An Unbiased Analysis Method
to Quantify mRNA Localization Reveals
Its Correlation with Cell Motility
Hye Yoon Park,1,2Tatjana Trcek,1Amber L. Wells,1Jeffrey A. Chao,1and Robert H. Singer1,2,*
1Department of Anatomy and Structural Biology
2Gruss Lipper Biophotonics Center
Albert Einstein College of Medicine, Bronx, NY 10461, USA
a large fraction of transcripts to restrict its translation
to specific cellular regions. Although current high-
resolution imaging techniques provide ample infor-
mation, the analysis methods for localization have
either been qualitative or employed quantification
in nonrandomly selected regions of interest. Here,
we describe an analytical method for objective
quantification of mRNA localization using a com-
bination of two characteristics of its molecular dis-
tribution, polarization and dispersion. The validity of
the method is demonstrated using single-molecule
FISH images of budding yeast and fibroblasts. Live-
cell analysis of endogenous b-actin mRNA in mouse
fibroblastsrevealsthatmRNApolarization has ahalf-
life of ?16 min and is cross-correlated with directed
cell migration. This novel approach provides insights
into the dynamic regulation of mRNA localization and
its physiological roles.
Cells achieve polarity in part by localization of mRNA, which
allows protein synthesis to be confined to a subcellular com-
partment (Meignin and Davis, 2010). In the studies of mRNA
localization in somatic cells, visualization has played a crucial
role since the first observation almost 30 years ago (Jeffery
et al., 1983; Lawrence and Singer, 1986). Although in situ hy-
bridization is still widely considered as the standard tool, a
variety of techniques in imaging and labeling have enabled
detection of single RNA molecules not only in fixed cells (Femino
et al., 1998) but also in live cells in real time (Bertrand et al.,
Although imaging techniques are highly sophisticated, anal-
ysis of mRNA localization has been mostly limited by the qualita-
tive interpretation. For instance, transcripts expressed during
Drosophila embryogenesis are classified into ?35 localization
categories (Le ´cuyer et al., 2007). A conventional method to
quantify RNA localization is based on manual counting of
the cells by two independent observers who are blind to the
experimental conditions. Although it may eliminate potential
bias in data selection and processing, ambiguities still remain
analysis, the importance of quantitative measurement has
been recognized for several reasons. First, localization can be
described as a continuous process rather than an ‘‘all-or-
nothing’’ occurrence (Zimyanin et al., 2008). Second, a metric
for localization could identify distinct populations that may be
missed by a binary analysis. Moreover, quantification could be
automated to facilitate high-throughput image analysis where
differential response of cells could be examined under manifold
A few quantification methods have been suggested in the
literature to analyze subcellular localization of mRNA. In one
approach, the most dense and least dense regions of the cell
were selected, and the ratio of the highest RNA density to
the lowest density was calculated (Lawrence and Singer,
1986). Latham et al. (1994) counted a cell as localizing if 80%
signal was concentrated in leading lamella area comprising
a quarter of the cell area. More recently, Yamagishi et al.
