Small but mighty: In mice, around ten million olfactory sensory neurons analyze incoming air for promising odors. The corresponding figure for humans has not been identified, but it is assumed to be significantly lower.

32 MaxPlanckResearch 3 | 09 FOCUS_Olfaction

The Genetic Story behind a Good Nose

Peter Mombaerts is as familiar with the world of molecules, genes and cellular signals as he is with the world of odors. The Belgian, who is now Director at the Max Planck Institute of Biophysics in Frankfurt/Main, is one of the researchers who have clarified what pathway odors take through the nose and brain – or at least the basic principles of it.

TEXT KLAUS WILHELM

he importance of the world of But many of the questions in olfactory through the nose and the brain. It is olfaction for humans, and research remain “baffling,” Mombaerts also surprising in view of the fact that even more so for other mam- notes. There are several reasons for this. women, in particular, strongly affirm mals like dogs, cats and mice, For one thing, the secrets of olfactory that their partners must “smell good” is immediately obvious to any- processing in mammals can be decod- to be attractive. There is hardly any- T one who thinks about it. But the genet- ed only by carrying out complex and thing more repellent to us than a bad ics and biochemistry that underlie the time-consuming animal experiments in smell. Biologists, physicians and psy- sense of smell are beyond most people’s which the olfactory receptor genes are chologists are well aware of just how imaginations. Depending on the mam- genetically modified in a specific way loaded natural smells are with informa- malian species in question, between 300 using increasingly finely-tuned meth- tion – for humans and, to an even and 1,200 genes form the basis of the ods. Faster experiments using cell cul- greater extent, for other mammals. ability to discriminate between the tures do not work. For another, only a When a dog sniffs the urine of another unquantifiable abundance of existing small, select group of researchers have member of its species, it can, without odors. These olfactory receptor genes taken up this difficult scientific chal- seeing the other animal, immediately contain the instructions for proteins lenge. Peter Mombaerts estimates that identify its gender. When a mouse that recognize the various structures of there are only “around a thousand peo- breaks out in a cold sweat, it sends out odorant molecules. Moreover – and in ple” working on olfaction throughout a warning to the other animals in the Mombaerts’s view this is the most signif- the world. group. The teat secretion of a mouse icant of his many findings – they also guides the sucking behavior of her control the navigation of axons from A SMELL SAYS MORE THAN pups. As Peter Mombaerts stresses, the olfactory sensory neurons to the A THOUSAND WORDS there is no doubt that “olfaction is the brain. The entire procedure of olfactory most elementary and fleeting of all processing takes a good 200 millisec- When we consider that Germans alone senses.” An ideal detector of body and onds. Researchers like Andreas Schäfer, spend billions of euros every year on environmental chemistry, the sense of who heads an Independent Junior Re- the pleasure of surrounding themselves smell is almost impossible to mislead. search Group at the Max Planck Insti- with scents like vanilla, rose and musk, Mombaerts became enamored with tute for Medical Research in Heidelberg, it may come as a surprise to learn the science of olfaction back in 1991. take a stopwatch in their hands when that so few scientists are involved in At that time, a new era in olfactory

Photo: fotolia they put mice on the trail of a smell. researching the pathway odors take research had just begun. Prior to this

3 | 09 MaxPlanckResearch 33 FOCUS_Olfaction

» Since the breakthrough of 1991, it has gradually become clear how an odor literally rushes to the brain and is assessed, decoded, and stored there.

date, the nose was viewed as the most genetic abundance corroborates the im- mysterious of the sensory organs. This mense significance of olfaction. With- was clearly due to the nature of the out it, the sense of taste would be beast: hearing, for example, is based on helpless – not only in mice, but also in a linear system of sound waves that can humans. The final say on all culinary be detected by a biological structure creations – be it tuna steak, hamburger, with relative ease and transformed into pasta, or wine – goes to the olfactory sensory impressions. Similarly, sight is epithelium. also based on the detection of wave- The family of olfactory receptor lengths that can be understood in genes is remarkable in a number of re- numerical ranges – for red, green and spects. Unlike almost all other genes, blue. But how can a mammal, with its these genes are not interrupted by so- limited genetic resources, discriminate called introns – DNA segments that do between the hundreds of thousands of not encode information about the odorant molecules with different chem- structure of a protein. This intronless ical structures? “From today’s perspec- gene structure probably made it easier tive, there were some bizarre ideas for the olfactory receptor genes to pro- floating around the scientific world liferate with new variants in the ge- back then,” remembers the Max Planck nome over the course of evolution and researcher, such as the theory relating to become such a huge family of genes. The images show the two olfactory bulbs of to the existence of seven “primary” Moreover, in all of the mammals stud- a mouse: The olfactory sensory neurons in odors, similar to the primary colors in ied so far, the genes appear to be locat- the olfactory epithelium are stimulated here by butenal. The degree of neuronal activity the visual system. ed seemingly haphazardly on all chro- is shown using different colors (bottom). mosomes. “There’s no identifiable logic A HUGE FAMILY OF OLFACTORY here,” says Mombaerts. The team work- RECEPTOR GENES ing with the scientist succeeded in demonstrating that the control areas An article published in 1991 revolu- for gene expression are “inconceivably tionized the field of olfactory research. small,” even in terms of genetic dimen- American biologists Richard Axel and sions. Such control areas determine Linda Buck, who have since been when and where a gene is activated and awarded a Nobel Prize, reported to their how its information is ultimately im- stunned colleagues that there are more plemented in a protein. than 1,000 genes for the detection of Since the breakthrough in 1991, it odors in the genome of rats. “The larg- has gradually become clear how an est family of genes in mammals over- odor literally rushes to the brain and is all,” says Mombaerts. Of the approxi- assessed, decoded, and stored there. All mately 30,000 genes in mice and rats, smelly living creatures or things release around 1,000 are involved in olfaction. volatile molecules that are almost al- This figure is lower in humans, but ways complex mixtures. Smells consist

