The following musings concern certain figures that warrant further scrutiny. Through these distillations, I attempt to draw some general conclusions about how the fly's control system operates and why it evolved this way. Some of the annotations also offer historical perspective.

Figure 2.1 One corollary issue (symbolized by the hourglass) is: how do cells measure time over periods longer than a mitotic cycle? Cells presumably need to do so in order to know when to stop dividing and start differentiating. Possible timekeeping devices include the “POU Hourglass,” which limits the number of mitoses in certain neuroblasts in the fly CNS. This clock gauges the declining amount of the POU-domain proteins Pdm-1 and Pdm-2. Mammalian oligodendrocytes use an “HLH Hourglass” that triggers differentiation when the amount of specific HLH-domain proteins drops below a critical threshold. An oscillator based on this sort of mechanism may be involved in vertebrate somitogenesis. Other protein clocks appear to count mitoses and meioses leading to sperm and egg differentiation. Another strategy involves using a cascade of transcription factors to trigger different events in different phases of the cascade. For RNA clocks, and for general discourses on (noncircadian) timekeeping.

A deeper issue concerns how structures are represented abstractly in the genome. Given that we know most of the genes involved in bristle development (cf. App. 3), we can begin to ask how “bristle” is “written” in “gene language.” The answer is not obvious.

The epidermis of a D. melanogaster adult has on the order of 500,000 cells, ∼5,000 of which (∼1%) make bristles. A priori, it would seem reasonable to expect bristles to sprout as randomly as the hairs on a human arm. However, even the most scattered bristles – the tergite microchaetes (mCs) – have fairly uniform spacing. At the other extreme of precision are the 40 macrochaetes (MCs) on the head and thorax, whose basic layout has been conserved for 50 million years (Fig. 3.1).

Except for the MCs, the bristles of each body region tend to vary in number and position from one fly to the next. Interestingly, most bristles are organized in rows that run parallel or perpendicular to axes of the body or limbs. Within such rows, the bristles are aligned more or less accurately and are spaced more or less evenly. Different rules govern different patterns. Thus, notal mCs form jagged rows along the anteriorposterior axis, while wing bristles form straight rows along the margin, eye bristles arise at alternating vertices of each ommatidium, and belly (sternital) bristles are spaced at intervals proportional to their shaft lengths.

Why do such patterns exist? Surely, some are adaptive. For example, flies use “brushes” (parallel transverse rows) on the legs to wipe dust from their eyes, and other patterns appear to map air currents, prevent wetting, or act as shock absorbers. However, many may simply be accidents of evolution.

Three of the 5 chief transduction pathways used by discs are outlined below and diagramed in Figure 5.6. The other two cardinal pathways–Notch and EGFR–are discussed in Chapters 2 (Fig. 2.2) and 6 (Fig. 6.12), respectively. Abbreviations: “MF” (morphogenetic furrow), “vh” (vertebrate homolog).

Hedgehog signaling pathway (see for overview). Although the agents below constitute the standard version of this pathway, notable deviations have been found. However, the heretical proposal that Hh controls target genes in Bolwig's organ without employing Cubitus interruptus has been refuted. Aside from the components below, the zinc-finger protein Combgap regulates the levels of Cubitus interruptus in leg, wing, and eye discs but acts in the Wingless pathway in optic lobes. Another possible player is oroshigane (unknown role), which acts upstream of patched. Other genes have been implicated (e.g., shifted), but how their products fit into the chain remains to be determined.

Hedgehog (Hh, named for the lawn of spiky denticles in LOF embryos) is a 471-a.a. polypeptide (nascent form) that cleaves itself (between residues 257 and 258) into C- and N-terminal moieties: “HhC” bears the catalytic active site, whereas “HhN” is the signaling fragment. HhC is soluble.

Tactile stimuli are hard for arthropods to detect through the armor of their rigid exoskeleton. To solve this problem, flies use bristles (Fig. 2.1). When a bristle is deflected, the pivoting of the shaft in its socket deforms the dendrite of a neuron attached to the shaft's base. The resulting depolarization sends an action potential to the central nervous system (CNS). Flies can pinpoint sensations because axons from different bristles get “wired” to different CNS target cells during metamorphosis, although much remains to be learned about the topology of these neurosensory maps (cf. Ch. 6).

Mechanosensory bristles are formed by 5 cells: 2 superficial cells that secrete cuticle (the shaft and socket cells) and 3 subepidermal cells that do not (the neuron, sheath, and glial cells). These 5 cells descend from a “sensory organ precursor” (SOP). The SOP divides to produce one daughter (IIa) that yields the outer cells, and another (IIb) that yields the inner cells. The sheath cell wraps the neuron's dendrite, while the glial cell wraps the axon. A sixth cell – the “bract cell” – is found in association with bristles on the distal leg and proximal wing. It secretes a thickened hair (“bract”) that is pigmented like the bristle shaft but much smaller. The bract cell is not part of the SOP clone. The way in which it is recruited from epidermal cells is discussed later.

