InteractiveFly: GeneBrief

pou domain motif 3: Biological Overview | References

Olfaction depends on the differential activation of olfactory receptor neurons (ORNs) and on the proper transmission of their activities to the brain. ORNs select individual receptors to express, and they send axons to particular targets in the brain. Little is known about the molecular mechanisms underlying either process. This study has identified a new Drosophila POU gene, pdm3, that is expressed in ORNs. Genetic analysis shows that pdm3 is required for odor response in one class of ORNs. This study shows that pdm3 acts in odor receptor expression in this class and that the odor response can be rescued by the receptor. Another POU gene, acj6, is required for receptor expression in the same class, and a genetic interaction was found between the two POU genes. The results support a role for a POU gene code in receptor gene choice. pdm3 is also expressed in other ORN classes in which it is not required for receptor expression. For two of these classes, pdm3 is required for normal axon targeting. Thus, this mutational analysis, the first for a POU class VI gene, demonstrates a role for pdm3 in both of the processes that define the functional organization of ORNs in the olfactory system (Tichy, 2008).

Each olfactory receptor neuron (ORN) makes two remarkable choices that underlie the sense of smell. ORNs select individual odor receptor genes to express, a process that determines which odors they detect, and they send axons to particular targets in the brain, which determines what behaviors are elicited by the odors. The molecular mechanisms underlying both choices are largely unknown (Tichy, 2008).

ORNs in Drosophila are contained within two organs, the antenna and the maxillary palp. The odor response spectrum of an ORN class is conferred by the expression of one or a small number of odor receptor (Or) genes. The organization of ORN classes is stereotyped, and depends on the proper selection of individual Or genes from among a large repertoire. ORNs send axons to individual glomeruli, which are spheroidal modules in the antennal lobe of the brain. ORNs that express the same receptor converge on the same glomerulus (Tichy, 2008).

In Drosophila, the expression of particular Or genes, and hence the odor specificity of the ORN, depends on a regulatory code of cis-acting elements (Ray, 2007; Ray, 2008). Transcription factors that act in the process of receptor gene expression include abnormal chemosensory jump 6 (Acj6), a POU transcription factor. The acj6 gene was identified by a defect in olfactory behavior. In a null mutant of acj6, some maxillary palp ORNs respond normally to odors, some are present but have lost response to all odors, and some undergo changes in odor specificity. Correspondingly, acj6 is required for the expression of a subset of Or genes (Tichy, 2008).

The present study identified a new POU gene, pdm3, a POU gene of class VI. pdm3 is expressed in ORNs. Loss of pdm3 results in loss of odor response in a class of ORNs in the maxillary palp. Correspondingly, the defective ORNs lose Or gene expression. The phenotypes can be rescued with a cDNA representing either pdm3 or the affected Or gene, indicating a surprising degree of specificity to the odor-sensitivity phenotype. A genetic interaction between pdm3 and acj6, and ORNs can be divided into three categories, those that depend on both of these POU genes, those that depend on acj6 alone, and those that depend on neither for the acquisition of odor specificity. These results are consistent with a combinatorial code of POU genes. pdm3 is expressed in ORN classes in which it is not required for odor response, and in at least two of these classes pdm3 is required for normal axon targeting. pdm3 thus functions in both of the processes that dictate the precise organization of the Drosophila olfactory system: the expression of individual odor receptors and the targeting of individual glomeruli (Tichy, 2008).

To identify a new POU gene, the Drosophila genome sequence was searched with POU domain sequences from all six POU gene classes. The new gene belongs to class VI, and was named pdm3 (POU domain motif 3) by analogy to the previously named pdm1 and pdm2 genes. This is the first POU class VI gene in Drosophila, and no other members of this class were found in the fly genome (Tichy, 2008).

The POU domain of pdm3 consists of a POU-specific domain, a POU-homeodomain, and a linker region that separates the two. It is located in the C-terminal part of the gene, as in other POU genes of class VI. pdm3 spans ~8 kb and contains 10 exons. The predicted POU domain is the most highly conserved region of the gene, showing 70%-90% identity with other class VI POU proteins and 40%-50% identity with POU proteins of other classes (Tichy, 2008).