at 350, significant nonetheless. This of hundreds of chemical components Photos: Peter Mombaerts (2)

34 MaxPlanckResearch 3 | 09 that waft into the nose and collide there with the olfactory epithelium. This patch of tissue is located within the olfactory mucosa and is more or less the size of a postage stamp in humans, but much larger in dogs and mice relative to their body size. The ol- factory epithelium consists of three cell types: the supporting cells, which pro- vide important assistance in olfaction, the olfactory sensory neurons, and the basal cells – adult stem cells that replace the olfactory sensory neurons and thus above A pioneer in the realm of the senses: Flemish scientist Peter Mombaerts is one of the researchers who have clarified what pathway odors take ensure “the strongest neurogenesis of through the nose and brain. all in the adult body,” says Mombaerts. below Cross-section of the vomeronasal organ of a mouse. The different colors show that the epithelium consists of two types of neurons. RECEPTORS FISH FOR ODORANT MOLECULES

Approximately ten million olfactory sensory neurons in the mouse “ap- praise” the incoming air; the number of neurons in humans is unknown. Around 20 fine sensory hairs, known as the cilia, protrude into the nasal mu- cosa. Their cell membrane houses all of the molecular components that ensure that humans can perceive several mil- lion smells, even in low concentrations, and can discriminate between thou- sands of them – despite having only 350 types of molecular receptors that are encoded by the olfactory receptor genes. In the olfactory epithelium of mice and dogs, around 1,200 different types of receptor proteins scan the incoming odorant molecules. “The olfactory receptors are the basis of olfactory perception in mam- mals,” stresses Mombaerts. The recep-

Photos: Christophe VanderEecken (top)/Tomohiro Ishii tor proteins, which consist of around

3 | 09 MaxPlanckResearch 35 FOCUS_Olfaction

320 amino acids, are similar in overall during their maturation, but ultimate- structure; they traverse the cell mem- ly opt for a single receptor type. Such brane of the olfactory sensory neurons details are important for Mombaerts, as seven times. Certain parts of the he wants to understand how and why receptors display the greatest diversity: a cell chooses only one olfactory recep- the binding pocket, the area where tor gene for expression and – while “in the interaction between the odorant biology nothing is really perfect” – why molecule and receptor takes place, is the mechanism appears to be so good. probably located there. Every olfactory Mombaerts’s team discovered in exper- 1 sensory neuron in the mouse has all iments with genetically modified mice 1,200 receptor genes in its genome, but that this selection process does not in- produces only a single receptor type. volve so-called DNA rearrangements. “Even though, for technical reasons, it’s Experts had favored this mechanism not really possible to provide definitive for a long time because the immune proof, and the evidence that exists is system uses it to ensure that it can somewhat weak, this hypothesis has recognize countless pathogens based now become dogma,” says Mombaerts. on just a few gene segments. The Frank- What is clear is that, in humans, tens Thus, for better or for worse, he works furt-based researcher believes that the of thousands of neurons of each olfac- on the assumption of the “one neuron- unusually short control elements in tory receptor type are distributed one receptor hypothesis.” front of the olfactory receptor genes throughout the olfactory epithelium. There is, however, some evidence determine the likelihood with which a Equipped in this way, humans can dif- that olfactory sensory neurons can ac- neuron chooses a particular receptor ferentiate between the smell of lemons tually produce several receptor types for expression. and oranges, for example. When a hu-

