6B), k
d1
can be
425
attributed to co-transcription degradation, and the lifetime of nascent mRNA can be
426
estimated as 40 min (1/k
d1
), much longer than the transcription time of 1.3 min. Hence,
427
co-transcriptional degradation of lacZ mRNAs is likely very rare in B. subtilis, like in E.
428
coli. The lack of co-transcriptional degradation can be explained by the membrane
429
localization of the main endoribonuclease performing mRNA degradation, RNase Y and
430
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
16
RNase E, in B. subtilis and E. coli, respectively
18,75,76
. We note that k
d2
in B. subtilis is low
431
relative to E. coli (see Fig. 5K), and lacZ mRNA levels do not return to the basal level
432
within 10 min (Fig. 6C). This is quite surprising because the amount of LacZ proteins
433
expressed from the 30-s induction was minimal according to the Miller assay using a
434
sensitive fluorogenic LacZ substrate (Fig. S6E), suggesting that the remaining lacZ
435
mRNAs do not support protein synthesis. Possibly, translation initiation is slow in this
436
strain, and/or functional inactivation of mRNAs precedes the chemical degradation of
437
mRNAs in B. subtilis.
438
In contrast to E. coli and B. subtilis, C. crescentus is known to have cytoplasmic RNase
439
E
24-26
. Hence, we investigated the possibility of co-transcriptional degradation of lacZ
440
mRNA in C. crescentus using a strain where lacZ is placed under the xylose-inducible
441
promoter in the chromosome (strain LS2370
77
; Fig. 6D). First, we measured the
442
transcription and translation times of lacZ by qRT-PCR and by Miller assay after induction
443
with xylose. The translation time was 2.5 ± 0.21 min (Fig. 6E and S6F), similar to the
444
transcription time of 2.3 ± 0.27 min (Fig. 6F), suggesting that transcription and translation
445
are coupled. To measure k
d1
and k
d2
of lacZ mRNAs, transcription was re-repressed with
446
rifampicin at t = 50 s after addition of xylose. qRT-PCR data show that Z5 decays very
447
quickly after rifampicin addition (Fig. 6G). Strikingly, Z3 does not increase above the basal
448
level, such that the time windows i and iii cannot be defined for fitting Z5 for k
d1
and k
d2
.
449
If we take T
3’
of 2.3 min from the induction-only experiment (Fig. 6F) to estimate the time
450
window i (blue box in Fig. 6G), we obtain k
d1
of 1.3 ± 0.13 min
-1
. The absence of 3’ lacZ
451
mRNA signals in the re-repression experiment (also minimal LacZ protein expression;
452
Fig. S6G) indicates significant premature transcription termination, which contributes to
453
high k
d1
.
454
Indeed, when we blocked the rho factor activity with BCM, the steady-state levels of
455
Z5 and Z3 increased, suggesting that rho-dependent premature termination affected lacZ
456
mRNA levels in non-treated cells (Fig. S6H vs. 6F). Repeating k
d1
measurement in BCM-
457
treated cells allowed us to obtain the true rate of co-transcriptional degradation, k
d1*
of
458
0.71 ± 0.093 min
-1
(Fig. S6I). Through mathematical modeling, we estimated that
459
prematurely terminated mRNAs are degraded at k
dPT
of 3.4 ± 0.61 min
-1
and the
460
probability of premature transcription termination (PT) to be 69 ± 4.4% (see method
461
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
17
details). The fast mRNA degradation likely originates from the cytoplasmic distribution of
462
RNase E in C. crescentus cells. Interestingly, RNase E in C. crescentus has been shown
463
to interact with the rho factor
78
. We speculate that the cooperation between the rho factor
464
and RNase E results in the high k
d1
and high probability of premature transcription
465
termination we observed in C. crescentus (Fig. 6G).
466
The high probability of premature transcription termination (~69%) agrees with the
467
absence of 3’ lacZ mRNA signal when lacZ transcription was induced only for 50 s (Fig.
468
6G). However, it seems incompatible with transcription-translation coupling concluded
469
based on the synchronized transcription and translation times (Fig. 6E-6F). We note that
470
the transcription and translation times were determined by the first (fastest) RNAPs and
471
ribosomes arriving at the 3’ end, and they can be the same even though only a small
472
fraction of RNAPs are coupled to a ribosome. Hence, it is likely that a significant fraction
473
of RNAPs is uncoupled with a ribosome during the transcription of lacZ in C. crescentus
474
and experiences premature transcription termination.
475
476
Discussion
477
Our findings have implications for gene regulation based on when and where mRNAs are
478
degraded within a bacterial cell. In bacteria, such as E. coli and B. subtilis, where major
479
ribonuclease and RNA degradosome are localized to the membrane, co-transcriptional
480
mRNA degradation is likely negligible for most genes, and mRNA degradation takes place
481
exclusively on the membrane once mRNAs are released from the gene loci (Fig. 7A-7B).
482
The lack of co-transcriptional degradation would be advantageous when more proteins
483
need to be made per transcripts.
484
Our data showing co-transcriptional degradation of lacY mRNA suggests an exception
485
to this rule for genes encoding inner membrane proteins (Fig. 7C). We note that the high
486
k
d1
of 5’ lacY mRNA measured in lacY2-lacZ-venus (SK575; Fig. 3I) and in full lacY mRNA
487
(SK564; Fig. S3E) likely reflects genuine co-transcriptional degradation, without
488
premature transcription termination because (1) the native (strong) lacZ RBS was used
489
and (2) full lacY transcript (SK564) showed 0% premature transcription termination (Fig.
490
S3F). In terms of the mechanism, membrane localization of nascent mRNAs may not be
491
the only reason that lacY mRNA is degraded co-transcriptionally. The lacZ sequence
492
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
18
within lacY2-lacZ-venus mRNA had similar k
d1
and k
d2
as those of the lacZ-alone case
493
(SK98) even though their localization (or the proximity to the membrane) were vastly
494
different (Fig. 3H). Hence, we speculate that there are additional features in the lacY2
495
sequence (i.e. the first two transmembrane segments) that promote its co-transcriptional
496
mRNA degradation. For example, the signal recognition particle (SRP) and SecYEG,
497
proteins involved in transertion of LacY
79,80
, might interact with RNase E to promote the
498
degradation of lacY2 sequence. Although this idea remains to be tested, such a
499
mechanism can also explain earlier results that translational fusion of a SRP signal
500
peptide to a random gene decreases the transcript’s lifetime
81
and that fast degradation
501
of ptsG mRNA encoding transmembrane glucose transporter (IIBC
glc
) requires the
502
transmembrane segment of its protein
82
. Also, this hypothesis predicts that the rate of co-
503
transcriptional mRNA degradation of lacY would decrease when RNase E is localized in
504
the cytoplasm. Indeed, previous genome-wide characterization of mRNA lifetimes in the
505
RNase E ΔMTS strain showed that many genes encoding inner membrane proteins are
506
preferentially stabilized in this cytoplasmic RNase E mutant
81,83
. These results suggest
507
that membrane localization of RNase E is important for differential regulation of
508
membrane protein expression in comparison to cytoplasmic proteins. Since membrane
509
surface area is limited (more than the cytoplasmic volume)
84
and since membrane
510
channel proteins (such as LacY) have a higher activity cost when expressed
84,85
, tight
511
regulation of membrane protein expression, by employing co-transcriptional degradation
512
mechanism, is likely beneficial for cellular fitness.