(2009) selected regions of interest (ROI), and calculated the
ratio of the mRNA concentration at the leading edge and in
the perinuclear region. For the transcripts that localize in the
vicinity of a certain cellular compartment, the distance between
the mRNA and the cellular objects may be used to quantify
localization (Jourdren et al., 2010). These varying analysis
methods could lead to significantly different conclusions, and
demonstrate the need for an objective quantitative analysis of
Here we demonstrate an analytical method applied to three
different cell types: budding yeast, chicken embryonic fibro-
blasts (CEF), and mouse embryonic fibroblasts (MEF). We intro-
duce two measures to characterize mRNA distribution, namely
polarization and dispersion. When mRNA distribution is asym-
metric in a cell, the centroid of the mRNA should deviate from
the centroid of the cell. Wedefine the displacement vector point-
ing from the center of the cell to the center of mRNA as the RNA
polarization vector (Figure 1A). The polarization index (PI) is
determined by dividing the size of the polarization vector by
the radius of gyration of the cell (Rgcell)
Cell Reports 1, 179–184, February 23, 2012 ª2012 The Authors 179
ðxRNA? xcellÞ2+ðyRNA? ycellÞ2+ðzRNA? zcellÞ2
where xRNA, yRNA, zRNAare the coordinates of the centroid for the
RNA, and xcell, ycell, zcellare those for the cell in a general three-
dimensional case. The radius of gyration Rgcellis calculated by
the root-mean-square distance of all pixels within the cell from
the centroid of the cell. The displacement between the two
centroids is divided by Rgcellin order to assess the polarization
normalized to the size and the elongation of the cell. The second
quantity, the dispersion of mRNA, ismeasured bycalculating the
second moment m2of RNA positions:
ðxi? xRNAÞ2+ðyi? yRNAÞ2+ðzi? zRNAÞ2o
where N is the total number of mRNA molecules, xi, yi, ziare the
coordinates of the ith mRNA, and s2
the positions. The second moment is dependent on the shape
and size of the cell as well as the RNA distribution. To normalize
the effect from the cell morphology, we divide the second
moment of mRNA m2by the second moment of the hypothetical
uniform distribution m0
generated and m0
pixel’s coordinates within the mask
zare the variances of
2. A binary mask image of each cell is
2is calculated as the second moment of each
ðXi? xRNAÞ2+ðYi? yRNAÞ2+ðZi? zRNAÞ2o
where M is the total number of pixels within the mask, and Xi, Yi,
Ziare the coordinates of the ith pixel. Then we define the disper-
sion index (DI) as
DI has by definition a value of 1 if the cell has an absolutely
uniform distribution of mRNA (Figure S1A). When mRNA is con-
centrated ina certain region, DIis lessthan 1(Figures S1B–S1D).
If mRNA is spread around the rim of the cell, DI becomes larger
than 1 (Figure S1E).
To test the two-quantity approach (PI and DI), we analyzed
localization of the budding yeast ASH1 mRNA. ASH1 expression
is cell-cycle regulated and reaches a peak during mitosis. In
wild-type (WT) cells, the majority of ASH1 transcripts are actively
localized to the bud (Long et al., 1997) (Figure 1B). Among
several proteins that regulate ASH1 localization, She2p is the
primary RNA-binding protein (Niessing et al., 2004). In the
absence of She2p, ASH1 mRNA becomes homogeneously dis-
tributed between the mother cell and the bud tip (Long et al.,
1997) (Figure 1D). We performed single-molecule fluorescence
in situ hybridization (FISH) (Femino et al., 1998) on WT and
DSHE2 cells as described previously (Trcek et al., 2012; Zenklu-
sen et al., 2008). Multiple fields were imaged and maximum
intensity projections of Z stacks were generated for two-dimen-
sionalanalysis.Byusingaleast-squares Gaussian fittingroutine,
weobtained the intensity and spatial information of each fluores-
cent particle and processed them to quantify polarization and
dispersion of RNA distribution. In order to analyze the posttran-
scriptional localization of mRNA, the multiple copies of mRNA at
the transcription sites were excluded from the analysis. In wild-
type strain, PI was 0.71 ± 0.04 (SEM) and DI was 0.50 ± 0.04.
In DSHE2 strain, PI was 0.16 ± 0.02 and DI was 0.82 ± 0.02.
Consistent with the human perception of the representative
images (Figures 1B and 1D), the objective metrics indicate that
the highly polarized localization of ASH1 mRNA in wild-type
cells was disrupted by the deletion of SHE2 gene. In order to
assess the relationship between the polarization and dispersion
of mRNA in each strain, we computed the correlation co-
efficient of the two metrics. The Pearson correlation coefficient
was ?0.72 in wild-type and ?0.13 in the DSHE2 strain. In wild-
type cells, higher polarization of ASH1 mRNA was strongly
associated with tighter confinement of the mRNA. This nega-
tive correlation between the polarization and the dispersion
Figure 1. Quantification of ASH1 mRNA
(A) A schematic showing the polarization vector
of mRNA distribution pointing from the centroid
of the cell to the centroid of the single mRNA
(B) Overlay image of ASH1 mRNA molecules
(magenta), and nuclei (blue) in a wild-type cell.
ASH1 mRNAs localize to the daughter bud tip.
