Microscopio Confocal

Oct. 2008 No. 30
CONFOCAL APPLICATION LETTER

reSOLUTION
Dye Separation

Introduction
Today a wide variety of fluorescent dyes and fluorescent proteins are available for multicolor fluorescence microscopy. Recorded signals from these fluorescent molecules provide complex information about multilabeled samples, often necessitating quantification or localization/colocalization analysis. Ifsignificant overlap of the excitation or emission spectra of multiple fluorophores occurs it becomes difficult to distinguish between the different signals. Consider a combination of the four fluorophores Alexa 488, Alexa 546, Alexa 568 and TOTO-3 (see Fig. 1). It is difficult to separate emission signals from these dyes due to their strong spectral overlap, which results in signals from multiple dyes in eachchannel. This phenomenon is termed crosstalk, or bleed-through. Interpreting multicolor images can be challenging in this case because they arise from a mixture of signals from multiple dyes.

Excitation Spectra ExcitationSpectra

400

450

500

550

600

650

700

Emission Spectra Emission Spectra

450

500

550

600

650

700

750

Fig.1: Excitation andemission spectra of Alexa 488, Alexa 546, Alexa 568 and TOTO-3.

Crosstalk There are different options to avoid and/or remove crosstalk of fluorophores for multi-labeled samples. For example, when using simultaneous scan mode there are acquisition strategies to minimize crosstalk. One way is to optimize the detection range to avoid crosstalk. Reducing the excitation light for each respective fluorophorewill also reduce the emission intensity, which in turn reduces the degree of crosstalk. But, if the degree of overlap is too strong (Alexa 546/Alexa 568 or Dapi/FITC) it is better to choose the sequential scan mode. However, sequential scan may not be the best choice when speed is important (i.e. for live cell imaging). Simultaneous detection of all dyes may be necessary; sequential scan mode maybe too slow. In addition, samples that are stained with multiple fluorophores that are excited by the same laser line (see example Fig. 1) will exhibit crosstalk despite using sequential scan. In these cases a mathematical restoration of dyes into separate channels may be necessary. This will be discussed in the following. Consider a FITC/TRITC double-labeled sample. See in Fig. 2a the emissionspectrum of only one dye. The total emission light collected from FITC will be distributed in both channels. Here the green channel collects about 3/4 of the entire green signal while 1/4 of the signal spills over into the red channel. For the red channel a similar situation exists (Fig. 2b). The total light collected from TRITC will be distributed in both channels. Here we estimate 4/5 of the redsignal is seen in the red channel and 1/5 of the signal goes into the green channel.

2 Confocal Application Letter

Dye Separation

Fig. 2a: Emission spectrum of the green channel We estimate here: 3/4 of all FITC emission goes into the green channel 1/4 of all FITC emission goes into the red channel

Fig. 2b: Emission spectrum of the red channel We estimate here: 1/5 of all TRITC emissiongoes into the green channel 4/5 of all TRITC emission goes into the red channel

In a double-stained sample (Fig. 2c), signals from both dyes will be present. In our example you will record 3/4 FITC + 1/5 TRITC in the green channel and 1/4 FITC + 4/5 TRITC in the red channel.

Fig. 2c: Emission signals of a double-labeled sample. The black curve represents the sum of the signals of bothfluorophores.

Fig. 2d: Removing crosstalk: 1/4 of all FITC emission has to be removed from red channel 1/5 of all TRITC emission has to be removed from green channel

Confocal Application Letter 3

The goal is now to separate the signals, so that each channel contains only the signal from one dye. This means that 1/5 of the TRITC signal has to be removed from the green channel, and 1/4 of...