513
Another important determinant of mRNA degradation is the timing that transcripts are
514
released from the gene. This timing can vary depending on the gene length and RNAP
515
speed. Also, in the case of polycistronic genes, mRNA processing in the intergenic
516
region
40-42
can release the promoter proximal gene first while promoter distal gene is
517
being transcribed. Adding to this list, our work highlights the important role played by RBS
518
sequences in permitting premature release of incomplete transcripts (Fig. 5J).
519
Based on our model (Fig. 1A), the mean mRNA lifetime in the steady state is affected
520
by the degradation rates of nascent (k
d1*
), fully-transcribed (k
d2
), and prematurely-
521
released (k
dPT
) transcripts because these three types of mRNA can have distinct
522
degradation rates. If ribosomes indeed protect mRNA from degradation
16,17
, each of these
523
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
19
rates may increase with lower RBS strength. However, our data suggests that these rates
524
do not vary much among the RBS mutant strains we examined; instead, the portion of
525
prematurely released transcripts varies significantly (Fig. 5J-5K), eventually yielding
526
different protein output for different RBS sequences (Fig. S5B-S5E). Considering that
527
premature transcription termination is a hallmark event in the absence of transcription-
528
translation coupling
86,87
(Fig. 5I), we identified RBS sequences as the key genetic feature
529
that can modulate the probability of transcription-translation coupling and subsequently
530
mean mRNA lifetimes across the genome.
531
The RBS sequences we tested cover a wide range of translation initiation strengths
532
observed in the genome
88
(Fig. S5G-S5H). If we compare the maximum translation
533
initiation strength we observed non-zero percentage of premature transcription
534
termination (strain SK420 and SK518 in Fig. S5G) with endogenous translation initiation
535
strengths across the E. coli genome
88
, we estimate that at least 58% of all genes may
536
experience some percentage of premature transcription termination due to compromised
537
transcription-translation coupling (Fig. 5I and S5H). This estimation is consistent with the
538
high percentage of 3’-end mRNAs detected at the 5’ UTRs and inside of genes in a recent
539
E. coli transcriptome analysis
89
. Collectively, these results support that transcription-
540
translation uncoupling arising from low translation initiation rate and the resulting
541
premature transcription termination are likely common across the E. coli genome.
542
T7 transcription systems in E. coli, often used for bioengineering and synthetic biology
543
field
90
, are known to experience transcription-translation uncoupling because T7 RNAP
544
outpaces the host ribosome (8-fold speed difference
91
), yet T7 RNAP does not
545
prematurely terminate
92
. Based on our model of mRNA degradation in E. coli, we predict
546
that transcripts made by T7 RNAPs are degraded once transcription is completed, as
547
opposed to experiencing co-transcriptional degradation as proposed previously
46
.
548
We found that gene expression in B. subtilis is analogous to the T7 system, such that
549
premature transcription termination is negligible even though RNAP and ribosome are
550
uncoupled (Fig. 6B). We note that a recent study showed that B. subtilis RNAPs can
551
prematurely dissociate from DNA during transcription of lacZ in a rho-independent
552
manner, especially when their speed is slow
71
. Hence, premature transcription
553
termination may be possible under certain conditions and in certain genes that are under
554
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
20
the control of riboswitches and attenuators
93
and help down-regulate protein expression
555
in B. subtilis.
556
In bacteria, such as C. crescentus, where major ribonuclease and RNA degradosome
557
are located in the cytoplasm, mRNA degradation may start during transcription (Fig. 7D).
558
We observed in C. crescentus, high rate of co-transcriptional degradation rate (k
d1*
) for
559
lacZ mRNA and significant premature transcription termination (PT = 69%). This high
560
premature transcription termination suggests that many RNAPs transcribing lacZ were
561
not coupled to a ribosome and points out that the equality between transcription and
562
translation times (Fig. 6E-6F) may not be a good indicator for the percentage of
563
transcription-translation coupling. Together with the fact that the rho factor physically
564
interacts with RNase E in C. crescentus
78
, our results imply that transcription, premature
565
transcription termination, and mRNA degradation are highly coupled in the cytoplasm of
566
C. crescentus.
567
In conclusion, our work overall identifies subcellular localization of RNase E (or its
568
homologue) and premature transcription termination in the absence of transcription-
569
translation coupling (arising from weak RBS sequences) as spatial and genetic design
570
principles by which bacteria have evolved to differentially regulate transcriptional and
571
translational coupling to mRNA degradation across genes and species. These principles
572
will serve the basis for quantitative modeling of protein expression levels across the
573
genome
24,81,94
and for comprehending the subcellular localization patterns of mRNAs
574
found for different genes and in different bacteria species
24,81,94
. In the future, it would be
575
interesting to investigate whether our findings are relevant to the coordination of
576
transcription, translation, and mRNA degradation in other contexts where there is a lack
577
of membrane-bound microcompartments, such as archaea
95
, chloroplast
36
, and
578
mitochondria
96
.
579
580
Acknowledgements
581
We thank Drs. Christine Jacobs-Wagner, Gene-Wei Li, Jason Peters, Jared Schrader,
582
Lucy Shapiro, and X. Sunney Xie for strains and materials, Dr. Nam Ki Lee for an RNA
583
extraction protocol, and Dr. Ido Golding for allowing us to use his real-time PCR machine
584
in the beginning of the project. We also thank Brooke Ramsey for helping with
585
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
21
experiments, Laura Troyer for illustrations, Dr. Marie Bao for editorial assistance, and the
586
members of the Kim lab for critical reading of the manuscript. This work was supported
587
by NSF Center for Physics of Living Cells (1430124), NSF Science and Technology
588
Center for Quantitative Cell Biology (2243257), NIH (R35GM143203), and Searle
589
Scholars Program.
590
591
Author contributions
592
Conceptualization, Se.K. and Sa.K.; Methodology and investigation, Se.K., Y.W., A.H.,
593
and Sa.K.; Writing and visualization, Se.K., Y.W., A.H., and Sa.K.; Supervision and
594
funding acquisition, Sa.K.