Scale bar represents 1 mm.
(C) Scatter plot of polarization index in x axis and
dispersion index in y axis for wild-type cells.
(D) Overlay image of ASH1 mRNA molecules, and
nuclei in a DSHE2 cell. Deletion of She2p causes
a complete delocalization of ASH1 transcripts.
Scale bar represents 1 mm.
(E) Scatter plot of polarization index versus dis-
persion index for DSHE2 cells. The red arrow
heads indicate the data points for the cells shown
in (B) and (D).
See also Figure S1.
180 Cell Reports 1, 179–184, February 23, 2012 ª2012 The Authors
disappeared in DSHE2 strain (Figures 1C and 1E). The cor-
relation coefficient describesthe characteristics ofRNAdistribu-
tion in a certain population of cells.
We next applied the analysis method for the distribution of
bution of b-actin mRNA and GAPDH mRNA in the same cell, we
investigated if localization is a general effect for any mRNA or
a specific effect for b-actin mRNA. Figure 2A shows a represen-
tative image of primary CEF cells in which b-actin mRNA is local-
ized to the leading edge (indicated with arrowheads), whereas
GAPDHmRNAismoreuniformly distributed.Because thedensi-
ties of these mRNA species are too high to distinguish indi-
vidual molecules, the distribution of pixel intensity values were
analyzed instead of the distribution of discrete mRNA particles.
where rijis the distance from the centroid of the cell to the pixel
(i,j) within the cell boundary, and Iijis the intensity value of the
pixel in a two-dimensional image. A representative cell shown
in the left side of Figure 2A exhibits polarized localization of
of GAPDH mRNA (PI = 0.33, DI = 1.17). In a randomly-chosen
population of CEF cells (n = 99), the mean PI was 0.47 ± 0.03
for b-actin mRNA and 0.29 ± 0.02 for GAPDH mRNA. In 82%
of the cells(in the upper triangle above the red dashed line in Fig-
ure 2B), the distribution of b-actin mRNA was more polarized
than GAPDH mRNA. However, the difference in the mean
DI was not as significant: 0.98 ± 0.05 for b-actin mRNA and
0.82 ± 0.04 for GAPDH mRNA. Unlike yeast ASH1 mRNA, the
mean values of PI and DI for b-actin mRNA indicate a moderate
polarization and a loose dispersion on average because of the
intrinsic cell-to-cell variation in primary CEF culture. Previously,
Latham et al. (1994) reported that only 30%–35% of primary
CEF localized b-actin mRNA to the cell periphery using a binary
counting method. These cells showing b-actin mRNA localiza-
tion may represent a subpopulation of motile cells (Kislauskis
et al., 1997). Although the cell population was heterogeneous,
the correlation coefficient between PI and DI for b-actin mRNA
was highly negative (r = ?0.61) in contrast to the low value for
GAPDH mRNA (r = ?0.27) (Figures S2C and S2D). This result
shows that b-actin mRNA tends to exhibit polarized localization
whereas GAPDH mRNA does not. The correlation coefficient
between PI and DI describes the distinct localization property
of each mRNA species in a population of cells.