In initial RT-PCR experiments, pdm3 was amplified from adult head RNA. Two splice forms were amplified, henceforth called the 'long' and 'short' forms. They differ in that the long form includes exon 7, which is missing from the short form. This exon encodes 18 aa that interrupt the POU domain. An optional exon has likewise been found at the same location in Rpf-1, a human ortholog of pdm3 (Zhou, 1996).

The fruit fly detects odors with two types of olfactory organ, the antenna and the maxillary palp. In situ hybridization experiments showed that pdm3 is expressed in a subset of cells in the antenna. Expression was also detected in the maxillary palp, where a double-label experiment was performed to determine whether expression was neuronal. It was found that cells that express pdm3 RNA are labeled by an antibody against Elav, which labels the nuclei of differentiated neurons (Tichy, 2008).

To characterize the expression of pdm3 in more detail, an antibody was generated, using as antigen a portion of the protein that did not contain POU domain sequences. The antibody labeled cells in the maxillary palp. All or almost all of these Pdm3⁺ cells are Elav⁺. Many, but not all, of the Elav⁺ cells are Pdm3⁺. In the Pdm3⁺ Elav⁺ cells, the labeling appeared coincident. Together, these results show that pdm3 is expressed in the nuclei of a subset of maxillary palp neurons (Tichy, 2008).

A mutant line was obteind that contains a transposon insertion in pdm3. The insertion lies within an intron between two exons that encode the POU domain. This insertion eliminated Pdm3 expression, indicating specificity of the antibody. The insertion did not markedly reduce the number of Elav⁺ cells, suggesting that the loss of pdm3 does not cause a loss of neurons (Tichy, 2008).

To determine whether pdm3 is required for the function of ORNs, ORN responses was measured to odors using single-unit electrophysiology. The antenna and maxillary palp contain a number of functional types of sensilla, as defined by electrophysiological and molecular analysis. Each sensillum type contains up to four ORNs, usually two, which are combined in a stereotyped configuration. The maxillary palp is the simpler organ, in that it contains only three functional types of sensilla, pb1, pb2, and pb3, each of which contains a pair of ORNs: pb1A and pb1B; pb2A and pb2B; pb3A and pb3B. Odorants pass through pores in the sensillum walls, traverse the internal lymph, and activate the ORNs. Seven Or genes are expressed in the maxillary palp: of the six ORN classes, each expresses one odor receptor, except that one class, pb2A, expresses two (Tichy, 2008).

Of the six ORN classes of the maxillary palp, one is severely defective in pdm3. The pb1A ORN, which in wild type responds to a pulse of E2-hexenal with a train of action potentials, does not respond in pdm3. pb1A is present and yields spontaneous action potentials, but does not respond to the odorant (Tichy, 2008).

Systematic measurement of neuronal responses, measured in action potentials/s, after stimulation with a diverse panel of odorants, revealed that pb1A failed to respond to any tested odorant. In contrast, in the same sensillum, the neighboring pb1B ORN responded normally to 4-methyl phenol and 4-propyl phenol, its most effective odorants. Neurons of the pb2 and pb3 sensilla gave very similar responses in pdm3 and control animals (Tichy, 2008).

As a more stringent test of the role of pdm3 in ORN response, the pdm3 mutation was tested in heterozygous combination with a deletion for the region. In this pdm3/Df heterozygote, all responses were very similar to those of the pdm3 homozygote, indicating that the pdm3 insertion is a null allele (Tichy, 2008).

The olfactory phenotype is not limited to the maxillary palp. The large basiconic sensilla of the antenna were examined, and one of the four ORNs in the ab1 sensillum, ab1A, is present but has lost odor response. Limited analysis of the ORNs in ab2 and ab3 sensilla revealed no abnormalities. The detection of an olfactory phenotype in some antennal ORNs is consistent with detection of pdm3 expression in a subset of antennal cells (Tichy, 2008).

Because pb1A is present in pdm3 and yields spontaneous action potentials, but does not respond to odors, the hypothesis was tested that pdm3 affects expression of its odor receptor. Or42a is the odor receptor that is expressed in pb1A (Tichy, 2008).

As a first test of this hypothesis, it was asked whether Pdm3 is expressed in pb1A in wild type. A double-label experiment showed that Pdm3 is in fact expressed in Or42a-expressing cells of wild type. It was then found that Or42a expression is undetectable in pdm3. In contrast, the neighboring pb1B cell shows normal expression of its receptor gene, Or71a, and all other maxillary palp odor receptor genes are expressed in pdm3. Thus these expression studies coincide precisely with the functional studies: the one class of ORN that has lost function is the one class of ORN that has lost expression of a receptor gene (Tichy, 2008).

