Exeter | ERASE


Characterisation of the fluorescent reporter iLOV


iLOV was introduced to the iGEM database as BBa_K6600004 in 2011 by Glasgow 2011, but so far has seen no further characterization. iLOV has a range of advantages over the more commonly used GFP and related fluorescent proteins, including:

  • Half the sequence length of GFP (halving DNA synthesis costs) [2]
  • Faster maturation of iLOV flurophore than GFP [2]
  • Stable over a wide pH range [3]
  • Recovers spontaneously after photo-bleaching [2]
  • Ability to work in anoxic conditions [1]

The aim of this study is threefold: the emmission spectra of iLOV will be explored over a range of excitation wavelengths, the response of the iLOV protein to exciting photons will be measured and the recovery rate of GFP and ILOV from photo-bleaching will be compared.

Experimental Method

Excitation and Emission Scanning

The scanning of the excitation and emission spectra would occur in two stages. First a broad range scan, 5nm spacing between emission scans, provides the general topography of florescence from the iLOV protein. Then areas around the peaks from the broad range scan would be explored in more detail with 1nm spaced emission scans.

The scans would be made using a TECAN florescence measuring machine. This machine is capable of measuring fluorescence across a 96 well plate, enabling multiple measurements of each iLOV growth culture. The pitfall of the machine's high through put design is that measurements of fluorescence are made by both exciting the protein and detecting emitted photons along the same axis (i.e. directly above the well). This results in photons emitted from the machine reflecting back from the bottom of the well and interfering in emission measurements. As such the TECAN cannot measure emissions from the protein that are closer than about 30nm from excitation wavelengths.

Figure 1: The TECAN microplate reader.

E. coli Top10 cells, with and without the iLOV expression plasmid (Part BBa_K660004) were grown overnight in liquid broth. 10 \ul of the overnight culture was added to 190 \ul of minimal media into each well of the 96 well plate. Minimal media removes the autofluorescence observed using Lysogeny Broth (LB).

The cultures where grown at 37$^o$C with constant shaking at 800rpm until they reached an optical density (OD) above 0.5 at 600nm. The Broad range emission scans at excitation intervals of 5nm from 350nm to and emission was measured from 400nm to 700nm in 5nm intervals. To reduce potential photo-bleaching the number of emission reads for each measurement was limited to five and the wells used for each measurements where alternated. Due to reflection from the excitation photons a portion of the measurements must be discounted where the excitation wavelength came within around 30nm of the emission measurement wavelength.

Fluorescence from each well was divided by the OD specific to that well to standardise fluorescence to cell density. The relative fluorescence of the control cells (Top10) was then subtracted from those cells expressing iLOV.

The second set of finer spaced emission scans, were carried out around peak identified . These scans were taken at 1nm excitation and emission intervals between points defined by the prior part of this experiment.

To minimise the risk of photo-bleaching the iLOV protein only three wells where used for each set of measurements and these were alternated over the course of the experiment. For the finer spaced measurements six well were used to improve the accuracy of the result.

Photo-Bleaching of iLOV and GFP

To test the effect of photo-bleaching on iLOV and GFP two measurements were carried out. The first to determine how exposure time to exciting photons affects protein fluorescence, the second will measure the ability to recover and recovery time from prolonged exposure.

Samples were grown as previously stated, in a 96 well plate and for this experiment E. coli expressing GFP was included as a direct comparison.

The measuring conditions from the detailed scan where be repeated, only without changing the excitation wavelength. This was only repeated for iLOV and Top 10 wells. The repeat would allow for any photo-bleaching effect to be observed minus the bias from changing wavelength. Three scans, in quick, succession, were completed at an excitation wavelength of 449nm and emission was measured from 460nm to 550nm in 2nm intervals to increase the exposure time to 4.5 ms per scan.

To determine how fluorescence is effected by exposure time the scan was repeated. Only three wells were scanned at a time, rotating there use to give the protein more time to recover between experiments. The number of reads per measured emission wavelength was decreased to 1, 20$\mu$s integration. This cut the exposure time by a factor of 10. Triplets of scans were then taken with increasing number of reads (up to 20) to increase the exposure time yet further.