Finally, we examined the performance of this method to quan-
tify the dynamics of b-actin mRNA localization in living mamma-
lian cells. Because the correlation between PI and DI for b-actin
mRNA was high, the polarization of mRNA distribution was used
as a single metric in live cell analysis. We visualized the endoge-
nous b-actin mRNA using the cells from the Actb-MBS mouse
that contains 24 repeats of MS2 binding site (MBS) cassette in
the 30untranslated region (UTR) of the b-actin gene (Lionnet
et al., 2011). Primary mouse embryonic fibroblasts (MEF) were
isolated from the mouse, infected with lentivirus that expresses
MS2 capsid protein fused with GFP (MCP-GFP) to allow fluo-
rescent labeling of the b-actin transcript, and stained with
membrane-permeable cytoplasmic dye. A nuclear localization
sequence (NLS) was added to the N-terminus of MCP-GFP so
that free NLS-MCP-GFPs preferentially localize in the nucleus
lowering the background in the cytoplasm (Bertrand et al.,
1998). As a negative control, we infected wild-type MEF with
NLS-MCP-GFP expressing lentivirus and observed little fluores-
cence inthe cytoplasm (FigureS3A). IntheActb-MBS MEF cells,
NLS-MCP-GFP binds to MBS-tagged b-actin mRNA and the
complex is exported out to the cytoplasm (Figure S3B). Using
time-lapse imaging, we monitored the localization pattern of
b-actin mRNA as the cell migrated on fibronectin-coated glass
was measured at 1-min interval, and defined as the protrusion
vector. We found that the mRNA polarization vector (yellow
arrows, Figure 3A) and the protrusion vector (red arrows, Fig-
ure 3A) were highly correlated in space and in time. In all of the
migrating cells that we monitored (n = 11), the autocorrelation
curve of protrusion vector decays faster than the one for
mRNA polarization vector, indicating that mRNA localization
is a slower process than random protrusions. From the mean
auto-correlation curves of polarization vector (black curve, Fig-
are estimated to be ?16 min for mRNA localization and ?4 min
for random protrusions. To quantify the relationship between
RNA localization and cell protrusion, we calculated the correla-
tion coefficients of two vectors with varying time lags. Within
our time resolution, the mean correlation was the highest at
zero lag time (Figure 3C). However, the skewed correlation in
the negative time lag suggests that cell protrusion precedes
b-Actin mRNA Distributions in Chicken
(A) FISH image of GAPDH mRNA (red), and b-actin
mRNA (green). Scale bar represents 10 mm.
(B) Scatter plot of polarization index for GAPDH
mRNA in x axis and b-actin mRNA in y axis.
(C) Scatter plot of dispersion index for GAPDH
mRNA in x axis and b-actin mRNA in y axis. Red
dashed lines indicate cells that have the same
index for GAPDH and b-actin mRNA.
See also Figure S2.
2. Comparisonof GAPDH and
Cell Reports 1, 179–184, February 23, 2012 ª2012 The Authors 181
mRNA localization in overall cell movement. It has been shown
that mRNA localization is not necessary for random protrusions,
but required for directed migration (Shestakova et al., 2001). We
found that the net migration distance in 2 hr is highly correlated
with the mean polarization index of the cell (r = 0.75) (Figure 3D).
This result supports that localization of b-actin mRNA has a
physiological role in directed cell migration.
In summary, the combination of polarization and disper-
sion of mRNA distribution enables an effective assessment of
mRNA localization. This approach will allow us to quantify subtle
changes in RNA distribution by gene deletion or mutation, and to
compare the localization characteristics of different transcripts.
Moreover, a quantitative analysis of a single cell in time-lapse
images willfacilitate studies onthe kinetics and thephysiological
role of mRNA localization.
We have demonstrated the utility of the new method in
budding yeast and primary chicken and mouse embryonic fibro-
blasts. There is a requirement for this analysis that the entire
cell area is imaged while maintaining sufficient sensitivity for
detecting RNA signals. In order to apply this method for larger
cells such as neurons(Bassell et al., 1998; Lyles et al., 2006),
Drosophila embryos (Le ´cuyer et al., 2007), and syncytial muscle
cells (Kislauskis et al., 1993; Sigrist et al., 2000), it may be neces-
sary to image multiple fields in three dimensions and stitch them
together to reconstruct the whole specimen (Preibisch et al.,
2009). In case of further application to the cells in vivo in the
tissue environment, it will be useful to outline the cells by cell-
surface markers or plasma membrane stains because the
surrounding tissues may obscure the regions of the cell for anal-
ysis. The simple unbiased analysis of intracellular localization
Figure 3. Localization of b-Actin mRNA in a Migrating Cell
(A) Time-lapse images of a mouse embryonic fibroblast. The color map shows the fluorescence intensity of MCP-GFP labeling endogenous b-actin transcripts
divided by the intensity of red fluorescent cytoplasmic dye. Yellow arrows show the polarization vector of mRNA distribution, and red arrows show the protrusion
vector of the cell.
(B) Mean auto-correlation curves of the polarization vector (black) and the protrusion vector (red) (n = 11 cells). The shaded areas indicate 95% confidence
(C) Mean cross-correlation curve of the polarization vector and the protrusion vector (black curve) and 95% confidence interval (gray area) for n = 11 cells.