595
596
Figure Legends
597
Figure 1 Two-phase mRNA degradation in bacteria. (A) Definition. mRNA degradation
598
can occur during transcription (on nascent transcripts) and after transcription (on released
599
transcripts) with different rates k
d1
and k
d2
, respectively. (B) Schematics of the time-
600
course assay, in which lacZ transcription is pulse-induced and its transcripts are
601
quantified with qRT-PCR primers amplifying the 530-660 nucleotide (nt) region (Z5) and
602
2732-2890 nt region (Z3) of lacZ (gene length = 3072 nt). The first and last RNAPs pass
603
the Z5 probe site at T
5’
and t
5’
, respectively, and they pass the Z3 probe site at T
3’
and t
3’
,
604
respectively. Blue and yellow shaded boxes indicate the time when k
d1
and k
d2
can be
605
measured by an exponential decay fit, respectively. (C) Anticipated result when mRNA
606
degradation takes place with k
d1
= 0.18 min
-1
and k
d2
= 0.42 min
-1
. (D) Time course data
607
of 5’ and 3’ lacZ mRNA (Z5 and Z3) after induction with 0.2 mM IPTG at t = 0 and re-
608
repression with 500 mM glucose at t = 75 s (strain SK98, grown in M9 glycerol at 30°C).
609
Error bars represent the standard deviation from three biological replicates.
610
611
Figure 2 Effect of RNase E localization and its C-terminal domain on k
d1
and k
d2
of lacZ
612
mRNA. (A) Schematic description of wild-type RNase E and mutants forming different
613
RNA degradosome complexes. The wild-type RNase E interacts with PNPase (green),
614
Enolase (yellow), and RhlB (purple) to form the RNA degradosome. Not drawn to scale.
615
(B) k
d1
and k
d2
of lacZ mRNA in RNase E localization mutant strains (strain SK98, SK339,
616
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
22
SK370, and SK369). Transcription of lacZ was induced with 0.2 mM IPTG at t = 0 and re-
617
repressed with 500 mM glucose at t = 75 s. Error bars represent the standard deviation
618
from three biological replicates. ** and * indicate p<0.01 and p<0.05, respectively (two-
619
sample t test).
620
621
Figure 3 Effect of transertion on k
d1
and k
d2
of lacZ mRNA. (A) Localization of 5’ lacZ
622
mRNA in the presence of active transcription of lacY or aadA from a constitutive promoter
623
P
con
. An example FISH image of 5’ lacZ mRNA when transcription was induced with 0.2
624
mM IPTG and re-repressed with 500 mM glucose at 75 s after induction. The example
625
image is from cells taken at t = 1 min. Scale bar = 1 µm. (B-C) 2D histogram of Z5
FISH
626
localization at different time points along the time-course assay. Colors denote the
627
probability of finding Z5
FISH
in a certain bin location. To minimize noise, the normalized
628
positions of foci along the cell long and short axes were calculated in the first quartile and
629
extended to the other three quartiles using mirror symmetry. The bin size is 70-80 nm.
630
The white ovals are cell outlines, and the white lines are the axes of symmetry. For each
631
histogram, over 5,000 foci were analyzed. The same color scale was used for both
632
histograms. (D) k
d1
and k
d2
of lacZ mRNA with different downstream genes: lacY (SK390),
633
aadA (SK435), or none (SK98). (E) A pair of strains to study the direct transertion effect
634
on lacZ mRNA degradation kinetics. (F-G) 2D histogram of Z5
FISH
localization in lacY2-
635
lacZ-venus (SK575, F) and lacZ (SK98, G). Transcription of lacZ was induced and re-
636
repressed the same way as described in panel A. For each histogram, over 2,000 mRNA
637
foci were analyzed. Except only 564 foci were used for SK98 t = 2 min. (H) k
d1
and k
d2
of
638
lacZ mRNA in SK575 and SK98. (I) Relative mRNA level of the lacY2 sequence in SK575
639
measured by qRT-PCR during the time course assay described in panel A. lacY2 is
640
probed by primers amplifying 80-222 nt region of lacY2 sequence. Blue and yellow
641
shaded boxes indicate the time windows i and iii for k
d1
and k
d2
of lacY2, respectively. In
642
all panels, error bars represent the standard deviation from three biological replicates.
643
Also, ns indicates a statistically nonsignificant difference (two-sample t test).
644
645
Figure 4 Degradation kinetics of araB mRNA and the effect of RBS sequences on k
d1
.
646
(A) Design of strains used in this figure. 5’-UTR sequences from the first base (+1) of the
647
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
23
transcript to the start codon (atg) are shown. SD elements estimated by an RBS
648
calculator
63
are underlined. We note that lacZ in the original lac locus was deleted when
649
lacZ was placed in the ara locus. (B) araB mRNA level change from induction with 0.2%
650
arabinose at t = 0 and re-repression with 500 mM glucose at t = 50 s. 5’ araB and 3’ araB
651
were probed by qRT-PCR primers amplifying 33-210 nt and 1536-1616 nt regions of araB.
652
Blue and yellow boxes denote the time windows where k
d1
and k
d2
are measured. (C) k
d1
653
measured in strains shown in panel A. P
ara
and P*
ara
were induced with 0.2% arabinose,
654
and P
lac
was induced with 0.2 mM IPTG. 500 mM glucose was added at t = 50 s (for araB)
655
or 75 s (for lacZ) to turn off the promoter. (D) LacZ protein expression measured by Miller
656
assay. LacZ expression was induced and re-repressed the same way as in the qRT-PCR
657
experiment (panel C), and the total LacZ protein produced from the pulsed induction were
658
calculated from each strain. In all panels, error bars indicate the standard deviations from
659
three or more biological replicates (except D, from two replicates). ***, **, and * indicate
660
p<0.001, 0.01, and 0.05, respectively, and ns indicates a statistically nonsignificant
661
difference (two-sample t test).
