The lesswright (lwr) or semushi gene encodes an enzyme, Ubc9, that conjugates a small ubiquitin-related modifier (SUMO). Since the conjugation of SUMO occurs in many different proteins, a variety of cellular processes probably require lwr function. This study demonstrates that lwr function regulates the production of blood cells (hemocytes) in Drosophila larvae. lwr mutant larvae develop many melanotic tumors in the hemolymph at the third instar stage. The formation of melanotic tumors is due to a large number of circulating hemocytes, which is approximately 10 times higher than those of wild type. This overproduction of hemocytes is attributed to the loss of lwr function primarily in hemocytes and the lymph glands, a hematopoietic organ in Drosophila larvae. High incidences of Dorsal (Dl) protein in the nucleus were observed in lwr mutant hemocytes, and the dl and Dorsal-related immunity factor (Dif) mutations were found to be suppressors of the lwr mutation. Therefore, the lwr mutation leads to the activation of these Rel-related proteins, key transcription factors in hematopoiesis. Also demonstrate was the fact that dl and Dif play different roles in hematopoiesis. dl primarily stimulates plasmatocyte production, but Dif controls both plasmatocyte and lamellocyte production (Huang, 2005). Similar results have been reported by Chiu (2005). Loss-of-function mutations in dUbc9 cause strong mitotic defects in larval hematopoietic tissues, an increase in the number of hematopoietic precursors in the lymph gland and of mature blood cells in circulation, and an increase in the proportion of cyclin-B-positive cells. In the larval fat body, dUbc9 negatively regulates the expression of the antifungal peptide gene Drosomycin, which is constitutively expressed in dUbc9 mutants in the absence of immune challenge. dUbc9-mediated drosomycin expression requires Dorsal and Dif. Although both studies are clear in concluding that Ubc9/Lesswright appears to function upstream of Cactus and Dorsal/Dif, the specific target is not yet known. Wild-type Ubc9/Lesswright hold the Toll pathway in check, and could target Cactus, Dorsal/Dif or another upstream component in the pathway (Huang, 2005; Chiu, 2005).

What is a possible molecular mechanism for the regulation of the Toll pathway? The Lwr protein can regulate activities of its target proteins by conjugating SUMO molecules. For example, the human Cact homologue, IκBα, is known to be sumoylated, and the sumoylated IκBα molecules have been shown to be resistant to ubiquitination because the ubiquitination sites on IκBα are occupied by SUMO molecules (Desterro, 1998). In other words, sumoylation stabilizes IκBα, which in turn represses NF-κB activity. Alternatively, their physical associations alone can control physiological functions of Lwr's targets. Recently, the Paired-like homeobox protein Vsx-1 was reported to physically interact with Ubc9 (the vertebrate Lwr homologue) and to require Ubc9 function for Vsx-1's nuclear localization. However, sumoylation of the Vsx-1 protein was not detected (Kurtzman, 2001). In this case, physical association with Ubc9 itself is a key for Vsx-1 function (Huang, 2005).

The Lwr protein may regulate activities of its target proteins by conjugating SUMO molecules or simply by physical association with its target. In the Tl (or Tl-like) pathway of Drosophila, either of these mechanisms are possible. The Dorsal (Dl) protein was shown to physically interact with Lwr and to be sumoylated in a transfection experiment using Schneider L2 cells. Sumoylated Dl proteins dissociated from Cactus (Cact) and entered the nucleus, where they exhibited an increased level of transactivation (Bhaskar, 2000; Bhaskar, 2002). However, this model does not fit observations of higher incidence of Dl nuclear localization in lwr mutant hemocytes than in wild type hemocytes (Huang, 2005).

Another possible target protein is Cact. As in IκBα, the Cact protein was shown to physically associate with Lwr (Bhaskar, 2000). Unlike IκBα, sumoylated Cact proteins were not detected in the same study. However, physical association of Lwr and Cact might stabilize Cact proteins in the cytoplasm by counteracting ubiquitination or phosphorylation, a prerequisite for ubiquitination of Cact. In this scenario, the loss of lwr function would make Cact be susceptible to degradation, which results in migration of Dl and Dif to the nucleus. Other components in the Tl pathway could be regulated similarly, although they have not been reported to interact with Lwr (Huang, 2005).

The innate immune system is evolutionarily ancient and common among most eukaryotes and consists of humoral and cellular components. In Drosophila, the humoral response primarily represents the production of antimicrobial peptides in the fat body. The cellular innate immunity is handled by circulating blood cells capable of recognizing and neutralizing foreign objects. These cells also work as scavengers of apoptotic cells in normal development. Because the balance of these functions is so important, the proliferation and differentiation of hemocytes must be tightly regulated (Huang, 2005).

