We and others showed previously that by using the Tetracycline-controlled Tet-on/Tet-off system, endogenous OR gene regulation mechanisms can be bypassed to express any tgOR
in a large number of OSNs 
. Here, we have engineered transgenic animals where two different tgORs
tagged with either an IRES-GFP or IRES-lacZ) were placed under the control of a single bidirectional Tet-operator
(). Mice carrying tg(bi)OR
transgenes were crossed into a background expressing the tetracycline dependent transactivator (TTA) in all mature olfactory neurons (OMP-TTA
mice (see Methods
for details). In one OMP-tg(bi)OR
line, Line 319A (referred to as the OiS-line), an octanal receptor (rI7
) was expressed in more than 90% of OSNs and an acetophenone receptor (M72
) in a further ~5% (, S1
). Thus, the expression of rI7
in OiS-mice resembles that of tgORs in the OMP-tgM71
lines that were reported earlier 
. The anatomy of the MOB in OiS-mice is similar to controls and other OMP-tgOR lines 
with most glomeruli filled with rI7-IRES-GFP
expressing primary afferents (GFP-fluorescence, ). Interestingly, although M72-IRES-LacZ
was expressed in about 5% of OSNs, we observed very few lacZ positive fibers in the MOB ().
Expression of rI7 and M72 in OiS-mice.
Previously it has been reported that mice with global expression of the OR M71 cannot detect acetophenone, the M71 cognate ligand, although they still are able to detect other odorants 
. Therefore we anticipated that the OiS-line would behave in a similar manner and began by testing mice in a standard habituation-dishabituation assay 
. Our results (Fig. S2
) suggest that this assay lacks sensitivity as control animals showed minimal responses to a range of odorants including octanal and acetophenone. Intriguingly, both control and OiS-animals spent no more time investigating octanal-scented filter paper than a filter paper spotted with mineral oil. Surprisingly, however, OiS mice regularly exhibited a dramatic and involuntary phenotype: out of 21 OiS-mice exposed to 1% octanal, three exhibited major tonic-clonic seizures, two exhibited moderate and two milder seizure-like effects (see below and Movies S1
). As expected, control mice (including mice carrying the OiS-bicistronic tgOR
but no TTA
driver) never exhibited seizure-like symptoms in the presence of octanal and OiS-animals never showed even mild effects that resembled seizure when exposed to a wide variety of other odorants including acetophenone (Movies S4 & S5). Finally, in trials where OiS-animals did not experience seizure, mice regularly investigated and even sniffed the filter paper (Movie S6).
Seizures followed a stereotyped progression that did not require mice to approach the odorant-scented filter paper. Strong seizures began within the first 15–20 s of a transgenic mouse being exposed to octanal, while weaker events were usually observed within the first minute of exposure. Symptoms were characterized by three stages: in the first, mice began to rapidly blink their eyes before retreating while losing control of their forelimbs; this phase typically lasted 10–15 s. The second stage began with mice rearing up onto their hind-legs while retracting their forepaws and evolved to mice foaming at the mouth over a period of 5–10 s. In the third phase, mice collapsed backwards before repeatedly rearing and collapsing again for a period of 15–30 s. After this, mice generally lay immotile for a period ranging from 10 s to 30 minutes before fully recovering. Animals exhibiting all these symptoms were classified as showing severe effects, whereas mild and moderate symptoms were ascribed to mice exhibiting only the first or first two stages of seizure, respectively.
We reasoned that if octanal was inducing seizure through the olfactory system, increasing the concentration of the odorant should increase the number and severity of events observed. Indeed, when 10% octanal was used as odorant both the fraction of mice displaying seizure and the severity of the symptoms increased (). We noted that individual mice exposed to this test multiple times did not show a particular pattern in their response: within the limits of our data all OiS-transgenic mice were equally likely to exhibit symptoms (and variable severity of seizure) each time they were exposed to octanal thus it is unlikely that genetic variability or level of transgene expression account for differences in response.
Table 1 Number of OiS-mice that experienced seizures when exposed to 1% or 10% octanal (see Movies S1, S2, S3, S4, S5, and S6 for examples of the different levels of seizure recorded here).
