To determine the mitotic role of Ran in vivo, proteins were injected that would perturb the Ran pathway into Drosophila embryos and their effects on MT organization and karyokinesis were assessed. Early embryonic Drosophila nuclei are contained within a syncytial cytoplasm and undergo mitosis synchronously for the first 14 nuclear cycles, followed by cellularization. At nuclear cycle 10, nuclei are at the embryo cortex where spindle assembly and nuclear division can be monitored by four-dimensional confocal microscopy. To reduce effects on Ran-dependent events during interphase (e.g., nuclear transport), embryos were injected just before mitotic entry, and cellular events were followed during the first mitosis after injection. However, it is still possible that a small percentage of defects arise from inhibiting Ran just before mitosis. Initially, when material is injected into an embryo, it will form a concentration gradient within the embryo with the highest concentration around the injection site. This concentration gradient of perturbant generates more severe effects proximal to the injection site and less severe effects distal to the injection site and has been described in other studies. This facet of the system results in different phenotypes depending on the amount of injection material reaching the spindle despite all the spindles sharing the same cytosol. This offers the current study a significant advantage because weak phenotypes can be seen that may be masked by more severe phenotypes when the Ran pathway is more severely inhibited (Silverman-Gavrila, 2006).
To perturb the Ran pathway, recombinant proteins were injected that are expected to either inhibit or activate the pathway. Multiple strategies were used to inhibit the Ran pathway. One approach was to inject a dominant negative allele of Ran, RanT24N, which is locked in the GDP bound form (Kornbluth, 1994). RanT24N could inhibit the Ran pathway by binding to and inhibiting RCC1, the nuclear exchange factor that generates RanGTP (Dasso, 1994), and/or by binding to importin ß (Hughes, 1998), although this binding is indirect and its affect on importin ß function is unknown (Lounsbury, 1996). To gain insight into which inhibitory process predominates during mitosis in Drosophila embryos, the localization of rhodamine-labeled GST-RanT24N was visualized. Rhodamine-labeled RanT24N localizes to condensed mitotic chromosomes throughout mitosis, as does RCC1 (Trieselmann, 2002). This localization pattern differs markedly from mitotic importin ß, which localized throughout the embryo with some concentration at the residual nuclear envelope (Trieselmann, 2002). Rhodamine-labeled GST-RanT24N binds equally well to RCC1 as unlabeled GST-RanT24N. Thus, the predominant effect of injected RanT24N will likely be the inhibition of RCC1, thereby preventing the continual generation of RanGTP at chromosomes. In addition to RanT24N, the Ran pathway can be inhibited by injecting RanGAP, which activates the intrinsic GTPase activity of Ran, thus reducing the level of RanGTP in the embryo. This inhibition should be less severe than that obtained with RanT24N, because RanGAP does not affect the production of RanGTP (Silverman-Gavrila, 2006).
In mitosis, the Ran pathway is proposed to function through RanGTP, which prevents the inhibition of NLS-containing SAFs via its NTR binding activity. Injecting RCC1 could elevate RanGTP levels, thereby up-regulating the pathway (Silverman-Gavrila, 2006).
Ran is involved in spindle assembly:
Inhibiting the Ran pathway in embryos disrupts spindle assembly. As expected, injection of RanT24N causes the greatest frequency of severe spindle assembly defects. The most dramatic phenotype was the complete inhibition of spindle assembly, where no spindle MTs were nucleated from either centrosomes or chromatin. When MTs were observed, they emanated around condensed chromatin rather than from centrosomes as in control and failed to resolve into spindles. These MTs persisted around condensed chromatin, which stained positive for phospho-histone H3, suggesting that it remained in a mitotic state. Although these effects were observed most frequently with the highest concentrations of RanT24N, they were also observed upon injection of Ran-GAP proximal to the injection site (Silverman-Gavrila, 2006).
To observe more subtle effects of Ran, lower concentrations of RanT24N and RanGAP that allowed some MT production were injected. Under these conditions, spindles initially formed, but they soon became unstable and disorganized. Bundles of centrosomally nucleated MTs grew outward, away from the chromatin. Often, these MTs encountered MTs from neighboring spindles causing the spindles to fuse. Bipolar spindles that did form had altered dimensions (Silverman-Gavrila, 2006).
Spindle pole organization was also disrupted: poles became broader and less focused and in extreme cases, monopolar spindles formed. Where nuclei possessed multiple centrosomes, multipolar spindles formed. Sometimes, poles split or centrosomes were released. Injection of recombinant RCC1 also disrupted spindles. Although many defects were similar to those caused by RanT24N, RCC1 did stimulate the formation of a higher proportion of multipolar spindles (Silverman-Gavrila, 2006).
These data suggest that Ran is required for spindle assembly in vivo in Drosophila embryos (Silverman-Gavrila, 2006).
Nuclear transport receptors inhibit spindle assembly:
Previous studies strongly suggest that the mitotic function of Ran stems from RanGTP, generated at chromosomes, binding to NLS-containing SAFs and inhibiting them (reviewed in Di, 2004). Therefore, to determine whether this mechanism applies in vivo, NTRs were microinjected to elevate their levels within the embryo. These additional NTRs would be predicted to counteract the action of RanGTP and inhibit RanGTP-dependent events. One method was to inject recombinant nuclear import receptors either individually (importin alpha or importin ß) or in combination, because they act together in a complex. In addition, a fragment of importin ß was used that lacks the amino-terminal domain, making it insensitive to RanGTP. Another way to elevate NTRs would be to inject RanBP1. RanBP1 can bind to RanGTP and stimulate the release of NTRs from Ran, in particular nuclear export factors and their cargo from RanGTP (Bischoff, 1997; Kehlenbach, 1999). Alternatively, RanBP1 could stimulate a decrease in RanGTP levels by enhancing RanGAP activity (Coutavas, 1993; Bischoff, 1995) or by inhibiting RCC1 (Nicolas, 1997). Although no homologue of RanBP1 exists in Drosophila, the Ran binding domain in human RanBP1 is 43.8% identical to those in Drosophila RanBP2 (NUP358) and would therefore act as a diffusible source of Ran binding domains to inhibit the Ran pathway (Silverman-Gavrila, 2006).