662
663
Figure 5 Origin of fast k
d1
observed in lacZ mRNA with a weak RBS. (A) 5’ UTR
664
sequences of native lacZ and a weak RBS mutant. mRNA sequences from the first base
665
of the transcript to the start codon (atg) are shown. SD sequences estimated by an RBS
666
calculator
63
are underlined. (B) k
d1
and k
d2
of lacZ mRNA measured by induction with 0.2
667
mM IPTG at t = 0 and re-repression with 500 mM glucose at t = 75 s. Error bars represent
668
the standard deviation from three biological replicates. *** denotes p<0.001, and ns
669
indicates a statistically nonsignificant difference (two-sample t test). (C-D) 2D histogram
670
of Z5
FISH
localization depending on the RBS sequence. After shifting the temperature to
671
43.5°C for 10 min, lacZ expression was induced with 0.2 mM IPTG at t = 0 and re-
672
repressed with 500 mM glucose at t = 50 s. In each case, over 25,000 mRNA foci were
673
analyzed. (E) Number of fluorescent Z5
FISH
spots detected per cell at t = 60 s during the
674
time-course experiment described in panel C-D. Error bars represent the standard error
675
from bootstrapping. (F-H) Z5 and Z3 levels after lacZ transcription was induced with 0.2
676
mM IPTG at t = 0. In (H), 100 µg/mL BCM was added 5 min before IPTG addition. Error
677
bars represent the standard deviation from two biological replicates. (I) Effect of RBS
678
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
24
strength on the fate of mRNA. (J-K) Relationship between k
d1
or k
d2
and the probability
679
of premature transcription termination in various RBS-lacZ mRNAs and P
ara
-araB mRNA.
680
See Table S4 for the list of strains used. The line fit is based on the equation (1). Error
681
bars for k
d1
or k
d2
represent the standard deviation from three replicates and those for PT
682
were calculated from the steady-state ratio of Z5 and Z3 in two replicates.
683
684
Figure 6 Degradation kinetics of lacZ mRNA in B. subtilis and C. crescentus. (A) IPTG-
685
inducible lacZ in the chromosome of B. subtilis. For qRT-PCR, we used the same Z5 and
686
Z3 primers used in E. coli lacZ. (B-C) Z5 and Z3 levels after induction with 5 mM IPTG at
687
t = 0, probed by qRT-PCR. To measure lacZ mRNA degradation rates in (C), transcription
688
was re-repressed with 200 µg/mL rifampicin at t = 30 s. The time windows used for k
d1
689
and k
d2
fitting are indicated as blue and yellow boxes. B. subtilis cells were grown in
690
MOPS media supplemented with maltose at 30°C. Error bars represent the standard
691
deviation from two (B) or three (C) biological replicates. (D) Xylose-inducible lacZ in C.
692
crescentus. For qRT-PCR, we used the same Z5 and Z3 primers used in E. coli lacZ. (E)
693
Translation kinetics of LacZ protein expression in C. crescentus after adding 0.3% xylose,
694
probed by Miller assay using MUG (3-methylumbelliferyl-beta-D-galactopyranoside) as a
695
sensitive LacZ substrate. Error bars represent the standard deviation from three biological
696
replicates. (F-G) Z5 and Z3 levels after induction with 0.3% xylose at t = 0, probed by
697
qRT-PCR. To measure lacZ mRNA degradation rates in (G), transcription was re-
698
repressed with 200 µg/mL rifampicin at t = 50 s. The time window used for on k
d1
fitting
699
for Z5 is indicated as blue box. C. crescentus cells were grown in M2G at 28°C. Error bars
700
represent the standard deviation from five (F) or three (G) biological replicates.
701
702
Figure 7 Generalizable model of mRNA degradation in bacteria. (A-C) Scenarios in E.
703
coli (and possibly other bacterial species having the main ribonuclease on the membrane)
704
for genes encoding cytoplasmic proteins with strong RBS (A) and weak RBS (B) and for
705
genes encoding inner membrane proteins (C). (D) A scenario in C. crescentus and
706
possibly other bacterial species having the main ribonuclease in the cytoplasm. The
707
cartoon is drawn to reflect that nucleoid takes a large area of the cytoplasm in C.
708
crescentus
97
.
709
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
25
References
710
1. Campos, M., and Jacobs-Wagner, C. (2013). Cellular organization of the transfer
711
of genetic information. Curr Opin Microbiol 16, 171-176.
712
2. Shine, M., Gordon, J., Schärfen, L., Zigackova, D., Herzel, L., and Neugebauer,
713
K.M. (2024). Co-transcriptional gene regulation in eukaryotes and prokaryotes.
714
Nat Rev Mol Cell Biol.
715
3. Li, G.-W., Burkhardt, D., Gross, C., and Weissman, Jonathan S. (2014).
716
Quantifying absolute protein synthesis rates reveals principles underlying
717
allocation of cellular resources. Cell 157, 624-635.
718
4. Scott, M., and Hwa, T. (2023). Shaping bacterial gene expression by physiological
719
and proteome allocation constraints. Nat Rev Microbiol 21, 327-342.
720
5. Nieuwkoop, T., Finger-Bou, M., van der Oost, J., and Claassens, N.J. (2020). The
721
ongoing quest to crack the genetic code for protein production. Mol Cell 80, 193-
722
209.
723
6. Carrier, T.A., and Keasling, J.D. (1997). Controlling messenger RNA stability in
724
bacteria: Strategies for engineering gene expression. Biotechnol Prog 13, 699-708.
725
7. Miller, O.L., Hamkalo, B.A., and Thomas, C.A. (1970). Visualization of bacterial
726
genes in action. Science 169, 392-395.
727
8. Landick, R., Carey, J., and Yanofsky, C. (1985). Translation activates the paused
728
transcription complex and restores transcription of the trp operon leader region.
729
Proc Natl Acad Sci USA 82, 4663-4667.
730
9. Proshkin, S., Rahmouni, A.R., Mironov, A., and Nudler, E. (2010). Cooperation
731
between translating ribosomes and RNA polymerase in transcription elongation.
732
Science 328, 504-508.
733
10. Irastortza-Olaziregi, M., and Amster-Choder, O. (2021). Coupled transcription-
734
translation in prokaryotes: An old couple With new surprises. Front Microbiol 11.
735
11. Blaha, G.M., and Wade, J.T. (2022). Transcription-translation coupling in bacteria.
736
Annu Rev Genet 56, 187-205.
737
12. Morikawa, N., and Imamoto, F. (1969). Degradation of tryptophan messenger: On
738
the degradation of messenger RNA for the tryptophan operon in Escherichia coli.
739
Nature 223, 37-40.
740
13. Morse, D.E., Mosteller, R.D., and Yanofsky, C. (1969). Dynamics of synthesis,
741
translation, and degradation of trp operon messenger RNA in E. coli. Cold Spring
742
Harbor Symp Quant Biol 34, 725-740.
743
14. Cannistraro, V.J., and Kennell, D. (1985). Evidence that the 5' end of lac mRNA
744
starts to decay as soon as it is synthesized. J Bacteriol 161, 820-822.
745
15. Chen, H., Shiroguchi, K., Ge, H., and Xie, X.S. (2015). Genome-wide study of
746
mRNA degradation and transcript elongation in Escherichia coli. Mol Syst Biol 11,
747
781.
748
16. Deana, A., and Belasco, J.G. (2005). Lost in translation: the influence of ribosomes
749
on bacterial mRNA decay. Genes Dev 19, 2526-2533.