The Drosophila larval hemocyte population consists of several cell types, some of which reside only in a blood-cell-forming organ, the lymph gland. Three cell types, plasmatocytes, crystal cells, and lamellocytes (the subject of this study), are present in hemolymph. The majority are plasmatocytes, which are able to phagocytose microbes and apoptotic cells. A small fraction (<5%) are crystal cells, which are involved in melanization reactions. Rarely found in circulation in normal circumstances, lamellocytes are key defensive players when larvae are infested by parasitoid wasp eggs. In addition to these cells, prohemocytes (hematopoietic stem cells) and secretory cells are found in the lymph glands. A small number of prohemocytes circulate in hemolymph. The population structure of hemocytes changes as larvae develop and undergo metamorphosis. Several genes have been found to direct differentiation of specific hemocytes. A Drosophila GATA homologue, serpent, is required for the expression of two lineage-specific genes, glial cell missing (gcm) and lozenge (lz). The gcm gene encodes a transcription factor and is required for the plasmatocyte lineage. A Drosophila acute myeloid leukemia-1 homologue, lz, and the Notch pathway are required for the crystal cell lineage. Recently, yantar and collier were found to direct lamellocyte differentiation. However, precise mechanisms of the proliferation and differentiation of hemocytes still remain elusive (Huang, 2005).

Three signal transduction pathways are known to influence the number of hemocytes in circulation. They are the Ras, JAK/STAT, and Tl (dealt with in this study) pathways. Mutations that activate these pathways increase the total hemocyte numbers 10- to 100-fold and often stimulate lamellocyte production in the absence of parasitoid wasp infestation. Ras functions in the MAP kinase pathway and stimulates cell division. It is one of the most common oncogenes found in many different human cancers. When the Drosophila ras1 gene is overexpressed in the lymph gland and hemocytes, total hemocyte counts increase nearly 100-fold. The JAK-STAT pathway is a conserved signal transduction pathway and plays a variety of roles in Drosophila and many other eukaryotes including humans. A dominant mutation of JAK, hopscotch^Tum-l, causes a drastic increase in plasmatocytes as well as lamellocytes in a temperature-dependent manner. Similarly, dominant mutations of Toll, Tl^10B, and Tl³, produce a large number of hemocytes. In this case, the lamellocyte population increases to 10%-20% of the entire hemocyte population. Several genes in the Tl signaling pathway also exhibit a similar defect when mutated. Although it is not clear if these signal transduction pathways possess lineage specific functions, they play some general role in Drosophila hematopoiesis (Huang, 2005).

The studies of the NF-κB pathway, discovered in 1986, have made a huge contribution to the understanding of mammalian immune systems, particularly in acquired (or adaptive) immunity. NF-κB is a dimer of members of the Rel-related transcription factors. A unique aspect in the NF-κB pathway is that Inhibitors of κB, IκBs, keep the transcription factor NF-κB in the cytoplasm from its action in the nucleus. Ubiquitination, thus degradation, of IκB is required for the activation of the NF-κB pathway. A Drosophila Rel-related gene, dl, was discovered to be an important component in the establishment of the embryonic dorsal-ventral axis and was found to be a part of Tl signaling. The activation of Tl signaling starts from the cell surface receptor Tl and ends at the entrance of the Dl transcription factor into the nucleus. The function of Tl signaling in innate immunity was later discovered indirectly. Sequence analysis of several antimicrobial peptide genes spotted κB sites, binding sites for NF-κB and Dl, in their 5′ regulatory regions, thjat connected the Tl pathway to humoral immunity. Discovery and subsequent studies of other Rel-related genes, Dif and Relish, further support a key role of the Tl pathway in Drosophila immunity. Furthermore, the homology between mammalian IκBs and Cactus (Cact) demonstrates that the parallelism extends to the regulatory mechanism of mammalian and Drosophila NF-κB activity. The presence of multiple Rel-related proteins and Tl-like receptors in Drosophila as well as in other eukaryotes makes the system more complex, but versatile in innate immune responses and hematopoiesis. In the last decade, remarkable progress has been made toward the understanding of Drosophila humoral immunity including the role of the Tl pathway. In contrast, the regulatory mechanisms of cellular immunity are much less understood (Huang, 2005).

Recently, ubiquitin's small cousins, SUMO molecules, were discovered to add another layer of complexity in NF-κB signaling. SUMO conjugation (sumoylation) regulates a variety of cellular functions (Melchior, 2000; Muller, 2001; Yeh, 2000). Proteins subjected to sumoylation are involved in oncogenesis, transcriptional regulation, nuclear transport, and others. Ubiquitin and SUMO molecules are biochemically quite similar to each other. Furthermore, they require a set of similar modifying enzymes, activating (E1), conjugating (E2), and ligating (E3) enzymes, all of which show high conservation among most eukaryotes. The similarity between ubiquitination and sumoylation creates an interesting situation where the same lysine residues on a given protein can be modified by both ubiquitin and SUMO. Thus, ubiquitination and sumoylation can counteract each other. This scenario is seen in the regulation of IκBα degradation (Desterro, 1998). In addition to NF-κB signaling, the activity of the JAK/STAT pathway might be regulated in part by sumoylation since a Protein inhibitor of activated STAT (Pias) possesses a RING finger motif, an E3 SUMO ligase signature. However, these possibilities have not yet been explored in Drosophila hematopoiesis (Huang, 2005 and references therein).

