FRET is suitable for TSA.

When enough protein is added and the proper temperature range is selected, SYPRO Orange gives a strong ang detectable signal at FRET mode.

It should be noted, that buffer composition can greatly influence on TSA results, as well as rate of sample heating.

Thanks Igor.

Why I'm hesitant on accepting these results is because FRET is a completely different type of technique & scanning mode and I can't seem to find the ex and em wavelengths for FRET. 

If I were to use an actual channel/filter on the CX96 (other than FRET), what would you or everyone would suggest?

Hello Shirley,

I'm not sure on the CX96 instrument exactly, but it appears from the BioRad website there is an option for a FAM reporter. The FAM reporter works quite well for SyPro Orange on my QuantStudio 3 instrument. The HEX reporter would also likely work.

I'd try collecting a uniform sample on all possible channels and looking to see what gives you the most robust shifts (reproducibility) and also the biggest delta RFUs between folded and unfolded (sensitivity). 

Good luck!

Hi Kyle, Thanks for your input. 

I have tried using FAM as well, it's giving me the same results as water. The bold line is H2O. Which leads me to think something must be wrong with the reagents. 

- Can't be my BSA as it was fresh and pure from an ampule I opened.

- Buffers all sterilized.

- 5X sypro orange was diluted from 5000x with H2O

- Filters/channels are all wrong and not calibrated!

Shirley,

I agree, I think at this point it is most likely your reagents went bad. Was the stock stored in DMSO? If you have an undiluted stock, you can dilute to 1000x in DMSO and try again with a gradient of buffer strength, [protein], and maybe even [SyPro Orange].

It's also possible it's a calibration issue, but all I can say is check the manual for that, as I'm unfamiliar with that particular instrument. Perhaps run whatever controls it suggests?

On my instrument I usually see >100,000 RFU shifts between folded and unfolded protein, minimally a few thousand to consider it robust - but this will all be instrument specific.

Finally, if you haven't been, consider incubating the dye with your protein samples for 15 minutes to an hour before running it so it can equilibrate with the system for a baseline level of fluorescence. Sometimes if you add it then immediately start the experiment you will have very high initial fluorescence that rapidly drops before then producing the melting curve.

Hope this helps.

As far as I know, CFX96 doesn't need in aditional calibration.

In my TSA experiments with thermostable proteins at given concentration of SyproOrange FRET mode gives the highest delta between folded/unfolded protein, while signal at FAM/HEX channels was much lower. I agree with Kyle, that root of problem may be in water solution of dye (it can aggregate) or in initial incubation of protein with dye before run. May be it also has sense to change model protein.

Update:

Bought new SYPRO and made dilutions with DMSO. 

This time around, I tackled it like this:

Titration of [NaCl], titration of [SYPRO] with all channels selected while incubating at RT for 30 min prior to run. BSA is still at 0.2mg/mL final conc.

Result: Crap. In pic, blue is FAM and orange is ROX. I only showed those two because I read around here that ROX worked as well, but not in this case! It is definitely not the amount of salt or dye that is affecting it. The only thing that I haven't changed yet is the protein and instrument. I don't understand why this instrument wouldn't work or detect it though. 

I'm getting more and more frustrated with this.

Hello Shirley,

That's unfortunate about the new reagents. Here's what I would try next, assuming all of your reagents are now fine:

1. Check the program is holding long enough between ramping steps. Sometimes if the run is 'continuous' the data points look funky because the system hasn't equilibrated with the the dye before the measurement is taken. If it's an option in your software, I'd recommend a 'step & hold' of around 0.5 C with a 10-20s hold after each ramping step.

2. Do a gradient of protein concentration. It's possible 0.2mg/mL for BSA is not enough to elicit a strong signal on your instrument. I don't know what would be a strong response on your instrument, but 1500 seems small, and so it's possible those curves are effectively flat and not responding at all.

3. Try different proteins if you have them, especially other cytosolic, well-behaved (in solution) proteins.

Hope this  helps. Good luck!

EDIT: I forgot to mention that I too have been told ROX works well for SyPro Orange and have yet to actually see it perform as well as FAM.

Thank you Kyle for your response. 

The program is definitely holding long enough as I have set it up to hold for 1 min before ramping up to the next temperature.

I will try and do a titration of [BSA] and probably throw in some other protein to add some more excitement. 

Surprisingly my actual data results gave me melting points of 55C, 56C, and 58C (somehow the program was able to decipher that from those curves and data points) and guess for which filter -- ROX! But I'm not sure if it's because these samples have the highest conc of SYPRO (50X). 

That's very interesting. From the plot above it seems very hard to see that, but perhaps if you only view those curves so the scaling isn't so harsh (from FAM) you'll see the features it is pulling out. 

I'm thinking if the 50X dye was the only set to yield reasonable numbers it's either A) because you have the most sensitivity for the protein (i.e. increasing [protein] may help), or B) it's just coincidence.  Let's hope it's A!

Also, shameless plug - I wrote some software for the analysis of TSA data that only requires you're able to format your data in a csv with temperature in the first column and all your subsequent wells as the following columns (with the first row being labels). If you'd like to use that to speed up your data analysis the link is: https://beamerlab.shinyapps.io/software/ and if you need more help figuring out how to use it let me know.

Hi Shirley,

Did you figure out how to fix your problem. I have been getting similar curves like you showed with a different dye.

Also, how do you program for a step and hold on the machine. As far as I know, I have programmed for ramping up temperature but no step and hold between change of temperatures.

Best,

Leepakshi Khurana

I am having similar issues did anyone decide a good filter for DSF with sypro orange? When I scanned all channels the only result that doesn't start with high RFU like Shirley's is using Texas Red.. I'm unsure if this filter is even one to consider but it shows expected results. Has anyone tried this or had similar results?

I tested the cfx96 recently and for my proteins, the fret channel turned out to be best for thermal shift with SYPRO Orange. I didn't like it though because there is no info about how it works. I'm not into using a black box. I think that BSA is not a good standard for this technique, because it normally binds free fatty acids in the serum and thus readily binds the probe. By the way, maybe it is possible to saturate the ffa binding sites of bsa with a large excess of a probe?

Rafał Hołubowicz Shirley Guan I am trying with cfx96 qPCR instrument and CPM dye with a membrane protein. Even using all channels to detect I don't get any output in the "view curves" section of the run.

Also how do you set the protocol for the run? Slowest ramp is 6 degrees per minute which is quite fast for protein melt scans. How do you programme the ramp and hold on the machine?