In a lab at UCLA, an eye has parted company with the fly it once looked out of and is now buried in a little bullet of plastic ominously clamped in a high-tech slicing machine. Tools of the eye-cutting trade are scattered about: razor blades, microscope slides, boxes of nameless bits and pieces, as well as modeling clay diverted from its serious scientific use to become little animals with paper-clip feet.
Click! goes the cutter, down comes the eye, and slice! goes the razor-sharp blade, again and again and again. Superthin slivers peel away from the plastic, each with its circular wafer of eye sitting like a lone red peppercorn inside a translucent slice of salami. A technician collects them one by one on a small flat stick and gently places them on a glass slide. Half an hour later a shimmering array of hexagons, 630 times larger than life, covers the microscope’s field of view. Within each hexagon huddles a group of seven cells clustered into a gray blue arrowhead. If this exquisitely ordered array weren’t so terribly small, it would make beautiful tiling for the wall of a mosque.
Beautiful it may be, but beauty is not what drives this lab work. It’s not that the dozens of scientists, both here and around the world, who ponder the mysteries of fly eyes are unmoved by the inherent elegance of their subject. But it is chiefly the biological process giving rise to such beauty that inspires them. How, they ask, does order arise out of chaos? How does a jumbled sheet of boring, undifferentiated cells attain the remarkable precision of the fly eye?
It’s almost as if you begin with a bunch of random tiles on the floor and then, starting on one side of the room, this beautiful parquet floor sort of spreads itself across, says Don Ready, a developmental biologist at Purdue in West Lafayette, Indiana. It’s like watching a crystal being formed--but with the eye, it’s not a hardwired crystal. It’s kind of a software crystal. It’s the logic of development that brings cells together into this striking structure.
The logic of development--this software, if you like--is genes, hundreds and hundreds of them, turning on and turning off at the right time and place, jostling some cells to become this, some to become that, some to kindly wait their turn. When all that jostling is over, the 750 hexagonal facets of the fly eye slot neatly together, each fitted out with 19 cells. Within a single facet are photoreceptors--light-sensing cells of various types--bound together like staves in a barrel. There are cone cells that form the surface of the facet, and from that surface sprouts a single bristle, so handy for sensing muck on the eye that should be wiped away. There’s the nerve that connects each facet to the fly’s brain. And finally, there are the pigment cells that give the eye its rich red color. If the fly didn’t isolate each facet from the next with this layer of light- absorbing pigment, light coming into one facet could bounce over into its neighbors and trigger the wrong photoreceptors to fire. Without knowing where each ray of light originated, the fly would have trouble building up a precise visual map of all it surveys. (Contrary to popular belief, the multiple facets don’t convey multiple images to the brain. Instead they organize the light coming into the eye into one coherent image, much as pixels on a tv screen organize a picture.)
To most fly eye aficionados, however, the function of these 19 cells is far less absorbing than the manner in which they choose their destinies. When Larry Zipursky, a developmental biologist at ucla, was recently asked by his eight-year-old daughter what he does all day in the lab, he didn’t tell her he was devoting his life to the eye of a fly. Instead he tried to explain that he wants to understand something much, much deeper: he wants to know how any animal--fly or frog or three-toed sloth or human--builds up an eye, a brain, a spleen, or indeed a whole body. For him the fly eye, and the genes that control its development, are just a convenient place to start. Name any organ, name any organism-- chances are the genes that put it together are strikingly similar. The fly eye, says Zipursky, is a terrific experimental system--one that you can manipulate and that allows you to ask fundamental questions about development.
Part of what makes the eye so terrific is simply that flies don’t need eyes to survive--at least, not in a lab--so researchers can tinker with fly eye genes and still have a living critter to work with. Moreover, the fly eye is a simple, regular structure--much simpler and more comprehensible than, say, a rodent brain. Most important, the fruit fly Drosophila melanogaster--darling of geneticists for nearly a century--is easy to keep, easy to breed, and intensely well studied. It even has its own genome project.
But even ignoring all this worthy usefulness, one simply has to hand it to an organ that can tempt a staid magazine like Science to emblazon its cover with something looking like a monster from a cheap horror movie. In the March 1995 issue, developmental geneticist Walter Gehring and his colleagues at the University of Basel, in Switzerland, published Frankensteinian photos of flies sprouting eyes on their wings, their legs, and their antennae. They reported creating flies with up to 14 eyes--eyes that could sense light, mind you.