Listed below are all the major models (a.k.a. hypotheses, metaphors, scenarios) and mysteries (a.k.a. enigmas, riddles, paradoxes) discussed in the text, with definitions in parentheses. Also listed are devices (a.k.a. gadgets, tricks) and epiphanies (a.k.a. principles, rules). Asterisks indicate names that were coined for convenience. Having a taxonomyof concepts – even a silly taxonomy – is a useful heuristic for thinking. Within each category, concepts are listed alphabetically. Numbers in bold are pages where the ideas are diagramed.

Few phenomena are as entrancing as the transformation of one thing into another. The ancients believed that sorcerers had such powers, and modern magicians can still fool children with illusions of this sort. A special class of mutations can actually accomplish this feat.

“Homeosis” means a transformation of one body part into another. The term was coined by William Bateson to describe deformities that are occasionally found in nature. In his classic 1894 monograph, Bateson cataloged 886 abnormal biological specimens, many of which exhibited homeosis. His intent was to investigate how anatomy varies as a way of comprehending how evolution works. This goal was obvious from the book's overtly Darwinian title: “Materials for the Study of Variation Treated with Especial Regard to Discontinuity in the Origin of Species.”

In an article entitled “Pattern formation in the embryo and imaginal discs of Drosophila: What are the links?”, Adam Wilkins and David Gubb posed a question that was on the minds of many researchers at that time (1991). The embryo's segmentation hierarchy was basically understood, but it was unclear what these various genes might be doing in discs. Wilkins and Gubb argued that segment-polarity genes supply the angular values of the Polar Coordinate (PC) Model, which until then had only been an abstract formalism.

Several predictions of this “Angular Values Conjecture” were soon put to the test by molecular genetics. Chief among them was the expectation that LOF and GOF alterations in segment-polarity genes should reorganize disc anatomy. This prophecy was indeed fulfilled, and the experimental probing uncovered a trove of insights into the machinery of disc patterning.

The conjecture itself, however, turned out to be wrong. Segment-polarity genes do not paint a pinwheel on each disc. Rather, they draw a few important lines – the compartment boundaries.

The Molecular Epoch of disc research was launched in 1991

The Wilkins-Gubb paper, with its clarion call for a molecular assault on discs, provides a convenient demarcation between the Cellular and Molecular Epochs of disc research. In the 1990s, many links were patiently forged between the blastoderm and the adult. During this process, several old puzzles at the cellular level were solved by clever experiments at the molecular level.

Imaginal discs are hollow sacs of cells that make adult structures during metamorphosis. They are so named because “imago” is the old term for an adult insect, and their shape is discoid (i.e., flat and round like a deflated balloon). They arise as pockets in the embryonic ectoderm and grow inside the body cavity until the larva becomes a pupa, at which point they turn inside out (“evaginate”) to form the body wall and appendages. In a D. melanogaster larva there are 19 discs (Fig. 1.1). Nine pairs form the head and thorax, and a medial disc forms the genitalia. The abdominal epidermis comes from separate cell clusters called “histoblast nests”. Unlike discs, histoblast nests remain superficial during larval life and do not grow until the pupal stage.

Given the diversity of cell types in the adult skin (e.g., bristles, sensilla, photoreceptors) and the commonality of their descent from one progenitor (the fertilized egg), it is natural to ask how cells specialize to adopt divergent roles. In principle, cells can acquire instructions from ancestors or contemporaries. More specifically, a cell can inherit predispositions from its mother (“intrinsic” mode), take cues from neighbors (“extrinsic” mode), or both. The predispositions could be gene states, while the cues could be diffusible ligands.

To the extent that fates are assigned intrinsically, there should be a rigid correspondence between (1) parts of the anatomy and (2) branches of the lineage tree.

All protein domains that were mentioned in the text or tables are inventoried below. For further information, consult PROSITE (www.expasy.ch/prosite) or the following reviews: domains in general, DNA-binding domains, scaffolding domains, extracellular domains, domain classification, domain evolution, protein-protein binding, protein-peptide binding, receptor-ligand binding, signal transduction, and an inventory of fly protein domains.

D. melanogaster proteins vary in size from 21 a.a. (L38, a ribosomal protein) to 5201 a.a. (Kakapo, a cytoskeletal component needed for intercellular adhesion). The domains listed below vary from 4 a.a. (WRPW) to ∼270 a.a. (PAS).