WHERE SMELLS TRIGGER EMOTIONS

The vomeronasal organ (VNO) looks like a tiny thread and onds, whether a female is available to mate via pheromone huddles against the base of the nasal mucosa. “We demon- messages, or whether urine originates from a strange male – strated for the first time in mice that the cells of the VNO something that triggers instant aggression in rodents. The detect pheromones with their special receptors,” says Peter approximately 300 receptors found in the VNO neurons differ Mombaerts of the Max Planck Institute for Biophysics in Frank- clearly from the olfactory receptors in the olfactory epithelium. furt. Furthermore, the VNO regulates sex and aggression in a And, as Mombaerts’ team discovered, the axons from the VNO very direct way through the perception of the pheromones of neurons with the same receptor do not project toward a single conspecifics. glomerulus in the olfactory bulb, but to 15 small glomeruli of The team working with the neurobiologist – who was the accessory olfactory bulb. based at The Rockefeller University in New York at that time – The scientist suspects “that the olfactory processing of deleted from the mouse genome 16 genes that are important pheromones takes place mainly in the accessory olfactory bulb.” for VNO neurons, and this genetic manipulation generated The accessory olfactory bulb does not send axons to the cere- defects in the processing of pheromones. Female mice barely bral cortex, but directly to the hypothalamus and the emotion- took care of their offspring and the sexual drive of the males al centers of the brain, which control the behavior of the mice declined. Male mice can usually recognize, in a matter of sec- directly through secreted hormones and messenger substances.

36 MaxPlanckResearch 3 | 09 2

man breathes in olfactory compounds, the receptors activate, via so-called G 1 Olfactory sensory neurons in the olfactory only the olfactory sensory neurons that proteins, an enzyme that produces epithelium that express the same receptor gene were dyed blue. The cell express the receptors for the cognate vast quantities of the messenger sub- bodies of the neurons are located in the odorant molecules are activated. Thus, stance cAMP. Positively charged ions epithelium (left); its axons (nerve fibers) dozens of receptor types are stimulated then stream into the cell through the run across the surface of the olfactory simultaneously by each odorant, but in membrane channels that are now bulb (right) and culminate in a neural different combinations. opened, and electrical signals, action knot, a glomerulus. “Olfactory perception is highly potentials, are generated. These prop- 2 General overview of the olfactory bulb combinatorial,” says Peter Mombaerts. agate along the nerve extension (axon) in which all mature sensory neurons were This logic is the only way humans can of the olfactory sensory neuron to the dyed blue. The small blue balloons in the center are the glomeruli toward identify thousands of different odors brain’s olfactory bulb (bulbus olfacto- which the nerve fibers converge. with ease, while mice and dogs can rius) where, in the mouse, 2,000 knots The bundle of nerves protruding from potentially identify hundreds of thou- made of neural extensions known as the left consists of the fibers of the sands. Because small structural chang- glomeruli are found. vomeronasal organ and projects toward the accessory olfactory bulb. es in odorant molecules alter the inter- action with the receptors only AN OLFACTORY BULB WITH gradually, the number of chemicals A SYSTEM that can be smelled is theoretically un- limited. A receptor protein recognizes This is precisely the interface at which a defined part of a molecule very spe- the initially indiscriminate receptor sig- cifically and thus interacts only with nals become more specific. The axons odorants that contain this part. In of olfactory sensory neurons do not higher concentrations, however, mol- project toward the olfactory bulb in a ecules with similar structures also acti- random fashion. Instead, axons of ol- vate the receptor. factory sensory neurons with the same The olfactory sensory neurons that receptors form bundles, which are are activated in this way transform the probably held together by adhesion olfactory stimuli into electrical signals. proteins, such that “their axons ulti- They do this through a chain of mately coalesce into the same glomer- biochemical reactions: When odorant ulus of the olfactory bulb,” says Peter molecules have found matching recep- Mombaerts. Thus, olfactory sensory tors on the surface of an olfactory sen- neurons with receptor A go to glomer-

Photos: Peter Mombaerts (2) sory neuron and have docked there, ulus A, olfactory sensory neurons with

3 | 09 MaxPlanckResearch 37 SPLIT-SECOND SMELLING

The brain wastes no time when it comes to processing odors – as Andreas Schäfer is only too well aware. The biophysicist heads an Independent Junior Research Group at the Max Planck Institute for Medical Research in Heidelberg, where he used a simple experiment to determine the “smelling speed” of mice: The animals first broke through a light barrier. At that moment, an odor was blown into their noses. The rodents quickly learned that, after sniffing the scent of bananas, they were allowed to lick some sugar water as a reward. After sniff- ing an apple scent, in contrast, they were left empty-handed – whereupon they retreated. “Only 230 milliseconds elapsed from the inhalation of the smell to the behavior ultimately triggered by the action,” says the researcher. “It all takes place at unim- aginable speed.” In subsequent experiments, the rodents were presented with very similar smells and, as a result, the processing time increased to 340 milliseconds – not even half a second. As part of his ongoing research, Schäfer’s team is using sophisticated techniques to switch certain groups of nerve cells in the olfac- tory bulb on and off to observe the resulting changes in behav- ior. This work is being carried out in cooperation with Thomas Kuner’s team at the University of Heidelberg. Initial findings indicate that the animals can better distinguish between Mouse sniffing the apple odor in the behavioral experiment. smells when the neurons are rendered more excitable.

