Transcriptional burst frequency and burst size are
equally modulated across the human genome
Roy D. Dara,b,c,1, Brandon S. Razookya,d,e,1, Abhyudai Singhd,2, Thomas V. Trimelonif, James M. McCollumf,
Chris D. Coxg,h, Michael L. Simpsonb,I,3, and Leor S. Weinbergera,d,j,3
aGladstone Institutes, San Francisco, CA 94158;
cDepartment of Physics and Astronomy, University of Tennessee, Knoxville, TN 37996;
at San Diego, La Jolla, CA 92093;
University of Tennessee, Knoxville, TN 37996;hDepartment of Civil and Environmental Engineering, University of Tennessee, Knoxville, TN 37996;
IDepartment of Materials Science and Engineering, University of Tennessee, Knoxville, TN 37996;
of California, San Francisco, CA 94158; and
bCenter for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, TN 37831;
dDepartment of Chemistry and Biochemistry, University of California
eBiophysics Graduate Group, University of California, San Francisco, CA 94158;
gCenter for Environmental Biotechnology,
jDepartment of Biochemistry and Biophysics, University
fDepartment of Electrical and Computer Engineering, Virginia Commonwealth University, Richmond, VA
Edited by Jonathan S. Weissman, University of California, San Francisco, CA, and approved September 13, 2012 (received for review August 8, 2012)
Gene expression occurs either as an episodic process, characterized
by pulsatile bursts, or as a constitutive process, characterized by
a Poisson-like accumulation of gene products. It is not clear which
mode of gene expression (constitutive versus bursty) predomi-
nates across a genome or how transcriptional dynamics are influ-
enced by genomic position and promoter sequence. Here, we use
time-lapse fluorescence microscopy to analyze 8,000 individual
human genomic loci and find that at virtually all loci, episodic
bursting—as opposed to constitutive expression—is the predomi-
nant mode of expression. Quantitative analysis of the expression
dynamics at these 8,000 loci indicates that both the frequency
and size of the transcriptional bursts varies equally across the
human genome, independent of promoter sequence. Strikingly,
weaker expression loci modulate burst frequency to increase
activity, whereas stronger expression loci modulate burst size to
increase activity. Transcriptional activators such as trichostatin A
(TSA) and tumor necrosis factor α (TNF) only modulate burst size
and frequency along a constrained trend line governed by the
promoter. In summary, transcriptional bursting dominates across
the human genome, both burst frequency and burst size vary by
chromosomal location, and transcriptional activators alter burst
frequency and burst size, depending on the expression level of
stochastic noise ∣ automated single-cell imaging ∣ human
immunodeficiency virus ∣ long terminal repeat promoter
classical view of gene expression as a constitutive, Poisson-like
accumulation of gene products (Fig. 1A) is supported by a com-
prehensive large-scale study in bacteria, demonstrating that >400
genes appear to follow constitutive (or Poisson-like) gene expres-
sion (1). Constitutive expression has also been reported for
subsets of human genes (2). Conversely, several elegant studies
showed that specific promoters in bacteria and yeast express
gene products in an episodic process (Fig. 1B), characterized by
pulsatile bursts in transcription (3–9). Given this conflicting evi-
dence, it remains unclear if episodic bursting is the predominant
mode of gene expression across a genome or just a highlighted
exception. If bursting is predominant, it is not clear if or how it
depends on genomic location.
To globally determine if constitutive Poisson-like expression
or episodic bursty expression dominates throughout the human
genome, we capitalize on a recently proposed theoretical frame-
work (10) for extracting the details of gene regulation from the
time-resolved structure of fluctuations (i.e., noise) in gene expres-
sion. This analysis quantifies time-lapse expression trajectories to
obtain three orthogonal measures of expression: the average
expression level, the magnitude of expression fluctuations (as
measured by the coefficient of variation squared, CV2), and the
here exists conflicting evidence over the predominant mode
of gene expression in both prokaryotes and eukaryotes. The
autocorrelation time of expression fluctuations (as measured
by the noise autocorrelation time at half of its initial value, τ1∕2)
(11, 12) (Fig. 1C). Although this three-dimensional noise space is
impractical to analyze directly, different two-dimensional projec-
tions of noise space allow the quantification of rate parameters
in gene-regulatory models and provide a convenient method to
differentiate between underlying gene-expression mechanisms,
such as constitutive versus bursty transcription (Fig. 1C). For
example, transcriptional bursting increases both noise magnitude
and noise autocorrelation time and shifts points in the CV2
versus τ1∕2plane to the upper right quadrant relative to a con-
stitutive expression model (Fig. 1C, Bottom Left). Conversely,
translational bursting shifts noise magnitude, but not the auto-
correlation time (10, 13).
