| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
A Special Series of Reviews: RNA Localization |
ZMBH, University of Heidelberg, Heidelberg, Germany
ABSTRACT
It has become evident over the past years that a large fraction of messenger RNAs is tightly associated with the cytoskeleton. Whereas microtubules are involved in RNA–cytoskeletal association in large cells like oocytes, neurons, or oligodendrocytes, microfilaments play the major role in smaller somatic cell types. Association of RNA with cytoskeletal filaments clearly is required for mRNA transport, but also appears to be crucial for efficient protein synthesis. Recent data now shed light on how mRNAs attach to the cytoskeleton. Messenger RNA seems to interact with microtubules or microfilaments in the form of large ribonucleoprotein particles, which in some cases also contain components of the protein synthesis apparatus. Recently, a number of RNA binding proteins have been identified in flies, amphibians, and mammals that are essential for the interaction of mRNA with cytoskeletal filaments or with microtubule- or actin-associated proteins. Such proteins include heterologous ribonucleoproteins, which are also involved in nuclear export of RNA.—Jansen, R.-P. RNA–cytoskeletal associations.
Key Words: translation • RNA localization • RNA binding proteins • microtubules • microfilaments
BACKGROUND
The high degree of organization in a eukaryotic cell depends on the association of cellular organelles and complexes with a stable interphase. This interphase is provided by the cytoskeleton, which is composed of three major filament systems: microtubules, intermediate filaments, and micro- (or actin-) filaments (1) . These filaments, together with a vast number of associated proteins, are involved in cell movement, cell shape formation, positioning of organelles, transport of vesicles and organelles, and chromosome segregation. They are also important for a (so far) less well understood degree of cellular organization, the positioning of messenger RNAs in the cytoplasm.
A substantial fraction of messenger RNAs, once they are exported from the nucleus, becomes nonrandomly positioned in the cytoplasm. ThesemRNAs include specific messages that are transported to certain sites in the cytoplasm like the anterior/posterior ends in oocytes or dendrites/axons in neurons 2-4) . However, it has become evident that many more RNAs are not randomly distributed in the cytoplasm, but are tightly associated with the cytoskeleton (4 , 5 ). The fraction of associated poly(A) RNA or ribosome/RNA complexes (polysomes) differs from 15–75%, depending on the experimental approach and the cell type analyzed (4 , 6 , 7 ).
Since RNA transport and localization have been reviewed extensively 2-4) and are the topic of several other reviews in this FASEB J. series on RNA localization, this article will only briefly discuss the mechanisms where it is required. Instead, I will concentrate on the steady-state interaction of mRNA and cytoskeletal filaments, on mechanisms of how RNAs might associate with the cytoskeleton, and on the putative function of this interaction.
Initial studies have concentrated on the association of the cytoskeleton with polyribosomes and/or total poly(A) RNA. Two lines of evidence—biochemical fractionation experiments and in situ analysis—suggest that many mRNAs are associated with the cytoskeleton.
Early biochemical cell fractionation experiments used low concentrations of non-ionic detergents to release soluble and membrane-bound macromolecules 8-11) . The unsoluble fraction, termed the cytoskeletal (CSK)2 fraction, was heavily enriched in all three forms of cytoskeletal filaments: microtubules, intermediate filaments, and microfilaments (actin filaments). In addition to cytoskeletal filaments, the CSK fraction from a variety of cell lines contained polysomes (10, 12; see also ref 4 and references therein), translation initiation, and elongation factors (13 , 14 ) and up to 75% of the cellular mRNA as judged by hybridization with probes against total poly(A) RNA (7) . Hybridization with probes against several specific mRNAs (15 , 16 ) have supported this idea. In fibroblasts, both polysomes and mRNAs remained associated with the cytoskeletal fraction even under conditions of high ionic strength, strongly implying physiological binding rather than a weak adsorbance of RNA to the cytoskeleton (15) .
There is some doubt, however, about the interpretation of these early findings. As Hesketh and others (17) have pointed out, the cytoskeletal fraction contains membranous fragments of the endoplasmic reticulum that might have collapsed and been trapped by the attached cytoskeletal fibers. Therefore, not only cytoskeletal-attached but also membrane-bound polysomes might be expected in the CSK fraction because they become trapped during the isolation procedure, leading to an overestimation of the fraction of cytoskeletal-bound RNAs. Indeed, polysomes containing specific mRNAs that code for membrane proteins have been detected in the CSK fraction (18) . Such mRNAs would be expected to cofractionate with the endoplasmic reticulum.
Studies using a different extraction procedure have nevertheless confirmed that a large fraction of mRNAs and polysomes (about 15–30%) are cytoskeletal associated (5, 6 and references therein). To avoid the contamination of CSK associated with membrane-associated polysomes, CSK-bound polysomes were specifically released by disruption of the cytoskeleton, using either drug treatment or salt extraction (7 , 17 , 19 ). Treatment of cells with the microfilament-disrupting drug cytochalasin or salt-induced actin depolymerization preceding polysome extraction resulted in a loss of polysomes and mRNA from the cytoskeletal fraction and an increase of free polysomes. These results emphasize the concept of cytoskeletal-associated mRNAs.
The second experimental line of evidence for mRNA–cytoskeletal association comes from microscopical approaches. Both polysomes and poly(A) RNA are found in the close vicinity of filamentous structures. Immunohistochemical and histochemical studies have demonstrated that ribosomes (20) as well as initiation and elongation factors (21 , 22 ) show an intracellular distribution pattern resembling that of the cytoskeletal network. In situ hybridization against poly(A) RNA showed that most of the mRNA remained associated with cellular filaments even after cytoplasmic extraction and solubilization of membranes with non-ionic detergent (7 , 23 ). This indicates a tight association of poly(A) RNA and cytoskeletal structures. The application of digital image analysis techniques (7 , 24 ) and ultrastructural studies support this idea and have extended the colocalization of filaments and RNA down to the nanometer range (25) . In fibroblasts, most of the detected poly(A) RNA was localized within a 5 nm distance from intermediate and microfilaments. In neurons, poly(A) RNA was detected close to microtubules (23) . These studies also provided evidence that RNA is not evenly spread along actin or microtubule core filaments, but is found in clusters at filament intersections (`vertices') (25) or between loosely bundled microtubules (23) . Other proteins that interact with mRNA, like elongation factors and polysomes, can also be found at these sites (23 , 25 ). The majority of cytoskeleton-associated poly(A) RNA colocalized with ribosomes. However, although suggested by early studies (13) , ribosomes are not essential for RNA–cytoskeletal association. Disruption of polyribosomes by puromycin treatment did not significantly reduce the amount of cytoskeletal-bound poly(A) RNA (7 , 23 ).