Injection of NTRs affected spindles in a similar way as the injection of moderate concentrations of RanT24N, resulting in a comparable array and degree of phenotype severity. These data suggest that Ran regulates spindle assembly through regulating the binding of NTRs to SAFs. Interestingly, upon injection of pairs of importins, many more MTs initially formed in prometaphase as bundles around condensed chromosomes. In contrast, MTs in control embryos nucleate from centrosomes and rapidly grew toward chromosomes. In NTR-injected embryos, bundles of MTs slowly reorganized into spindle-like structures but never fully resolved into stable spindles (Silverman-Gavrila, 2006).
Injection of RanBP1 also caused similar defects to injected RanT24N and NTRs. However, uniquely some spindles exhibited consecutive defects. First centrosomes were released, then the spindle elongated. Because one function of RanBP1 is to stimulate the release of nuclear exported cargo from exportins on the cytoplasmic face of the nuclear pore complex (Kehlenbach, 1999), tests were performed to see whether a nuclear exportin has a role in spindle assembly. Injection of CAS, the nuclear export factor for importin alpha, resulted in defects similar to those caused by low doses of RanT24N. However, there were differences: CAS caused the highest frequency of multipolar spindles (Silverman-Gavrila, 2006).
Together, these data suggest that Ran, through its modulation of NTR binding to SAFs, regulates multiple aspects of spindle assembly in syncytial Drosophila embryos (Silverman-Gavrila, 2006).
MT targeting to chromosomes:
A common phenotype seen in all injections that inhibit the Ran pathway (RanT24N, RanGAP, NTRs, and RanBP1) was a failure of centrosomally nucleated MTs to grow rapidly and directly toward chromosomes. On inhibition of the Ran pathway, bundles of MTs, still focused at the poles, began to grow in random directions, probing their environment by repeated cycles of growth and shrinkage. This is in stark contrast to control embryos where centrosomally nucleated MTs grew rapidly toward chromosomes. These data suggest that Ran is required for correct MT targeting to chromosomes by preventing NTRs binding to and inhibiting SAFs (Silverman-Gavrila, 2006).
Chromosome organization:
To analyze the role of the Ran pathway in mitotic chromosome organization, a similar strategy was used to that described above, using a Drosophila line expressing GFP-histone to visualize the chromatin. In a small number of nuclei, inhibiting the Ran pathway prevented nuclear division. Chromatin masses seemed to stay in an interphaselike state (decondensed) or early prophase-like state (slightly condensed) and did not divide, suggesting that either the Ran pathway has a role in chromosome condensation or Ran has a role in mitotic entry. Some nuclei persisted at the cortex, and others exhibited nuclear fallout. However, the DNA in most nuclei condensed and chromosomes segregated. In some cases, even where chromosomes condensed normally, congression to the metaphase plate was disrupted causing a broadening (~3-fold) of the metaphase plate in the centrosome-to-centrosome axis. These data suggest that the Ran pathway is required for chromosome congression to the metaphase plate (Silverman-Gavrila, 2006).
Chromosome movement and segregation:
Inhibiting the Ran pathway using factors that directly affect Ran (RanT24N, RanGAP, and RCC1) or downstream factors of Ran (NTRs), disrupted chromosome segregation to different degrees compared with uninjected and control-injected embryos. Frequently, inhibition of the Ran pathway resulted in the formation of anaphase bridges. Although some bridges resolved, others did not, leading to the formation of nuclei with a 4N DNA content (judged by a doubling of the fluorescence intensity compared with adjacent nuclei that had normal division). In less severe cases, individual chromosomes lagged behind the main chromosome mass. Occasionally, disjoined chromosomes fused with their sister chromosomes resulting in nuclei with a 4N DNA content. In extreme cases, separated chromosomes from two or more nuclei fused, leading to the formation of large chromosomal masses or two or more nuclei fused (Silverman-Gavrila, 2006).
Other chromosome segregation phenotypes stemmed from apparent defects in cell cycle progression as chromosomes remained in a metaphase-condensed state. Occasionally, there seemed to be an override of the spindle checkpoint, as evidenced from the observation that metaphase chromosomes decondensed without disjunction and transited into an interphaselike state. This was also the most common effect upon injection of RCC1, which should up-regulate the Ran pathway (Silverman-Gavrila, 2006).
Distal to the injection site on morphologically normal spindles inhibition of the Ran pathway caused up to a 50% reduction in the velocity of chromosome movement to the spindle poles. In contrast, up-regulation of the pathway upon the injection of RCC1 did not alter the velocity of chromosome segregation. Together, these data demonstrate that the Ran pathway is required for multiple steps in chromosome segregation during anaphase (Silverman-Gavrila, 2006).
Midbody organization:
The effect of inhibiting the Ran pathway was not restricted to spindle assembly. Distal to the injection site spindles formed normally and progressed to anaphase but were unable to form normal midbodies. In some instances, midbody formation was completely inhibited, but when midbodies did assemble they had unbundled MTs, were bent or were narrow. These data demonstrate a role for the Ran pathway after spindle assembly and metaphase, suggesting that the Ran pathway is required throughout mitosis for correct MT organization (Silverman-Gavrila, 2006).
Perturbing the Ran pathway in vivo has dramatic effects on spindle assembly and function. Ran could achieve this by directly regulating the activity of a large number of SAFs, by regulating signal transduction pathways, or both. One candidate target is the Aurora A kinase, which in vitro is suggested to be downstream of Ran in the Ran spindle assembly pathway (Trieselmann, 2003; Tsai, 2003, 2005). Aurora A is recruited to centrosomes in interphase by Centrosomin where it is activated by ajuba and HEF-1. During mitosis in somatic mammalian cells Aurora A relocates to spindle MTs in a TPX2-dependent manner (Silverman-Gavrila, 2006 and references therein).
To address whether Aurora A could be affected by Ran, Aurora A localization was assessed in control and RanT24N-injected embryos. In control embryos, Aurora A localized to centrosomes in interphase and prophase, but in prometaphase it began to redistribute along spindle MTs. On injection of RanT24N, 83.3% of spindles showed mislocalization of Aurora A, which now concentrated at centrosomes and did not localize to spindle MTs (Silverman-Gavrila, 2006).
To determine whether Aurora A targeting to MTs is regulated by NTRs as demonstrated in vitro (Trieselmann, 2003), DeltaN importin ß was injected and Aurora A localization was examined. DeltaN importin ß injection also prevented Aurora A redistribution to spindle MTs in 92% of spindles. This strongly suggests that in vivo Aurora A function is regulated by a NTR sensitive SAF (Silverman-Gavrila, 2006).