The energy of the photons was not changed as the highest energy available bleached the sample only marginally more than 449nm photons. This is shown in figure 6, the result of preliminary test to determine whether it was best to change the energy of the photons or the exposure time. This test was completed using the same samples used in the intensity experiment, changing the wavelength of the emitted photons and measuring the florescence drop. Using 449nm photons to excite the protein produces the maximum fluorescence output whilst bleaching the sample.

To measure the recovery rate of the iLOV the protein was exposed to exciting photons for 40 ms. The fluorescence of the protein was then measured with 5 20$\mu$s reads at an excitation wavelength of 449nm and emission wavelength of 494nm. This only exposed the protein for 100 $\mu$s at a time, hopefully not affecting the recovery rate greatly. The fluorescence was measured in triplets with 10 second intervals between measurements.

To expose the difference between the effect of photo-bleaching on iLOV and GFP, the two previous experiments were repeated for the E. coli containing a GFP sequence. The emission of GFP at a 449 nm excitation wavelength is higher than that of native Top 10 cells allowing for the measurement of photo-bleaching despite not using the peak excitation wavelength for GFP.

Results, Analysis and discussion

Excitation and Emission Scanning

The fluorescence from the iLOV protein is found using equation 1,

\begin{equation} \operatorname{F} = \frac{1}{N}\sum_{i}^{N} \frac{f_i}{O_i} - \frac{1}{M}\sum_{j}^{M} \frac{g_j}{O_j}. (1) \end{equation}

The fluorescence from each of the wells is divided by the OD in that well ($O_{i/j}$). The average fluorescence per OD of the Top 10 cells ($g_j$) is then subtracted from the average fluorescence per OD of the Top 10 cells with iLOV expressing sequences ($f_i$). This leaves only the fluorescence due to the iLOV protein.

Figure 2 shows the emission topography of the iLOV protein in E.coli. It is clear that there is one large peak of emission centered around an excitation wavelength of 440nm which causes the iLOV protein to emit at around 500nm. The plane of zero fluorescence at the top left of figure 2 is caused by reflection from the excitation photons.

Figure 2: The florescence response of the protein iLOV to a range of excitation wavelengths. The peak response occurred at an excitation wavelength of 440nm and emission measurement at 500nm. Emission measurements are taken at 5nm intervals whilst excitation wavelengths are spaced at 5nm intervals

Following from this result the area between 420nm and 480nm excitation wavelengths were chosen to be scanned at a higher resolution. The emitted fluorescence would be measured at 2nm intervals between 460nm and 550nm for excitation wavelengths spaced 1nm apart.

Figure 3 shows the result of the higher resolution scan. It can immediately be seen that there is some artificial effect at work. The striation of the results is most probably caused by photo-bleaching, where the exciting photons have changed the conformation of the protein. The altered state of the protein may have a higher or lower rate of reemission of the excitation photons.

The striation of the results invalidates them from determining the peak emission and excitation wavelength as it cannot be determined whether it is the rest conformation of iLOV that is being measured.

Figure 3: The florescence response of the protein iLOV to a range of excitation wavelengths. The peak response could not be determined from this measurement due to the photo bleaching of the iLOV (Though it occurred at an excitation wavelength of 444nm and emission measurement of 500nm). Emission measurements are taken at 2nm intervals whilst excitation wavelengths are spaced at 1nm intervals.

Photo-bleaching was observed in the high resolution scan but not in the broad range scan. This may be because during the broad range emission scans the proteins are exposed to exciting photons for 1.8 ms per scan, whilst in the high resolution scan the protein are exposed for 4.5ms per scan. Three scans are taken in quick succession.

To determine whether the effect is indeed artificial, the fluorescence emission at 500nm was plotted across the excitation wavelengths. This can be seen in figure 4. The triplets of successive scans are coloured. Measurements taken from the the broad range scan are plotted for reference.

Fugure 4: The fluorescence emission of the protein iLOV at 500nm for a range of excitation wavelengths. The first measurement of each triplet is always notable higher than the trend.

From figure 4 it can be seen that the first measurement is notably higher than the trend for the next two measurements in each triplet. This would suggest that the protein has been photo-bleached as the emission response decreased after exposure to excitating photons. The difference in fluorescence between the high resolution scan and the broad range scan is greatest at excitation wavelengths of 420nm and 440nm which correspond to measurements taken in the final scan of a triplet. This effect will be explored in more detail in the photobleaching section.