Gehring’s group created their monster flies with the help of eyeless, a sexy new gene on the eye research block. (Genes are often named for the things that go wrong when they’re damaged--no prizes, then, for guessing what eyeless mutants look like.) Two years ago, when Gehring’s lab cloned the gene responsible, the researchers noticed that it carried the distinctive genetic sequence of a protein that tells body parts what they’re going to be by turning other genes on and off. What they had, in other words, was a control gene, which when you screw it up, produces a fly with no eyes. Could eyeless, they asked, be some kind of master switch, a gene that could tell a tissue--perhaps any tissue--to be an eye?
If so, they reasoned, you could turn eyeless on in the wrong place and get eyes in the wrong place, too. This is just the kind of finagling that’s easy to do with modern molecular biology, so the researchers promptly engineered their flies, then sat down to wait. It was a fantastic experience, says Gehring. Things happened gradually. First we saw only a little red pigment. A few days later we saw the first few real eye facets. And then we saw real big eyes appearing on the wings. By the end of the experiment, when they’d seen eyes popping up in all kinds of places on the fly, Gehring and colleagues were convinced they were right: eyeless did indeed seem to be a gene that could prompt unlikely tissues to form eyes.
But even if eyeless triggers the formation of an eye, that doesn’t mean that the developmental problem is solved. Far from it. Though eyeless may set eye formation in motion, events of staggering complexity follow. Maybe 5,000 genes act together to build up an eye, and researchers are only beginning to figure out which these are, and how and when they do what they do.
It’s like trying to put together a 5,000-piece jigsaw puzzle, says Gerry Rubin, a developmental geneticist at the University of California at Berkeley. You’ve got your four or five pieces and you’re very happy because you just found the pieces fit together down here in this little piece of farmhouse. But there’s the whole rest of the landscape to fill in. And a lot of blue sky. So far we’ve clearly just dealt with maybe 5 or 10 percent of the puzzle. We’re working hard to fill in the rest.
That hard work, for the most part, starts with finding flies with screwed-up eyes, each fly a mutant that’s missing just one crucial eye- making gene. Once you have a mutant fly, you isolate the mutant gene, then its normal counterpart. And once you’ve got the normal gene, you can try to figure out how it fits into the jigsaw puzzle of eye development.
The first step is easy: mutant, substandard eyes are easy to spot. Imagine a drunken worker piecing together that beautiful parquet floor. Gone is the smooth, seamless tiling of cells. In its place is a rough, jumbly raspberry of an eye, with facets laid down at random.
There are scores of such mutant genes, with names like rough, rough eye, roughened, roughish, roughex, roughoid, and rough deal--even roughest and roughestlike--which testify to the textural problems they cause. There are other mutant flies with eyes the color of claret and burgundy, oranges and lemons, white chocolate and dark chocolate; eyes that are red on the top half but white on the bottom; eyes shaped like bars or lozenges or sporting weird square facets instead of sensible hexagonal ones. To say nothing of shrunken eyes, eyes turned into wings or feet, or-- as with eyeless--eyes just plain gone.
But sometimes you need the help of a microscope to see what’s gone wrong with an eye. A fly bearing the mutant gene sevenless, for example, looks perfectly normal. Stare closely into its eye, though, and you’ll see that it lacks a single photoreceptor, the one called number seven.
Seymour Benzer’s group at Caltech first spotted it back in the 1970s. Today you can tell that Benzer, a neurogeneticist, still loves eyes: the walls of his office are covered with fly eye photos and sketches, including a surrealistic painting by Benzer himself. He eagerly points out one image of an eye that he’d like to turn into flamboyant wallpaper. (He may very well do so. Benzer has already put stained-glass fly eyes in the front door of his home.)
Two decades ago Benzer wasn’t particularly interested in eyes. I’d become fascinated by the question of how the information of the gene could build up behaviors, he recalls. I was already persuaded that behavior had a genetic basis, from experience with my own children. So I thought, ‘Okay, what’s a simple behavior to work on?’ I chose the animal’s response to light. It seemed simple: light goes into the fly eye, and behavior--in the form of movement toward light-- comes out.
To find their misbehaving mutants, Benzer’s team plied millions of fruit flies with noxious chemicals to damage their genes, then screened their offspring for sadly inappropriate reactions to light. Not surprisingly, some of these young flies simply had poor eyesight--you can’t, after all, react correctly to light if you can’t sense it properly. Sevenless (a fly lacking that gene) was one of these.