Whether the Ran pathway is required for MT-independent mitotic events was investigated. One such process is the assembly of a matrix as defined by one of its components, Skeletor (Walker, 2000). RanT24N disrupted skeletor distribution within the embryo in 80.9% of spindles. Skeletor was asymmetrically distributed outside the area of disrupted spindles, persisting in the vicinity of condensed chromatin rather than around the residual spindle. These data suggest that Skeletor organization is dependent upon the Ran pathway (Silverman-Gavrila, 2006).
Many of the spindle defects observed upon disruption of the Ran pathway could stem from misregulation of motor proteins. Indeed, previous in vitro studies showed that the Ran pathway affects the function of Eg5, a tetrameric kinesin of the kinesin-5 family (Lawrence, 2004) involved in spindle assembly (Wilde, 2001), of Kid (Trieselmann, 2003), a chromokinesin of the kinesin-10 family (Lawrence, 2004), and XCTK2 (Ems-McClung, 2004), a mitotic kinesin of kinesin-14 family (Lawrence, 2004). Therefore, the consequence were examined of disrupting the Ran pathway on the localization of KLP61F (the Drosophila homologue of Eg5) and KLP3A, a Drosophila chromokinesin from the kinesin-4 family (Lawrence, 2004; Silverman-Gavrila, 2006 and references therein).
After the injection of RanT24N, KLP61F distribution was disrupted in 91.8% of spindles. In control embryos, phospho-KLP61F (phosphorylated at T933 and the form of KLP61F that is recruited to spindles) (Sharp, 1999) has a cell cycle-dependent distribution: it is nuclear in interphase and localizes to spindle MTs in prometaphase and metaphase, with a small proportion at centrosomes. In anaphase and telophase, KLP61F localizes to the central spindle (Sharp, 1999). However, upon RanT24N injection KLP61F was barely detectable on spindles with disorganized MTs. Furthermore, an absence of KLP61F staining at centrosomes correlated with metaphase spindles that were shorter and had broader poles. In addition, in RanT24N injected embryos KLP61F did not localize to 93.7% of midbodies (Silverman-Gavrila, 2006).
KLP3A has a cell cycle-dependent distribution localizing to the nucleus in interphase and relocating to the central spindle region, including chromosomes, from prometaphase to anaphase. In telophase, KLP3A persists on the central midbody MTs (Kwon, 2004). After injection of RanT24N, KLP3A distribution was disrupted in 96.2% of spindles: KLP3A localized to chromosomes but not to central spindle MTs. Together, these data demonstrate that one function of the Ran pathway is to regulate the targeting of mitotic motors in vivo (Silverman-Gavrila, 2006).
Segregation Distorter (SD) is a meiotic drive system in Drosophila that causes preferential transmission of the SD chromosome from SD/SD+ males owing to dysfunction of SD+ spermatids. The Sd locus, which is essential for distortion, encodes a truncated RanGAP (Ran GTPase activating protein), a key nuclear transport factor. This study shows that Sd-RanGAP retains normal enzyme activity but is mislocalized to nuclei. Distortion is abolished when enzymatic activity or nuclear localization of Sd-RanGAP is perturbed. Overexpression of Ran or RanGEF (Ran GTPase exchange factor) in the male germline fully suppresses distortion. It is concluded that mislocalization of Sd-RanGAP causes distortion by reducing nuclear RanGTP, thereby disrupting the Ran signaling pathway. Nuclear transport of a GFP reporter in salivary glands is impaired by SD, suggesting that a defect in nuclear transport may underlie sperm dysfunction (Kusano, 2001).
Segregation Distorter (SD) is a meiotic drive system in Drosophila that causes preferential transmission of the SD chromosome from SD/SD+ males owing to the induced dysfunction of SD+ spermatids. The key distorter locus, Sd, is a dominant neomorphic allele encoding a truncated, but enzymatically active, RanGAP (RanGTPase-activating protein) whose nuclear mislocalization underlies distortion by disrupting the Ran signaling pathway. This study shows that even wild-type RanGAP can cause segregation distortion when it is overexpressed in the male germ line or when the gene dosage of a particular modifier locus is increased. Both manipulations result in substantial nuclear accumulation of RanGAP. Distortion can be suppressed by overexpression of Ran or Ran guanine nucleotide exchange factor (RanGEF) in the male germ line, indicating that the primary consequence of nuclear mislocalization of RanGAP is reduction of intranuclear RanGTP levels. These results prove that segregation distortion does not depend on any unique properties of the mutant RanGAP encoded by Sd and provide a unifying explanation for the occurrence of distortion in a variety of experimental situations (Kusano, 2002).
A fundamental principle of Mendelian genetics is the equal transmission of both homologues or alleles from a heterozygous pair. Nonetheless, meiotic drive systems, in which this principle is regularly violated by the preferential transmission of a particular chromosome or allele at the expense of its partner, exist in nature. Segregation distorter (SD) is a naturally occurring meiotic drive system located on the second chromosome of Drosophila melanogaster. SD/SD+ males transmit the SD chromosome to almost 100% of the progeny. Distortion at full strength requires not only the primary locus, Sd, but also several upward modifiers including Enhancer of SD [E(SD)], Modifier of SD [M(SD)], and Stabilizer of SD [St(SD)]. The target of Sd and the upward modifiers is the Responder (Rsp) locus: chromosomes that carry a sensitive (Rsps) or supersensitive (Rspss) allele of Responder are sensitive to the action of SD, whereas those carrying an insensitive allele (Rspi) are resistant. The basic mechanism of distortion is sperm dysfunction, which involves a failure of chromatin condensation in SD+-bearing spermatids, leading to subsequent defects in spermatid elongation and maturation (Kusano, 2002 and references therein),
Sd, a dominant neomorphic mutation, encodes a truncated RanGTPase-activating protein (RanGAP) lacking 234 aa at the C terminus (Sd-RanGAP; Merrill, 1999). Ran is a small GTPase located predominantly in the nucleus. Along with its cofactors, RanGAP and RanGEF (Ran guanine nucleotide exchange factor), Ran is essential for nuclear transport (Görlich, 1999) as well as for other nuclear functions, including cell cycle regulation, mitotic spindle formation, and postmitotic nuclear envelope assembly. The cytoplasmic localization of RanGAP and the nuclear localization of RanGEF establish a concentration gradient of RanGTP across the nuclear envelope that is critical for proper function of the Ran signaling pathway. Sd-RanGAP retains essentially normal enzymatic activity but is mislocalized to nuclei. Both enzymatic activity and nuclear localization of Sd-RanGAP are required for distortion (Kusano, 2001). Overexpression of either Ran or RanGEF in the male germ line suppresses distortion, suggesting that Sd-RanGAP causes distortion by diminishing the concentration of nuclear RanGTP, thereby disrupting Ran-dependent functions. In particular, nuclear transport may be impaired (Kusano, 2001; Kusano, 2002).
