|
|
||||||||
Department of Biochemistry and Molecular Biology, University of Massachusetts Medical School, 377 Plantation Street, Worcester, Massachussetts 01605, USA
1Correspondence: Department of Biochemistry and Molecular Biology, University of Massachusetts Medical School, 377 Plantation St., Worcester, MA 01605, USA. E-mail: thoru.pederson{at}umassmed.edu
| ABSTRACT |
|---|
|
|
|---|
Key Words:
| INTRODUCTION |
|---|
|
|
|---|
I would like to present some of the work my colleagues and I have been undertaking over the past ~7 years to understand how RNA molecules get from their DNA birth sites to their correct functional locations throughout the cell. This is a formidable molecular traffic problem from the chemical point of view, and we know (so far) of no Golgi-like sorting apparatus to facilitate the process. Although the proper spatial localization in the cell of the various kinds of RNA is central to gene readout, this aspect of eukaryotic gene expression has, until recently, been much more opaque than our more comprehensive and detailed understanding of transcription, RNA processing, and the various functions of RNA. This state of affairs has been because of, in large part, the lack of ways to study RNA movements inside living cells.
| MATERIALS AND METHODS |
|---|
|
|
|---|
| RESULTS AND DISCUSSION |
|---|
|
|
|---|
In collaboration with Long-guang Cao, the graduate student in Yu-li
Wangs lab, Jin Wang fluorescently labeled several pre-mRNAs and, with
Caos collaboration, microinjected them through glass needles into the
nuclei of rat NRK fibroblasts. (They did not tell me of these
experiments until a few weeks later, when the first suite of them had
been successfully done.) What Jin Wang and Long-guang Cao found was
that the fluorescent pre-mRNA introduced into the nucleus congressed
over a period of 1530 min into a discrete array of sites peppered
throughout the nucleoplasm (1)
. These sites were already
known to be where certain small nuclear ribonucleoprotein (snRNP)
particles are concentrated in the nucleus, (13)
and
because these snRNPs are cofactors for pre-mRNA splicing, the
fluorescent pre-mRNA localization results my student had obtained rang
true (1)
. As an extension of these studies, he fixed the
fluorescent pre-mRNA containing cells and then carried out
immunocytochemistry using monoclonal antibodies to known snRNP
proteins. This revealed a spatial concordance between the sites at
which the pre-mRNA had become localized in the living state and the
nuclear regions of concentrated snRNP proteins (1)
.
Because the fluorescent pre-mRNAs (including ones for the ß chain of
human hemoglobin, a rat neuropeptide pre-mRNA, and an adenovirus
pre-mRNA) were transcribed in vitro from cloned DNAs, it was
possible to study mutant pre-mRNAs. This revealed that a specific
region near the 3' end of the pre-mRNA intron was required for pre-mRNA
localization at the discrete nucleoplasmic sites (1)
. This
was the first demonstration of a sequence-specific localization of
exogenously delivered RNA within the cell nucleus.
The new-cleolus
We then sought to expand this fluorescent RNA cytochemistry
approach. A postdoctoral fellow in my laboratory, Marty Jacobson,
embarked on studies of the nucleolar localization of two small RNAs
using this method and identified nucleolar localization sequences in
one of these RNAs (2
, 3)
. Marty also elected to study the
intranuclear localization of microinjected wild-type and mutant signal
recognition particle (SRP) RNA. This study revealed that SRP RNA
rapidly shuttles to the nucleolus (within 3060 s) and that this
nucleolar localization involves discrete domains within the SRP RNA
molecule (7)
. This finding suggests that the biosynthesis
of the SRP, a cytoplasmic translational arrest machine, involves a
nucleolar station in its intranuclear transit and constitutes part of
the evolving evidence that the nucleolus is the site of various RNA
processing and ribonucleoprotein assembly events beyond its classically
envisioned role in ribosome biosynthesis (14)
. Three of
the RNAs that have been recently linked to the nucleolus [i.e., SRP
RNA (7)
, telomerase RNA (14)
, and U6 RNA
(14)
] are, interestingly, parts of catalytic
ribonucleoprotein machines. The SRP is a GTPase, telomerase is a
reverse transcriptase, and U6 RNA is an essential part of the catalytic
center of the spliceosome. Our fluorescent RNA cytochemistry approach,
particularly as applied to SRP RNA (7)
, was a key factor
in leading me to rethink the nucleolus (14
, 15)
, an
example of how new techniques can revise ones conceptual perspective.