N.B.: Customarily, “domain” denotes a motif in proteins, while “box” refers to DNA. Thus, for example, the homeobox encodes the homeodomain. “Repeat” does not connote identity within a protein (e.g., only 6 of the 38 residues are invariant among Notch's 36 EGF-like repeats), nor does it imply interchangeability. For instance, the LIM domains of Lim3 can replace those of Apterous in wing development but not in the CNS. Likewise, Cactus's ankyrin repeats cannot functionally substitute for those of Notch. The specificity of repeats is epitomized by the “arm” domain:

How embryos “self-assemble” has fascinated thinkers for millennia. Among the ancient Greeks, Aristotle (384–322 bce) made copious observations and coined the term “morphogenesis,” which is still in use today. For the past century, the science of “developmental mechanics” has hammered at this problem relentlessly, but it is only in the last decade that the core mysteries have finally cracked. The deepest secrets have come from a fairylike fly named Drosophila melanogaster, probably the same species of “gnat” that Aristotle himself noticed hovering over vinegar slime. Unfortunately, these insights can only be fully appreciated in the arcane language of fly genetics. Hence this book full of runes and rules.

This book concerns cuticular patterns, the cellular machinery that makes them, and the genetic circuitry that runs the machinery. Although it is mainly a survey, it is also a narrative that traces the roots of our knowledge. The story that it tells – albeit in condensed form – rivals the Iliad in scope (legions of researchers devoting decades to attacking thousands of genes) and the Odyssey in wonderment (monstrous mutants posing riddles that challenge even the most clever explorer-heroes). Indeed, truth is often stranger than a fairy tale in the realm of the fly. Believe it or not, there are even remote islands where giant drosophilids with dappled wings and feathery legs have been spied dancing and fighting in the misty forests.

Compound eyes have ∼750 facets, with 8 photoreceptors per facet

A fly's face is dominated by its eyes (Fig. 7.1). Each of the two compound eyes is a honeycomb matrix of ∼750 “ommatidial” subunits. Each subunit, in turn, has 8 photoreceptors or “R” cells (R1–R8) for a total of ∼6,000 receptors per eye. At this pixel density, flies see grainier images than humans, who have ≥ 100,000 receptor cells in the fovea alone. Because fly and human eyes appear to have had a common evolutionary origin, the obvious “One Eye or Many? Riddle” is: Did our common ancestor have a simple or a compound eye? If the former, then why/how did insects multiply it? If the latter, then why/how did chordates reduce it to a solitary remnant? Of course, there is a third possibility. Our common ancestor might have had only a primitive light detector, and we chordate or arthropod descendants then built our own versions of eyes based on the genes that were active at those spots on our face.

The epithelium of the eye disc is a monolayer (as is true for all discs; cf. Ch. 4), but the epithelium of the adult eye is stratified. Above the bundle of 8 R cells, each adult ommatidium has 4 “cone” cells that secrete the lens (no relation to vertebrate cones). Between the bundles are pigment cells that prevent blurring by absorbing scattered photons: 2 PPCs, 6 SPCs, and 3 TPCs (primary, secondary, and tertiary pigment cells) per ommatidium.

The A-P axis is governed by Hh and Dpp but not by Wg

Like the leg disc, the wing disc uses Hedgehog to set up a border zone just ahead of the A/P compartment boundary (cf. Fig. 5.7), but its zone emits only one long-range morphogen – namely, Dpp. Wingless is irrelevant for the wing's A-P axis and instead functions along its D-V axis. Both morphogens are essential: wings fail to develop when the disc is deprived of either Dpp or Wg.

Topologically, the wing is like a squashed leg (Fig. 6.1). Its D and V faces are apposed, and its veins run along its length like the leg's bristle rows. However, while the prospective bristle rows converge centrally in the leg disc (cf. Fig. 5.1), the primordia of veins 2–5 are parallel to one another and intersect a perpendicular line (the future margin). Thus, it is unclear whether the wing has a true “tip” like the leg. Certainly, the expression of Dll in a band along the wing margin (Fig. 6.2) differs from the circle of Dll in the leg disc (cf. Fig. 5.4).

The stripe where dpp is expressed in a mature disc is ∼5 cells wide.

Insect imaginal discs are barely visible to the naked eye, so detailed observations had to await the invention of adequate magnifying lenses. Discs were first described by the great naturalist Jan Swammerdam (1637– 1680), a contemporary of Leeuwenhoek's, who applied his training in human anatomy to the study of insect morphology. In his Book of Nature (printed in English in 1758), Swammerdam waxes lyrical about the metamorphosis (which he calls “mutation”) of appendages (“horns” are antennae) in hymenopteran larvae (“worms” of bees):