By analogy with results of recent studies on yeast RanGAP (Feng, 1999), nuclear mislocalization of Sd-RanGAP can be explained if the usual cytoplasmic distribution of RanGAP is the outcome of a dynamic mechanism in which RanGAP shuttles in and out of nuclei. Normally, this process favors nuclear export over nuclear localization. Shuttling of yeast RanGAP is mediated by a nuclear localization signal (NLS) and two nuclear export signals (NESs) that are located in evolutionarily conserved regions of the protein. Amino acid alignment between yeast and Drosophila RanGAP suggests that these NLSs and NESs are conserved in Drosophila RanGAP. The C-terminal deletion in Sd-RanGAP removes one of the putative NESs, which could shift the equilibrium in favor of a predominant nuclear localization (Kusano, 1999). If this idea is correct, then there are no intrinsic functional differences between Sd-RanGAP and wild-type RanGAP other than their contrasting subcellular distributions. Thus, even wild-type RanGAP might be capable of causing segregation distortion if its subcellular distribution were perturbed such that there was significant accumulation inside nuclei. Alternatively, distortion could depend on entirely novel attributes of the truncated Sd-RanGAP that could not be mimicked in any simple way by the wild-type protein (Kusano, 2002).
This study demonstrates that wild-type RanGAP is capable of causing segregation distortion when it is overexpressed in the male germ line, a manipulation that results in aberrant nuclear accumulation of this enzyme. Moreover, presence of an extra dose of E(SD) causes segregation distortion even when Sd is absent. In this case, too, there is substantial accumulation of wild-type RanGAP inside nuclei. Distortion in these cases can be suppressed by overexpression of Ran or RanGEF in the male germ line, arguing that, as in the case of SD, a reduction in the level of intranuclear RanGTP is the primary cause of distortion. These results prove that nuclear localization of RanGAP activity is sufficient to cause distortion and provide a unifying explanation for the occurrence of distortion in a variety of experimental situations (Kusano, 2002).
Thus, overexpression of wild-type RanGAP in the male germ line causes strong segregation distortion. This distortion exhibits all of the properties characteristic of that associated with native SD chromosomes: it requires RanGAP enzymatic activity, it depends on the presence of upward modifiers and a sensitive Responder (Rsp) target, and it is eliminated by known suppressors of segregation distortion as well as by overexpression of Ran or RanGEF in the male germ line. Most strikingly, overexpression of wild-type RanGAP results in substantial mislocalization of the protein to nuclei in primary spermatocytes, similar to what has been observed for normal levels of expression of Sd-RanGAP. This result suggests that wild-type RanGAP is capable of entering nuclei and most likely does so to some degree even at endogenous levels of expression. Although this conclusion is contrary to the prevalent view that RanGAP is localized exclusively in the cytoplasm, it is in agreement with studies in yeast indicating that the normal cytoplasmic localization of RanGAP in the outcome of a dynamic equilibrium in which RanGAP shuttles in and out of the nucleus (Feng, 1999). This process is thought to be mediated by the activities of an NLS and two NESs that are located in evolutionarily conserved portions of the protein (Feng, 1999). Amino acid alignments with Drosophila RanGAP suggest that these elements are also present in the Drosophila protein. It has been argued that the truncated RanGAP encoded by Sd deletes one of the NESs, thereby biasing the subcellular distribution from cytoplasm to nucleus. In the present case, 10- to 20-fold overexpression of wild-type RanGAP is apparently sufficient to alter the usual equilibrium such that an excessive amount of RanGAP accumulates inside nuclei (Kusano, 2002).
The notion that nuclear mislocalization of enzymatically active RanGAP is responsible for distortion can also account for the distortion caused by two doses of E(SD) (Temin, 1991). Increased dosage of the modifier element E(SD) results in a marked accumulation of wild-type RanGAP inside spermatocyte nuclei even though the protein is expressed at endogenous levels. On the basis of this observation, it seems likely that E(SD) encodes a factor that facilitates nuclear import of RanGAP, inhibits its export, or in some other way enhances the nuclear localization of RanGAP. Because distortion associated with wild-type RanGAP is elicited by two doses of E(SD) but not one, it is imagined that the physiological effect of this protein is dose dependent and that a certain threshold amount is required to accumulate a sufficient level of wild-type RanGAP in germ-line nuclei to produce distortion. A single dose of E(SD) substantially increases the strength of distortion caused by Sd-RanGAP (Ganetzky, 1977; Merrill, 1999). The results suggest that E(SD) may do so by further increasing the propensity of Sd-RanGAP to localize to nuclei. Because E(SD) behaves as a dominant neomorphic mutation, what function is normally subserved by the wild-type protein remains an intriguing question. Unfortunately, the location of E(SD) within the heterochromatin of the second chromosome will greatly complicate molecular identification and characterization of this gene (Kusano, 2002).
The results argue that, aside from its altered subcellular distribution, there is nothing intrinsically unusual about the truncated Sd-RanGAP compared with the wild-type enzyme. Although the physical basis for nuclear mislocalization of RanGAP differs for Sd-RanGAP, ß2-RanGAP, and increased dosage of E(SD), the underlying mechanism of distortion is apparently the same in all of these cases. In the appropriate genetic background, all that is required to cause segregation distortion in Drosophila is an excess of enzymatically active RanGAP in nuclei of the male germ line. The information gained from this study provides a unifying explanation for the occurrence of distortion and will be essential in elucidating the remaining mechanistic details of this phenomenon (Kusano, 2002).