Importantly, analysis of the τ1∕2axis is critical to fully parame-
terize two-state transcriptional bursting models (Fig. 1C), which
always include at least three unknown parameters: the rate of
transition to a transcriptionally active state (kon), the rate of tran-
sitioning to a transcriptionally inactive state (koff), and the rate
of transcription once in the active state (km) (13, 14). Analyses
of a single two-dimensional plane (e.g., CV2versus expression
level) cannot fully determine these three rate parameters. Con-
versely, analyses of CV2versus expression level and τ1∕2versus
expression level allow the determination of these three para-
meters, and analysis of CV2-versus-τ1∕2facilitates direct compar-
isons of data containing widely varying expression levels, because
it removes the reciprocal dependence of noise magnitude on
expression level (10).
The ability to accurately quantify these transcriptional rate
parameters is essential for answering basic questions about the
mechanisms that regulate transcription. Previous studies ele-
gantly applied flow cytometry (6, 7) and time-lapse microscopy
(1, 15, 16) to analyze gene-expression noise in large subsets of
genes. However, a tedious experimental bottleneck of subcloning
and expansion of isogenic populations necessarily limits the
throughput of these noise-analysis approaches. Here, we circum-
vent this subcloning requirement to globally apply the analytical
Author contributions: R.D.D., B.S.R., C.D.C., M.L.S., and L.S.W. designed research; R.D.D.,
B.S.R., T.V.T., and J.M.M. performed research; R.D.D., B.S.R., C.D.C., M.L.S., and L.S.W.
contributed new reagents/analytic tools; R.D.D., B.S.R., A.S., T.V.T., J.M.M., M.L.S., and
L.S.W. analyzed data; and R.D.D., B.S.R., C.D.C., M.L.S., and L.S.W. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1R.D.D. and B.S.R. contributed equally to this work.
2Present address: Department of Electrical and Computer Engineering, University of
Delaware, Newark, DE 19716.
whomcorrespondence may beaddressed:E-mail: email@example.com
This article contains supporting information online at www.pnas.org/lookup/suppl/
17454–17459 ∣ PNAS ∣ October 23, 2012 ∣ vol. 109 ∣ no. 43www.pnas.org/cgi/doi/10.1073/pnas.1213530109
framework of noise space across the human genome and quantify
these transcriptional rate parameters. The analysis addresses two
specific questions regarding transcriptional regulation in human
cells. (i) Does constitutive (Poisson-like) expression or episodic
(bursty) expression dominate throughout the human genome?
(ii) Does genomic location influence either mode of expression?
For example, if bursting is operant, does genomic location influ-
ence burst size or burst frequency, and which is predominantly
influenced? Recent studies (17–19) have tackled these questions
for specific genes, but no clear and broad consensus has yet
To globally apply the analytical framework of noise space to
screen for constitutive versus bursty expression across the human
genome, we capitalized on the semirandom pattern of integration
exhibited by the HIV-1 lentivirus, where the majority of integra-
tions (approximately 69%) occur within transcriptionally active
regions (20, 21). Jurkat T lymphocytes were infected with HIV-
based lentiviral vectors encoding a short-lived, two-hour half-life
version of green fluorescent protein (referred to as d2GFP), to
generate a library of cells in which the vector is integrated at a
different genomic position in each individual cell (i.e., a poly-
clonal library) (Fig. 2A). To focus on measuring the intrinsic fluc-
tuation dynamics of the genomic region surrounding the vector-
integration site, we utilized a vector encoding the HIV-1 LTR
promoter, which is relatively weak and heavily influenced by the
expression dynamics of the local chromatin environment (22).
Initially, the analytical framework for distinguishing bursty
expression was applied to five isoclonal populations: Each popu-
lation was grown from a single parent cell, and all daughter
cells, therefore, share the same LTR genomic integration site
(Fig. 2 A–D). Cell fluorescence was then imaged for 18 h, and the
resulting fluorescence intensity trajectories were used to con-
struct a three-dimensional noise space with three axes: noise mag-
nitude as measured by the coefficient of variation squared (CV2),
noise autocorrelation represented by the half-autocorrelation
time (τ1∕2), and mean expression level (hGFPi) (Fig. 2 B–C and
SI Appendix). Analysis of these trajectories in the noise space
allows comparisons between isoclonal populations and differen-
tiation between isoclones that exhibit constitutive transcription
versus episodic bursts of transcription (Fig. 2D). To generate an
initial baseline origin for noise-space analysis of the single-cell
data, we identified isoclones that were the most Poissonian in
their expression fluctuation profiles from a library of isoclonal
populations (17) (SI Appendix). Two isoclonal populations exhi-
biting the fastest fluctuation autocorrelation decays (i.e., shortest
autocorrelation times) were selected as the most Poissonian and
used to establish an origin of the CV2-versus-τ1∕2noise map
(Fig. 2D, clones 1 and 2). The isoclone heat map represents the
probability of where a randomly chosen single cell would land on
the noise map (Fig. 2D).