(D) Migration distance as a function of the time-average of the polarization index over 2 hr. The correlation coefficient between the polarization index and the
migration distance is 0.75.
See also Figure S3 and Movie S1.
182 Cell Reports 1, 179–184, February 23, 2012 ª2012 The Authors
presented in this work may considerably enrich studies on the
local regulation and function of many mRNA species.
FISH for Yeast
Yeast cells were grown in rich media until they reach early log phase with
OD600 ?0.5 (Table S1). Cells were fixed by adding 8 ml of 32% paraformalde-
hyde to 42 ml of culture for 45 min at room temperature. The rest of the spher-
single stranded DNA probes each of which was 50-nucleotide-long and
labeled with four Cy3 dyes (Table S2). The labeling efficiency of each probe
was above 90%, indicating efficient coupling of fluorescent dye with each
probe. Cells in G2 and mitosis were selected using morphological markers
and analyzed for ASH1 mRNA localization (Trcek et al., 2011).
FISH for Chicken Embryonic Fibroblasts
Primary chicken embryonic fibroblast (CEF) cells were obtained from Charles
River Laboratories (Wilmington, MA). Cells were plated on 10 cm culture
dishes and grown at 37?C in MEM supplemented with 10% FCS in an atmo-
sphere of 95% air/5% CO2for 1–2 days. Cells were trypsinized and seeded
on coverslips at a density of 1 3 105cells/mL. After overnight incubation, cells
were washed with DPBS and fixed in 4% paraformaldehyde in PBS for 20 min.
The coverslips were stored in 70% ethanol at 4?C for a few days and pro-
cessed for fluorescence in situ hybridization as described in Singer lab proto-
cols (http://singerlab.org/protocols/). Slides were imaged using an Olympus
BX-61 microscope equipped with an X-cite 120 PC lamp (EXFO), a UPlanApo
1003 1.35 NA oil immersion objective (Olympus) and a CoolSNAP HQ CCD
camera (Photometrics). We used Chroma filter set 31000 (DAPI), 41007a
(Cy3), and 41008(Cy5). Cells were imaged with0.2 mmZ steps ineach channel
using MetaMorph software (Molecular Devices).
Live Cell Imaging of Mouse Embryonic Fibroblasts
Primary MEFs were cultured from 14-day-old embryos isolated from a preg-
nant female of Actb-MBS mouse (Lionnet et al., 2011). The head and dark
cardiac tissue was removed from the embryo and the rest of the body was
digested with Trypsin EDTA for 20 min. After adding media (DMEM, 10%
FBS, 1% pen/strep), we selected fibroblasts by plating the cells on 10 cm
culture dishes for 1 hr and washing off the unattached cells with fresh media.
Thenextday,cells wereplated onfibronectin-coated MatTek dishes (MatTek),
infected with lentivirus, and incubated for 48 hr to express NLS-MCP-GFP.
We stained the cells with 5 mM CellTracker Orange CMRA (Invitrogen) and
replaced the culture media with L-15 media containing 10% FBS, 1% pen/
strep, and 1% oxyrase (Oxyrase) prior to the experiment. Time-lapse images
were taken on an Olympus IX-71 inverted microscope equipped with a
UApo/340 403 1.35 NA oil immersion objective (Olympus), an MS-2000 XYZ
automated stage (ASI) and an iXon electron-multiplying charge-coupled
device (EMCCD) camera (Andor). The temperature was kept at 37?C with
60% humidity inan environmental chamber (Precision Plastics). The excitation
sources were 488 nm line from an argon ion laser (Melles Griot) and a 561 nm
diode-pumped solid state laser (Cobolt). Multiple fields of images were
acquired every 1 min using MetaMorph (Molecular Devices).