This study reports that the Lwr protein, a SUMO conjugase, plays an important role in regulation of larval hematopoiesis in Drosophila melanogaster. Recessive as well as dominant negative mutations of the lwr gene result in the overproliferation of hemocytes in larvae. Loss of lwr function led to the accumulation of the Rel-related protein Dl in the nuclei of circulating hemocytes, and dl and Dif mutations are suppressors of the lwr mutation. Taken together, these results indicate that the lwr function is inhibitory to dl and Dif activities. Furthermore, it is demonstrated that dl and Dif possess different properties in hematopoiesis, particularly in lamellocyte production (Huang, 2005).

Thus mutation of the lwr gene, which encodes a SUMO-conjugating enzyme, causes the activation of Rel-related proteins, Dl and Dif, and leads to overproduction of larval hemocytes. Immunochemical and genetic analyses indicate that the lwr mutation results in the activation of Dl and Dif primarily in plasmatocyte and prohemocytes, leading to overproliferation of plasmatocytes and lamellocytes in larvae. Dl nuclear localization in the lwr mutant hemocytes was observed as frequently as those in the Tl^10B mutant hemocytes. The high hemocyte counts of the lwr mutant larvae are almost completely suppressed by removing both dl and Dif functions from the lwr mutant background. Therefore, it is proposed that the wild type lwr function negatively regulates the activity of the Tl pathway and/or the pathways of Tl-like genes (Huang, 2005).

It has been demonstrated that lwr function is required for the Ran-dependent nuclear transport system (Epps, 1998). This raises a question of how efficiently Dl and Dif proteins are transported to the nucleus in lwr mutant hemocytes. Nuclear import of Dl and Dif was not overly impaired. These contrary observations can be explained by functional redundancy of lwr with 25 other Drosophila E2 enzymes. Since functions of most E2 enzymes have not been clearly defined, some of them may be SUMO conjugases. This possibility was also suggested in the embryo, particularly in the posterior half of the embryo (Epps, 1998). Alternatively, the residual lwr function may be sufficient to mediate nuclear import of Dl proteins because hypomorphic lwr alleles were used in this study. In any case, a possible negative effect on nuclear transport appeared minimal in lwr mutant hemocytes (Huang, 2005).

The loss of dl function in the lwr mutant background resulted in a reduction of plasmatocyte, not lamellocyte counts, and the overexprssion of dl with the CgGAL4 driver stimulated plasmatocyte production only. Therefore, the role of dl in hematopoiesis is limited to plasmatocyte production. Although dl plays a role in plasmatocyte production in the wild type background, the contribution of dl to plasmatocyte production seems minimal when Dif is overexpressed. In other words, Dif can replace dl function in Drosophila hematopoiesis (Huang, 2005).

In contrast, Dif is capable of stimulating the production of both plasmatocytes and lamellocytes. Because the loss of Dif function in the lwr mutant background led the lamellocyte population to a level very close to those of wild type strains, Dif is most likely to play an essential and sole role in lamellocyte production when the Tl signaling is activated by the lwr mutation. Furthermore, when Dif was overexpressed by the CgGAL4, the levels of plasmatocytes and lamellocytes were similar to those reached when Tl^10B was overexpressed by the same driver. Taken together, it is concluded that Dif is necessary and sufficient to promote the production of plasmatocytes and lamellocytes in larvae in the Tl (Tl-like) pathway. This also means that Dl-responsive genes can be regulated by Dif (Huang, 2005).

It is hypothesized that the genes for plasmatocyte production can be divided into at least two classes. The first class consists of genes regulated primarily by Dl. The genes in this class can be stimulated by Dif in the absence of Dl. It is known that Dl proteins bind to more strictly defined κB sites than Dif, and that Dif can bind more broadly to κB consensus sequences including Dl-binding sites. Several humoral immunity genes were found to belong to this class. Thus, it is possible that such genes exist among genes essential for plasmatocyte production. The other class includes genes regulated only by Dif. In these genes, consensus κB sites are probably responsible for their expression via the Dif protein (Huang, 2005).