The NF-kappaB/Rel pathway functions in the establishment of dorsal-ventral polarity and in the innate humoral and cellular immune response in Drosophila. An important aspect of all NF-kappaB/Rel pathways is the translocation of the Rel proteins from the cytoplasm to the nucleus, where they function as transcription factors. A new protein, Tamo, has been identified that binds to Drosophila Rel protein Dorsal, but not to Dorsal lacking the nuclear localization sequence. Tamo does not bind to the other Drosophila Rel proteins, Dif and Relish. The Tamo-Dorsal complex forms in the cytoplasm and Tamo also interacts with a cytoplasmically orientated nucleoporin. In addition Tamo binds the Ras family small GTPase, Ran. Tamo functions during oogenesis and, based on phenotypic analysis, controls the levels of nuclear Dorsal in early embryos. It further regulates the accumulation of Dorsal in the nucleus after immune challenge. It is concluded that Tamo has an essential function during oogenesis. Tamo interacts with Dorsal and proteins that are part of the nuclear import machinery. It is proposed that tamo modulates the levels of import of Dorsal and other proteins (Minakhina, 2003).
The Bj1 gene encodes the Drosophila homolog of RCC1, the guanine-nucleotide exchange factor for RanGTPase. This study provides the first phenotypic characterization of a RCC1 homolog in a developmental model system. Bj1 (dRCC1) was identified in a genetic screen to identify mutations that alter central nervous system development. Zygotic dRCC1 mutant embryos exhibit specific defects in the development and differentiation of lateral CNS neurons although cell division and the cell cycle appear grossly normal. dRCC1 mutant nerve cords contain abnormally large cells with compartmentalized nuclei and exhibit increased transcription in the lateral CNS. Since RCC1 is an important component of the nucleocytoplasmic transport machinery, it was found that dRCC1 function is required for nuclear import of nuclear localization signal sequence (NLS)-carrying cargo molecules. Finally, it was shown that dRCC1 is required for cell proliferation and/or survival during germline, eye and wing development and that dRCC1 appears to facilitate apoptosis (Shia, 2004).
Three of the four independently induced KetelD dominant-negative female sterile mutations that identify the Drosophila importin-ß gene, originated from a C4114-->T transition and the concurrent replacement of Pro446 by Leu (P446L). CD spectroscopy of representative peptides with Pro or Leu in the crucial position revealed that upon the Pro-->Leu exchange the P446L mutant protein loses flexibility and attains most likely an open conformation. The P446L mutation abolishes RanGTP binding of the P446L mutant form of importin-ß protein and results in increased RanGDP binding ability. Notably, the P446L mutant importin-ß does not exert its dominant-negative effect on nuclear protein import and has no effect on mitotic spindle-related functions and chromosome segregation. However, it interferes with nuclear envelope formation during mitosis-to-interphase transition, revealing a novel function of importin-ß (Timinszky, 2002).
The fact that in three of the four independently isolated KetelD mutations the same C-->T transition and the concurrent replacement of Pro446 by Leu is the basis of dominant female sterility underlines the importance of Pro446 in importin-ß function. It is assumed that during a nuclear import cycle, whereas importin-ß interacts with the NLS containing protein (directly or through importin-ß), nucleoporins and RanGTP, the conformation of importin-ß changes significantly. In fact the region around HEAT repeat 10 was suggested, based on X-ray crystallography, to be a flexible point during switching between the IBB- and the Ran-bound forms. Pro446 resides in the linker region connecting HEAT repeats 10 and 11 and, as described in this study, plays a crucial role in enduring flexibility of importin-ß. CD spectra of model peptides representing the wild-type and the P446L mutant proteins reveal loss of flexibility upon Pro446-->Leu replacement. The lost flexibility is most likely the consequence of fusion of the small alpha-helix in the linker region with the alpha-helix of HEAT 10B. Computer 3D modeling of the P446L protein structure, based on results of CD spectroscopy, shows altered molecular structure: the P446L molecule takes on an open conformation such that its inner hydrophobic surface becomes exposed to water, explaining the reduced hydrophilic nature of the P446L protein (Timinszky, 2002).
The significant conformational change due to the exchange of Pro446 to Leu in the Ketel protein is further supported by the S317T suppressor mutation that restores Ketel gene function. In human importin-ß the corresponding Ser311 (in the linker region between HEAT repeats 7 and 8) and Pro441 are 32.5 Å apart and yet the Ser-->Thr exchange in the Drosophila homologue restores function of importin-ß. The 10 Å area surrounding serine is hydrophobic. The stronger hydrophobicity of threonine compared with serine does perhaps increase apolar interactions and bend the molecule back to its functional structure (Timinszky, 2002).
Experiments with digitonin-permeabilized HeLa cells show that, to a reduced extent, the P446L proteins do participate in formation of the nuclear import complexes and in their docking to the cytoplasmic surface of the NE; however, they do not support import of the complexes into the nuclei in the presence of Ran, energy source, RanGAP and RanBP1. In fact, the import complexes do not form upon the addition of the latter components. Apparently the main structural domains of the P446L protein are intact (binds importin-alpha, NPC and Ran) but the interaction with Ran is altered. Indeed, the binding of wild-type and P446L Ketel proteins to Ran are very different: the P446L cannot bind to RanGTP, to which the wild-type importin-ß binds strongly, but shows elevated affinity to RanGDP, to which the wild-type protein shows very little affinity. It is noteworthy that a single amino acid exchange outside the classical Ran-binding domain can change Ran binding ability dramatically. The change in Ran-binding ability is most likely the source of the KetelD-associated dominant female sterility. However, the KetelD-associated dominant-negative effect is not manifested via nuclear protein import but rather through the prevention of cleavage nuclei formation: revealing a novel importin-ß function (Timinszky, 2002).
Injection experiments into wild-type cleavage embryos revealed that the P446L mutant protein does not inhibit nuclear protein import: when co-injected with P446L, a fluorescent nuclear substrate readily enteres the nuclei. Furthermore, although the cleavage nuclei enter mitosis and the chromosomes segregate normally, intact NE never forms in the presence of P446L mutant importin-ß. Failure of NE assembly in the presence of P446L is revealed by the following observations; 1) the homogenous distribution of a fluorescent nuclear substrate, the high molecular weight dextrane and the GFP-tubulin; 2) the absence of the nuclear lamina lining. Thus the mutant P446L importin-ß reveals a novel importin-ß function required during the mitosis-to-interphase transition, a function distinct from the already known functions of importin-ß in nuclear protein import and in mitotic spindle assembly (Timinszky, 2002).