Image segmentation and quantification were performed using custom
designed MATLAB programs. If the cell was larger than the field of view, the
2D/3D Stitching Plugin available through Fiji was used to create a tiled image
(Preibisch et al., 2009). To analyze the fixed-cell data, cells were segmented
using the DIC or auto-fluorescence image. A binary mask was generated
and the centroid of the cell was calculated by the mean value of x- and
y-coordinate of the pixels within the mask region. For single-molecule FISH
images, a maximum intensity projection of Z stacks was used to quantify
mRNA distribution. In budding yeast images, each fluorescent spot was fit
with a two-dimensional Gaussian function, yielding the amplitude and the
position of the fluorescent particle. From the histogram of fluorescence inten-
sity, the majority of the detected spots were identified as single probes bound
nonspecifically. The average intensity of single probes was used to calculate
the number of probes bound in each fluorescent spot inside a cell. Only the
bright spots binding more than four probes were selected to analyze the
number and the position of the target mRNA. The mean and the variance of
the x- and y-coordinates of the RNA molecules except for the ones at the tran-
scription sites were used to calculate the polarization and dispersion indices.
For FISH images of CEFs, the background was determined from the median
intensity value within each cell boundary. After background subtraction, the
intensity-weighted centroid and the second moment were calculated.
To analyze the live-cell data, masks were generated that define the cell and
the nucleus from the CellTracker dye image and the NLS-MCP-GFP image,
respectively. The center of the cell was identified by the intensity-weighted
centroid of the cytoplasmic dye image. After background subtraction, the
NLS-MCP-GFP image was divided by the CellTracker image to obtain RNA
distribution image normalized by the cytoplasmic volume. The center of RNA
was determined by the weighted centroid of the resulting image excluding
thenuclear region. Foreach timepoint, the RNA polarizationvector wascalcu-
lated as the vector pointing from the center of the cell to the center of RNA.
The protrusion vector was defined as the instantaneous velocity of the cell
calculated every 1 min. The net migration distance was determined to be the
distance between the beginning and ending points of the cell trajectory.
Auto-correlation and cross-correlation of any two sets of vectors were
method demonstrated by Weiger et al. (2010). The half-lives of mRNA polari-
zation and cell protrusion were determined at the correlation coefficient of
0.5 from the autocorrelation curves.
Supplemental Information includes three figures, two tables, and one movie
and can be found with this article online at doi:10.1016/j.celrep.2011.12.009.
This is an open-access article distributed under the terms of the Creative
Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported
Microscopy equipment for the live cell imaging experiments was provided by
the Gruss Lipper Biophotonics Center. This work was supported by NIH
GM57071, GM84364 and GM86217 to R.H.S. H.Y.P. was supported by
National Research Service Awards F32-GM087122.
Received: August 30, 2011
Revised: November 22, 2011
Accepted: December 23, 2011
Published online: February 16, 2012
Bassell, G.J., Zhang, H., Byrd, A.L., Femino, A.M., Singer, R.H., Taneja, K.L.,
Lifshitz, L.M., Herman, I.M., and Kosik, K.S. (1998). Sorting of beta-actin
mRNA and protein to neurites and growth cones in culture. J. Neurosci. 18,
Bertrand, E., Chartrand, P., Schaefer, M., Shenoy, S.M., Singer, R.H., and
Long, R.M. (1998). Localization of ASH1 mRNA particles in living yeast. Mol.
Cell 2, 437–445.
Femino, A.M., Fay, F.S., Fogarty, K., and Singer, R.H. (1998). Visualization of
single RNA transcripts in situ. Science 280, 585–590.
Jeffery, W.R., Tomlinson, C.R., and Brodeur, R.D. (1983). Localization of actin
messenger RNA during early ascidian development. Dev. Biol. 99, 408–417.
Cell Reports 1, 179–184, February 23, 2012 ª2012 The Authors 183
Jourdren, L., Delaveau, T., Marquenet, E., Jacq, C., and Garcia, M. (2010). Download full-text
CORSEN, a new software dedicated to microscope-based 3D distance
measurements: mRNA-mitochondria distance, from single-cell to population
analyses. RNA 16, 1301–1307.
Kislauskis, E.H., Li, Z., Singer, R.H., and Taneja, K.L. (1993). Isoform-specific
30-untranslated sequences sort alpha-cardiac and beta-cytoplasmic actin
messenger RNAs to different cytoplasmic compartments. J. Cell Biol. 123,
Kislauskis, E.H., Zhu, X., and Singer, R.H. (1997). beta-Actin messenger
RNA localization and protein synthesis augment cell motility. J. Cell Biol.