To summarize these results, as judged by both in situ and biochemical approaches, a large fraction of mRNAs is tightly associated with cytoskeletal structures.
CYTOSKELETAL FILAMENTS AND mRNA ASSOCIATION
What cytoskeletal filament systems are involved in RNA association? Of the three major filamentous systems, the role of microfilaments (or actin filaments) and microtubules is well established whereas the role of intermediate filaments is contradictory.
Regions rich in intermediate filaments have been found to be devoid of polyribosomes (12) , and polysome-rich areas did not contain intermediate filaments (10) . However, the 50 kDa cap binding protein that is part of the translational initiation complex colocalizes with intermediate filaments (26) . The mRNA encoding glial fibrillary acidic protein colocalizes with intermediate filaments in astrocytes (27) . Furthermore, in situ analysis of poly(A) distribution in fibroblasts suggests that up to 30% of cellular mRNA is found in close proximity to intermediate filaments (25) . The apparently contradictory results perhaps reflect two different states of poly(A) RNA. Translationally active poly(A) RNA colocalizing with polysomes might be absent from intermediate filaments. On the other hand, either translationally inactive poly(A) mRNA or translation factors not actively involved in protein synthesis could be associated with intermediate filaments. Some evidence for such an interpretation comes from the distribution of elongation factor EF-2 in growth-arrested vs. dividing fibroblasts (22) . In dividing fibroblasts with a highly active protein synthesis, EF-2 associates with microfilaments whereas quiescent cells display a codistribution of EF-2 with intermediate filaments. An alternative explanation for the large number of poly(A) RNA found to be associated with intermediate filaments could be that this poly(A) RNA reflects the distribution of a subset of intermediate filament-specific mRNAs rather than that of total poly(A) RNA.
Microfilaments play the major role in RNA–cytoskeletal association and mRNA localization in most cell types, including somatic cells like fibroblasts, myoblasts, and neuroblasts (7 , 15 , 25 , 28 , 29 ), amebas (30) , algae (31) , and yeast (32 , 33 ). Most of these data rely on the effect of actin-depolymerizing drugs such as cytochalasin or latrunculin on RNA–cytoskeletal binding or RNA localization. A drawback to using cytoskeleton-perturbing drugs is their putative pleiotropic effect. Therefore, such experiments must be well controlled and their results interpreted with care. A number of experiments have shown that either some cells of a given population or certain cells of an organism are more resistant to drug treatment than others (23 , 29 ). Furthermore, subsets of cytoskeletal filaments might be less susceptible to some drugs than others, obscuring experimental interpretation (34) . However, because most of the model systems used to study mRNA–cytoskeletal association are not amenable to genetic approaches, filament-disrupting drugs are frequently the only experimental tool available. Only a few reports have addressed the question of RNA–actin interaction using mutants that disrupt or disturb the actin network. In Drosophila, localization of oskar mRNA to the posterior pole of the oocyte requires the function of two actin binding proteins: profilin and tropomyosin 35-37) . Tropomyosin, profilin, actin, and a set of specific genes (SHE1-SHE5) that include at least two other actin-interacting proteins are essential to localize ASH1 mRNA in budding yeast (32) .
As already discussed, ultrastructural in situ hybridization showed a very tight association of actin filaments and poly(A) RNA in fibroblasts (25) , where more than 70% of detected poly(A) RNA are closely associated with actin filaments. The prominent actin stress fibers do not contain much poly(A) RNA. The majority are associated with a branched filament network, especially at the intersections (Fig. 1 ).
|
Microtubules do not play any significant role in RNA binding or localization in the above-mentioned cell types. However, they are important in neurons, oligodendrocytes, and oocytes. Here, localization of mRNAs involving transient interaction of RNA with the cytoskeleton during transport as well as steady-state attachment of RNAs both depend on a proper microtubule network. Paradigms for mRNAs that are localized in such large cells are Drosophila oskar (38 , 39 ), Xenopus Vg1 (40) , or myelin basic protein-1 (MBP1) RNA found in oligodendrocytes (24) . Steady-state studies of RNA–cytoskeleton association in neurons demonstrated that poly(A) RNA became detached from the cytoskeleton upon colchicine but not upon cytochalasin treatment, indicating that microtubules are essential for this association (23) . On the ultrastructural level, poly(A) RNA can be found in close proximity to microtubules. Similar to the case of RNA–microfilament association in fibroblasts (25) , intersecting or loosely bundled microtubules, not tightly bundled filaments, colocalize with RNA in neurons (Fig. 1) .
Why are microtubules essential for RNA–cytoskeletal association in some cells while microfilaments are required in others? Microtubules appear to play the major role in large cell types like oocytes and neurons. In addition, a significant number of RNAs are localized in these cells. Microtubules might serve as the major railway system to allow fast transport over large distances to the destination sites. Smaller cells like yeast or somatic cells might not require a very fast transport system for RNA localization. Therefore, they could rely on the extensive actin network for transport and/or anchoring of RNAs. Some cells show both microtubule–RNA and microfilament–RNA interactions. Eggs of the sea urchin Paracentrotus contain at least three maternal messages—bep1, bep3, and bep4—which are asymmetrically distributed and bound to the cytoskeleton. Association of bep3 RNA with the cytoskeleton is sensitive both to actin- and microtubule-disrupting drugs, implying a requirement of both filament types for RNA binding to filaments (41 , 42 ). In the frog Xenopus, translocation of Vg1 RNA to the vegetal pole of oocytes depends on microtubules, whereas subsequent anchoring requires a functional cortical actin network (40) . Another vegetally localized mRNA, Xcat-2 RNA, becomes tightly associated with the actin-rich cortex once it has reached it in an actin-independent manner (43) . Transport of oskar mRNA to the posterior end in Drosophila oocytes requires microtubules but also cytoplasmic tropomyosin II, an actin binding protein (36 , 37 ). In this case, tropomyosin might function during retention of the localized RNA at the actin-rich posterior cortex (44) . These examples suggest a job sharing of actin and microtubule filaments in some cells. Whereas microtubules might be the transport railway, actin could be required for docking and further positioning of RNAs. Such a collaboration of both filament systems has recently been described for vesicular transport in melanocytes (45 , 46 ).