Importantly, nontranscriptional phenomena that influence
noise behavior (e.g., protein and mRNA lifetimes, GFP matura-
tion, extrinsic noise) are already embedded in the noise of these
isoclonal populations, specifically the most Poissonian reference
clones. Thus, comparison between these necessarily precludes the
possibility that noise-map shifts are due to these nontranscrip-
tional phenomena. In addition, the high-frequency noise-proces-
sing technique minimizes extrinsic noise effects (SI Appendix)
(12). Because it is likely that these two isoclones are somewhat
bursty in their expression (SI Appendix), comparisons to these iso-
clones represent a highly conservative assay for bursty expression.
Nevertheless, the results show that even in a small panel of five
clones, there are marked changes in both noise magnitude and
autocorrelation time of approximately 1.5-fold in normalized
space (Fig. 2D). Clones 4 and 5 show significant changes in auto-
correlation time with smaller differences in noise magnitude.
These differences in noise magnitude were validated against con-
ventional flow cytometry measurements (SI Appendix, Fig. S1).
The agreement between data from the 12–18 h microscopy
experiments and flow cytometry demonstrates that this CV2-ver-
sus-τ1∕2analysis has the fidelity to differentiate transcriptional
dynamics between different isoclones.
We next extended the analysis to image polyclonal populations
—consisting of thousands of integration sites (Fig. 3A)—to glob-
ally apply the analytical framework of noise space to screen for
constitutive versus bursty expression across the human genome.
We analyzed more than 8,000 distinct genomic loci with three
different promoters integrated throughout the genome by ima-
ging cells for 12–18 h (Fig. 3B). The constitutive expression origin
of the CV2-versus-τ1∕2noise map previously determined for
the most-Poissonian isoclones (Fig. 2D and SI Appendix) was
compared to the 8,000 loci. To control for LTR-specific or vector-
specific artifacts, we also tested self-inactivating lentiviral vectors
models of transcription across the genome. (A and B) Schematics of the con-
stitutive, Poisson-expression model and the episodic, bursty gene-expression
model, together with three expression trajectories from hypothetical geno-
mic loci. Sites that exhibit constitutive (i.e., Poisson) expression exhibit small
and relatively fast fluctuations in gene products over time. Alternatively,
loci that exhibit episodic expression bursts generate large, slow fluctuations
in gene expression. (C) The principle of noise space. The three-dimensional
noise space consists of noise magnitude, noise autocorrelation, and mean
expression level. Small, fast fluctuations have a small noise magnitude and
short autocorrelation times and thus cluster (after normalization) at the
origin of the noise magnitude-autocorrelation plane (gray region, Lower
Left). Large, slow (i.e., bursty) fluctuations have expanded noise magnitude
and extended autocorrelation times (red ovals). The three-dimensional
space can be decomposed into two additional two-dimensional projections
of noise magnitude and noise autocorrelation versus mean expression level
(Lower Center and Lower Right). For episodic-bursty expression, a trajectory’s
noise-space coordinates are invariably shifted away from the constitutive
model into the burst model space depending on changes to their transcrip-
tional parameters (10).
Fluctuations in gene expression to differentiate between alternate
Dar et al.PNAS
October 23, 2012
that encode either promoters for human elongation factor 1α
(EF1A) or human ubiquitin C (UBC) that in turn drive d2GFP.
UBC and EF1A are promoters that drive essential cellular
housekeeping genes: UBC promotes the ubiquitinization cascade
by marking proteins for proteosomal degradation, and EF1A
promotes the GTP-dependent binding of an aminoacyl-tRNA to
ribosomes. UBC and EF1A are among the most abundant pro-
teins in eukaryotic cells, and their promoters exhibit robust high-
level expression across integration sites in different cell types
(23, 24). Virtually all examined genomic loci for all promoters
exhibit noise-map shifts to the upper right (Fig. 3B), indicating
significant bursting in gene expression at virtually all genomic loci.
Similar widespread bursting is also observed in the THP-1 human
monocyte cell line and compared to a constitutive simulation (SI
Appendix, Fig. S2). Both synchronized and unsynchronized cells
exhibit similar shifts in noise space (SI Appendix, Fig. S3), indicat-
ing that transcriptional bursts appear to be common throughout
the cell cycle, as reported (25).
The UBC and EF1A promoters showed markedly lower CV2
values than the LTR promoter (Fig. 3B and SI Appendix, Fig. S4),
which is consistent with our previous study (17) showing that the
LTR promoter displays relatively higher levels of noise than other
eukaryotic promoters in yeast (6, 7). This shift in noise magnitude
is consistent with the well-known transcriptional elongation stall
that characterizes LTR expression (26). This stall results in
vary with genomic location. (A) Cells are infected with
a lentiviral vector expressing a 2-h half-life GFP repor-
ter (d2GFP) at a low multiplicity of infection (moi) to
ensure a single semirandom integration in each cell.
Individual single cells are isolated, grown (creating
isoclone populations), and imaged by time-lapse
fluorescence microscopy. (B and C) Single-cells are
tracked for 12–18 h, and an individual cell’s mean ex-
pression level, variance (σ2), and autocorrelation time
(τ1∕2) are extracted from the time trace (e.g., the
green circle represents a single cell’s noise space coor-
dinate). A constitutive model of gene expression that
displays abundance dependence (bold red arrows
from black model lines) was used to normalize each
cell’s noise magnitude (CV2) and autocorrelation
(τ1∕2). The normalized noise magnitudes and autocor-
relations are plotted in a ΔlogCV2− Δlogτ1∕2noise
map (Left). (D) Consistent shifts to the Upper Right
quadrant in ΔlogCV2− Δlogτ1∕2 space observed
for three LTR isoclones (clones 3, 4, and 5), are indica-
tive of transcriptional bursting relative to the least
bursty isoclones (clones 1 and 2). Bursting dynamics
varies between different clones as evidenced by
shifts in both noise autocorrelation and magnitude.
The isoclonal signature is taken from 18 h trajectories
of 400 cells.
Extracting transcriptional parameters from
To create the polyclonal population, cells are infected with a lentiviral vector
expressing d2GFP so that each cell represents a unique clone harboring a sin-
gle semirandom integration of reporter. (B) Resultant noise maps for over
8,000 individual cell trajectories for the HIV-1 LTR promoter, EF1A promoter,
and UBC promoter. The constitutive origin is derived from Fig. 2D (18 h).
Episodic-bursty expression dominates across the human genome. (A)
www.pnas.org/cgi/doi/10.1073/pnas.1213530109 Dar et al.
delayed switching to the transcriptional ON state in a two-state
transcription model (17) and predicts that noise frequency (not
only noise magnitude) (27) is modulated in different genomic or
chromatin environments (SI Appendix). In agreement with this
prediction, the distribution of points in the LTR noise map indi-
cates significant differences from the UBCand EF1A noisemaps.
In further support, an alternate representation of the noise-map
distribution that converts the distributions to centroids (with
error bars) can be used to conveniently visualize these differences
between promoters (SI Appendix, Fig. S4) (11).
Given conflicting reports on whether burst frequency varies
with genomic location (17–19, 28), we next determined if tran-
scriptional burst size, burst frequency, or both changed across the
genome (Fig. 4A). As mentioned above, transcriptional bursting
can be quantified by a two-state model of transcription (13, 14) in
which switching between the two states occurs at rates konand
koff, and transcription only occurs in the ON state with a rate
km(Eqs. 1–3 and Materials and Methods). The burst size, or num-
ber of mRNAs generated per activity pulse, is typically defined
as km∕koffand, in the limit of koff≫ kon, the burst frequency
is defined as kon(Fig. 4A) (17). To directly test if transcriptional
burst frequency changes across genome location, we analyzed the
polyclonal three-dimensional noise-space data to fit values for
km, koff, and kon. LTR polyclonal trajectories are subclustered
into groupings of approximately 60 cells, so that each subcluster
represents cells in a specific range of gene-expression levels (SI
Appendix, Fig. S5), and average noise autocorrelation is calcu-
lated for each subcluster by autocorrelation analysis (12). To
validate this subclustering approach, we verified that CV2values
from subclustered polyclonal trajectories agree with CV2from
conventional flow cytometry data from isoclonal populations
(Compare Fig. 4C to SI Appendix, Figs. S1 and S6) . Thus, each
polyclonal subcluster corresponds to an isoclonal population in
terms of average expression level (SI Appendix, Fig. S6). Strik-
ingly, this genome-wide data demonstrate that autocorrelation
time first increases with increasing expression and once an ex-
pression threshold is reached (gray line, Fig. 4B), autocorrelation
time decreases as expression level increases (Fig. 4B). This pat-
tern of concavity is inconsistent with constant burst frequency
(i.e., constant kon) across genomic locations and provides a
genome-wide measurement of konand koffchanges that is inde-
pendent of km(10).
Conventional approaches to quantify transcriptional burst
kinetics analyze noise magnitude (CV2-versus-hGFPi) (6, 7) and
the polyclonal data from Fig. 3 can also be analyzed in terms of
noise magnitude on the CV2-versus-hGFPi plane of noise space
(Fig. 4C). This noise magnitude analysis shows a strong initial
decrease in CV2at low expression levels (Fig. 4C). Then when
an expression threshold is reached (gray line, Fig. 4C) a leveling
off of CV2is observed for higher expression levels (Fig. 4C).
However, noise magnitude is insufficient to uniquely parameter-
ize the two-state model. Because burst size couples kmand koff,
it is only through the τ1∕2measurement, which is not influenced
by km, that the two parameters can be differentiated from one
another. Note, that the gray line in Fig. 4 B and C correspond to
the same expression threshold.
Fitting of the two-state model in the polyclonal three-dimen-
sional noise map space shows a strong initial increase in burst
frequency at low expression levels, whereas burst size remains
almost constant (Fig. 4 D and E). Upon reaching a threshold
expression level (gray vertical line in Fig. 4D), a switch in burst
dynamics occurs, and burst size increases, whereas burst fre-
quency remains constant (Fig. 4 D and E). The fold change in
transcriptional burst size and burst frequency values reveals that
both vary equally across genomic loci (Fig. 4F and SI Appendix,
Fig. S6). In addition, the measured burst-size range predicts an
average mRNA level of 110 molecules per cell, which is consis-
tent with previous measurements that used single-molecule
mRNA fluorescent in situ hybridization (29, 30). The success of
fitting the three-dimensional noise space is reflected in the close
agreement between a simulated autocorrelation curve and the
experimental trend (Fig. 4B) with the fit model parameters. This
fit shows that the assumed two-state model is sufficient to de-
scribe the measured system (SI Appendix). Much like LTR,
UBC, and EF1A display similar fold changes in burst size and
frequency and exhibit a similar pattern of increasing burst fre-
quency, followed by increasing burst size (Fig. 4G). These data
indicate that integration site influences burst kinetics, irrespective
of promoter type (i.e., cis sequence). However, UBC and EF1A
exhibit almost constant τ1∕2at the highest expression levels indi-
cating increases in only kmat these levels (SI Appendix, Fig. S7).
Interestingly, these two strong promoters individually span the
range of burst frequencies recently reported for a variety of
mammalian genes (18) (SI Appendix, Fig. S7), whereas the LTR
functions at much lower burst frequencies (Fig. 4).
To test how transcriptional activators influence burst dynamics
(Fig. 4), transcription was perturbed with transcriptional
activators, including the histone deacetylase inhibitor (HDAC)
trichostatin A (TSA), and the cell-signaling molecule tumor ne-
crosis factor α (TNF). TNF enhances expression by stimulating
the genome and are strongly dependent on expression level. (A) Schematic
of the two-state model of transcriptional bursting, where the promoter
switches between ON and OFF states at rates konand koffand transcribes at
rate kmin the ON state. Transcriptional dynamics are modulated through
changes in burst size, burst frequency, or both. (B) Noise autocorrelation,
noise magnitude (C), burst frequency (D), and burst size (E) versus abundance
for polyclonal subclusters of 2,000 12-h Ld2G single-cell trajectories. Low and
high abundance domains are separated by a solid gray threshold line which
indicates the changes in the trends of noise autocorrelation, noise magni-
tude, and hence burst size and burst frequency is observed. (F and G) As a
function of hGFPi, fold changes in burst size and frequency are comparable,
with an initial increase of frequency in all promoters investigated.
Transcriptional burst frequency and burst size vary equally across
Dar et al.PNAS
October 23, 2012
recruitment of a p50-RelA heterodimer to nuclear factor κB
(NFκB) binding sites (31) and the HIV-1 LTR encodes multiple
NFκB binding sites and is potently activated by TNF (22). We
previously reported that TNF only changes burst frequency of
the LTR while conserving burst size in a limited number of
isoclones (17), and were interested to see how widespread this
phenomena was across the genome. The τ1∕2-versus-hGFPi ana-
lysis of TNF stimulation shows a significant decrease in τ1∕2with
increasing expression level (Fig. 5A). Here, τ1∕2decrease with
increasing abundance is a direct indication of kinetic changes
and demonstrates that increasing expression level cannot be
explained solely by modulations of km(SI Appendix, Eq. S1)
Fitting of the different two-dimensional planes of three-
dimensional noise space upon TNF induction (Fig. 5 A and B)
demonstrates that both burst frequency and size significantly
increase as expression levels increase, with burst frequency
increasing at low expression levels and burst size increasing at
higher expression levels (Fig. 5 C and D). Interestingly, there
appears to be a threshold in expression level, above which kon
plateaus to values observed before adding TNF, and koffappears
to decrease. These data and analysis are corroborated by con-
ventional flow cytometry measurements of 35 isoclonal popula-
tions (SI Appendix, Fig. S6). Overall, these results suggest that
TNF induces expression from the LTR along existing burst trends
(Fig. 5C and D), and the use of TSA, which induces expression
through a different mechanism than TNF (32), corroborates this
observation (SI Appendix). This observed decrease in koffwith
TNF induction leads to extended duration of bursts and is con-
sistent with the reported inhibition of p50-HDAC1 repressive-
complex formation at LTR NFκB sites by p50/RelA heterodimers
(33)—the successful formation of HDAC1 leads to weakened
recruitment of RNA polymerase II and weakened transcriptional
initiation (34). The observed increases in konare also consistent
with increased recruitment of RNA polymerase II to the LTR
promoter NFκB sites induced by TNF (35, 36). Fitted parameter
estimates of LTR residency time in the presence of TNF were
used to represent an average over the first 12 h of stimulation
given the dynamic nonlinear nature of the NFκB response (37,
38). Collectively, these results enable estimation of LTR residency
time in the transcriptional ON and OFF states and show that TNF
extends duration in the ON state up to eightfold (Fig. 5E).