The P446L mutant importin-ß possesses altered Ran-binding properties: it does not bind RanGTP but shows elevated affinity to RanGDP. A series of experiments showed that altered RanGTP-RanGDP balance leads to a similar phenotype in yeast (i.e., arrest in mitosis-to-interphase transition). Results of enzyme assays described in the present paper show that the altered Ran-binding ability of P446L importin-ß does not interfere with the GTP hydrolysis and nucleotide exchange on Ran and thus it is unlikely that the KetelD-related defects are consequences of distorted Ran metabolism. Most probably importin-ß is a downstream effector of Ran during mitosis-to-interphase transition, as in nuclear protein import and mitotic spindle assembly (Timinszky, 2002).
Although several functions of Ran and importin-ß during the cell cycle have been described, the exact molecular mechanisms are still missing. This study describes a novel function of Drosophila importin-ß during mitosis-to-interphase transition where it is involved in the formation of intact nuclear envelope (NE). There seem to be three feasible explanations for the P446L-associated defects. First, since the P446L importin-ß shows higher affinity to RanGDP than wild-type importin-ß, a possible explanation may be the depletion of significant amounts of RanGDP that is required for NE reassembly at the end of mitosis. Removal of RanGDP by P446L may lead to the failure of cleavage nuclei formation. This explanation is thought not to be very likely for the following reasons. (1) Binding and nucleotide exchange assays revealed that the affinity of the P446L to RanGDP is low and hence depletion of a significant fraction of Ran from the cytoplasm is rather unlikely. (2) Interestingly, defects do not evolve in nuclear protein import or in spindle formation and chromosome segregation following injection of P446L despite the fact that both nuclear protein import and spindle formation have been shown to be Ran dependent. Ran's involvement in NE assembly has also been described but since none of the aforementioned Ran-related processes were disturbed, the P446L protein does not seem to disturb the Ran cycle. A second possible explanation of the P446L-related defects is perhaps the inability of the P446L protein to bind RanGTP and, consequently, the inability to release factors required for proper chromatin decondensation and/or NE assembly. In this case the role of importin-ß in the above processes would resemble its function in mitotic spindle formation, where it is thought to be required for the release of factors needed for spindle assembly [e.g. NuMA, TPX2. A third possibility is that the P446L-related defects are not associated with the change in Ran-binding ability. The P446L mutation may disturb the association of thus far unidentified factors (e.g. nucleoporins). In the case of the second and third possibilities the factor(s) required for the newly described importin-ß-related functions remain to be identified (Timinszky, 2002).
During early development in Drosophila, pseudocleavage furrows in the syncytial embryo prevent contact between neighboring spindles, thereby ensuring proper chromosome segregation. This study demonstrates that the GTPase Ran regulates pseudocleavage furrow organization. Ran can exert control on pseudocleavage furrows independently of its role in regulating the microtubule cytoskeleton. Disruption of the Ran pathway prevents pseudocleavage furrow formation and restricted the depth and duration of furrow ingression of those pseudocleavage furrows that form. Ran is required for the localization of the septin Peanut to the pseudocleavage furrow, but not anillin or actin. Biochemical assays revealed that the direct binding of the nuclear transport receptors importin α and β to anillin prevents the binding of Peanut to anillin. Furthermore, RanGTP reverses the inhibitory action of importin α and β. On expression of a mutant form of anillin that lacks an importin α and β binding site, inhibition of Ran no longer restricts the depth and duration of furrow ingression in those pseudocleavage furrows that form. These data suggest that anillin and Peanut are involved in pseudocleavage furrow ingression in syncytial embryos and that this process is regulated by Ran (Silverman-Gavrila, 2008).
During cytokinesis, the ingressing plasma membrane physically divides the mother cell into two daughter cells. Membrane ingression during cell division is both temporally and spatially regulated, ensuring that membrane scission occurs (1) only after the chromosomes have fully segregated and (2) between the two chromosomal masses. The signals within the cell that determine cytokinetic furrow positioning are complex, reflecting the strict control needed to ensure that cytokinesis is successful. Signals from astral microtubules, the spindle midbody, the nucleus, and the membrane itself direct the assembly of the contractile ring to the equatorial cortex of the plasma membrane. The contractile ring is an actomyosin-based structure that constricts and generates the force needed to drive membrane ingression. As the membrane ingresses, it is remodeled and stabilized (Silverman-Gavrila, 2008).
Other membrane ingression events share many of the same features and involve many of the same proteins as cytokinetic furrows. In the syncytial Drosophila embryo before cellularization, up to 6000 closely packed nuclei exist in a common cytosol close to the cortex. To ensure faithful chromosome segregation during the rapid nuclear divisions, nuclei are isolated from one another to prevent neighboring spindles from contacting and fusing. To achieve this, plasma membrane ingressions form transiently between nuclei during the rapid nuclear cycles before cellularization. These membrane ingressions, termed pseudocleavage or metaphase furrows, are organized by the actin cytoskeleton and bear a close resemblance to cytokinetic cleavage furrows. First actin caps form at the plasma membrane above each nucleus. Then during interphase, as the centrosomes migrate to either side of the nucleus, the actin caps expand correspondingly. In prophase the cap reorganizes to drive membrane ingression into the embryo such that nuclei and newly forming spindles are separated from one another. Toward the end of metaphase, the furrows begin to retract and dissipate by anaphase. This process is repeated from the tenth through the thirteenth nuclear cycles. During the fourteenth nuclear cycle, the syncytial embryo cellularizes to form 6000 columnar epithelial cells. In this instance the cleavage furrows extend down into the embryo, before growing transversally and fusing to form a single layer of nucleated cells (Silverman-Gavrila, 2008).
Most components required for furrow ingression are conserved between cytokinetic furrows (during conventional mitosis) and pseudocleavage furrows. However, there are some differences. Notably pseudocleavage furrows are membrane ingressions that do not meet and therefore do not lead to membrane fusion. Instead they extend into the embryo, perpendicular to the cortex, and then retract back toward the embryo cortex after the chromosomes have begun to segregate. In addition, there is a difference in the stage of the cell cycle when the furrow components assemble. Although the cytokinetic furrow begins to assemble during anaphase and is required to divide a cell in two, the syncytial embryo pseudocleavage furrows begin to assemble in prophase and serve to prevent neighboring spindles from contacting one another (Silverman-Gavrila, 2008).