What part of the mRNA is involved in cytoskeletal attachment and mRNA localization? Although earlier studies suggested that the poly(A) tail was important (7) , several lines of evidence converge on the importance of sequences in the 3' untranslated region (3' UTR) of mRNAs. Poly(A) sequences seem to be less important for cytoskeletal association since polysomes containing histone or reoviral mRNAs that lack poly(A) sequences can also bind to cytoskeletal filaments (18 , 47 ). Messenger RNAs lacking a 5' cap bind to the cytoskeleton, too, demonstrating the dispensability of the cap structure for mRNA–cytoskeletal association (47) . On the other hand, all cytoplasmically localized mRNAs characterized so far contain cytoskeletal-dependent localization signals in their 3' UTR (2–4, 48, 49 and references therein). Some of these signals have been mapped in closer detail. They vary in sequence and length, and can be as short as 21 nucleotides (24) and over 600 nucleotides long (50 , 51 ). However, so far no common structural motif involved in mRNA localization has been identified, although it has been suggested that RNA stem loop structures are an essential part of the signal (24 , 51 , 52 ).
TRANSIENT RNA–CYTOSKELETAL INTERACTION: mRNA TRANSPORT
Most experimental evidence to date suggests RNA localization to be an active, directional process (53 and references therein). Passive diffusion of large RNA–protein complexes would be hindered by the viscosity and exclusion limit of the cytoplasm, which allows only RNAs smaller than 1.6 kb to freely diffuse (54) .
Since RNA localization and transport will be the topic of other reviews in this series on RNA localization, I will only briefly discuss some aspects of the molecular machinery involved in RNA transport.
Active transport requires cytoskeletal-dependent motor proteins. These motors contact the filaments and generate the force required for movement. Depending on the cell type and the localized RNA studied, all three motor protein families—actin-interacting myosins as well as microtubule-interacting dyneins and kinesins—have been implicated in RNA translocation. So far, no specific microtubule-dependent motor has been reported that is part of an RNA transport complex. However, a large body of indirect evidence indicates that kinesins are involved. Kinesins are microtubule plus end-directed motors consisting of two heavy chains and several light chains (55 and references therein). In Drosophila oocytes, a fusion protein of kinesin and ß-galactosidase localizes to the plus ends of microtubules pointing to the posterior of the oocyte (56) . This correlates both in time and space with the posterior localization of maternal mRNAs like oskar (38 , 39 ). Supporting this observation, reorganization of the oocyte's microtubule network in Drosophila protein kinase A mutants leads to mislocalization of both posterior RNAs and kinesin-ß-Gal to the middle of the oocyte (57) .
Antisense oligonucleotides against kinesin mRNA have been used to study the effect of kinesin depletion on MBP RNA transport in oligodendrocytes (58) . Similar to the effect of microtubule disruption, depletion of the kinesin heavy chain resulted in a block of MBP RNA transport. In addition, single RNA granules (see below) could be observed during movement in oligodendrocytes and neurons. In both cell types, granules move with a speed of 0.1–0.2 µm/s, consistent with kinesin-driven transport (24 , 59 , 60 ).
Actin-dependent transport of RNA requires myosins as motors. So far the only myosin involved in RNA transport is the yeast Myo4p. Myo4p is an unconventional class V myosin required for the localization of ASH1 mRNA (32 , 33 ). Myo4p seems to be the best candidate for a motor directly involved in RNA transport. Disruption of the MYO4 gene has no other described effect than a defect in ASH1 mRNA localization (32 , 61 ). At the time of ASH1 localization, Myo4p colocalizes with ASH1 RNA (R. Jansen, unpublished results). ASH1 RNA-containing granules move with a speed that corresponds to the observed velocity of type V myosins (62; R. Long, personal communication). Taken together, Myo4p could in fact direct ASH1 RNA translocation in yeast.
Kinesins and myosins are modular proteins that use their motor domains to contact the corresponding cytoskeletal filament. But how do they interact with the transported RNA? With the exception of the kinesin KIF4 (55 , 63 ), no motor protein contains any sequences similar to DNA/RNA binding proteins. In fact, KIF4 most likely binds to chromosomes and not to RNA. A direct interaction of motor protein and RNA is less likely given the large number of localized RNAs. The general idea is that additional proteins are required that recognize the RNA target and bridge the gap between RNA and motor protein. A paradigm for such modular motor–target complexes is the case of retrograde vesicle transport by dyneins. Dynein, a minus end-directed motor, and at least six other proteins associate to form the complex that recognizes the target proteins on the vesicle to be transported(55) .
To summarize, only one cytoskeletal-associated motor protein has been identified to date that is directly involved in RNA transport. Especially the microtubule-dependent motor proteins involved in RNA translocation remain unknown. How such motors interact with the transported RNAs is also far from being understood.
RIBONUCLEOPROTEIN GRANULES
There is sufficient evidence that messenger RNAs can be found as integrated components of higher order structures throughout their lifetime. Such ribonucleoprotein (RNP) particles or granules have been detected both in RNA transport and while they are associated with cytoskeletal filaments.
Yeast ASH1 RNA can be seen as particles or RNA clusters at the final destination site, the daughter cell tip (32 , 33 ), and as `pearls on a string' throughout the cell that presumably represent transport intermediates (32) . The localization pattern of these putative transport particles resembles the pattern of actin cables, suggesting that these particles are associated with actin filaments during transport. This idea is consistent with the notion that ASH1 mRNA localization depends on a proper actin cytoskeleton and on the function of the unconventional myosin Myo4p.
In Xenopus oocytes, the vegetally localized messages Vg1 and Xcat-2 are also a part of granules (64 , 65 ). However, whether these granules actually represent transport intermediates or storage particles is not known. Likewise, an association of these granules with cytoskeletal structures has not been demonstrated.
