Insects exchange respiratory gases via tracheal systems –highly branched, gas-filled tubes. Numerous experiments over the past 50 years have shown that insects use information about concentrations of oxygen and carbon dioxide in the tracheal system to respond – behaviourally,morphologically, and physiologically – to changing metabolic demand or environmental oxygen levels. Although a few studies have suggested that the oxygen-sensing mechanism resides in the central nervous system, the molecular details of the mechanism have remained elusive. Now work by David Morton appears to have solved the mystery.

A clue came from recent work on Caenorhabditis elegans showing that nematodes lacking a functional copy of an atypical guanylyl cyclase(GCY-35) no longer congregated at preferred oxygen concentrations. Guanylyl cyclases are enzymes that synthesize cGMP, a ubiquitous and important intracellular messenger. GCY-35 and others are termed `atypical' because,unlike most soluble guanylyl cyclases, they bind oxygen and are not stimulated by nitric oxide. The mutant-nematode experiment directly implicated atypical cyclases as oxygen sensors.

Building on these findings, Morton decided to see if atypical cyclases in insects could also sense oxygen levels. He transfected mammalian cells with combinations of three atypical guanylyl cyclases from Drosophila and measured cellular cGMP content after exposing the cells to different oxygen levels. In anoxia, cells containing all combinations of atypical cyclases showed rapid several-fold increases in cGMP compared with cells in normal oxygen levels. By contrast, anoxia did not stimulate cGMP synthesis in cells transfected with a conventional cyclase. These results show that in Drosophila the atypical, but not the conventional, guanylyl cyclases can sense oxygen and are activated by low oxygen levels.

What additional characteristics should an oxygen sensor have, and do atypical guanylyl cyclases exhibit them? In engineering terms, a reasonable all-purpose sensor would give off a signal that is linearly proportional to the variable sensed – because highly nonlinear outputs would provide little discrimination over some portions of the sensed variable's range and too much over others. Morton showed that cells transfected with atypical cyclases exhibit just this sort of linear behaviour in response to graded oxygen concentrations. Cells accumulated cGMP at low levels in normoxia (21%oxygen) and at progressively higher levels in increasingly severe hypoxia. The response was not all-or-nothing.

A final experiment suggested that oxygen interacts with atypical cyclases by binding to their haem groups, the iron-containing groups that, in another incarnation, lug oxygen around our circulatory systems. Morton abolished haem function by incubating transfected cells with a soluble guanylyl cyclase inhibitor that oxidizes iron in the haem group. When he exposed these cells to anoxia, Morton found no accumulation of cGMP. These data suggest that, in normoxia, oxygen bound to the haem group inhibits guanylyl cyclase activity;in hypoxia or anoxia, oxygen dissociation from the group activates guanylyl cyclase activity.

Much remains to be done, not least demonstrating that atypical cyclases work in vivo in insects. Nonetheless, in this insect respiratory whodunit, the apprehension of a molecular suspect is an arresting development indeed.

Morton, D. B. (
). Atypical soluble guanylyl cyclases in Drosophila can function as molecular oxygen sensors.
J. Biol. Chem.