To determine whether the loss of pb1A odor response is attributable solely to loss of Or42a expression or whether other essential components are lost as well, expression of Or42a was drived in pb1A neurons of pdm3 using the GAL4-UAS system. Response was tested to the two odorants that elicit the strongest response from pb1A and it was found that responses to both were completely or largely restored when Or42a was driven by Elav-GAL4. These results suggest a surprising degree of specificity to the pdm3 phenotype at the molecular level (Tichy, 2008).

Next, the ability of pdm3 splice forms to rescue the phenotype was tested. Of the two splice forms, only the long form was amplified from adult maxillary palps in RT-PCR experiments. When expression of a long form cDNA was driven using the GAL4-UAS system, the phenotype of pb1A was rescued completely or largely in 7 of 29 pb1 sensilla, tested with three odorants. Expression of Or42a was also restored in a number of cells. In this experiment, an acj6-GAL4 driver was used, which drives expression in all or almost all maxillary palp ORNs. A faithful driver could not be constructed for pdm3, whose 5' end is separated from the next annotated upstream gene by a long region; it is possible that a pdm3 driver might restore the pb1A phenotype in a greater fraction of pb1 sensilla. Expression of a short form cDNA rescued in 1 of 19 pb1 sensilla tested (Tichy, 2008).

It is noted that in addition to the rescued pb1A cells, a number of pb1A cells were found that acquired an odor response profile different from that of wild type. Further testing revealed that this odor response profile is very similar to that conferred by a larval odor receptor, Or85c. Moreover, Or85c expression is detected in several cells of the 'rescued' maxillary palp. The proportion of cells with this response profile is greater when pdm3 is driven with a promoter that is expected to drive later expression in ORNs. Further testing beyond the scope of this study would be required to determine whether these results reflect the presence of different pdm3 interaction partners at different times at the promoters of Or42a and Or85c (Tichy, 2008).

Because pb1A is affected by both pdm3 and another POU gene, acj6, the relationship between them was investigated. Although the concept of a combinatorial code of transcription factors has been previously invoked in discussion of receptor gene choice in Drosophila (Ray, 2007; Ray, 2008), the interaction of two transcription factors within the same ORN has not yet been examined in Drosophila (Tichy, 2008).

It was found first that neither POU gene appears to regulate the other in the maxillary palp: pdm3 is expressed in a null mutant of acj6, and acj6 is expressed in a null mutant of pdm3. Although both pdm3 and acj6 are fully recessive, the transheterozygote shows a reduced response of pb1A. The response of pb1B is normal in this genotype. These results indicate a genetic interaction between the two POU genes in pb1A (Tichy, 2008).

Interestingly, two putative Pdm3 binding sites, i.e., sequences corresponding to binding sites for rat and zebrafish orthologs [(Brn-5) and (Pou[c]), respectively], lie near an Acj6 site, ~320 bp upstream of Or42a. One putative Pdm3 site overlaps with the Acj6 site, and the other is immediately adjacent to it. POU genes have been shown to heterodimerize on DNA, and this cluster of POU sites may bring Pdm3 and Acj6 in a position to do so upstream of a gene whose expression depends on both (Tichy, 2008).

From the initial analysis of pdm3 expression in the maxillary palp, it is clear that the number of cells expressing pdm3 is greater than the number of pb1A cells. To determine in which ORN classes pdm3 is expressed, a series of double-label experiments was performed using ORN class-specific probes and an anti-Pdm3 antibody. It was found that Pdm3 is expressed in four of the six maxillary palp ORN classes: pb1A, pb1B, pb3A, and pb3B. Expression was not detected in pb2A or pb2B. The expression of pdm3 in three ORN classes that do not require its expression for odor response suggested the possibility that pdm3 might have a different function in these cells (Tichy, 2008).

ORNs send axons to the antennal lobe of the brain, and axons of an individual ORN class converge precisely on an individual unit of the antennal lobe called a glomerulus. Thus, a spatial map of olfactory information is created in the antennal lobe by the stereotyped pattern of connectivity. There are 43 glomeruli in the antennal lobe, and both receptor-to-ORN and ORN-to-glomerulus maps have been constructed. A number of genes required for normal ORN axon targeting have been identified (Tichy, 2008).