receptor B go to glomerulus B, and so overlap. Even an individual odorant Mombaerts and his team recently iden- on. The Belgian scientist gave olfactory molecule can evoke a complex pattern tified a second control mechanism in research this fundamental insight because, in most cases, several different olfactory processing that imposes order through his pioneering experiments receptors respond. The brain constant- on the cornucopia of smells at a high- with genetically modified mice. When ly registers which types of olfactory re- er level. The researchers started by he replaced the gene for a certain recep- ceptors are activated at the same time. showing that the olfactory mucosa can tor type with a different receptor type In the example of the typical camem- be divided into two zones with differ- in the genome of the mouse, the axons bert pattern, the brain then generates ent classes of receptors. The top one, of these olfactory sensory neurons the matching olfactory image. referred to as “D” for dorsal, contains did not end in their usual glomeruli. Mombaerts is fascinated by the dual olfactory sensory neurons that express Instead, new glomeruli were formed. function of the olfactory receptors. On class I and II receptors (class I receptors the one hand, they recognize odorants. probably mainly recognize water-solu- OLFACTORY PATTERNS On the other hand, “the receptors also ble odorant molecules). The lower IN THE NOSE guide the axons to the right place in the zone, referred to as “V” for ventral, con- olfactory bulb – a wonderfully simple tains almost exclusively class II recep- The result of this process is a character- solution from an evolutionary perspec- tors. A mix of the cell bodies of the two istic and complex activation pattern – tive,” explains the olfactory researcher. cell classes is found in the dorsal olfac- a kind of glomerular map – that, con- At least one new glomerulus emerged tory epithelium. On the way to the versely, displays the mix of odorants an in the bulb with each new gene for an brain, the axons of the class I neurons animal has smelled. A “camembert pat- olfactory receptor. Glomeruli, in turn, fasciculate together via a mechanism tern” differs clearly from a “lemon pat- send their axons to the cerebral cortex that is not dependent on the receptor tern.” If individual chemical compo- where the final olfactory impression that is produced. When, through genet- nents are found in both smells, the and the associated links and memories ic manipulation, the researchers forced

patterns of the activated glomeruli are produced. dorsal olfactory sensory neurons that Photo: MPI for Medical Research – Rolf Sprengel

38 MaxPlanckResearch 3 | 09 FOCUS_Olfaction

500

400

300

200

100

0 Processing time (ms) Processing Simple Difficult (different odors) (similar odors)

Odor identification in mice is a fast process, but it depends on the similarity of the smells: the more similar the smells, the longer it takes.

normally produce a class I receptor to This will have to be followed by some cal properties of many odorant mole- instead produce a class II receptor, the big steps: “We still haven’t worked out cules. “As a result, we are a long way axons of these neurons still aligned the entire logic of the system,” Mom- from being able to predict how a mol- themselves with axons of neurons that baerts admits. Why do we perceive the ecule can best be altered in the test express a class I receptor. Therefore, one molecule phenylethyl ethanol as the tube to create, for example, the ulti- of the first phases in sorting axons on scent of a rose? “No idea,” says Mom- mate rose scent.” Or how a scent can their way to the olfactory bulb appears baerts. Although his team has de- be produced more inexpensively and to depend only on the class of the scribed some of the odorant molecules simply. The perfume industry would olfactory sensory neuron. Again, as that bind to the olfactory receptors in give almost anything for this magical Mombaerts notes, “such a small step recent years, the olfactory researchers code – the Holy Grail of olfactory to unravel the odor code.” still know too little about the chemi- research.

GLOSSARY

Intron Olfactory epithelium cAMP Glomeruli Pheromone A non-coding section of a gene’s DNA that A tissue comprising three Cyclic adenosine monophos- Spherical neural A messenger sub- divides neighboring exons (the coding cell types; it is specialized phate, a biological molecule knots through which stance that serves sections). Introns are transcribed, but then in the detection of smells that is derived from adenosine olfactory information biochemical spliced out of the RNA transcript before it is and is located in humans triphosphate (ATP) and that coming from the communication exported out of the nucleus for translation in the regio olfactoria of acts as a messenger substance olfactory sensory between members into the corresponding amino acid sequence. the nasal mucosa. in cellular signal transmission. neurons is processed. of a species. Illustration: designergold based on material provided by the MPI for Medical Research

3 | 09 MaxPlanckResearch 39