Transport and association of RNAs during Drosophila embryogenesis also involves RNP granules. Injection of bicoid RNA into early embryos results in sequestering of endogenous Staufen protein and the formation of large RNP granules (66) . These particles can associate with mitotic spindles formed by endogenous microtubules. There is further evidence that bicoid RNA transport takes place in the form of particles. Exuperantia protein is essential for bicoid movement from the nurse cells into the oocyte early in oogenesis. Fusions of Exuperantia with green fluorescent protein are detected as particles that enrich at the connections of nurse cells and the oocyte (67) . Although it has not been shown that bicoid RNA is actually part of the particles, they imitate the movement of bicoid RNA.
The best evidence for a RNA–cytoskeletal interaction in the form of RNP particles comes from studies of RNA localization in neurons and oligodendrocytes. In the latter case, both endogenous and injected MBP RNA become integrated into granules (24) . Injected MBP RNA can form such granules in the cytoplasm within minutes. The size of the granules is extremely large, with a mean particle radius of 0.7 µm (68) . Surprisingly, granules form independently of the injected RNA, indicating that, at least in oligodendrocytes, granule formation is a general feature of mRNAs (24) . However, only MBP RNA granules can align on tracks that presumably resemble cytoskeletal filaments. This is suggestive of a cytoskeletal association signal present in MBP RNA but absent from other messages. After the initial assembly and filamentous alignment, RNA granules are rapidly transported along microtubules from the cell body into oligodendrocytic processes. Both 3-dimensional immunofluorescence microscopy and drug interference studies have proved that microtubules are indeed the transport tracks (24) . As in oligodendrocytes, large RNA granules have also been detected in neurons (60) . Such granules contain poly(A) RNA, ribosomal subunits, and elongation factors. Recently, a specific mRNA, ß-actin RNA, has been shown to colocalize with a subset of granules in a microtubule-dependent manner (69) .
As a short summary, RNA containing particles or granules can be detected in a variety of different cells. It is tempting to speculate that the formation of these granules is important for the interaction of mRNA with the cytoskeleton.
What proteins are such RNA granules composed of? Granules in oligodendrocytes and neurons contain ribosomes and components of the translation apparatus (60 , 68 ). In fact, these granules are large enough to accommodate several hundred ribosomes. However, ribosomes are not an essential structural component of the particles, since RNA-containing (but ribosome-free) particles have been detected that show a similar structural organization (68) . This suggests that other proteins are important for the formation of RNA granules. Recently, a group of six 35 to 42 kDa proteins has been identified that bind specifically to the MBP RNA targeting signal (70) . One of these proteins is a heterologous nuclear protein, hnRNP A2. hnRNP proteins are involved in several steps of RNA maturation and localization (71) . However, so far it is not clear whether hnRNP A2 is in fact involved in the formation of MBP RNA granules, in attachment to the cytoskeleton, or in their targeting.
In human neuronal cells, ELAV proteins have been reported to associate
with mRNA in the form of granules (72)
. ELAV proteins
contain three highly conserved RNA recognition motifs (RRM). Human
ELAV-like neuronal protein Hel-N1 can be detected in cells as small
granules that colocalize with microtubules (72)
. These
granules appear to be RNP complexes since they also contain poly(A)
RNA. However, not all poly(A) RNA that can be detected as granular
material colocalizes with ELAV protein granules, which suggests the
presence of other ribonucleoparticles. ELAV-RNPs exist in two forms:
complexes containing poly(A) RNA as well as at least 20 different
polypeptides, and ß complexes made up from complexes and
polysomes. Disruption of polysomes causes loss of ß complexes and a
clustering of complexes. These clustered complexes still
associate with microtubules, suggesting that binding of ELAV-RNPs to
microtubules is independent of ribosomes.
THE CONNECTORS: RNA BINDING PROTEINS ASSOCIATED WITH THE CYTOSKELETON
The RNA 3' UTR sequences that play a crucial role in RNA localization or RNA binding to cytoskeletal structures do not bind directly to the core filaments. Ultrastructural in situ hybridization experiments show poly(A) mRNA in close vicinity to, but not exactly colocalizing with, core actin filaments (25) , suggesting that actin-associated components are involved in linking mRNAs and filaments. Similar observations have been made for microtubule-attached poly(A) RNA (23) . In addition, direct binding of RNA to filaments would have only limited capacity to specify the localization of different messages to different cell compartments.
The identification of proteins that bind to targeting sequences of specific RNAs (52 , 66 , 70 , 73-77 ) has shed some light on mRNA–cytoskeletal association. Some of these RNA binding proteins also bind to cytoskeletal filaments, which is consistent with the notion that mRNA targeting is cytoskeleton dependent. Such proteins may provide the connection between mRNAs and filaments.
Several proteins have been isolated and initially characterized on the basis of their ability to be cross-linkable to specific RNAs. However, the molecular nature of these proteins is not known.
TB-RBP (`testis/brain RNA binding protein') is a phosphoprotein that binds to specific mRNAs containing the so-called Y and H elements (76) . The protein attaches translationally repressed RNAs to microtubules; Colcemid, a microtubule-depolymerizing drug, disrupts this association. In the presence of TB-RBP, Y element-containing transcripts including the 3' UTR of oligodendrocyte MBP RNA can bind to microtubules in vitro (76) . However, although association of TB-RBP with mRNA has been demonstrated convincingly, direct binding to microtubules has not been demonstrated. Spnr (`spermatid perinuclear RNA binding protein') is a microtubule-associated RNA binding protein that localizes to the manchette in developing spermatids (77 , 78 ). The RNA target of spnr in vivo is unknown but has previously been suggested to be protamine 1 mRNA (78) . Endogenous spnr as well as a recombinant version can be pelleted with murine testis microtubules. These results suggest that spnr, in addition to its function as an RNA binding protein, might be a microtubule-associated protein (77) .
