I've written about this fascinating Drosophila gene, bicoid, several times before. It's a maternal effect gene, a gene that is produced by the mother and packaged into her eggs to drive important early events in development, in this case, establishing polarity, or which end of the egg is anterior (bicoid specifies which end of the egg will form the fly's head). Bicoid is also a transcription factor, or gene that regulates the activity of other genes. We also see evidence that it is a relatively new gene, one that is taking over a morphogenetic function that may have been carried out by several other more primitive genes in the ancestral insect.

For those who haven't committed to memory my prior posts on this gene, here's the short summary. Bicoid RNA (in blue, below) is transcribed by maternal genes, and packaged and localized to the anterior end of the egg before it is laid. The RNA is translated into the Bicoid protein (in red and pink) which diffuses through the egg, setting up a concentration gradient, high at the front end, low in back. The Bicoid protein binds to the DNA of the cells in those locations, and differentially activates later genes, called gap genes (in green). Where the Bicoid concentration is high, for instance, genes called orthodenticle and hunchback are turned on, while where Bicoid is low, genes like knirps can be activated.

bicoid and Bcd-target gene expression along the anteroposterior axis of the Drosophila embryo. Localization of the bcd mRNA is shown in blue. The Bcd protein gradient is represented in red and the expression of otd, hb and kni represented in green. Embryos are shown anterior to the left and dorsal up.

The classic summary of this process can be found in The Making of a Fly: The Genetics of Animal Design^{(amzn/b&n/abe/pwll)} by Peter Lawrence, which is unfortunately a bit out of date. Bicoid is one of the genes that was discovered early in the fusion of molecular genetics and developmental biology, and the literature on it has grown immense and somewhat indigestible. I was very pleased to discover a review by McGregor that neatly summarizes all the details, from sequence to localization to binding to evolution, in one tidy place.

I won't try to cover all of the molecular details here, but will focus on just one interesting question: where did bicoid come from? As we've done more and more comparative work, moving beyond just Drosophila, it has become clear that bicoid is special and unique, present in only some flies, but not all, and not present at all in other organisms, like us humans.

Bicoid is only found in one specific but familiar and successful clade of flies, the Cyclorrhapha. The fly "kind" is a very rich collection of diverse animals from mosquitos to houseflies, containing about 125,000 species (one of the crimes of creationists is the way they belittle diversity, reducing this distinguished lineage to "just" flies). Only the Cyclorrapha, which includes Drosophila and the common housefly, Musca domestica, have a bicoid gene. Here's a quick refresher of the different fly groups:

(from FLYTREE)

Here is a closer look at the fly clades that use bicoid (in red):

Phylogenetic relationships of dipterans. All species shown are brachycerans except Clogmia albipunctata (Nematocera). cyclorrhaphan species are shown in red. The approximate origin of bcd is indicated by an arrow. Contrast inverted images of Drosophila buscki (left) and Musca domestica (right) embryos are shown to scale. Black bar shown for scale indicates approximately 500 µm.

When I say that a fly has a bicoid gene, that doesn't mean they are all identical. The homeodomain of these proteins in Musca and Drosophila differ by 5 amino acids (out of 60), for instance, and Megaselia and Drosophila differ by 18. These differences cause subtle, species-specific variations in the binding properties of the protein for DNA. Musca Bicoid is close enough to Drosophila that it can be used to rescue Drosophila embryos that lack the gene, but Megaselia Bicoid doesn't do a thing for Drosophila.

If non-Cyclorrhaphan flies lack bicoid, where did it come from? We don't expect that it magically appeared out of nowhere. All of these flies have a closely related gene called Hox3, another homeobox gene. In the non-Cyclorrhaphans, Hox3 is expressed in the maternal tissues as well, and is packaged and deposited in the eggs; however, it isn't localized to just the anterior pole, and spreads uniformly throughout the egg. Hox3 is also activated zygotically in the developing embryo, and is expressed in extraembryonic tissues.

In the stem Cyclorrhaphans, Hox3 was duplicated, and gave rise to two new genes: bicoid and a closely related sister gene, zerknüllt. These two genes specialized; zerknüllt retained the specific zygotic function, and is important in setting up a thin dorsal strip of extraembryonic membrane. Bicoid's expression was confined to just the maternal subset of Hox3's job. Evolution and simple genetics made the process more complicated, using two genes to do a task formerly handled by one.

This process, of increasing the number of components to carry out a function, is common in evolution. The first step of duplication simply increases redundancy, with two genes doing the same thing. The second step is a loss of parts of the function, such as the bicoid copy losing it's role in extraembryonic tissues, and the zerknüllt copy losing it's role in maternal expression. More parts doing the same thing; that's a reasonable definition of increased complexity.

Freed of the job of handling extraembryonic membranes, bicoid evolved new properties. One was localization to just one end. This may have been a neutral event—flies had mechanisms in place to suppress Hox3 expression in the posterior end, so initially a mutation that made bicoid "sticky" and stay in place where it was secreted would have just made axis formation a little more robust. From there, there has been increasing elaboration of other genes that lock bicoid in place.

Wait—if bicoid in Drosophila is now so essential in defining the anterior end of the egg, what did the ancestral flies and what do modern non-Cyclorrhaphans use? The evidence suggests that the primary regulators in the non-Cyclorrhaphans are the gap genes, hunchback and orthodenticle. These genes are also maternally expressed in many species, and may be the primitive anterior specification genes. What happened, then, is the insects started with a simpler scheme in which the mother secreted a gene product like hunchback into the egg to set up the anterior pole…

…to a point where two genes are secreted…

…to the current situation, in which hunchback is secondary and regulated by bicoid.

These are events that require only natural, commonly observed genetic processes to occur, are supported by the known molecular relationships of these genes, and which explain some peculiarities of Drosophila development, such as the maternal production of hunchback, which only makes sense as a historical relic. It is also an excellent example of how evolutionary processes can ratchet up complexity, replacing a one step process with a two step process by inserting a new event at the beginning. Complexity is a consequence of chance and happenstance, not design.

McGregor AP (2005) How to get ahead: the origin, evolution and function of bicoid. BioEssays 27:904-913.