Another simple explanation for why OiS-mice exhibit variable responses to octanal could be that different proportions of OSNs are activated by odorant. However, we found no evidence supporting such a scenario when we examined c-fos
expression as a measure of neuronal activity 
. Octanal exposure induced (predominantly weak) c-fos
expression in a relatively large subset of OSNs of control animals (compare ), consistent with octanal activating multiple different endogenous ORs in the mouse 
. As expected, OiS-mice exhibited much more pronounced OSN activity after octanal exposure. However, there was no detectable difference in the number of responding OSNs or the strength of c-fos
expression between mice that experienced a seizure and those that showed no adverse response to octanal (). Indeed, in all cases, essentially the entire population of OSNs now expressed high levels of c-fos.
These results suggest that the variation in symptoms that we observed is more dependent on differences in signal processing in the MOB and higher brain centers than the overall extent of MOE-activation.
Octanal induced activation of OSNs in the MOE.
We next examined neuronal activity in the MOB again using c-fos induction to allow assessment of changes in activity throughout the bulb. Although immediate early gene expression provides a powerful approach to study large scale and global changes in neuron firing, its temporal resolution and sensitivity coupled with appreciable background mean that subtle changes in activity may be missed. Control animals exhibited appreciable neuronal activity in both periglomerular (PG) and granule cells both after exposure to carrier (mineral oil) and odorant (octanal, ). A similar pattern of c-fos expression was observed in OiS-mice exposed just to mineral oil (). Notably, however, OiS-mice showed differences in c-fos induction that paralleled their seizure symptoms when challenged with octanal. Mice that showed no response did not experience a seizure were characterized by little or no increase in bulbar c-fos expression after octanal exposure (). In contrast, mice exhibiting strong seizures displayed increased expression of c-fos in M/T cells () as well as very prominent activity in the vast majority of granule cells () but not in the inhibitory PG cells ().
Neuronal activation in the MOB after octanal stimulation.
Since seizures result from an over-excitation of neurons and the spread of this excitation in the brain, we anticipated that mice exhibiting odor-induced seizures would show increased neuronal activity in the piriform cortex, the next major center in the olfactory pathway. The piriform cortex of control animals revealed scattered c-fos
expression in a distributed subset of neurons after octanal exposure (), consistent with previous results 
. As expected, OiS-mice that showed no signs of seizure when exposed to octanal exhibited a similar pattern of scattered c-fos
expression in piriform cortex (). In marked contrast, OiS mice that experienced strong seizures displayed robust c-fos
expression in the piriform cortex and other regions of the forebrain () as well as the hippocampus (), reminiscent of previous reports on kindled seizures 
The spread of neuronal activity corresponds to severity of seizures in OiS-mice.
Odorant induced seizures in OiS-mice are tied to pronounced neuronal activity in the hippocampus.
In OiS-mice, octanal exposure always induced strong OSN activation (c-fos expression, ) but neural activity in the olfactory cortex and forebrain was tightly correlated with seizure symptoms ( & ). Why do different trials produce such different outcomes? We reasoned that the sense of smell needs mechanisms to separate salient and rapidly varying input from background and perhaps selectively inhibits widespread but slowly changing olfactory activity. To test this idea, we used an olfactometer to apply a controlled flow of octanal to OiS-mice. Remarkably, seizures were never observed (0/7 mice) when a 0–5% octanal gradient was delivered over a period of 4 minutes. In contrast, rapid exposure to 5% octanal induced symptoms in all animals within the first 20 s (7/7 mice; 2 strong, 4 moderate, 1 mild).
We also examined c-fos
expression in mice from the two groups of mice that were exposed to octanal in the olfactometer. demonstrates that in the olfactometer, both rapid and gradual exposure of OiS mice to octanal induces c-fos
expression in almost every OSN. Interestingly, the level of c-fos
expression induced by the octanal gradient was much lower than that observed after rapid delivery of 5% octanal indicating that peripheral desensitization occurs during this regime of gradually increasing odorant delivery. Importantly, however, both treatments activated a similar number of OSNs. Graded octanal delivery did not increase c-fos
expression in the piriform (). In contrast, rapid exposure of OiS mice to 5% octanal activated M/T and granule cells in the MOB as well as piriform neurons and other regions of the forebrain (, S3
). Thus controlled octanal delivery closely recapitulates the activity patterns observed when mice exhibited seizures after passive exposure to octanal and, as predicted, a rapid increase in global OSN activation can overcome powerful inhibitory processes that otherwise suppress widespread olfactory activity.
c-fos expression in OiS mice under controlled octanal exposure.