Problems and pitfalls
My students and postdocs have always cringed to hear me say after
reading their fellowship proposals, "Its okay, but what about the
problems and pitfalls?" It is such a vivid (and accurate) term. When
I was a postdoc at Albert Einstein College of Medicine (19681971) my
mentor was a member of the NIH Cell Biology Study Section and asked me
to join him in reading grant applications. I got the sense that the
requested section on problems and pitfalls was a creative challenge
and important opportunity for the applicant, perhaps even the
intellectual moment of truth as regards winning the reviewers
confidence. As a NIH grant applicant soon thereafter and a subsequent
Cell Biology Study Section member (19751979), I had this notion
convincingly confirmed at both ends. What are the problems and
pitfalls of the fluorescent RNA microinjection approach?
The first and most obvious point to consider is that the microinjected
RNAs are tagged by covalently attached fluorescent groups and must
therefore be considered derivatized. This issue is formally the same as
the one that arises in the use of fluorescent proteins for cell
microinjection studies. In our experiments, rhodamine is typically
attached to the RNA at an average spacing of 1 molecule every 3060
nucleotides (1
, 4)
. This balances fluorescent signal
strength with retaining RNA function as defined by demonstrably correct
ribonucleoprotein assembly in vitro (1
2
3)
. A
second issue is whether the volume of introduced fluorescent RNA
(usually 0.010.1 pl) or the fluorescent RNA itself impairs cell
physiology. The fact that many studies have now shown that
nucleus-microinjected RNAs move to the same intranuclear sites as are
occupied by their endogenous RNA counterparts (2
3
4
, 6
, 7
, 16
17
18
19)
indicates that this aspect of cell function, namely RNA
traffic, is not impaired by the presence in the cell of a few thousand
fluorescent RNA molecules (4)
. We have also observed that
cells microinjected with fluorescent RNA divide throughout the
subsequent several hours of observation, indicating that the presence
of fluorescent RNA does not impair progression through the cell cycle.
Thus, fluorescently tagged RNA, as we prepare it (vide
supra), appears to behave properly and does not seem to be
deleterious to cells. Given the fact that many proteins have been found
to behave properly as regards intracellular localization when fused to
an ~27,000 molecular weight tag, namely jellyfish green fluorescent
protein, the rhodamine labeling of RNA we use is comparatively minor on
a molecular mass basis. In certain instances, it may be desirable to
keep a certain functional region of a given RNA nonlabeled. This can be
accomplished by fluorescently labeling the RNA only at one (or both) of
its ends (with an obvious sacrifice of fluorescence intensity per RNA
molecule) or by ligating fluorescent and nonfluorescent blocks of RNAs
that have the desired nucleotide sequence.
Biophysical and physical chemical aspects of intranuclear RNA
motion
The most advanced aspects of this work address the
microscale movement of RNA within the nucleus, captured in
vivo in real time. Fluorescent correlation spectroscopy (FCS) is a
method of fluctuation analysis used to measure molecular transport and
chemical kinetics by tracking the number of molecules entering and
leaving a very small interrogation volume over successive time
intervals (20)
. This method of measuring molecular motion
has recently been integrated with avalanche photodiode detection
systems, having dead times on the order of only 25 ns and advanced
confocal optics and algorithms that model the simultaneous diffusion of
as many as three components moving at different rates
(20
21
22
23)
. These advances have led to the advent of FCS as
a cell biological tool to study molecular movement in living cells. The
Carl Zeiss Company (Jena, Germany) has developed such an
instrument, and we were pleased that Zeiss selected the Worcester
Foundation as one of the test sites. This fortunate circumstance
coincided with our burgeoning interest in more quantitatively measuring
RNA movement in cells.
In my laboratory, Joan Politz had previously developed a method, while
a postdoctoral fellow in the laboratory of Robert Singer (then at the
University of Massachusetts Medical School), to detect poly(A) RNA in
cells by an in vivo nucleic acid hybridization procedure
(24)
. Her method involves incubating rat myoblasts with
oligo(dT) and then detecting the intracellular sites at which the
oligo(dT) has hybridized to the poly(A) tracts of RNA by fixing the
cells and conducting an in situ reverse transcriptase
reaction in which the oligo(dT) that is hybridized to poly(A) primes
synthesis of a complementary DNA (cDNA). The cDNA products are detected
in situ by their incorporation of deoxynucleotides
tagged with digoxigenin or other suitable detectors. In collaboration
with Elizabeth Browne in the laboratory of David Wolf at the Worcester
Foundation, Joan embarked on an investigation of the intranuclear
movement of fluorescent oligo(dT) in rat myoblasts using FCS. This
study provided a number of very interesting findings (25)
.