put method to dynamically profile gene-regulatory mechanisms
and the effects of perturbations on gene expression. A significant
methodological advantage of analyzing three dimensions of noise
space is the ability to more accurately constrain two-state transcrip-
tional burst models and the polyclonal nature of the approach en-
ables shotgun mapping of gene regulation dynamics on a genome-
The resulting genome-wide data demonstrate that constitutive
transcription is rare across the human genome. Instead, the
overwhelming majority of human genomic loci appear to stochas-
tically fire in episodic bursts. Analysis of noise space demon-
strated that both transcriptional burst frequency and burst size
vary in roughly equal degree across the human genome
(Fig. 4 D–G). Intriguingly, there appears to be a threshold expres-
sion level below which integrations modulate only transcriptional
burst frequency and above which only burst size is modulated
(Fig. 4 B–G and SI Appendix, Figs. S6 and S7). This transition
could result from recently reported refractory periods inherent to
bursting kinetics (18, 39). Burst frequency can be increased at
loci where transcriptional bursts are infrequent, but as frequency
increases, the refractory period temporally precludes further
increases in frequency. Therefore, once this frequency ceiling is
reached, the only way to increase expression is to increase the
transcription rate or extend the duration of each burst.
As proposed (40), widespread episodic bursting may allow
limited transcriptional resources within the cell to be efficiently
allocated to achieve high-level transcription across large numbers
of loci. This efficient allocation of resources may be the biological
analog of time-domain multiplexing approaches used to effi-
ciently transmit data in signal processing applications.
Materials and Methods
Lentiviral Vectors. Lentiviral vectors were cloned as described (41) and used to
infect 5 × 105Jurkat cells at a multiplicity of infection <0.1, resulting in
25,000–50,000 infected cells each with a unique integration site. Cells were
then sorted by FACS and fluorescentlyimaged on glass-bottom dishes in RPMI
medium 1640 with 10% fetal calf serum and 1% penicillin-streptomycin.
Imaging. Imaging was performed in humidified conditions at 37°C and 5%
CO2for 12–24 h with a 40X (1.2 N.A.) oil-immersion objective on an Olympus
DSU microscope equipped with an automated linear-encoded X-Y stage, as
described in refs. (12) and (42). Image processing and cell tracking were
performed in Matlab with an in-house algorithm (12) and a single 12-h
experiment could generate up to 1,000 trajectories for analysis.
Calculations. For each trajectory, noise autocorrelation (τ1∕2) and noise mag-
nitude (CV2) were calculated using an established noise-processing algorithm
(11, 12). A reported theory (10, 13) of the two-state transcriptional bursting
model yields analytical expressions for both the autocorrelation of the noise,
τ1∕2, and the noise magnitude, CV2(see SI Appendix, Materials and Methods).
Detailed discussion of the noise-mapping approach, including analytical
arguments and stochastic simulations (of Poissonian gene expression) by
the Gillespie algorithm (43, 44), are described in detail in SI Appendix.
Transcriptional burst dynamics are quantified by deriving analytical ex-
pressions for burst size and burst frequency with formulations from previous
analyses (13, 14, 17) and low promoter activity assumptions where koff≫ kon,
koff≫ km, koff≫ γpand km≫ ðγmþ γpÞ):
CV2¼C1ð1 þ BSÞ
hpi ¼BF · BS · kp
tional activators. (A–D) TNF addition (filled red circles) shifts the measured
integration sites to the higher abundance and burst dynamic domain along
the nondrug curve (empty circles). Large autocorrelation shifts implicate
changes in burst kinetics. (E) Estimated residence times in the active (ON)
and inactive (OFF) states.
Transcriptional burst size and frequency are altered by transcrip-
www.pnas.org/cgi/doi/10.1073/pnas.1213530109Dar et al.