The localization of bicoid and oskar mRNAs in the Drosophila oocyte depends on the activity of the Staufen protein (2, 49, 73 and references therein). Staufen is an RNA binding protein that contains five double-stranded RNA (dsRNA) binding domains, at least one of them binding to dsRNA in vitro (79) . Whereas oskar mRNA is localized in a staufen-dependent manner to the posterior of the oocyte, bicoid is localized to the anterior and anchored to the anterior cytoskeleton. This anchoring requires Staufen (66) . Localization of oskar and localization/anchoring of bicoid RNAs require a proper microtubule network (2 , 80-82 ). Injection experiments in early embryos have shown that exogenous bicoid RNAs can recruit endogenous Staufen protein and the resulting complex associates with microtubules. However, it is not anteriorly localized (66) . Ferrandon et al. (66) have proposed that the bicoid/Staufen complex but not Staufen itself is recruited to microtubules because the bicoid RNA might induce a conformational change in Staufen that is essential for its association with microtubules. Whether the bicoid/Staufen complex binds directly or indirectly to microtubules is unclear.
In addition to its role in microtubule-dependent RNA localization in oocytes, Staufen is also involved in RNA localization in somatic cells. In neuroblasts, Prospero protein and prospero RNA are both asymmetrically distributed during mitosis due to their localization only to the basal side of the dividing neuroblast 83-86) . Asymmetric distribution of prospero RNA requires a functional microfilament system but is independent of microtubules (29) . RNA localization also requires Staufen protein (85 , 86 ). Staufen interacts directly with the 3' UTR of prospero mRNA and with a protein called Miranda (87 , 88 ). The three partners might form a complex that localizes to the basal half of the cell cortex. Formation of this complex probably requires the function of another asymmetrically localized protein, Inscuteable (85 , 89 ). Asymmetric localization of Inscuteable as well as anchoring of Miranda and Staufen to the cell cortex requires actin filaments but not microtubules (29) . Taken together, these data suggest that Staufen is involved in different RNA localization pathways using different cytoskeletal filament systems.
Vg1 mRNA is transported to the vegetal pole of Xenopus oocytes in a microtubule-dependent manner (40) . It cofractionates with endogenous oocyte microtubules, suggesting a tight association with microtubules (90) . A 69 kDa protein, Vg1 RBP (`Vg1 RNA binding protein'), binds to the essential targeting sequences in the 3' UTR of Vg1 mRNA and to microtubules (90) . Protein depletion experiments showed that Vg1 RNA attachment to microtubules is Vg1 RBP dependent, implying that Vg1 RBP is the link between cytoskeletal filaments and Vg1 mRNA. Vg1 RBP (or Vera = `Vg1 binding and ER association') contains five RNA binding domains. One amino-terminal RRM domain (`RNA recognition motif', 91) and four KH domains (`hnRNP K homology', 91) comprise almost 40% of the entire protein sequence (74 , 75 ). However, no sequences with any similarity to microtubule binding proteins have been identified. In addition, Vg1 RBP/Vera has been shown independently to cofractionate and colocalize with membranes of the ER (92) . It has been suggested that Vg1 RBP/Vera might attach Vg1 mRNA to a specific subcompartment of the oocyte's ER and that this ER subcompartment binds to microtubules (75) . Whether Vg1 RBP/Vera binds directly to microtubules in vivo remains to be proved.
Whereas Vg1 mRNA is transported in a microtubule-dependent manner, chicken fibroblast ß-actin mRNA localization is dependent on microfilaments. A 54 nucleotide `zip code' sequence in the 3' UTR of ß-actin mRNA is essential for this localization (93) . Using the proximal half of this sequence as an affinity matrix, a 68 kDa protein called ZBP-1 (`zip code binding protein 1') and at least two other proteins 53 kDa and 120 kDa in size were purified due to their ability to bind the zip code sequence (52) . Although it is not clear whether any of these three proteins can connect the RNA to actin filaments, other proteins that were also purified using this affinity approach were gelsolin and fibroblast tropomyosin, two bona fide actin binding proteins, which implies that the zip code binding complex is in close contact with actin binding proteins. However, the role of ZBP-1 is far from being understood. Additional work will have to show whether the proposed ZBP complex exists in vivo, colocalizes with the ß-actin mRNA and microfilaments, and can directly or indirectly bind to actin.
ZBP-1 and Vg1 RBP/Vera show striking sequence similarity. Like Vg1 RBP/Vera, ZBP-1 contains one RRM and four KH domains. Both proteins are 78% identical and 84% similar over the entire length and more than 95% identical in their RNA binding domains (74) . Yet they bind different RNA sequences and help these various RNAs to associate with different parts of the cytoskeleton. In that neither ZBP-1 nor Vg1 RBP/Vera contains any actin or microtubule binding domain, it is tempting to speculate that these proteins function as connectors that attach the RNA to other proteins that can directly interact with filaments.
How can such similar proteins recognize different targets, namely, actin- or microtubule-associated proteins? Even more to the point: How can a single protein, Staufen, localize one mRNA via microtubules and another via a pathway that involves microfilaments? There are three explanations (see also Fig. 2 ). First, different domains of the protein(s) can recognize actin- or microtubule-associated complexes. For example, Staufen protein contains two RNA binding domains that do not recognize RNA in vitro and have been suggested to fulfill a different function (87) . In fact, deletion of one of these regions, dRBD2, has no effect on prospero RNA localization in neuroblasts and association of Staufen with Miranda. However, it is absolutely essential for localization of oskar mRNA in the oocyte (87 , 94 ). The other region, dRBD5, binds directly to Miranda and is essential for prospero mRNA localization. Second, the RNA binding protein will attach its target RNA to any cytoskeletal adaptor present or to the major cytoskeletal system of the cell. Given that, it would depend on the presence/absence of filaments or of cytoskeletal adaptor proteins as to which filament an RNA–RNP protein complex attaches. Support for this idea comes from the fact that ß-actin mRNA associates with actin filaments in fibroblasts but its localization depends on the microtubule network in neurons (69) , which represents the major RNA transport system in these cells. Yet another alternative explanation would be that not only the RNA binding protein but the complex of RNA and RNA binding protein(s) determines targeting to the correct filament system. Again, in the case of Staufen, injection of exogenous bicoid RNA into early Drosophila embryos recruits endogenous Staufen protein and leads to assembly of RNP granules that colocalize with microtubules (66) . Injection of prospero RNA recruits Staufen into RNP granules (87) as well. However, these granules do not associate with microtubules, suggesting that either some information on bicoid RNA or another bicoid-specific factor is missing in the case of Staufen-prospero RNPs that is required for RNP–microtubule association.
|
Apart from specific RNA binding proteins that mediate the association of specific mRNAs with filaments, the high proportion of poly(A) RNA found associated with the cytoskeleton implicates the presence of other proteins that mediate the general association of mRNA with the cytoskeleton. What are these factors?