The errors in figure 4 are calculated via equation,

\begin{equation} \delta F = \sqrt{\sum_{i}^{N}\left(\frac{f_i\delta O_i}{O_i^2 N}\right)^2+\sum_{j}^{M}\left(\frac{g_j\delta O_j}{O_j^2 M}\right)^2}, (2) \label{Error} \end{equation}

where $\delta O_i$ is the error in the OD for each well, $f_i$, $g_i$ are the fluorescence measurements from iLOV and Top 10 wells respectively and N, M are the number of wells used.

Due to the striation of the fine-scale results a previous measurement was used to confirm the peak of fluorescence. This result can be seen in figure 5.

This result was not used originally as measurements where carried out in Lysogeny Broth (LB) which has a higher autofluorescence than the minimal media used previously. The higher autofluorescence of the LB resulted in a masking of the fluorescence topography from the Top 10 wells. This resulted in the Top 10 cells without iLOV sequences reading as having a higher fluorescence than those with the iLOV sequence for certain ares of the experimental space. The fluorescence topography of the iLOV protein is still visible despite this and matches the topography observed in minimal media.

The peak excitation wavelength is 449$\pm$1 nm and the peak emission corresponding to this excitation wavelength is 494$\pm$2nm. The peak has a lower apparent fluorescence than previous measurements, this is due to a lower manual gain on these measurements. The manual gain is the factor that the measured intensity is multiplied by. It was lowered for these measurements in an effort to take emission measurements closer to the excitation wavelength. However this was unsuccessful, only resulting in some reflection measurements being read as within the acceptable fluorescence range.

Figure 5: The florescence response of the protein iLOV to a range of excitation wavelengths. The peak response occurred at an excitation wavelength of 449nm and emission measurement at 494nm. The fluorescence of the iLOV protein is calculated to be less than 0 for some areas of this plot. This artifact is likely due to interference caused by the Lysogeny broth that the experiment was carried out in.

No photo-bleaching was observed in this measurement despite the protein being exposed for 5ms per scan. This is likely due to the greater opacity of LB in comparison to minimal media.

Photo-Bleaching of iLOV and GFP

Figure 5 shows the fluorescence response of iLOV to exciting photons from three consecutive 4.5 ms scans each taken as a fraction of the first. The measuring conditions of the scans was repeated from the detailed scans used in the previous section. This was done to expose the cause of the striation seen in the results. It is clear to see that the fluorescence drops after the first scan then remains constant between the second and third scan. This drop is caused by photo-bleaching of the iLOV protein.

The rate of photo-bleaching across the scans is constant. This is shown by the slope of the trend lines being close to zero. It is interesting that there is no drop in fluorescence between scans two and three despite the rate of photo-bleaching during scan two being the same as scan one. The lack of a drop suggests that the rate of photo bleaching during each scan is zero and that any drop in fluorescence occurs after exposure. The absence of a drop between scans two and three also suggests that some maximum bleaching effect is reached.

Figure 5: The fluorescence drops after the first scan. This is the action of photo-bleaching. The second and third scan remain at the same, decreased, fluorescence level despite further exposure.

The 449nm wavelength used for the triplet of scans in figure 5 corresponds to an energy ($E$) of,

\begin{equation} E = \frac{hc}{\lambda e} = 2.761 [\text{eV}]. \end{equation}

Electron volts (eV) are used in the following section as they are appropriate to the energy scale used in the experiment. For reference molecular bonds have energies of the order 1 to 10 eV.

Figure 6 shows the decrease in fluorescence caused by increasing the photon energy. The figure also shows fluorescence loss for the same energy photons at a higher exposure time. The TECAN plate reader was capable of emitting photons at a maximum energy of 5.39 electron volts . At this maximal energy the decrease in fluorescence is minimal compared to the decrease that could be obtained by increasing the exposure time. As the exposure time could be increased over a far greater range it was decided that this should be the variable parameter in the photo-bleaching experiment.

The energy was not changed to the maximum for the photo bleaching experiment as the greatest emission from iLOV was observed at an excitation wavelength of 449nm. The larger emission minimised the error in measurements of the fluorescence decrease.