A key protein involved in cytokinetic furrow function is anillin, which has multiple domains allowing it to bind and bundle actin filaments, target septins to the plasma membrane, and interact with components of the microtubule-bound centralspindilin complex (Gregory, 2008). Consequently anillin is thought to act as a scaffold for the correct assembly of the contractile ring (Piekny, 2007). It is not fully understood how the role of anillin in cytokinesis is regulated. However, its role in remodeling the actomyosin contractile ring in somatic cells is in part regulated by its differential spatial positioning in the cell during the cell cycle. In interphase anillin localizes to the nucleus where it cannot interact with actin and myosin at the plasma membrane. However, in mitosis upon nuclear envelope breakdown, anillin is released from the nucleus and is targeted to the cortex of the plasma membrane and later to the equatorial cortex of the plasma membrane in a RhoGTP-dependent manner. The spatial regulation of anillin during the cell cycle contributes to the restriction of its function to mitosis. However, in Drosophila syncytial embryos anillin is cytosolic, localizing to pseudocleavage furrows throughout the nuclear cycle, suggesting that it may be regulated by other mechanisms (Silverman-Gavrila, 2008).
One function of anillin is to target septins to the contractile ring. Septins are a family of GTP-binding proteins that can assemble into filaments. Septins have been attributed multiple roles: as membrane diffusion barriers, as stabilizers of the furrow, in membrane trafficking, and as a scaffold. In Drosophila there are five septins: Peanut, Sep1, Sep2, Sep4, and Sep5. Peanut, Sep1 and Sep2 have been isolated as a stoichiometric complex that in vitro can polymerize into filaments. In contrast, Xenopus laevis Sept2 can self assemble into filaments, suggesting that septins may function independently (Silverman-Gavrila, 2008 and references therein).
The GTPase Ran is a key positive regulator of mitosis (Ciciarello, 2007). RanGTP regulates a number of mitotic factors that are sequestered in the nucleus by nuclear transport receptors during interphase. In mitosis RanGTP antagonizes the binding of nuclear transport receptors to these proteins and thereby promotes their activity. RanGTP is at its highest concentration around the chromosomes, where RCC1 the nucleotide exchange factor for Ran is localized. Consequently, RanGTP has been proposed to act as a spatial cue by only activating these mitotic proteins close to the chromosomes (Caudron, 2005; Kalab, 2006). In so doing RanGTP is thought to specify where certain mitotic processes occur in the cell. For example, it could specify that spindle assembly only occurs around chromosomes. The full extent to which this mechanism regulates the mitotic cell is not known and continues to expand (Silverman-Gavrila, 2008).
This study demonstrates a new role for Ran in regulating pseudocleavage furrow ingression, a membrane invagination process in early Drosophila embryos. The Ran pathway regulates the interaction between anillin and the septin Peanut, thereby regulating furrow stability (Silverman-Gavrila, 2008).
A cytological screen was carried out to identify mitotic processes regulated by the Ran pathway. Inhibitors of the Ran pathway were injected into GFP-α-tubulin-expressing embryos just before mitotic entry, and then microtubule organization was monitored by time-lapse microscopy. One phenotype, the fusion of neighboring spindles, occurred more frequently upon the injection of inhibitors of the Ran pathway compared with control injections. In control injected embryos 0.2% of observed spindles fused to a neighboring spindle. In contrast, inhibition of the Ran pathway by injecting either the dominant negative allele of Ran, RanT24N, or importin α resulted in 8.4 and 7.8% of observed spindles fusing to neighboring spindles, respectively (Silverman-Gavrila, 2008).
Peanut is recruited to ingressing furrows by anillin, a multifunctional protein required for cytokinesis that interacts with myosin II, actin, and septins. Septins bind to the carboxy-terminus of anillin, which includes a pleckstrin homology (PH) domain. Drosophila anillin has three potential nuclear localization signals (NLS) that could bind to the nuclear transport receptors importin α and β. Two of the NLS motifs are located in or directly adjacent to the PH domain (Silverman-Gavrila, 2008).
To determine if the carboxy-terminus of anillin could bind to importin α and β, a fusion was constructed between GST and the carboxy-terminus of anillin (amino acids 815-1201, anillin-CT, and its ability to bind to recombinant importin α and β was analyzed. Both importin α and β bound to anillin-CT, and this binding was reversed in the presence of RanQ69L, a point mutant of Ran locked in the GTP-bound state. Of the two potential NLS motifs, the one located between amino acid residues 989 and 999, bares the closest resemblance to an archetypal bipartite NLS and is found in the same region of human anillin (amino acids 887-898). Mutation of lysines 997-999 to alanine (3A-anillin-CT) abrogate both importin α and β binding to this region of anillin, suggesting that amino acids 989-999 constitute a nuclear transport receptor-binding site (Silverman-Gavrila, 2008).
It was next asked if the anillin-CT could interact with Peanut. GST-anillin-CT was incubated with 0-3-h Drosophila embryo extract and then isolated using glutathione agarose beads. Anillin-CT copurified with Peanut and another septin, Sep2. However, the addition of exogenous importin α and importin β inhibited the binding of Peanut to anillin-CT in a concentration- and NLS-dependent manner. This inhibition was specific to Peanut, because Sep2 binding to anillin-CT was not inhibited by importins (Silverman-Gavrila, 2008).
To determine if the in vivo targeting of Peanut and Sep2 to the pseudocleavage furrows is differentially regulated, importin α was injected into syncytial embryos and GFP-Sep2 localization was determined by time-lapse microscopy. Consistent with in vitro results, GFP-Sep2 localization was not perturbed upon interfering with the Ran pathway. Furthermore, in fixed GFP-Sep2-expressing embryos in which the Ran pathway has been perturbed, Peanut fails to localize to nascent furrows, whereas GFP-Sep2 does localize to nascent furrows. These data suggest that Peanut and Sep2 are differentially regulated by Ran and that Sep2 can localize to pseudocleavage furrows independently of Peanut (Silverman-Gavrila, 2008).