The Dl protein can function as a repressor in a context-dependent manner. A good example in the Drosophila immunity is the Cecropin gene, whose expression depends heavily on Dif. Co-expression of Dl expression along with Dif strongly represses the expression of the Cecropin gene. However, no significant dominant negative effect of dl on plasmatocyte production was observed in overexpression experiments. At the same time, no synergistic effect on plasmatocyte production was observed when both dl and Dif were co-expressed. Therefore, regulatory mechanisms of plasmatocyte-specific genes are not so complex (Huang, 2005).

The collagen IV (Cg25C) gene, whose promoter was used for the CgGAL4 driver, is expressed in embryonic and larval hemocytes (Yasothornsrikul, 1997). These larval hemocytes are most likely plasmatocytes, although the authors of the study did not distinguish plasmatocytes and prohemocytes. Nonetheless, the current results show that these Cg25C-expressing hemocytes are capable of proliferating and differentiating into plasmatocytes when Dl and Dif is activated (Huang, 2005).

It was surprising that the CgGAL4 driver induced a relatively large number of lamellocytes because this driver is not capable of inducing GAL4 in lamellocytes and because the collagen IV gene is expressed primarily in plasmatocytes. This can be interpreted in two different ways. The first one is that the Cg25C gene is expressed in lamellocyte precursors in the lymph gland, where these precursor cells can divide. In addition to different cell types distinguished by their morphologies, the cells in the lymph gland can be classified based on expression of different genes. For example, the Posterior Signaling Center, which is important for crystal cell differentiation, was defined as Serrate-expressing cells in the first lobe of the lymph gland. Thus, it is conceivable that the Cg25C gene might be expressed in the cells that have not been clearly defined in the lymph gland. The second interpretation is that plasmatocytes can differentiate into lamellocytes when Dif are highly activated. This interpretation supports a classical view of lamellocyte differentiation. Further investigations are necessary to answer these questions (Huang, 2005).

GENE STRUCTURE

cDNA clone length - 1259

Bases in 5' UTR - 246

Exons - 2

Bases in 3' UTR - 533

PROTEIN STRUCTURE

Amino Acids - 159

Structural Domains

The yeast UBC9 and hus5 gene products have been identified as putative E2 members of the ubiquitin-conjugating enzyme (UBC) family and have been shown to play an essential role in cell cycle progression. A Drosophila Ubc9/Hus5 homologue (termed dUBC9) has been identified in an attempt to identify proteins that interact with the amino-terminal transcriptional repression domain of the Groucho corepressor. The predicted dUBC9 protein consists of 159 amino acids and shows 85%, 68%, and 54% amino acid sequence identities with human UBC9 homologue, Schizosaccharomyces pombe Hus5, and Saccharomyces cerevisiae Ubc9 proteins, respectively. Expression of dUBC9 cDNA complements a temperature-sensitive ubc9-1 mutation of S. cerevisiae to fully restore normal growth, indicating that the dUBC9 protein can act as a substitute for the yeast Ubc9 protein. The dUBC9 transcripts were about 1.2 kb and were detected at all stages of Drosophila development and in ovaries and Schneider cells. However, an increased level was observed in early embryos and ovaries. The dUBC9 gene is present as a single copy in the genome and localized in segment 21C-D on the left arm of the second chromosome (Joanisse, 1998; Ohsako, 1999).

The interaction between small ubiquitin-related modifier SUMO and its conjugating-enzyme Ubc9 (E2) is an essential step in SUMO conjugation cascade. However, an experimental structure of such a transient complex is still unavailable. A structural model of SUMO-3-Ubc9 complex has been obtained with HADDOCK, combining NMR chemical shift mapping information. Docking calculations were performed using SUMO-3 and Ubc9 structures as input. The resulting complex reveals that the complementary surface electrostatic potentials contribute dominantly to the specific interaction. At the interface, similar numbers of oppositely-charged conserved residues are identified on the respective binding partners. Hydrogen bonds are formed in the vicinity of the interface to stabilize the complex. Comparison of the structure of SUMO-3-Ubc9 complex generated by HADDOCK and the experimental structures in free form indicates that SUMO-3 and Ubc9 maintain their respective fold as a whole after docking. However, the N-terminal helix alpha1 and its subsequent L1 loop of Ubc9 experience sizeable changes upon complex formation. They cooperatively move towards the hydrophilic side of the beta-sheet of SUMO-3. These observations are consistent with the data from previous Ubc9 mutational analysis and conformational flexibility studies. Together, it is proposed that the SUMO-3-Ubc9 interaction is strongly electrostatically driven and the N terminus of Ubc9 shifts to SUMO-3 to facilitate the interaction. The NMR-based structural model, which provides considerable insights into the molecular basis of the specific SUMO-E2 recognition and interaction, implicates the general interaction mode between SUMO-3 and Ubc9 homologues from yeast to humans (Ding, 2005).

date revised: 22 June 2006

Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.