750
17. Dreyfus, M. (2009). Killer and protective ribosomes. Prog Mol Biol Transl Sci 85,
751
423-466.
752
18. Miczak, A., Srivastava, R.A.K., and Apirion, O. (1991). Location of the RNA-
753
processing enzymes RNase III, RNase E and RNase P in the Eschenchia coli cell.
754
Mol Microbiol 5, 1801-1810.
755
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
26
19. Liou, G.-G., Jane, W.-N., Cohen, S.N., Lin, N.-S., and Lin-Chao, S. (2001). RNA
756
degradosomes exist in vivo in Escherichia coli as multicomponent complexes
757
associated with the cytoplasmic membrane via the N-terminal region of
758
ribonuclease E. Proc Natl Acad Sci USA 98, 63-68.
759
20. Khemici, V., Poljak, L., Luisi, B.F., and Carpousis, A.J. (2008). The RNase E of
760
Escherichia coli is a membrane-binding protein. Mol Microbiol 70, 799-813.
761
21. Mackie, G.A. (2013). RNase E: at the interface of bacterial RNA processing and
762
decay. Nat Rev Microbiol 11, 45-57.
763
22. Roggiani, M., and Goulian, M. (2015). Chromosome-membrane interactions in
764
bacteria. Annu Rev Genet 49, 115-129.
765
23. Libby, E.A., Roggiani, M., and Goulian, M. (2012). Membrane protein expression
766
triggers chromosomal locus repositioning in bacteria. Proc Natl Acad Sci USA 109,
767
7445-7450.
768
24. Montero Llopis, P., Jackson, A.F., Sliusarenko, O., Surovtsev, I., Heinritz, J.,
769
Emonet, T., and Jacobs-Wagner, C. (2010). Spatial organization of the flow of
770
genetic information in bacteria. Nature 466, 77-81.
771
25. Al-Husini, N., Tomares, D.T., Bitar, O., Childers, W.S., and Schrader, J.M. (2018).
772
α-proteobacterial RNA degradosomes assemble liquid-liquid phase-separated
773
RNP bodies. Mol Cell 71, 1027-1039.e1014.
774
26. Bayas, C.A., Wang, J., Lee, M.K., Schrader, J.M., Shapiro, L., and Moerner, W.E.
775
(2018). Spatial organization and dynamics of RNase E and ribosomes in
776
Caulobacter crescentus. Proc Natl Acad Sci USA 115, E3712-E3721.
777
27. Yarchuk, O., Jacques, N., Guillerez, J., and Dreyfus, M. (1992). Interdependence
778
of translation, transcription and mRNA degradation in the lacZ gene. J Mol Biol
779
226, 581-596.
780
28. McCormick, J.R., Zengel, J.M., and Lindahl, L. (1994). Correlation of translation
781
efficiency with the decay of lacZ mRNA in Escherichia coli. J Mol Biol 239, 608-
782
622.
783
29. Wagner, L.A., Gesteland, R.F., Dayhuff, T.J., and Weiss, R.B. (1994). An efficient
784
Shine-Dalgarno sequence but not translation is necessary for lacZ mRNA stability
785
in Escherichia coli. J Bacteriol 176, 1683-1688.
786
30. Arnold, T.E., Yu, J., and Belasco, J.G. (1998). mRNA stabilization by the ompA 5'
787
untranslated region: two protective elements hinder distinct pathways for mRNA
788
degradation. RNA 4, 319-330.
789
31. Duviau, M.-P., Chen, F., Emile, A., Cocaign-Bousquet, M., Girbal, L., and Nouaille,
790
S. (2023). When translation elongation is impaired, the mRNA is uniformly
791
destabilized by the RNA degradosome, while the concentration of mRNA is altered
792
along the molecule. Nucleic Acids Res 51, 2877-2890.
793
32. Richards, J., Luciano, D.J., and Belasco, J.G. (2012). Influence of translation on
794
RppH-dependent mRNA degradation in Escherichia coli. Mol Microbiol 86, 1063-
795
1072.
796
33. Zhu, M., Mori, M., Hwa, T., and Dai, X. (2019). Disruption of transcription–
797
translation coordination in Escherichia coli leads to premature transcriptional
798
termination. Nat Microbiol 4, 2347-2356.
799
34. Johnson, G.E., Lalanne, J.-B., Peters, M.L., and Li, G.-W. (2020). Functionally
800
uncoupled transcription–translation in Bacillus subtilis. Nature 585, 124-128.
801
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
27
35. Saito, K., Green, R., and Buskirk, A.R. (2020). Translational initiation in E. coli
802
occurs at the correct sites genome-wide in the absence of mRNA-rRNA base-
803
pairing. eLife 9, e55002.
804
36. Aït-Bara, S., and Carpousis, A.J. (2015). RNA degradosomes in bacteria and
805
chloroplasts: classification, distribution and evolution of RNase E homologs. Mol
806
Microbiol 97, 1021-1135.
807
37. Carpousis, A.J., Campo, N., Hadjeras, L., and Hamouche, L. (2022).
808
Compartmentalization of RNA degradosomes in bacteria controls accessibility to
809
substrates and ensures concerted degradation of mRNA to nucleotides. Annu Rev
810
Microbiol 76, 533-552.
811
38. Adesnik, M., and Levinthal, C. (1970). The synthesis and degradation of lactose
812
operon messenger RNA in E. coli. Cold Spring Harbor Symp Quant Biol 35, 451-
813
459.
814
39. Schwartz, T., Craig, E., and Kennell, D. (1970). Inactivation and degradation of
815
messenger ribonucleic acid from the lactose operon of Escherichia coli. J Mol Biol
816
54, 299-311.
817
40. Lim, L.W., and Kennell, D. (1979). Models for decay of Escherichia coli lac
818
messenger RNA and evidence for inactivating cleavages between its messages.
819
J Mol Biol 135, 369-390.
820
41. Cannistraro, V.J., Subbarao, M.N., and Kennell, D. (1986). Specific
821
endonucleolytic cleavage sites for decay of Escherichia coli mRNA. J Mol Biol 192,
822
257-274.
823
42. McCormick, J.R., Zengel, J.M., and Lindahl, L. (1991). Intermediates in the
824
degradation of mRNA from the lactose operon of Escherichia coli. Nucleic Acids
825
Res 19, 2767-2776.
826
43. Mudd, E.A., Krisch, H.M., and Higgins, C.F. (1990). RNase E, an
827
endoribonuclease, has a general role in the chemical decay of Escherichia coli
828
mRNA: evidence that rne and ams are the same genetic locus. Mol Microbiol 4,
829
2127-2135.