The axon targeting was examined of two ORN classes that express pdm3 but that do not require pdm3 for odor response: pb1B, which expresses the Or71a receptor, and pb3A, which expresses Or59c. It was found that in a pdm3 mutant, the axonal projections of both ORN classes are abnormal. In the wild type, in both cases the projections converge on a single, discrete glomerulus in each antennal lobe. In pdm3, the convergence is less precise and the boundaries of the targeted areas are less discrete. In some cases, the axon tracts appear to bifurcate, with labeled axons occupying more than one region. In most cases, the axons appear to terminate in the general vicinity of the wild-type target, but not in the precise positions of the wild-type glomerular targets. In summary, pdm3 has effects on ORN axon targeting, in addition to its role in specifying odor response and receptor gene expression (Tichy, 2008).

This study has identified a new Drosophila POU gene, pdm3, that acts in both receptor gene expression and axonal targeting. Pdm3 is a POU transcription factor of class VI. Vertebrate members of class VI are expressed in brain and spinal cord. Notably, a mouse ortholog of pdm3, Emb, is expressed in the olfactory bulb (Tichy, 2008 and references therein).

The mechanism of ORN fate determination has been elegantly examined in Caenorhabditis elegans. The AWA, AWB, and AWC cells express multiple olfactory receptors and sense distinct but overlapping sets of odors. The specification of AWA fate requires lin-11, a LIM homeodomain gene. lin-11 regulates ODR-7, a nuclear hormone receptor that promotes expression of AWA-specific genes and represses AWC-specific genes. The Aristaless-related homeodomain gene alr-1 is also required for specification of AWA. ceh-37, an Otx homeodomain gene, regulates lim-4, another LIM homeobox gene, to promote AWB neuron fate. Another Otx gene, ceh-36, is required for AWC neuron fate (Tichy, 2008 and references therein).

The first olfactory phenotype detected for pdm3 was its defective odor response in an electrophysiological assay. The phenotype is striking in its specificity, at both the cellular and molecular levels. At the cellular level, only one ORN class in the maxillary palp, pb1A, was defective in a null mutant of pdm3. The pb1B neuron, which neighbors pb1A in the same sensillum, and the other four ORN classes all appeared to yield normal odor responses. At the molecular level, the pb1A physiological phenotype appeared to be attributable largely if not completely to the loss of a single gene: Or42a, the odor receptor gene normally expressed in this neuron. Moreover, Or42a is the only maxillary palp Or gene affected by pdm3 (Tichy, 2008).

Although the specificity of the molecular and cellular phenotypes was surprising, the specificity of one phenotype is in good agreement with the specificity of the other. The cascade of molecular steps between the presentation of an odor molecule and the production of an action potential is poorly understood but presumably requires the agency of a number of different genes. It seems likely, however, that the signaling pathway used by one maxillary palp ORN class is the same as that used by the others, with the exception of the odor receptor. Thus the ability to rescue the pb1A odor response with its receptor gene alone supports a model in which the rest of the signaling pathway does not depend on pdm3. In this case, pdm3 would affect odor response only in those maxillary palp ORNs whose receptor genes it affects, i.e., only in pb1A (Tichy, 2008).

The specificity of pdm3 is in contrast to that of acj6, which affects the odor response of four ORN classes in the maxillary palp (Clyne, 1999). Moreover, in acj6 some of the unresponsive ORNs were not detected electrophysiologically, either because of a defect that made them physiologically silent or because they were absent. Thus acj6 appears to affect more ORNs and to affect some of them more severely than pdm3. Noted is the formal possibility that pdm3 acts in some ORNs other than pb1A but is functionally redundant in them (Tichy, 2008).

The requirement of both pdm3 and acj6 in pb1A led to an investigation of whether they act together or independently. A genetic interaction was found between these two POU genes in the analysis of pb1A odor response. The location of overlapping putative binding sites for both transcription factors upstream of Or42a suggests the possibility of a biochemical interaction between them. Heterodimerization, one possible means of interaction, has been shown previously for other POU proteins. The genetic interaction and coexpression of acj6 and pdm3 in ORNs is in contrast to the relationship between acj6 and another POU gene, drifter, in projection neurons (PNs), the postsynaptic partners of ORNs. Although both acj6 and drifter act in PNs, they are expressed in mutually exclusive subsets of PNs (Tichy, 2008).

