One of those difficult misconceptions that is hard to root out of people's heads is the idea that individual genes 'make' something. We all have this bias, this tendency to reify the gene into something concrete—scientists do it all the time, too. You can see it in the list of gene names at OMIM, for instance; many are named after diseases or their consequences in adults. The message of which I try to always remind myself (not always successfully) is that genes don't make things, interactions between collections of genes and the environment make things. Biology arises out of the processes, not the structures; it's the reactions, not the end-product.

A paper in the latest BioEssays reminds me of this. It's a short review of Hox genes and insect wing formation that carries the same message, that morphology is a consequence of patterns of gene interactions.

First, some basics. The insect thorax is made of three segments, the prothorax, mesothorax, and metathorax, or T1, T2, and T3. Each segment has a pair of ventral appendages, the legs, and may have a dorsal appendage, the wings and wing derivatives. The dorsal appendages are highly modified. They are completely turned off in T1, and in many insects, T2 and T3 carry two pairs of wings. In butterflies and dragonflies, for instance, the forewings are on T2 and the hindwings are on T3.

Sclerotised dorsal appendages in two different insects. A: The beetle Calosoma semilaeve, with left elytron and wing expended. B: a male strepsipteran Eoxenos laboulbenei.

Some insects modify this latter arrangement, however. Flies have only one pair of wings, the forewings on T2; the hindwings on T3 have been modified into a small balancer organ called the haltere. Beetles do something different, and have retained the T3 hindwings for flying, but have modified the T2 forewings into hard wing covers, called elytra. The Strepsiptera have reversed the Dipteran pattern, and fly with their T3 hindwings and have turned their forewings into balancing organs.

In flies, we have mutations that change these dorsal appendages. For example, mutations that disrupt a gene called Ultrabithorax (Ubx for short) transform the haltere into a forewing, producing four-winged flies. There is the temptation to say to oneself, "Aha! So Ubx in normal flies is responsible for making halteres—it is the haltere gene!" Resist it. It isn't.

Ubx is a gene that is expressed in T3, and weakly in the posterior part of T2. Other genes with roles in segment identity are Antennapedia (Antp) in T2, and Sex combs reduced (Scr) in T1. We can very roughly say the genes are associated with segments in the order Scr-Antp-Ubx, corresponding to T1-T2-T3, as diagrammed below. It's a little messier than that, because they overlap and because early development is by parasegments, which are a half-segment out of register with segments (really, let's not get into that now, unless you desperately want a headache).

A schematic sketch of Scr, Antp and Ubx Hox genes' expression in the three insects of interest, the fly Drosophila (Diptera), the beetle Tribolium (Coleoptera) and the butterfly Junonia (Lepidoptera). Expression domains do not strictly correspond to segmental, but rather parasegmental units. Dark grey: strong expression; light grey: weak expression.

The key thing here is the text at the bottom of the diagram, which shows something fascinating: Scr, Antp, and Ubx are expressed in exactly the same pattern in the Dipteran Drosophila, the beetle Tribolium, and the butterfly Junonia. The expression of Ubx only leads to haltere formation in flies; in beetles and butterflies, the T3 segment expresses Ubx, and the animals make wings.

Obviously, Ubx does not make halteres; Ubx in the context of many other downstream genes, some of which differ in flies and beetles, contributes to the specification of a unique identity for the T3 segment. Here's the author's conclusion:

The present data lead us to adjust our point of view on Hox genes' function. It is common to say that the Hox genes determine the identity of body parts. What does identity mean? From our own human perspective, most of us thought that it meant a precise morphology for a group of cells expressing a given Hox selector. Genetic analysis reveals the genome's point of view, that of the organism itself in its evolution. It now appears that the precise morphology of an organ or a segment does not matter, what matters for the Hox genes is relative morphology. In other words, identity, determined by Hox genes, means no more than difference between a certain body part and its neighbours along the AP axis.

We can thus view the Hox genetic programme as a meta- programme, versatile enough to accommodate changes in the underlying genetic programmes specifying the precise morphology of individual parts. In summary, the Hox genes do not make the difference between the three thoracic segments, after all. They might just ensure that they are (and maybe have to be) different from one another.

Evo-devo is really moving fast to leave the ghosts of molecular preformationism behind, and our vision of how developmental biology works is becoming progressively more strange and abstract. Give us a few more years, and developmental biology is going to be as weird and mind-bending as modern physics.

Deutsch J (2005) Hox and wings. BioEssays 27:673-675.