BS ¼CV2· hpi Download full-text
BS · C1· C2
where BS is the burst size, BF is the burst frequency, kmis the transcription
rate, kpis the translation rate, γmand γpare the mRNA and protein decay
rates, respectively, hPi is the mean protein abundance, and b is the transla-
tional burst rate. hPi, or the mean number of GFP molecules in the measure-
ments, is assumed to be directly proportional to hFLi, the mean fluorescence
intensity. Eqs. 2 and 3 reveal that measurements of CV2and hFLi are sufficient
to quantify burst size and burst frequency within a constant, which is only
dependent on the translation rate and decay rates of mRNA and protein.
Assuming these remain constant, while varying integration site or promoter
sequence, an abundance-dependent burst size and frequency trend can be
Note. Full decomposition of koff, kon, and kmcan only be performed through
the use of the full 3D noise space.
ACKNOWLEDGMENTS. We thank Hana El-Samad, Ido Golding, Jim Kadonaga,
Laurie Boyer, Alex Hoffmann, John Cooke, David Karig, and members of the
Weinberger and Simpson labs for helpful comments. B.S.R. was supported by
National Science Foundation Graduate Research Fellowship Grant 1144247
and by National Institutes of Health (NIH) Molecular Biophysics Training
Grant GM08326. L.S.W. acknowledges support from the Pew Scholars Pro-
gram in the Biomedical Sciences and the Alfred P. Sloan Research Fellowship
Program. This work was supported by the NIH Director’s New Innovator
Award (OD006677) (to L.S.W.), the in-house research program at the Center
for Nanophase Materials Sciences at Oak Ridge National Laboratory (spon-
sored by the Office of Basic Energy Sciences, US Department of Energy)
(to R.D.D. and M.L.S.), and by the National Institute of General Medical
Sciences National Systems Biology Centers at University of California at
San Diego (P50 GM085764) and University of California, San Francisco (P50
1. Taniguchi Y, et al. (2010) Quantifying E. coli proteome and transcriptome with single-
molecule sensitivity in single cells. Science 329:533–538.
2. Yunger S, Rosenfeld L, Garini Y, Shav-Tal Y (2010) Single-allele analysis of transcription
kinetics in living mammalian cells. Nat Methods 7:631–633.
3. Golding I, Paulsson J, Zawilski SM, Cox EC (2005) Real-time kinetics of gene activity in
individual bacteria. Cell 123:1025–1036.
4. Pedraza JM, Paulsson J (2008) Effects of molecular memory and bursting on fluctua-
tions in gene expression. Science 319:339–343.
5. Cai L, Friedman N, Xie XS (2006) Stochastic protein expression in individual cells at
the single molecule level. Nature 440:358–362.
6. Newman JR, et al. (2006) Single-cell proteomic analysis of S. cerevisiae reveals the
architecture of biological noise. Nature 441:840–846.
7. Bar-Even A, et al. (2006) Noise in protein expression scales with natural protein
abundance. Nat Genet 38:636–643.
8. So LH, et al. (2011) General properties of transcriptional time series in Escherichia coli.
Nat Genet 43:554–560.
9. Blake WJ, M KÆrn, Cantor CR, Collins JJ (2003) Noise in eukaryotic gene expression.
10. Cox CD, McCollum JM, Allen MS, Dar RD, Simpson ML (2008) Using noise to probe and
characterize gene circuits. Proc Natl Acad Sci USA 105:10809–10814.
11. Austin DW, et al. (2006) Gene network shaping of inherent noise spectra. Nature
12. Weinberger LS, Dar RD, Simpson ML (2008) Transient-mediated fate determination in
a transcriptional circuit of HIV. Nat Genet 40:466–470.
13. Simpson ML, Cox CD, SaylerGS (2004) Frequency domain chemical Langevin analysis of
stochasticity in gene transcriptional regulation. J Theor Biol 229:383–394.
14. Kepler TB, Elston TC (2001) Stochasticity in transcriptional regulation: Origins,
consequences, and mathematical representations. Biophys J 81:3116–3136.
15. Sigal A, et al. (2006) Variability and memory of protein levels in human cells. Nature
16. Cohen AA, et al. (2008) Dynamic proteomics of individual cancer cells in response to a
drug. Science 322:1511–1516.
17. Singh A, Razooky B, Cox CD, Simpson ML, Weinberger LS (2010) Transcriptional
bursting from the HIV-1 promoter is a significant source of stochastic noise in HIV-1
gene expression. Biophys J 98:L32–L34.
18. Suter DM, et al. (2011) Mammalian genes are transcribed with widely different
bursting kinetics. Science 332:472–474.