830
44. Carpousis, A.J., Luisi, B.F., and McDowall, K.J. (2009). Endonucleolytic initiation
831
of mRNA decay in Escherichia coli. Prog Mol Biol Transl Sci 85, 91-135.
832
45. Hui, M.P., Foley, P.L., and Belasco, J.G. (2014). Messenger RNA degradation in
833
bacterial cells. Annu Rev Genet 48, 537-559.
834
46. Iost, I., and Dreyfus, M. (1995). The stability of Escherichia coli lacZ mRNA
835
depends upon the simultaneity of its synthesis and translation. EMBO J 14, 3252-
836
3261.
837
47. Petersen, C. (1993). Translation and mRNA stability in bacteria: a complex
838
relationship. In Control of mRNA stability, J.G. Belasco, and G. Brawerman, eds.
839
(Academic Press), pp. 117-145.
840
48. Massé, E., Escorcia, F.E., and Gottesman, S. (2003). Coupled degradation of a
841
small regulatory RNA and its mRNA targets in Escherichia coli. Genes Dev 17,
842
2374-2383.
843
49. Strahl, H., Turlan, C., Khalid, S., Bond, P.J., Kebalo, J.-M., Peyron, P., Poljak, L.,
844
Bouvier, M., Hamoen, L., Luisi, B.F., and Carpousis, A.J. (2015). Membrane
845
recognition and dynamics of the RNA degradosome. PLOS Genet 11, e1004961.
846
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
28
50. Carpousis, A.J. (2007). The RNA degradosome of Escherichia coli: An mRNA-
847
degrading machine assembled on RNase E. Annu Rev Microbiol 61, 71-87.
848
51. Lopez, P.J., Marchand, I., Joyce, S.A., and Dreyfus, M. (1999). The C-terminal half
849
of RNase E, which organizes the Escherichia coli degradosome, participates in
850
mRNA degradation but not rRNA processing in vivo. Mol Microbiol 33, 188-199.
851
52. Leroy, A., Vanzo, N.F., Sousa, S., Dreyfus, M., and Carpousis, A.J. (2002).
852
Function in Escherichia coli of the non-catalytic part of RNase E: role in the
853
degradation of ribosome-free mRNA. Mol Microbiol 45, 1231-1243.
854
53. Kleppe, K., Övrebö, S., and Lossius, I. (1979). The bacterial nucleoid. Microbiol
855
112, 1-13.
856
54. Norris, V. (1995). Hypothesis: chromosome separation in Escherichia coli involves
857
autocatalytic gene expression, transertion and membrane-domain formation. Mol
858
Microbiol 16, 1051-1057.
859
55. Woldringh, C.L. (2002). The role of co-transcriptional translation and protein
860
translocation (transertion) in bacterial chromosome segregation. Mol Microbiol 45,
861
17-29.
862
56. Kim, S., Beltran, B., Irnov, I., and Jacobs-Wagner, C. (2019). Long-distance
863
cooperative and antagonistic RNA polymerase dynamics via DNA supercoiling.
864
Cell 179, 106-119.
865
57. Yang, S., Kim, S., Kim, D.-K., Jeon An, H., Bae Son, J., Hedén Gynnå, A., and Ki
866
Lee, N. (2019). Transcription and translation contribute to gene locus relocation to
867
the nucleoid periphery in E. coli. Nat Commun 10, 5131.
868
58. Steitz, J.A., and Jakes, K. (1975). How ribosomes select initiator regions in mRNA:
869
base pair formation between the 3' terminus of 16S rRNA and the mRNA during
870
initiation of protein synthesis in Escherichia coli. Proc Natl Acad Sci USA 72, 4734-
871
4738.
872
59. Shine, J., and Dalgarno, L. (1974). Terminal sequence of Escherichia coli 16S
873
ribosomal RNA: Complementarity to nonsense triplets and ribosome binding sites.
874
Proc Natl Acad Sci USA 71, 1342-1346.
875
60. Jacques, N., and Dreyfus, M. (1990). Translation initiation in Escherichia coli: old
876
and new questions. Mol Microbiol 4, 1063-1067.
877
61. Ringquist, S., Shinedling, S., Barrick, D., Green, L., Binkley, J., Stormo, G.D., and
878
Gold, L. (1992). Translation initiation in Escherichia coli: sequences within the
879
ribosome-binding site. Mol Microbiol 6, 1219-1229.
880
62. Salis, H.M., Mirsky, E.A., and Voigt, C.A. (2009). Automated design of synthetic
881
ribosome binding sites to control protein expression. Nat Biotechnol 27, 946-950.
882
63. Seo, S.W., Yang, J.-S., Kim, I., Yang, J., Min, B.E., Kim, S., and Jung, G.Y. (2013).
883
Predictive design of mRNA translation initiation region to control prokaryotic
884
translation efficiency. Metab Eng 15, 67-74.
885
64. Terpe, K. (2006). Overview of bacterial expression systems for heterologous
886
protein production: from molecular and biochemical fundamentals to commercial
887
systems. Appl Microbiol Biotechnol 72, 211-222.
888
65. Choi, P.J., Cai, L., Frieda, K., and Xie, X.S. (2008). A stochastic single-molecule
889
event triggers phenotype switching of a bacterial cell. Science 322, 442-446.
890
66. Khlebnikov, A., Datsenko, K.A., Skaug, T., Wanner, B.L., and Keasling, J.D. (2001).
891
Homogeneous expression of the PBAD promoter in Escherichia coli by constitutive
892
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
29
expression of the low-affinity high-capacity AraE transporter. Microbiol 147, 3241-
893
3247.
894
67. Hansen, M.T., Bennett, P.M., and von Meyenburg, K. (1973). Intracistronic polarity
895
during dissociation of translation from transcription in Escherichia coli. J Mol Biol
896
77, 589-604.
897
68. Stanssens, P., Remaut, E., and Fiers, W. (1986). Inefficient translation initiation
898
causes premature transcription termination in the IacZ gene. Cell 44, 711-718.
899
69. Vogel, U., Sørensen, M., Pedersen, S., Jensen, K.F., and Kilstrup, M. (1992).
900
Decreasing transcription elongation rate in Escherichia Coli exposed to amino acid
901
starvation. Mol Microbiol 6, 2191-2200.
902
70. de Smit, M.H., Verlaan, P.W.G., van Duin, J., and Pleij, C.W.A. (2009). In vivo
903
dynamics of intracistronic transcriptional polarity. J Mol Biol 385, 733-747.
904
71. Zhu, M., Mu, H., Han, F., Wang, Q., and Dai, X. (2021). Quantitative analysis of
905
asynchronous transcription-translation and transcription processivity in Bacillus
906
subtilis under various growth conditions. iScience 24, 103333.