In visible light, sevenless flies navigated normally. Ultraviolet light was where they went wrong. Imagine a simple maze with two paths, one leading toward a green light, and the other leading toward uv light. Normal flies will head for uv light. But sevenless flies buzzed off to the green. The reason they spurned uv, it turned out, was simple enough: cell seven, and only cell seven, contains a light-gleaning molecule crafted to sense the short wavelengths that make up the ultraviolet part of the light spectrum. That cell is never made in the mutant.
To spot just how the blunder occurs, you must walk back in time and watch the eye form in the pale, wriggling maggots that will eventually become adult flies. Being elbow deep in maggots is part and parcel of working with Drosophila, and researchers don’t seem to mind it. A fly lab is chock-full of bottles, each with its dollop of the yeasty brown paste that fly larvae find so scrumptious. Take a close look and you’ll see them plunging their bodies through the glop, slurping five times their weight each day as they tunnel. Just four days after hatching, a fly larva will have grown from less than a millimeter to nearly four millimeters. Then, after deciding it has eaten its fill, it slithers up the side of the bottle. Time to find a nice, high spot to form a pupa.
These mountaineering larvae are what Amanda Pickup, a graduate student in fly researcher Utpal Banerjee’s ucla lab, is after. The feat she’s about to perform is not for the fainthearted or the blunt-tweezered. Out onto a fluid-filled dish spill the larvae; there they lie, flicking their little bodies from side to side, bright in the dissecting scope’s spotlight. As quick as a wink, Pickup’s two tweezers grab the mouthparts and the middle of a grub, then slowly tug apart, ripping the body asunder. Tiny white entrails flop out into the dish: waving fronds of salivary gland, twisted coils of gut, blobby bits of fat, and--most important of all--18 small round disks floating among the coils and blobs.
These small, nondescript clusters of cells, buried deep within the grub, are silently dividing and making secret decisions about the larva’s future. The disks will make up much of the future fly, and two in particular are already well on the way to becoming eyes. Lying there in the dish, though, they don’t look anything like the splendiferous tiling of the adult fly eye. Granted, the disks are round in shape, like eyes. But where are the clusters of light-sensing cells, the cone cells and pigment cells, the sheen of the lenses and that bright, luscious red?
A lot of the cellular choreography that assembles the eye comes later, during pupal life. That’s when bristles form, as do pigment cells. The eye, under its thin pupal covering, turns from white to yellow to orange to red, like a fruit ripening on the vine. There are leftover cells, too--cells that must be killed off lest they fatally disrupt the orderly array of facets. At the eleventh hour these cells commit suicide, and the eye neatens up like some windswept hairdo straightened out with a comb.
But already, in the seemingly uncomplicated larva, things are happening--things that Don Ready, when he was a student with Benzer back in the 1970s, spent many hours chronicling. Peering at the developing disk, he noticed a long, thin furrow in the tissue, running from top to bottom. On one side of the furrow, cells were coming together in row after row of orderly clusters, each cluster destined to become a single fly eye facet. On the other side of the furrow, cells were still in a disorderly jumble. Ready dubbed this oddity the morphogenetic furrow. Over the next two days, as the furrow swept from one side of the disk to the other, the eye’s precise cellular architecture emerged.
The cells take on their fates, it’s now known, in a definite order. First, right in the furrow, photoreceptor cell number eight makes its appearance--a long line of eights, in a nice, neat row. As the furrow sweeps onward, cells two and five take shape next to each cell eight. Then three and four join the bunch, and then one and six. Last, just before the cone cells join the cluster, comes cell number seven--except in sevenless flies. The problem in little sevenless larvae, as Ready’s group discovered in 1986, is that those wayward cells turn into cone cells. Clearly, a small but crucial piece of the developmental puzzle is missing in sevenless mutants.
With that key finding, it began to look as if fly eyes might be perfect for studying cell signaling--the science of how cells are programmed to tell one another what they should do and become. Within a year, researchers determined what the sevenless mutants were missing. The normal gene encodes a protein that sits in the membrane of the cell, waiting for the signal to become cell seven. Removing the gene for that protein is like yanking out the phone at someone’s house: no matter how doggedly you call them thereafter, they’ll never, ever know that you did.
So who’s trying to call cell seven? A gene called bride of sevenless. Bride of sevenless also codes for a protein that sits in cell membranes. But this protein isn’t receiving a signal, it’s sending one--and it’s sending it from cell number eight. Go on, it says to the sevenless protein, tell your cell to become number seven. What follows is an intricate sequence of protein chatting to protein in a long, long line-- proteins with names like Son of sevenless, ras, raf, and more--until, finally, the message reaches the cell’s nucleus. Genes turn on and off, and presto! The cell becomes what every self-respecting cell seven should be. It makes the uv-sensing pigment and wires itself to the brain in the perfect, cell-seven way.