Recently, an RNA binding complex has been identified from neuronal cells that binds to RNA in a sequence-independent but size-dependent manner (95) . Besides a 160 kDa protein subunit that binds to RNA, this complex contains at least one cytoskeletal protein, MAP 1A. MAP 1A (`microtubule-associated protein 1A') is one of the major microtubule-associated proteins and is widely distributed among different cell types (96) . It is localized between microtubules rather than along bundled filaments. Poly(A) RNA, which is associated with microtubules in dendrites, shows a similar distribution pattern (23) . The RNA is frequently found between loosely bundled microtubules but is absent from tight bundles. A MAP 1A-containing RNP complex might meditate attachment of mRNAs to so far unknown cytoskeletal structures between microtubules rather than directly to microtubules. MAP 1A light chain 3 can bind to AU-rich elements in the 3' UTR of fibronectin mRNA and enhances its translation in smooth muscle cells (97) . This implicates a general role for this microtubule-associated protein in mRNA physiology.
Another protein that probably binds both cytoskeletal filaments and poly(A) RNA is mrnp41 from HeLa cells (98) . This protein can be cross-linked to RNA in vivo and subsequently copurifies with poly(A) RNA even under denaturing conditions, indicating a direct interaction of poly(A) RNA and mrnp41. Immunostaining revealed that mrnp41 is present in the nucleus, at the nuclear pores, and in a filamentous meshwork-like pattern in the cytoplasm. Indirect evidence points to a colocalization of this meshwork and poly(A) RNA. Disruption of either microtubules or microfilaments resulted in a loss of the filamentous staining, suggesting that cytoskeletal structures are involved in the formation of this meshwork.
It is surprising that mrnp41 is found in both the nucleus and cytoplasm, a distribution that has also been described for two proteins of the hnRNP family. The hnRNP A1-like hrp36 protein can be found associated with poly(A) RNA inside the nucleus and with ribosomes in the cytoplasm (99) . hnRNP A2 (70) , as discussed above, binds to the targeting sequence in the 3' UTR of myelin basic protein mRNA and is probably part of the large MBP RNA transport granules that have been observed (24) . This hnRNP protein can be detected in the nucleus as well. It appears that mrnp41, hrp36, and hnRNP A2 are involved in nuclear export and probably cytoplasmic targeting or attachment to the cytoskeleton.
RNA–CYTOSKELETAL ASSOCIATION AND TRANSLATION
What might be the reason for the tight association of so many translational components with the cytoskeleton? Obviously, a functional cytoskeleton is not obligatory for translation. However, the efficiency of protein synthesis benefits from a close spatial association of all factors involved in translation. Indeed, actively translated mRNA is often found in cytoskeletal-associated polysomes. During viral infection, host cell protein synthesis is shut down and the majority of newly synthesized proteins are of viral origin. Concomitantly, polysomes containing host cell mRNAs are released from the cytoskeleton and viral RNA-containing polysomes become cytoskeletal associated (11 , 12 , 47 ). In sea urchin eggs, polysomes become attached to the cytoskeleton upon egg activation when translation of stored messages is induced (100) .
In addition, there is evidence for a supramolecular organization of the
translation machinery. Metabolic labeling experiments have shown that
aminoacyl-tRNAs are channeled from the aminoacyl-tRNA synthetase via
elongation factors directly to the ribosome without dissociation into
the fluid cytoplasm (101
, 102
). The
supramolecular organization of the protein synthesis machinery and
efficient translation is concomitantly lost upon F-actin disruption
(103)
, strongly suggesting actin to be essential for this
organization of translational complexes. Additional data indicate that
many macromolecular components of the protein synthesis apparatus are
not freely diffusible in the cell (104)
. Poly(A) RNA,
polysomes as well as elongation factor alpha (EF1) have been
localized on an ultrastructural level to intersections of actin
filaments. EF1 might have a unique role for translation. It can
cross-link fibroblast actin filaments in a specific way, excluding
other actin binding proteins. EF1 might thereby generate a special
microcompartment important for mRNA anchoring and/or cytoskeletal
association of protein synthesis. EF1 from Xenopus
oocytes can sever microtubules by a so far unknown mechanism
(105)
and may play a similar role. Although suggested
originally, the idea that the elongation factor might be required for
the actual binding of the translational machinery to the cytoskeleton
was challenged by the result that EF1's ability to bind actin and
aminoacyl-tRNA is mutually exclusive (106)
. Actin-bound
EF1 is released by an increase in pH, which could reflect a change
in the environment of the translational microcompartment. Such a
release would make EF1 more accessible for aminoacyl-tRNA binding.
The release of sequestered EF1 should supply a high local
concentration of translationally active elongation factor, thereby
facilitating efficient protein synthesis (Fig. 3
). The association of other translational components like ribosomal
subunits or initiation factors (20
, 21
) with
the cytoskeleton might instead be transient and required in order to
increase the local concentration of translation factors.
|
Further evidence for a close spatial organization of translational
components with mRNAs comes from ultrastructural studies in
oligodendrocytes. Here, large RNP granules not only contain mRNAs,
including MBP mRNA and ß-actin mRNA, but also ribosomes,
aminoacyl-tRNA synthetase, and EF1 (60
,
68
). These granules presumably also represent
supramolecular translational complexes. Such granules apparently are
actively transported in oligodendrocytes from the cell body to cell
processes, suggesting that in order to increase the efficiency of
localized protein synthesis, not only mRNA localization but also
localization of the synthesis apparatus could be instrumental.
CONCLUSIONS AND PROSPECTS
Association of messenger RNA with the cytoskeletal framework is a crucial part of its metabolism. The cytoskeleton serves as solid-state interphase to position mRNAs and other components of the protein synthesis apparatus. Thus, it is ensured that a critical mass of translation factors and RNA comes in close enough contact with each other to allow an efficient translation. Certain cytoskeletal microdomains—for example, vertices generated by intersecting actin filaments or loosely bundled microtubules—appear to be the sites of such higher order translation structures.