907
72. McClure, W.R., and Cech, C.L. (1978). On the mechanism of rifampicin inhibition
908
of RNA synthesis. J Biol Chem 253, 8949-8956.
909
73. Mosteller, R.D., and Yanofsky, C. (1970). Transcription of the tryptophan operon
910
in Escherichia coli: Rifampicin as an inhibitor of initiation. J Mol Biol 48, 525-531.
911
74. Campbell, E.A., Korzheva, N., Mustaev, A., Murakami, K., Nair, S., Goldfarb, A.,
912
and Darst, S.A. (2001). Structural mechanism for rifampicin inhibition of bacterial
913
RNA polymerase. Cell 104, 901-912.
914
75. Lehnik-Habrink, M., Newman, J., Rothe, F.M., Solovyova, A.S., Rodrigues, C.,
915
Herzberg, C., Commichau, F.M., Lewis, R.J., and Stülke, J. (2011). RNase Y in
916
Bacillus subtilis: A natively disordered protein that is the functional equivalent of
917
RNase E from Escherichia coli. J Bacteriol 193, 5431-5441.
918
76. Durand, S., Gilet, L., Bessières, P., Nicolas, P., and Condon, C. (2012). Three
919
essential ribonucleases—RNase Y, J1, and III—control the abundance of a
920
majority of Bacillus subtilis mRNAs. PLOS Genet 8, e1002520.
921
77. Meisenzahl, A.C., Shapiro, L., and Jenal, U. (1997). Isolation and characterization
922
of a xylose-dependent promoter from Caulobacter crescentus. J Bacteriol 179,
923
592-600.
924
78. Aguirre, A.A., Vicente, A.M., Hardwick, S.W., Alvelos, D.M., Mazzon, R.R., Luisi,
925
B.F., and Marques, M.V. (2017). Association of the cold shock DEAD-box RNA
926
helicase RhlE to the RNA degradosome in Caulobacter crescentus. J Bacteriol
927
199, e00135-00117.
928
79. Zhang, X., and Shan, S.-o. (2014). Fidelity of cotranslational protein targeting by
929
the signal recognition particle. Annu Rev Biophys 43, 381-408.
930
80. Macfarlane, J., and Müller, M. (1995). The functional integration of a polytopic
931
membrane protein of Escherichia coli is dependent on the bacterial signal-
932
recognition particle. Eur J Biochem 233, 766-771.
933
81. Moffitt, J.R., Pandey, S., Boettiger, A.N., Wang, S., and Zhuang, X. (2016). Spatial
934
organization shapes the turnover of a bacterial transcriptome. eLife 5, e13065.
935
82. Kawamoto, H., Morita, T., Shimizu, A., Inada, T., and Aiba, H. (2005). Implication
936
of membrane localization of target mRNA in the action of a small RNA: mechanism
937
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
30
of post-transcriptional regulation of glucose transporter in Escherichia coli. Genes
938
Dev 19, 328-338.
939
83. Hadjeras, L., Poljak, L., Bouvier, M., Morin-Ogier, Q., Canal, I., Cocaign-Bousquet,
940
M., Girbal, L., and Carpousis, A.J. (2019). Detachment of the RNA degradosome
941
from the inner membrane of Escherichia coli results in a global slowdown of mRNA
942
degradation, proteolysis of RNase E and increased turnover of ribosome-free
943
transcripts. Mol Microbiol 111, 1715-1731.
944
84. Dykhuizen, D., and Davies, M. (1980). An experimental model: Bacterial
945
specialists and generalists competing in chemostats. Ecology 61, 1213-1227.
946
85. Eames, M., and Kortemme, T. (2012). Cost-benefit tradeoffs in engineered lac
947
operons. Science 336, 911-915.
948
86. Adhya, S., and Gottesman, M. (1978). Control of transcription termination. Annu
949
Rev Biochem 47, 967-996.
950
87. Yanofsky, C. (1981). Attenuation in the control of expression of bacterial operons.
951
Nature 289, 751-758.
952
88. Taniguchi, Y., Choi, P.J., Li, G.-W., Chen, H., Babu, M., Hearn, J., Emili, A., and
953
Xie, X.S. (2010). Quantifying E. coli proteome and transcriptome with single-
954
molecule sensitivity in single cells. Science 329, 533-538.
955
89. Adams, P.P., Baniulyte, G., Esnault, C., Chegireddy, K., Singh, N., Monge, M.,
956
Dale, R.K., Storz, G., and Wade, J.T. (2021). Regulatory roles of Escherichia coli
957
5' UTR and ORF-internal RNAs detected by 3' end mapping. eLife 10, e62438.
958
90. Wang, W., Li, Y., Wang, Y., Shi, C., Li, C., Li, Q., and Linhardt, R.J. (2018).
959
Bacteriophage T7 transcription system: an enabling tool in synthetic biology.
960
Biotechnol Adv 36, 2129-2137.
961
91. Iost, I., Guillerez, J., and Dreyfus, M. (1992). Bacteriophage T7 RNA polymerase
962
travels far ahead of ribosomes in vivo. J Bacteriol 174, 619-622.
963
92. Epshtein, V., Dutta, D., Wade, J., and Nudler, E. (2010). An allosteric mechanism
964
of Rho-dependent transcription termination. Nature 463, 245-249.
965
93. Dar, D., Shamir, M., Mellin, J.R., Koutero, M., Stern-Ginossar, N., Cossart, P., and
966
Sorek, R. (2016). Term-seq reveals abundant ribo-regulation of antibiotics
967
resistance in bacteria. Science 352, aad9822.
968
94. Nevo-Dinur, K., Nussbaum-Shochat, A., Ben-Yehuda, S., and Amster-Choder, O.
969
(2011). Translation-independent localization of mRNA in E. coli. Science 331,
970
1081-1084.
971
95. Clouet-d'Orval, B., Batista, M., Bouvier, M., Quentin, Y., Fichant, G., Marchfelder,
972
A., and Maier, L.-K. (2018). Insights into RNA-processing pathways and
973
associated RNA-degrading enzymes in Archaea. FEMS Microbiol Rev 42, 579-
974
613.
975
96. Cartalas, J., Coudray, L., and Gobert, A. (2022). How RNases shape mitochondrial
976
transcriptomes. Int J Mol Sci 23, 6141.
977
97. Gray, W.T., Govers, S.K., Xiang, Y., Parry, B.R., Campos, M., Kim, S., and Jacobs-
978
Wagner, C. (2019). Nucleoid size scaling and intracellular organization of
979
translation across bacteria. Cell 177, 1632-1648.e1620.