Latham, V.M., Jr., Kislauskis, E.H., Singer, R.H., and Ross, A.F. (1994). Beta-
actin mRNA localization is regulated by signal transduction mechanisms. J.
Cell Biol. 126, 1211–1219.
Lawrence,J.B., and Singer, R.H.(1986).Intracellularlocalizationof messenger
RNAs for cytoskeletal proteins. Cell 45, 407–415.
Le ´cuyer, E., Yoshida, H., Parthasarathy, N., Alm, C., Babak, T., Cerovina, T.,
Hughes, T.R., Tomancak, P., and Krause, H.M. (2007). Global analysis of
mRNA localization reveals a prominent role in organizing cellular architecture
and function. Cell 131, 174–187.
Lionnet, T., Czaplinski, K., Darzacq, X., Shav-Tal, Y., Wells, A.L., Chao, J.A.,
Park, H.Y., de Turris, V., Lopez-Jones, M., and Singer, R.H. (2011). A trans-
genicmouse for in vivo detection of endogenous labeled mRNA. Nat. Methods
Long, R.M., Singer, R.H., Meng, X., Gonzalez, I., Nasmyth, K., and Jansen,
R.P. (1997). Mating type switching in yeast controlled by asymmetric localiza-
tion of ASH1 mRNA. Science 277, 383–387.
Lyles, V., Zhao, Y., and Martin, K.C. (2006). Synapse formation and mRNA
localization in cultured Aplysia neurons. Neuron 49, 349–356.
localization. Curr. Opin. Cell Biol. 22, 112–119.
Niessing, D., Hu ¨ttelmaier, S., Zenklusen, D., Singer, R.H., and Burley, S.K.
(2004). She2p is a novel RNA binding protein with a basic helical hairpin motif.
Cell 119, 491–502.
Preibisch, S., Saalfeld, S., and Tomancak, P. (2009). Globally optimal stitching
of tiled 3D microscopic image acquisitions. Bioinformatics 25, 1463–1465.
Shestakova, E.A., Singer, R.H., and Condeelis, J. (2001). The physiological
significance of beta -actin mRNA localization in determining cell polarity and
directional motility. Proc. Natl. Acad. Sci. USA 98, 7045–7050.
Sigrist, S.J., Thiel, P.R., Reiff, D.F., Lachance, P.E., Lasko, P., and Schuster,
C.M. (2000). Postsynaptic translation affects the efficacy and morphology of
neuromuscular junctions. Nature 405, 1062–1065.
Trcek, T., Chao, J.A., Larson, D.R., Park, H.Y., Zenklusen, D., Shenoy, S.M.,
and Singer, R.H. (2012). Single-mRNA counting using fluorescent in situ
hybridization in budding yeast. Nat. Protoc. 7, 408–419.
Trcek, T., Larson, D.R., Moldon, A., Query, C.C., and Singer, R.H. (2011).
Single-molecule mRNA decay measurements reveal promoter regulated
mRNA stability in yeast. Cell 147, 1484–1497.
Weiger, M.C., Ahmed, S., Welf, E.S., and Haugh, J.M. (2010). Directional
persistence ofcell migrationcoincides with stability ofasymmetric intracellular
signaling. Biophys. J. 98, 67–75.
Yamagishi, M., Shirasaki, Y., and Funatsu,T. (2009). Size-dependent accumu-
lation of mRNA at the leading edge of chicken embryo fibroblasts. Biochem.
Biophys. Res. Commun. 390, 750–754.
Zenklusen, D., Larson, D.R., and Singer, R.H. (2008). Single-RNA counting
reveals alternative modes of gene expression in yeast. Nat. Struct. Mol. Biol.
Zimyanin, V.L., Belaya, K., Pecreaux, J., Gilchrist, M.J., Clark, A., Davis, I., and
St Johnston, D. (2008). In vivo imaging of oskar mRNA transport reveals the
mechanism of posterior localization. Cell 134, 843–853.
184 Cell Reports 1, 179–184, February 23, 2012 ª2012 The Authors