In addition, microtubules and microfilaments serve as a railway system to transport certain mRNAs to distinct compartments in the cell where they promote localized protein synthesis. During transport, the RNA is probably packed into particles or granules and is in a translationally repressed state so as to avoid inappropriate protein synthesis. In some cases, such granules contain not only RNA and specific RNA binding proteins, but also components of the translational machinery that might have to be delivered to the same sites as the translocated RNA.
A number of proteins bind either to specific mRNAs or to poly(A) RNA. An increasing fraction of these proteins also binds to cytoskeletal filaments. However, so far there is insufficient evidence that a single protein can serve as the connection between microtubules or microfilaments and mRNA. Instead, the observation that some RNA binding proteins like Staufen or Vg1 RBP and its chicken ortholog ZBP-1 can attach certain RNAs to microtubules and others to actin-interacting proteins suggests that such RNA binding proteins provide flexible connectors that can associate both with microtubule- or microfilament-bound factors. How the specificity of the RNA binding protein is regulated in the presence of both microtubules and microfilaments remains elusive and requires more future work.
An important question for future research will be the problem of when and how mRNAs associate with the cytoskeleton once they have been exported from the nucleus. Additional analysis of proteins such as the hnRNP proteins hrp36 and hnRNP A2, as well as mrnp41, could provide important clues. These proteins are involved in nuclear export and most likely in cytoskeletal attachment or transport of mRNA. It would not be surprising if proteins that accompany mRNAs during nuclear export are also important for their targeting and attachment. Indeed, further evidence for such a double function comes from studies of the viral HIV-1 Rev protein (107) . Rev is required not only for export of a class of incompletely spliced viral RNAs like gag mRNA, but also for their colocalization with cytoplasmic microfilaments, strongly suggesting a role for Rev in cytoskeletal attachment.
ACKNOWLEDGMENTS
I would like to thank all colleagues for sending me manuscripts and other material prior to publication. Special thanks go to John Hesketh and Robert Singer. R.-P.J. is a recipient of a grant from the Deutsche Forschungsgemeinschaft (JA696).
Note added in proof: The data mentioned as `R. Long, personal communication' have now been published: Bertrand, E., Chartrand, P., Schaefer, M., Shenoy, S. M., Singer, R. L., and Long, R. M. (1998) Localization of ASH1 mRNA particles in living yeast. Mol. Cell 2, 437–445.
FOOTNOTES
1 Correspondence: ZMBH, Im Neuenheimer Feld 282, D-69120 Heidelberg, Germany. E-mail:r.jansen{at}mail.zmbh.uni-heidelberg.de 
2 Abbreviations: CSK, cytoskeletal; ds, double-stranded;
EF1, elongation factor alpha; hn, heterologous nuclear; KH, hnRNP K
homology; MAP 1A, microtubule-associated protein 1A; MBP, myelin basic
protein; RNP, ribonucleoprotein; RRM, RNA recognition motifs; Spnr,
spermatid perinuclear RNA binding protein; TB-RBP, testis/brain RNA
binding protein; UTR, untranslated region; Vera, Vg1 binding and ER
association; ZBP-1, zip code binding protein 1.
REFERENCES
. Nature (London) 1990;347:494-496.[Medline]
from interaction with aminoacyl-tRNA in a pH-dependent reaction. J. Cell Biol. 1996;135:953-963.This article has been cited by other articles:
|
|
 |
|
  P. J. Santangelo and G. Bao Dynamics of filamentous viral RNPs prior to egress Nucleic Acids Res., June 28, 2007; 35(11): 3602 - 3611. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Masuda, T. Toda, R. Miyamoto, T. Haraguchi, and Y. Hiraoka Modulation of Alp4 function in Schizosaccharomyces pombe induces novel phenotypes that imply distinct functions for nuclear and cytoplasmic {gamma}-tubulin complexes Genes Cells, April 1, 2006; 11(4): 319 - 336. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Hori, Y. Taguchi, S. Uesugi, and Y. Kurihara The RNA ligands for mouse proline-rich RNA-binding protein (mouse Prrp) contain two consensus sequences in separate loop structure Nucleic Acids Res., January 12, 2005; 33(1): 190 - 200. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. X. Chuong, A. G. Good, G. J. Taylor, M. C. Freeman, G. B. G. Moorhead, and D. G. Muench Large-scale Identification of Tubulin-binding Proteins Provides Insight on Subcellular Trafficking, Metabolic Channeling, and Signaling in Plant Cells Mol. Cell. Proteomics, October 1, 2004; 3(10): 970 - 983. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. W. Khandjian, M.-E. Huot, S. Tremblay, L. Davidovic, R. Mazroui, and B. Bardoni Biochemical evidence for the association of fragile X mental retardation protein with brain polyribosomal ribonucleoparticles PNAS, September 7, 2004; 101(36): 13357 - 13362. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Aronov and J. E. Gerst Involvement of the Late Secretory Pathway in Actin Regulation and mRNA Transport in Yeast J. Biol. Chem., August 27, 2004; 279(35): 36962 - 36971. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Gilbert, A. Loranger, and N. Marceau Keratins Modulate c-Flip/Extracellular Signal-Regulated Kinase 1 and 2 Antiapoptotic Signaling in Simple Epithelial Cells Mol. Cell. Biol., August 15, 2004; 24(16): 7072 - 7081. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Coldwell, L. Hashemzadeh-Bonehi, T. M. Hinton, S. J. Morley, and V. M. Pain Expression of fragments of translation initiation factor eIF4GI reveals a nuclear localisation signal within the N-terminal apoptotic cleavage fragment N-FAG J. Cell Sci., May 15, 2004; 117(12): 2545 - 2555. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Vitour, P. Lindenbaum, P. Vende, M. M. Becker, and D. Poncet RoXaN, a Novel Cellular Protein Containing TPR, LD, and Zinc Finger Motifs, Forms a Ternary Complex with Eukaryotic Initiation Factor 4G and Rotavirus NSP3 J. Virol., April 15, 2004; 78(8): 3851 - 3862. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. CHAN, S. N. KHAN, I. HARVEY, W. MERRICK, and J. PELLETIER Eukaryotic protein synthesis inhibitors identified by comparison of cytotoxicity profiles RNA, March 1, 2004; 10(3): 528 - 543. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Malagon, A. H. Tong, B. K. Shafer, and J. N. Strathern Genetic Interactions of DST1 in Saccharomyces cerevisiae Suggest a Role of TFIIS in the Initiation-Elongation Transition Genetics, March 1, 2004; 166(3): 1215 - 1227. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Hamada, K. Ishiyama, S.-B. Choi, C. Wang, S. Singh, N. Kawai, V. R. Franceschi, and T. W. Okita The Transport of Prolamine RNAs to Prolamine Protein Bodies in Living Rice Endosperm Cells PLANT CELL, October 1, 2003; 15(10): 2253 - 2264. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ventela, J. Toppari, and M. Parvinen Intercellular Organelle Traffic through Cytoplasmic Bridges in Early Spermatids of the Rat: Mechanisms of Haploid Gene Product Sharing Mol. Biol. Cell, July 1, 2003; 14(7): 2768 - 2780. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. M. Jockusch, S. Huttelmaier, and S. Illenberger From the Nucleus Toward the Cell Periphery: a Guided Tour for mRNAs Physiology, February 1, 2003; 18(1): 7 - 11. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Audibert, D. Weil, and F. Dautry In Vivo Kinetics of mRNA Splicing and Transport in Mammalian Cells Mol. Cell. Biol., October 1, 2002; 22(19): 6706 - 6718. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Gillespie, P. Boevink, S. Haupt, A. G. Roberts, R. Toth, T. Valentine, S. Chapman, and K. J. Oparka Functional Analysis of a DNA-Shuffled Movement Protein Reveals That Microtubules Are Dispensable for the Cell-to-Cell Movement of Tobacco mosaic virus PLANT CELL, June 1, 2002; 14(6): 1207 - 1222. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. S. W. Ma, K. Moran-Jones, J. Shan, T. P. Munro, M. J. Snee, K. S. Hoek, and R. Smith Heterogeneous Nuclear Ribonucleoprotein A3, a Novel RNA Trafficking Response Element-binding Protein J. Biol. Chem., May 10, 2002; 277(20): 18010 - 18020. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. V. Zeenko, L. A. Ryabova, A. S. Spirin, H. M. Rothnie, D. Hess, K. S. Browning, and T. Hohn Eukaryotic Elongation Factor 1A Interacts with the Upstream Pseudoknot Domain in the 3' Untranslated Region of Tobacco Mosaic Virus RNA J. Virol., May 3, 2002; 76(11): 5678 - 5691. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Percipalle, A. Jonsson, D. Nashchekin, C. Karlsson, T. Bergman, A. Guialis, and B. Daneholt Nuclear actin is associated with a specific subset of hnRNP A/B-type proteins Nucleic Acids Res., April 15, 2002; 30(8): 1725 - 1734. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Boyko, J. A. Ashby, E. Suslova, J. Ferralli, O. Sterthaus, C. M. Deom, and M. Heinlein Intramolecular Complementing Mutations in Tobacco Mosaic Virus Movement Protein Confirm a Role for Microtubule Association in Viral RNA Transport J. Virol., March 19, 2002; 76(8): 3974 - 3980. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-P. Wang, I. Krits, S. Bai, and B. S. Lee Regulation of Enhanced Vacuolar H+-ATPase Expression in Macrophages J. Biol. Chem., March 8, 2002; 277(11): 8827 - 8834. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Zhang, K. Buder, C. Burkhardt, B. Schlott, M. Gorlach, and F. Grosse Nuclear DNA Helicase II/RNA Helicase A Binds to Filamentous Actin J. Biol. Chem., January 4, 2002; 277(1): 843 - 853. [Abstract] [Full Text]
|
|
|
|
|
|
M. Mallardo, S. Schleich, and J. Krijnse Locker Microtubule-dependent Organization of Vaccinia Virus Core-derivd Early mRNAs into Distinct Cytoplasmic Structures Mol. Biol. Cell, December 1, 2001; 12(12): 3875 - 3891. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. ANAND, L. VALENTE, A. CARR-SCHMID, R. MUNSHI, O. OLAREWAJU, P.A. ORTIZ, and T.G. KINZY Translation Elongation Factor 1 Functions in the Yeast Saccharomyces cerevisiae Cold Spring Harb Symp Quant Biol, January 1, 2001; 66(0): 439 - 448. [Abstract] [PDF]
|
|
|
|
|
|
M. R. Tock, A. P. Walsh, G. Carroll, and K. J. McDowall The CafA Protein Required for the 5'-Maturation of 16 S rRNA Is a 5'-End-dependent Ribonuclease That Has Context-dependent Broad Sequence Specificity J. Biol. Chem., March 17, 2000; 275(12): 8726 - 8732. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Dvorak, E. S. Morgan, L. M. Lichtenstein, P. F. Weller, and R. P. Schleimer RNA Is Closely Associated with Human Mast Cell Secretory Granules, Suggesting a Role(s) for Granules in Synthetic Processes J. Histochem. Cytochem., January 1, 2000; 48(1): 1 - 12. [Abstract] [Full Text]
|
|
|
|
 |
|
 L. D. ETKIN and H. D. LIPSHITZ RNA localization FASEB J, March 1, 1999; 13(3): 419 - 420. [Full Text]
|
|
|
|
|
|
S. Frey, M. Pool, and M. Seedorf Scp160p, an RNA-binding, Polysome-associated Protein, Localizes to the Endoplasmic Reticulum of Saccharomyces cerevisiae in a Microtubule-dependent Manner J. Biol. Chem., May 4, 2001; 276(19): 15905 - 15912. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. D. X. Chuong, R. T. Mullen, and D. G. Muench Identification of a Rice RNA- and Microtubule-binding Protein as the Multifunctional Protein, a Peroxisomal Enzyme Involved in the beta -Oxidation of Fatty Acids J. Biol. Chem., January 18, 2002; 277(4): 2419 - 2429. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Huttelmaier, S. Illenberger, I. Grosheva, M. Rudiger, R. H. Singer, and B. M. Jockusch Raver1, a dual compartment protein, is a ligand for PTB/hnRNPI and microfilament attachment proteins J. Cell Biol., November 26, 2001; 155(5): 775 - 786. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