This study has identified RanGTP as a regulator of the interaction between Peanut and anillin. This mechanism operates directly and independently of Ran's well-characterized role in regulating the mitotic microtubule cytoskeleton (Silverman-Gavrila, 2008).
Studies suggest that anillin is required for the recruitment of septins to the furrow. By perturbing the Ran pathway, this study has demonstrated that the recruitment of the septins Peanut and Sep2 is differentially regulated, consistent with previous observations that Sep1 recruitment to furrows is dependent on Peanut but Sep2 is not. Anillin lacking the importin binding site between residues 997 and 999 can bind to Peanut in the presence of importins, suggesting that importins directly block the anillin-Peanut interaction rather than disrupting the Peanut, Sep1, and Sep2 complex. These data suggest that although Peanut, Sep1, and Sep2 can exist in a single complex, they may be able to function independently of one another as has been demonstrated in vitro for a Xenopus septin (Silverman-Gavrila, 2008).
Perturbing the Ran pathway destabilizes pseudocleavage furrows. One mechanism for this is through the regulation of the anillin-Peanut interaction. In embryos that expressed an anillin mutant lacking the importin-binding site, Peanut recruitment to pseudocleavage furrows occurs even in the presence of exogenous importins, and furrows demonstrate wild-type dynamics. These data suggest that Peanut is required for pseudocleavage furrow stability. This role for anillin and Peanut is consistent with the observed role for these proteins in stabilizing the cellularization furrow later in Drosophila development. These findings may at first appear to contradict those studies, in which embryos lacking Peanut protein progressed through the syncytial nuclear divisions only showing the first defects during cellularization. However, these studies only analyzed syncytial furrows from the top, apical view and not from the lateral view to observe ingression dynamics. Therefore, these studies would not have detected changes in furrow ingression dynamics that were observed upon inhibition of Ran, which correlated with a failure to recruit Peanut to the furrow (Silverman-Gavrila, 2008).
The Ran pathway regulates pseudocleavage furrow ingression directly by regulating importin binding to anillin. It was previously shown that in Drosophila syncytial embryos the importin β, whose injection causes similar effects as importin α, is released from the nucleus upon nuclear envelope breakdown and becomes diffuse throughout the cytosol during the rest of mitosis. During this period pseudocleavage furrows begin to retract. Therefore, as importin β is cytosolic during metaphase and anaphase it could act to prevent the interaction of Peanut and anillin. In turn this would lead to furrow instability and retraction (Silverman-Gavrila, 2008).
It cannot unequivocally be ruled out that some of the defects caused by perturbing the Ran pathway are due to a disruption of microtubule cytoskeleton. Indeed, one microtubule-dependent furrow phenotype, the formation of pseudocleavage furrows that encompassed a small area of cytosol around a nucleus, was observed. This phenotype has also seen in another study upon depolymerization of microtubules in embryos. However, microtubule depolymerization when instigated in interphase does not cause a failure in pseudocleavage furrow formation, a finding consistent with a previous study (Silverman-Gavrila, 2008).
Another mechanism through which Ran could affect pseudocleavage furrows is by disrupting nuclear trafficking. Indeed it was observed that nuclear trafficking can be reduced by up to 50% upon disruption of the Ran pathway. However, it seems unlikely that the changes in nuclear import kinetics in these experiments disrupted the function of anillin because anillin is a cytosolic protein in the syncytial embryo and localizes to the leading edge of the ingressing furrow during interphase. It is not understood how anillin is retained in the cytoplasm of syncytial embryos because it is imported into nuclei in other developmental stages. However, this phenomenon is not unique to anillin and is also exhibited by the kinesin Pavarotti, another protein involved in pseudocleavage furrow organization (Silverman-Gavrila, 2008).
These studies suggest that Ran regulates multiple factors involved in pseudocleavage furrow ingression, because embryos expressing the mutant anillin still exhibit a failure to form all the expected pseudocleavage furrows. Failure to fully suppress the phenotype could be due to the continued presence of endogenous anillin or reflect that other Ran pathway-sensitive factors are involved in pseudocleavage furrow formation. Regulation through the Ran pathway could define a spatial cue concentrated around chromosomes and extending to the cortex. Such a spatiotemporal regulatory mechanism could be involved in promoting cytokinetic furrows in other cells. A recent study in oocytes finds that Ran regulates myosin II, whose activity is required for cytokinetic cleavage furrows. In addition importin α is required for ring canal organization during oogenesis. Ring canals form as a result of incomplete cytokinesis, and many proteins involved in cytokinesis both localize to and are required for their formation, including anillin and septins (Silverman-Gavrila, 2008).
The data suggest that the anillin-Peanut interaction, which is inhibited by importins must occur in regions of the cell where there are low levels of importins or high levels of RanGTP. Recent studies have visualized a RanGTP-importin β gradient and found that it persists from the chromosomes to the centrosomes, a distance similar to that between the metaphase plate and the cortex. Thus, RanGTP could play an important role in positioning the plane of cleavage by defining on the cell cortex where furrow proteins interact (Silverman-Gavrila, 2008).
Although there are clear differences between cytokinetic and pseudocleavage furrows, anillin and septins are involved in both. Therefore, this study suggests that Ran could also have a role in regulating cytokinetic furrows. Whether chromosomes play a significant role in cytokinesis remains controversial. However, studies where nuclei or chromosomes are asymmetrically positioned within a cell show that furrow ingression coincided with the region of the cell that contained the chromosomes, suggesting that signals from the nucleus and in particular the chromosomes had a role in specifying furrow ingression. Similarly, enucleated sea urchin eggs are able to duplicate their centrosomes and generate astral arrays of microtubules, but fail to form stable cleavage furrows. The current study proposes a molecular mechanism to explain, at least in part, these observations, suggesting that RanGTP generated around the chromosomes is a diffusible signal that facilitates multiple processes required for furrow formation. Whether RanGTP is required early in cytokinesis to 'prime' the cortex for a future ingression or acts directly later during the ingression process is unclear. Testing these hypotheses is not straightforward, since Ran is also required for organizing the mitotic microtubule cytoskeleton, which is required for cytokinesis. Taken together these findings suggest an additional mechanism involved in regulating cytokinesis that is dependent on signals from chromosomes in addition to those stemming from the different organizational states of the mitotic microtubule cytoskeleton (Silverman-Gavrila, 2008).
Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.