Figure 6: The decrease in fluorescence caused by increasing photon energy to the maximum available to the TECAN is marginal compared to increasing the exposure time. Keeping the photon wavelength at 449nm allows for any photo-bleaching effect to be measured over a greater range of emission.

Figure 7 shows the fluorescence decrease of iLOV and GFP as a function of photon exposure time. Values are expressed as a fraction of the initial fluorescence measurement. The fluorescence decreases as exposure time increases flattening off at 0.955 times the initial measurement. This flattening is probably the results of the experimental method. As multiple wells where measured, increasing the exposure time also increased the recovery time between exposures. Using a single well for each experiment would have reduced this systemic error but would not have removed it.

Little difference in the loss of fluorescence between GFP and iLOV is seen in figure 7. The fluorescence of GFP appears to flatten out at a higher fraction of the starting fluorescence than iLOV, this may be because GFP is a larger protein and so has a area to dissipate energy over.

Figure 7 suggests that the loss of fluorescence is proportional to the amount of energy put into the system. The protein must dissipate this energy at some rate. As such a better parameter to change with regards to observing photo bleaching would have been the intensity of the photon source. In the future this would mean using a laser source.

Figure 7: The florescence decreases with exposure time plateauing after around 25 ms. A higher intensity source of phtons (A laser) would be needed to reduce the fluorescence further. There is little difference between iLOV and GFP for these expose times.

Figure 8 shows the recovery of GFP and iLOV after a 40ms exposure to 449nm photons. 40ms was the maximum exposure time that could be programmed on the TECAN plate reader. The maximum exposure time was used for this experiment as only one scan of its duration was needed, further scans exposed the proteins for just 100 $\mu$s. Both GFP and iLOV can be seen to be recovering from the extended exposure, however there was a large deviation amongst the measurements and the correlation was weak. Only five reads where used to measure the fluorescence in an attempt to limit any additional bleaching.

The true recovery rate of iLOV and GFP cannot be determined from the experiment. This is because the recovery rate could not be measured independently of additional photo-bleaching events. By t = 5:02 the fluorescence of iLOV and GFP had reached the lower limit of the bleaching effect seen in figure 7, beyond this point the additional exposure would have certainly disrupted measurements.

Figure 8: The recovery could not be measured without further photo-bleaching of the sample.

Whilst this result reveals very little about the recovery rate of iLOV, an estimate of its upper limit can be made from the fluorescence topography scans. Wells where used alternately, spending 6 minutes inactive between uses. The fluorescence appeared to have returned to normal levels at the start of each triplet of scans (See figure 4). This suggests that the maximum recovery time from three 4.5 ms scans is 6 minutes.


This experiment characterised the fluorescence topography of iLOV. Photo-bleaching of the iLOV protein by a low power light source was also explored. These results should help future iGEM teams analyze data they collect whilst using the iLOV part.

The iLOV protein gave the greatest fluorescence response when excited with photons of wavelength 449$\pm$1nm whilst detecting photon emission at a wavelength of 494$\pm$2nm.

After a 400 $\mu$s exposure the fluorescence of iLOV was found to drop around 5%. This was approximately the same for GFP. To bleach the proteins any more than this a higher power light source would be required.

The upper limit of iLOV's recovery time from a triplet of 4.5ms exposures was 6 minutes. When mapping the topography of iLOV it was seen that accurate measurements of iLOV's fluorescence could be made with exposure times as little as 100 $\mu$s. Therefore photo-bleaching can be avoided when using iLOV as a reporter with the condition that the protein is allowed to rest for at least 6 minutes between measurements.

To better characterise the photo-bleaching effect on iLOV, a laser source could be used. If the power of the source is high enough irreversible bleaching in GFP could be shown along with proof of the photo-reversibility of the iLOV protein.


  2. The phtoreversible fluorescent protein iLOV outperforms GFP as a reporter of plant virus infection. Chapman, et al.
  3. Swartz TE, et al. (2001) The photocycle of a flavin-binding domain of the blue light photoreceptor phototropin.


Previous: Detection of Xenobiotics

Next: Our Parts

Exeter | ERASE