A major problem in olfactory system biology is how individual ORNs select which of a large family of odor receptor genes to express. There are 60 Or genes in the genome of Drosophila melanogaster. The expression of particular Or genes, and hence the odor specificity of the ORN, has recently been found to depend on a regulatory code of cis-acting elements. Positive and negative regulatory elements named Dyad-1 and Oligo-1 are required for the selection of Or genes in the correct olfactory organ (Ray, 2007). Within the maxillary palp, additional elements act positively to promote expression of individual Or genes in a subset of ORN classes, whereas other elements act negatively to restrict expression of individual Or genes to a single ORN class (Ray, 2007, 2008). Evidence was found that a combinatorial code of transcription factors underlies the problem of receptor gene choice (Ray, 2007). Transcription factors that act in this process include Lozenge, a Runx domain-containing protein that is required for the expression of two Or genes (Ray, 2007), and Scalloped, which mediates repression (Ray, 2008). Another subset of maxillary palp Or genes depends on the POU gene acj6 (Tichy, 2008).

The current finding that two POU genes, acj6 and pdm3, are required in the same ORN for receptor gene expression suggests that the combinatorial action of POU genes may be an important part of such a code. The Or genes of the maxillary palp can be divided into three classes: those that require both POU genes (Or42a), those that require only acj6 (Or85e, Or33c, Or46a, and Or59c), and those that require neither (Or71a and Or85d). Heterodimer formation and alternative splicing could expand the number of components that act in selecting individual receptors from the entire family of 60 Or genes in the entire olfactory system, including the ORNs of the antenna (Tichy, 2008).

It was found that pdm3 is also required for axon targeting of pb1B and pb3A cells. Interestingly, acj6 is also required for axon targeting of these ORNs, but it is not known whether the two POU proteins act together on common transcriptional target genes or independently on different genes required for axonal wiring (Tichy, 2008).

The relationship between receptor expression and ORN axon targeting has been a topic of great interest. In vertebrates, there is evidence that the odor receptor acts in both processes. In Drosophila, the receptor does not appear to be required for normal axon targeting. Likewise, the effects of pdm3 on receptor expression and axon targeting appear to be separable, in that pb1B and pb3A show apparently normal receptor gene expression but abnormal targeting. A related issue is the relationship between ORN activity and axon targeting. In vertebrates, there is evidence that ORN activity is necessary for the establishment or maintenance of correct ORN axon targeting. It was not possible to ask whether in pdm3 the lack of odor response in pb1A correlates with a failure in axon targeting, for lack of a pb1A GAL4-driver that functions normally in the absence of pdm3. However, it was found that the normal odor responses of pb1B and pb3A ORNs are not sufficient for normal axon targeting (Tichy, 2008).

This is the first mutational analysis of a class VI POU gene and demonstrates the essential role that pdm3 plays in the development of a highly complex and precisely organized sensory system. Further study of pdm3 may uncover critical roles in other systems as well (Tichy, 2008).

Search PubMed for articles about Drosophila Pdm-3

Clyne, P. J., et al. (1999). The odor specificities of a subset of olfactory receptor neurons are governed by Acj6, a POU-domain transcription factor. Neuron 22(2): 339-47. PubMed Citation: 10069339

Ray, A., van der Goes van Naters, W., Shiraiwa, T. and Carlson, J. R. (2007). Mechanisms of odor receptor gene choice in Drosophila. Neuron 53: 353-369. PubMed Citation: 17270733

Ray, A., van der Goes van Naters, W. and Carlson, J. R. (2008). A regulatory code for neuron-specific odor receptor expression. PLoS Biol 6: 1069-1083. PubMed Citation: 18846726

Tichy, A. L., Ray, A. and Carlson, J. R. (2008). A new Drosophila POU gene, pdm3, acts in odor receptor expression and axon targeting of olfactory neurons. J. Neurosci. 28(28): 7121-9. PubMed Citation: 18614681

date revised: 5 November 2008

Home page: The Interactive Fly © 2008 Thomas Brody, Ph.D.