First, a substantial fraction of the intranuclear oligo(dT), as well as
most of the control oligonucleotideoligo(dA), which is not expected
to hybridize appreciably to nuclear RNAwas observed to move within
the nucleus at mean translation times close to those measured for these
oligonucleotides in aqueous solution by the same FCS method (diffusion
coefficients = ~4x10-7 cm2/s). Second,
a large portion (~45%) of the intranuclear oligo(dT), but not
oligo(dA), moved more slowly (diffusion coefficient =
~1x10-7 cm2/s). The amount of this
slower-moving oligo(dT) fraction was substantially reduced if the
oligo(dT) was prehybridized in solution to oligo(dA) before
introduction into the cells, presumably because the oligo(dT) was
unable to subsequently hybridize to poly(A) RNA in the cell. Third,
this slower-moving oligo(dT), which we interpret as the fraction
hybridized to poly(A) RNA, had the same diffusion coefficient as
oligo(dT) hybridized to a 7.5-kb polyadenylated RNA in aqueous
solution. Fourth, we also detected a small amount of intranuclear
oligo(dT), ~15% of the total, that moved considerably more slowly
than the poly(A) RNA-hybridized fraction. The average diffusion
coefficient of this latter, very slow fraction of oligo(dT),
1 x 10-8 cm2/s, is consistent with it being
hybridized to poly(A) RNA that is itself associated with large
macromolecular complexes. The diffusion coefficient of the oligo(dT)
fraction that we interpret as hybridized to poly(A) RNA
(~1x10-7 cm2/s) is also close to that
calculated from Stokes law for a 7.5-kb poly(A) RNA molecule in the
form of a nuclear hnRNP particle (25)
, namely
2.29.2 x 10-8 cm2/s, assuming that the
ribonucleoprotein particle has a protein:RNA ratio of 4:1 (26
, 27)
and treating the particle as either a spherical or ellipsoid
structure (28)
.
In these studies (25)
we also used a different
biophysical method to investigate the intranuclear movement of
oligo(dT), namely fluorescence recovery after photobleaching (FRAP).
Unlike FCS, which is based on the fluorescence intensity fluctuations
of molecules darting in and out of a small intracellular (in our case,
intranuclear) sampling volume, FRAP measures the rate at which and
extent to which a population of fluorescent molecules moves into a
region where the existing molecules have been rendered nonfluorescent.
Also unlike FCS, FRAP as generally used does not resolve individual
kinetic components having characteristically different diffusion
coefficients. When we measured the intracellular movement of oligo(dT)
by FRAP, the average diffusion coefficient was 1.2 x
10-7 cm2/s (25)
. This value is
close to the intranuclear diffusion coefficient of oligo(dT) measured
by FCS, if the data are treated as a 1-component model, namely
~8.7 x 10-8 cm2/s. This is a
remarkable concordance of diffusion coefficients obtained by two
independent methods.
The overall conclusion from our FCS and FRAP experiments is that
both nonhybridized oligonucleotides and oligo(dT) hybridized to poly(A)
RNA move within the nucleus of living mammalian cells with diffusion
coefficients similar to those measured for the two respective classes
of molecules in aqueous solution. This addresses the physical milieu of
the cell nucleus and suggests that the amount of free volume is
substantial, irrespective of the existence (or nonexistence) of a
nucleoplasmic ground substance or other structural elements that have
been proposed to occupy the interchromatin space (29)
.
Having used the method at its infancy, my colleagues and I think
that FCS is a powerful new tool in the cell biologists kit. Its
promise has been recently reviewed and some of the potential problems
cited (23)
have been overcome in our studies
(25)
. In particular, the level of autofluorescence in the
nucleus of the cells we have used is quite low and has not been
problematic. In addition, the relatively low irradiance used in our FCS
studies, ~100 µJ over the 0.23 µ2 average
cross-sectional area in the Zeiss instruments confocal volume
(25)
, is not thermally invasive. Indeed, similar levels of
irradiance have been previously used (in non-FCS studies) on cells in
Drosophila embryos with no detectable adverse effects
(30)
.
| CONCLUSION |
|---|
|
|
|---|
Perspective
To cell biologists of my generation (I was born in 1941),
Keith Porter carried that all too prosaic label electron
microscopist. But that was by no means the whole story of his talent.
Keith was a passionate advocate of using living cells and indeed was
skilled at the nascent art of tissue culture when it was far from the
routine practice it is today. He also had a most impressive knowledge
of optimal living material among many phyla for revealing important
principles of cell biology. (In this respect, Keith was perhaps to the
cytoplasm what Joe Gall has been to chromosomes and the nucleus as
regards very intelligent selection and adroit exploitation of
exceptionally favorable biological material.) I think Keith would not
have been at odds with my goal of trying to follow RNA in living cells,
though undoubtedly he would have had some penetrating comments. But
they would have been delivered with his trademarka smileby which he
conveyed the shared joy of this endeavor we call cell biology.
| ACKNOWLEDGMENTS |
|---|
Note added in proof: Since this manuscript was submitted, an additional study has confirmed the diffusion-like movement of poly(A) RNA in the nucleus: Politz, J. C., Tuft, R. A., Pederson, T. and Singer, R. H. (1999) Movement of nuclear poly(A) throughout the interchromatin space in living cells. Current Biol. 9, 285291
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |