BIOLOGICAL OVERVIEW

Morphological complexity of neurons contributes to their functional complexity. How neurons generate different dendritic patterns is not known. The sequoia mutant was identified from a screen for dendrite mutants. Sequoia is a pan-neural nuclear protein containing two putative zinc fingers homologous to the DNA binding domain of Tramtrack. sequoia is required for specification of a subset of neurons such as the es neurons in the embryo; however, sequoia also functions in regulating dendrite and axon development of every neuronal type examined, including es and other PNS neurons, photoreceptors, and motoneurons in the CNS. In support of sequoia as a specific regulator of neuronal morphogenesis, microarray experiments indicate that sequoia may regulate downstream genes that are important for executing neurite development rather than altering a variety of molecules that specify cell fates (Brenman, 2001).

The Drosophila peripheral nervous system (PNS) is a good assay system for the genetic study of dendrite development because it is small enough, containing only 44 neurons per hemisegment, yet most multiple dendritic (md) neurons exhibit complex dendritic branching patterns. Having developed a system using green fluorescent protein (GFP) to visualize dendrite outgrowth in real time in living embryos (Gao, 2000), it was found that dendrite development follows a stereotyped spatial and temporal pattern. Dendritic outgrowth begins after an initial 'bud' site(s) has been chosen. Following primary dorsal dendrite extension, additional lateral dendritic processes elaborate and retract with only a subset becoming stabilized to form secondary dendritic branches (Brenman, 2001).

By monitoring the morphogenesis of these PNS neurons, attempts were made to isolate mutants that would help to dissect the developmental processes in dendrite formation. Some of the questions that might be addressed include: (1) What controls the distinct dendrite morphology of neurons within the same lineage? Several models have been proposed to describe possible lineages that give rise to md neurons. Despite differences in the exact relationship between external sensory (es) and md neurons, these models predict that some of the md and es neurons are derived from a common precursor cell. The dendritic morphologies of the two neuronal types produced by the same precursor are drastically different. The es neuron possesses a single unbranched dendrite, while most md neurons have multiple, highly branched dendritic processes. What causes these neuronal descendents of a common precursor cell to adopt distinct dendrite morphology is unknown. (2) How is the dendritic field of any given neuron specified? Are there 'universal' regulators involved in controlling the morphology of most and possibly all neurons? In the screen, a mutant, sequoia, was identified that provides insight to both of these questions. sequoia alters both the degree of branching and the resulting dendritic field. Dendrites from sequoia mutant embryos display an excessive outgrowth phenotype that is caused by precocious dendrite development as well as an inability to stop dendrite elongation at the appropriate time during embryonic development. sequoia mutants exhibit severe axon breaks in both the PNS and CNS. sequoia also plays a role in cell fate decisions since mutant embryos exhibit excess md neuron production at the expense of es neurons. Thus, similar to prospero, sequoia belongs to an emerging class of genes that affect the cell fate of a subset of neurons and regulate neuronal process formation in a much larger population of neurons. sequoia has more widespread effects on neuronal morphology than other pan-neural genes studied to date. As a likely transcription regulator, sequoia appears to preferentially regulate genes that control morphogenesis rather than genes known for directing cell fates in the nervous system (Brenman, 2001).

The screen utilized a Gal4 enhancer trap line driving GFP expression in md but not es neurons in the dorsal cluster of the embryonic PNS. In sequoia mutant embryos, a range of excessive dorsal dendrite outgrowth was observed for GFP-positive md neurons that fall into two groups. Neurons in one group occupy the same position as that in wild-type embryos. Neurons in the other group are located more dorsally at sites normally occupied by es neurons. The second group of neurons tend to fasciculate their dendritic processes together and extend fewer lateral branches. In either group of neurons, dendrites are abnormally long at late embryonic stages so that dendrites from dorsal clusters in the two hemisegments cross the dorsal midline and intermingle with each other. This is in stark contrast to dendrites from wild-type embryos that are significantly shorter and never reach the dorsal midline in newly hatched larvae (Brenman, 2001).

The excessively long dendrites observed in the mutant could arise from premature growth initiation, faster dendritic growth, and/or an inability to halt growth at the appropriate time. To examine these possibilities, multiple staging criteria were used as well as time-lapse analysis to determine when dendrite growth started and stopped, as well as the rate of growth. Immediately before the formation of cuticle that is impermeable to antibody (stage 16), dendrites labeled with 22C10 antibody were longer in sequoia mutants than those in wild-type embryos. Time-lapse analysis reveals that mutant embryos did not extend dendrites at a faster rate than that in wild-type embryos. Rather, mutant embryos initiate dendritic outgrowth precociously. Immunostaining of embryos at 11-12 hr after egg laying (AEL) revealed significant md dendrite extension in >85% of sequoia mutants but only in <10% of wild-type embryos. At this stage, es dendrites can be seen in wild-type embryos. The es dendrite is morphologically distinct from md dendrites; it has a characteristic dendrite 'squiggle' morphology and a dendritic cap that innervates a cuticular structure of the es organ and exhibits intense staining with the 22C10 antibody. Such es dendrites were not readily discerned in mutant embryos. In addition to cuticle formation, a second independent staging criterion was used to demonstrate precocious md dendrite extension in sequoia mutants. In sequoia mutants, md dendrites were detected as early as stage 14, when the midgut undergoes characteristic cell movements and constrictions. However, wild-type embryos that had progressed beyond this stage of gut development were typically devoid of md dendrites, in contrast to the precocious dendritic outgrowth in sequoia mutants (Brenman, 2001).

Not only did sequoia mutants initiate dorsal dendrite outgrowth abnormally early, they also failed to halt dorsally oriented outgrowth later in embryonic development. Time-lapse analysis revealed that in wild-type embryos, once the dorsal dendrites reach a certain length (at 16-17 hr AEL), dorsally oriented dendritic growth stops and only lateral branches increase in length (Gao, 2000). In sequoia mutants, the dorsally oriented dendritic tips continue growing toward the dorsal midline and eventually approach the dendritic tip from the contralateral homologous cluster. Failure to stop dorsal dendrite elongation is evident for all neurons in the mutant dorsal clusters. For both the md neurons with lateral branches and the extra md neurons that fasciculate their dendrites together and occupy positions of es neurons, their dorsal dendrites extend into the dorsal most zone normally devoid of dendrites at this developmental stage (Brenman, 2001).

Is the extensive overgrowth of dendrites in md neurons associated with any axonal defects? Characterization of the PNS stained with 22C10 antibody reveals that, instead of having normal axon fascicles that have entered the CNS, axons emerging from the dorsal cluster of sequoia mutants were fewer in number and frequently misrouted or stopped altogether in their path of outgrowth toward the CNS. Axons terminating prematurely were observed not only for md neurons but also for all other PNS neurons, resulting in complete axon breaks between PNS clusters. es neurons with the characteristic 'squiggle' morphology were frequently missing in sequoia mutants. Compared to wild-type embryos of the same developmental stage, shortened axons were routinely detected in sequoia mutant exhibiting abnormally elongated dendrites from PNS dorsal cluster md neurons. Immunostaining of sequoia mutants with monoclonal antibody 1D4 shows an axonal phenotype in the CNS similar to that in the PNS. Large fascicles of longitudinal axon tracts evident in the CNS of wild-type embryos are absent in mutant embryos. In addition, motoneuron projections to the periphery are also greatly reduced (Brenman, 2001).

A priori, one could imagine two extreme scenarios to generate the numerous subtypes of neurons, glia, and support cells. At one extreme, every specialized cell type has its own 'master regulator' molecule necessary and sufficient to generate a particular cellular identity. In this case, any cell fate transformation must result in morphological changes, and morphological changes must always be accompanied with cell fate changes. However, it is known from the examples of prospero and Notch mutants that this is not always the case. Cell fate phenotypes can often be separated from morphological phenotypes. This dichotomy clearly illustrates multiple functions for a single molecule. More likely, evolution hones genes and pathways to execute a particular task that will be used repeatedly in multiple developmental processes, including both cell fate specification and morphogenesis. As a consequence, there need not always be a link between them. For instance, a combination of Notch alleles and UAS-Notch transgenes has been used to produce cell fate transformation without axon defects or axon defects without cell fate transformation (Brenman, 2001).

sequoia is similar to the prospero locus in many respects. Both molecules are required to achieve specific cell fates in a small subset of the Drosophila nervous system. However, mutations of both genes affect many other cells besides those that exhibit clear fate changes. Both prospero and sequoia mutants display axon and dendrite phenotypes in addition to cell fate alterations. Clearly, genes involved in regulating neuronal morphology can have independent effects on cell fate and morphology. The prospero mutations affect cell fate in the CNS but not PNS and alter neuronal morphology in both. sequoia thus appears to be similar to prospero but has even broader roles in regulating neuronal morphogenesis (Brenman, 2001).

In sequoia mutant embryos, extra md neurons are generated at the expense of es neurons in the PNS dorsal cluster. This is highly reminiscent of Notch mutant embryos where extra neurons in the PNS are generated, and all of them appear to be md neurons by both enhancer trapping and morphological criteria. Moreover, these extra md neurons in Notch mutant embryos provide an example of generating md neurons at the expense of es neurons. Taken together, these observations support the notion that Sequoia could function in the Notch pathway. Not only are there phenotypic similarities, but Sequoia shares sequence homology (~55%) to the DNA binding zinc finger domains of Tramtrack. Tramtrack has been shown to function downstream of Notch in specifying cell fate in the nervous system. Unlike tramtrack, however, sequoia can alter dendrite morphology without changing cell fate. Whereas the es neurons normally have a single straight dendrite, extra dendritic 'arbors' of es neurons are occasionally observed in adult loss-of-function sequoia mutant clones even though they are clearly associated with external sensory structures typical of es organs. This sets sequoia apart from tramtrack and other mutants that affect cell fate in the nervous system without producing aberrant neuronal morphologies. Another example of this latter group is cut. cut mutants have es organs transformed into chordotonal organs, yet these transformed organs still display dendrites characteristic of their altered identity and axons that enter the CNS. In summary, sequoia resembles prospero and Notch as a molecule that can regulate cell fate independent of morphogenesis (Brenman, 2001).

During Drosophila embryogenesis, dorsal dendrite outgrowth of md neurons stops several hours before the larvae hatches, leaving a large uninnervated region flanking the dorsal midline. This area is eventually covered with dendrites by the second instar larval stage. Studies of the formation of dorsal dendritic territories (Gao, 2000) suggest that both intrinsic and extrinsic factors participate in defining dendritic fields. Several mutants identified in a screen (including sequoia described here) display a dorsal dendrite overextension phenotype, wherein the dorsal dendrites from the two hemisegments meet and/or intermingle at the dorsal midline in the embryo. In theory, this phenotype could be due to earlier dendrite outgrowth, faster dendrite outgrowth, and/or failure to halt dorsally oriented dendrite outgrowth. Mutants such as flamingo fail to stop dorsally oriented outgrowth (Gao, 2000), while sequoia mutants not only start dendrite outgrowth abnormally early but also fail to stop outgrowth later in embryogenesis. The nature of the dorsal dendrite outgrowth 'stop' signal in embryos remains unknown but is unlikely to be governed by competition from homologous neuronal dendrites, since these dendrites do not normally meet during embryogenesis. One could imagine different ways in which developing dendrites recognize 'stop' signals. In one scenario, an inherent 'ruler' predetermines the final dendrite length. This inherent ruler is either absent or reset to a longer length in sequoia mutant embryos. In another scenario, developing dendrites detect a 'stop' signal (e.g., a ligand present at the dorsal midline), and sequoia is required to generate the components necessary to respond to and transduce this 'stop' signal (Brenman, 2001).

If cytoskeletal or other components are required for assembly of both axons and dendrites, could it be that mutants that result in dendrite 'rich' structures may come at the expense of axonal structures and vice versa? Several of the mutants isolated in prior mutant screens display 'extra' dendritic branching or dorsally oriented dendrite outgrowth. These include tumbleweed, flamingo, kakapo, and sequoia. Interestingly, all of these mutants exhibit axonal aberrations. Specifically, some of the axons from these mutants fail to extend fully to reach their targets. Conversely, mutants such as shrub and shrinking violet that display a decrease in dendritic structure do not exhibit shortened axons. It remains to be seen whether these correlations between dendrites and axons reflect a true competition for resources between them or simply the involvement of the same genes in both axon and dendritic outgrowth. Alternatively, the presence of different molecular machinery for the elaboration of axons and dendrites could explain why the same gene would lead to very different effects on axon versus dendrite outgrowth (Brenman, 2001).

sequoia appears to be a universal regulator of morphology in Drosophila neurons. No another pan-neural gene, when removed, causes as widespread and profound defects in neuronal morphogenesis. Both phenotypic and microarray analyses indicate that sequoia regulates multiple aspects of neuronal morphogenesis, including dendrite arborization, length of neurites, and axon guidance. Preliminary microarray analysis further indicates that sequoia regulates, directly or indirectly, the expression of several genes involved in axon guidance as well as novel genes potentially important for neurite outgrowth, without affecting transcript levels for the vast majority of genes examined. This includes genes that specify cell fates and many genes expected to be required by all neurons. Further mechanistic insight may be gleaned from future studies of the downstream genes of sequoia at the cellular level for their potential roles in controlling neuronal morphology (Brenman, 2001).

GENE STRUCTURE

cDNA clone length - 3560

Bases in 5' UTR - 546

Exons - 4

Bases in 3' UTR - 392

PROTEIN STRUCTURE

Amino Acids - 873

Structural Domains

The sequence of a full-length EST reveals that sequoia encodes a predicted protein of 873 amino acids containing two zinc fingers and numerous opa, polyglutamine-rich repeats. Sequoia protein has homology to the zinc finger DNA binding domains of Tramtrack showing greater than 55% identity in this region (59/102 amino acids). Furthermore, the spacing between cysteines and between histidines in the zinc fingers are conserved (Brenman, 2001).

date revised: 20 December 2001

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

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