Even before this final choice is made, the future cell seven has already rejected all but two career options--becoming cell seven or becoming a cone cell. (It’s the signal from bride of sevenless that sends it inexorably down the cell-seven path.) Lots of molecular chattering probably goes on even earlier. Scientists suspect that cell number eight is the casting director for cluster formation: as cells jumble together at the furrow, cell number eight sends all kinds of signals to the other future members of each cluster so that they all take on their roles in the proper, regimented manner.
But it takes far more than one or two genes to assemble an eye. Whole hordes of them lend a hand in getting the furrow moving and in nudging cell eight into being--in helping create all the other frills and furbelows that go into making a fully functional eye. Learning how these genes work may provide clues about the diverse signals that control cell growth. And understanding that would have major implications not just for developmental biology but for medicine: many of the proteins in the sevenless cascade, for example, are the crucial proteins that tell cells-- cells of any kind--when to divide. Damage to the genes encoding these proteins could lead to cancer. For example, some 30 percent of human cancers involve damaged ras genes.
Fly eye genes are touching so many facets of biology these days that many researchers in the business have gained new interest in them. In particular, they’re reconsidering the question of how eyes evolved. Eyes of wildly different shapes and forms exist, from the compound eyes of insects and crustaceans to the single-chambered eyes of squid and vertebrates. Each is superbly adapted for the needs of the beast it resides in. And that has resulted in dramatically different designs. We humans, for instance, sense light on a concave surface: light enters the pupil and forms an inverted image on the retina, the inner surface of our eyeball. Flies do it the opposite way. They monitor the world on a bulging, convex retina, dotted with photoreceptors that harvest light from different points in space.
No wonder, then, that those who muse about evolution have thought long and hard about just how the eye evolved, and how many times it did so. In the textbooks--including my own--you’ll read that a prototype eye arose independently some 40 times, says Gehring. Many a detailed study of eyes went into that conclusion, yet what it means precisely has never been clear. Did different eyes evolve completely independently, each time? After all, the light-sensing pigment rhodopsin is basically the same in all animals--it’s not very likely that it evolved 40 times. Thus, many people suspect that somewhere in our deep past there was a rudimentary eye that great-great-grandfathered all the different kinds of eyes we see today.
With the discovery of the eyeless gene, the notion gains support. Not only does eyeless play a pivotal role in fly eyes, but suspiciously similar genes reside in everything from flatworms, squid, and sea squirts to mice and humans. Everywhere we look, we’re finding it, says Gehring.
What’s more, in people and mice at least, the gene’s role seems awfully close to its role in flies. There’s a genetic condition in humans called aniridia, in which the retina, iris, lens, and cornea are flawed. The imperfections stem from damage to the human counterpart to eyeless. Deleting one copy of the eyeless counterpart in mice--a gene called pax-6-- causes similar problems; deleting both copies leads to mice with no eyes, no nose, and an incomplete nervous system. Gehring even did an experiment in which he inserted the mouse pax-6 gene into a fly and turned it on inappropriately. Eyes--fly eyes--formed all over. In both vertebrates and invertebrates, the genes really do seem to work the same way.
If you think of the mouse eye and the Drosophila eye, they really couldn’t be more different from each other, says Stanford neurobiologist Russell Fernald. And yet now we see that some part of the developmental process is homologous, because it’s done with the same gene. That points to a common, ancestral eye. It might not have been much of an eye--perhaps just a sheet of light-sensing cells, the kind that exist in primitive creatures like flatworms. And lots and lots of independent evolution happened later. But there’s no reason to talk anymore about eyes evolving dozens of times from scratch.
You want to see something really beautiful? asks Benzer. He’s pointing to a stunning black fossil on his desk--a five-inch-long trilobite, ancient cousin of the fly, complete with body segments, legs, and gorgeous compound eyes. Already, 500 million years ago, that intricate eye was in place--and only a wink or so earlier in time, our ancestor and the ancestor of flies and trilobites headed off down their separate evolutionary roads. With the help of paleontologists, Benzer is hoping to trace a family tree of eyes, from the very first, light-sensing cells to the sophisticated specimens we see, and see out of, today. Organisms have worked out many different designs to achieve the same effect, says Benzer. But what seems to be common is the molecular components they use. They’re like molecular Legos you can put together in very different ways.