19. Skupsky R, Burnett JC, Foley JE, Schaffer DV, Arkin AP (2010) HIV promoter integration
site primarily modulates transcriptional burst size rather than frequency. PLoS Comput
20. Mitchell RS, et al. (2004) Retroviral DNA integration: ASLV, HIV, and MLV show distinct
target site preferences. PLoS Biol 2:1127–1137.
21. Schroder AR, et al. (2002) HIV-1 integration in the human genome favors active genes
and local hotspots. Cell 110:521–529.
22. Jordan A, Defechereux P, Verdin E (2001) The site of HIV-1 integration in the human
genome determines basal transcriptional activity and response to Tat transactivation.
EMBO J 20:1726–1738.
23. Kim DW, Uetsuki T, Kaziro Y, Yamaguchi N, Sugano S (1990) Use of the human elonga-
tion factor 1 alpha promoter as a versatile and efficient expression system. Gene
24. Ramezani A, Hawley TS, Hawley RG (2000) Lentiviral vectors for enhanced gene
expression in human hematopoietic cells. Mol Ther 2:458–469.
25. Harper CV, et al. (2011) Dynamic analysis of stochastic transcription cycles. PLoS Biol
26. Kao SY, Calman AF, Luciw PA, Peterlin BM (1987) Anti-termination of transcription
within the long terminal repeat of HIV-1 by tat gene product. Nature 330:489–493.
27. Sakane N, et al. (2011) Activation of HIV transcription by the viral Tat protein requires
a demethylation step mediated by lysine-specific demethylase 1 (LSD1/KDM1). PLoS
28. Lo MY, Rival-Gervier S, Pasceri P, Ellis J (2012) Rapid transcriptional pulsing dynamics
of high expressing retroviral transgenes in embryonic stem cells. PLoS One 7:e37130.
29. Raj A, Peskin CS, Tranchina D, Vargas DY, Tyagi S (2006) Stochastic mRNA synthesis in
mammalian cells. PLoS Biol 4:1707–1719.
30. Singh A, Razooky B, Dar RD, Weinberger LS (2012) Dynamics of protein noise can
distinguish between alternative sources of gene-expression variability. Mol Syst Biol
31. Vallabhapurapu S, Karin M (2009) Regulation and function of NF-kappaB transcription
factors in the immune system. Annu Rev Immunol 27:693–733.
32. VanLint C, Emiliani S, Ott M, Verdin E (1996) Transcriptional activation and chromatin
remodeling of the HIV-1 promoter in response to histone acetylation. EMBO J
33. Hayden MS, Ghosh S (2004) Signaling to NF-kappaB. Genes Dev 18:2195–2224.
34. Williams SA, et al. (2006) NF-kappaB p50 promotes HIV latency through HDAC recruit-
ment and repression of transcriptional initiation. EMBO J 25:139–149.
35. West MJ, Lowe AD, Karn J (2001) Activation of human immunodeficiency virus
transcription in Tcells revisited: NF-kappaB p65 stimulates transcriptional elongation.
J Virol 75:8524–8537.
36. Barboric M, Nissen RM, Kanazawa S, Jabrane-Ferrat N, Peterlin BM (2001) NF-kappaB
binds P-TEFb to stimulate transcriptional elongation by RNA polymerase II. Mol Cell
37. Williams SA, Kwon H, Chen LF, Greene WC (2007) Sustained induction of NF-kappa B
is required for efficient expression of latent human immunodeficiency virus type 1.
J Virol 81:6043–6056.
38. Hoffmann A, Levchenko A, Scott ML, Baltimore D (2002) The IkappaB-NF-kappaB
39. Suter DM, Molina N, Naef F, Schibler U (2011) Origins and consequences of transcrip-
tional discontinuity. Curr Opin Cell Biol 23:657–662.
40. Cai L, Dalal CK, Elowitz MB (2008) Frequency-modulated nuclear localization bursts
coordinate gene regulation. Nature 455:485–490.
41. Weinberger LS, Burnett JC, Toettcher JE, Arkin AP, Schaffer DV (2005) Stochastic
gene expression in a lentiviral positive-feedback loop: HIV-1 Tat fluctuations drive
phenotypic diversity. Cell 122:169–182.
42. Weinberger LS, Shenk T (2007) An HIV feedback resistor: Auto-regulatory circuit
deactivator and noise buffer. PLoS Biol 5:67–81.
43. Gillespie DT (1976) General method for numerically simulating stochastic time evolu-
tion of coupled chemical-reactions. J Comput Phys 22:403–434.
44. McCollum JM, Peterson GD, Cox CD, Simpson ML, Samatova NF (2006) The sorting
direct method for stochastic simulation of biochemical systems with varying reaction
execution behavior. Comput Biol Chem 30:39–49.
Dar et al.PNAS
October 23, 2012