980
981
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
60
40
20
0
6004002000
Time after induction (s)
Z5
Z3
Figure 1
RNAP
Ribosome
A
E. coli chromosome
B
Nascent mRNA
Released
mRNA
Z5 Z3
lacZ
Time
On
Off
lacZ mRNA
RNAP
+IPTG
+glucose
RT PCR probe sites:
(i)
(ii)
(iii)
5’ lacZ mRNA
degradation:
C
k
d1
k
d2
Ribonuclease
D
T
5’
T
3’
t
5’
t
3’
Z5
Z3
Time after induction (s)
0 200 400 600
0
20
40
60
Relative mRNA level (AU)
LacI
Relative mRNA level (AU)
t
5’
t
3’
T
5’
T
3’
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
Figure 2
A B
Inner membrane
RNase E ∆MTS
RNase E (1-529)
RNase E (1-592)RNase E
WT
RNase E
∆MTS
RNase E
(1-592)
RNase E
(1-529)
lacZ mRNA degradation rate (min
-1
)
k
d1
k
d2
Cytoplasm
0.6
0.4
0.2
0.0
**
*
*
*
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
t = 1 min
t = 2 min
t = 3 min
5’ lacZ mRNA in SK575
t = 1 min
t = 2 min
t = 3 min
5’ lacZ mRNA in SK98
0
0.02
0.04
Figure 3
A
D
k
d1
k
d2
lacZ mRNA degradation rate (min
-1
)
A constitutively expressed gene
added downstream of lacZ
F
lacZ
lacY (SK435)
or
aadA (SK390)
lacZ mRNA
FISH probes for 5’ lacZ
B C
P
con
lacY
aadA
None
H
I
lacZ mRNA degradation rate (min
-1
)
G
SK575
SK98
x
y
5’ lacZ mRNA in SK435
5’ lacZ mRNA in SK435
0
0.02
0.04
5’ lacZ mRNA in SK390
t = 1 min
t = 2 min
t = 3 min
t = 1 min
t = 2 min
t = 3 min
lacZ
lacY2 (1-74 aa)
venus
SK575
lacZ
SK98
E
Time after induction (s)
0 200 400 600
10
20
30
40
Relative mRNA level (AU)
0
lacY2
Z3
lacY2-lacZ-venus (SK575)
LacI
LacI
LacI
0.6
0.4
0.2
0.0
ns
ns
ns
ns
k
d1
k
d2
ns
ns
0.6
0.4
0.2
0.0
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
Figure 4
B
C
D
P
ara
-araB (SK472)
A
araB or lacZ mRNA
degradation rate, k
d1
(1/min)
E. coli
chromosome
ara locus
lac locus
Native araB: ACCCGTTTTTTTGGATGGAGTGAAACGatg (SK472, SK477)
Native lacZ: ATTGTGAGCGGATAACAATTTCACACAGGAAACAGCTatg (SK499)
Synthetic: ACCCGTTTTTTTGTAAGGCAGGATATTatg (SK613)
AraC
araB
AraC
lacZ
lacZ
AraC
lacZ
P
ara
-araB (SK472)
P
ara
-lacZ (SK477)
P
lac
-lacZ (SK499)
P
*
ara
-lacZ (SK613)
5’ UTR sequences:
P
ara
-araB
(SK472)
P
ara
-lacZ
(SK477)
P
lac
-lacZ
(SK499)
P*
ara
-lacZ
(SK613)
P
ara
-lacZ
(SK477)
P
lac
-lacZ
(SK499)
P*
ara
-lacZ
(SK613)
LacZ protein level (AU)
60
40
20
0
LacI
5’ araB
3’ araB
Time after induction (s)
0 200 400 600
Relative mRNA level (AU)
0
20
40
60
80
ns
**
0.6
0.4
0.2
0.0
P
lac
-lacZ
(SK98)
ns
ns
***
*
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
Figure 5
A
C
F
RNAP
Ribosome
RNase E
Strong RBS
Weak RBS
lacZ mRNA degradation rate (min
-1
)
D
I
Native lacZ: ATTGTGAGCGGATAACAATTTCACACAGGAAACAGCTatg
Weak RBS: ATTGTGAGCGGATAACAATTTCACACGGTTGCCAGCTatg
B
E
G
SK98 and SK519 (for rne3071)
SK421 and SK591 (for rne3071)
t = 30 s
t = 60 s
t = 90 s
5’ lacZ mRNA in SK519
t = 30 s
t = 60 s
t = 90 s
5’ lacZ mRNA in SK591
0
0.02
0.04
J K
k
d1
(min
-1
)
araB locus
lacZ locus
lacZ mRNA araB mRNA
araB locus
Z5
Z3
Time after induction (s)
0 300 600 900
0
100
200
300
Relative mRNA level (AU)
1200
400
Time after induction (s)
0 300 600 900 1200
Z5
Z3
0
5
10
15
Relative mRNA level (AU)
20
25
Native RBS (SK98)
Weak RBS (SK421)
H
k
d2
(min
-1
)
800400
Time after induction (s)
0 1200
0
5
10
15
Relative mRNA level (AU)
20
Z5
Z3
Weak RBS + BCM
Rho factor
5’ lacZ mRNA at t = 60 s
Number of spots per cell
Probability
0
0.2
0.4
0.6
1 2 3 4
SK519
SK591
k
d1
k
d2
***
ns
0.8
0.6
0.4
0.2
0.0
RBS: Native Weak
(SK98) (SK421)
0.8
0.6
0.4
0.2
0.0
100806040200
Premature termination (%)
0.8
0.6
0.4
0.2
0.0
100806040200
Premature termination (%)
araB locus
lacZ locus
lacZ mRNA araB mRNA
araB locus
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
Figure 6
A
Z5
Z3
lacZ
P
spank
-lacZ (B. subtilis)
B
D
XylR
lacZ
P
xylX
xylX-lacZ (C. crescentus)
E F
Z5
Z3
xylX
G
500
400
300
200
100
0
Relative mRNA level (AU)
12008004000
Time after induction (s)
LacI
Z5
Z3
30
20
10
0
Relative mRNA level (AU)
6005004003002001000
Time after induction (s)
C
5
4
3
2
1
Relative mRNA level (AU)
3002001000
Time after induction (s)
80
60
40
20
0
LacZ protein expression (AU)
6004002000
Time after induction (s)
IPTG
IPTG Rifampicin
Xylose
Xylose Rifampicin
Xylose
Z5
Z3
4
3
2
1
Relative mRNA level (AU)
3002001000
Time after induction (s)
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
RNAP
Ribosome
RNase E
Figure 7
A
B
C
D
C. crescentus
Nucleoid area
DNA
Premature
mRNA
Full-length
mRNA
Co-transcriptional
degradation
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 18, 2024. ; https://doi.org/10.1101/2024.04.18.588412doi: bioRxiv preprint
