Hi Roger,

we also had hideous problems with this system and I'm sorry to say we eventually gave up on it and I think I'm correct in saying we would not go down this route again. You could try some triton in the buffer and also blocking with BSA. You might just get lucky. My post doc did the work on this so you may want to email her for more details (Sarah Parry-Morris ).

Hi Roger,

We saw this problem too but just in few cases. Try different salt concentration, might help. Do you use SA or SSA sensors ? They behave differently.

Dear Roger,

We have done a lot of kinetic assays on the Octet using immobilized biotinylated peptides and different antibodies in solution. We are routinely using 5ug per ml of the peptide for immobilization. Than we block with biocytin and subsequently with casein block solution. This PBS casein solution is than also used for associatin and dissociation steps. This protocoll works very well for our system. If it dors for your protein I do not know of course...

Best, Ulli.

Dear Roger,

I recently experienced a very similar case with high aspecific binding of a non-glycosylated peptide to the biosensors where no biotinylated protein was loaded. Blocking the 'empty' biosensor with biotin didn't help, and increasing the BSA and/or tween-20 concentration in the buffer didn't help neither.

Personally I was considering your proposed solutions 1 and 4 as my next move to avoid aspecific binding.

But it is not the first time that I encounter the problem of non-specific binding with octet biosensors. While the Protein A biosensors seem to work rather fine, I had troubles with significant non-specific binding anti-his and anti-GST biosensors. I start to think that it is rather a more general problem of octet biosensors, rather that a problem with streptavidin itself.

I will let you know if I find a satisfying solution for my experimental set-up.

Hi Roger,

We have the Octet Red 96 in our lab and I admit I love the system. We observed more non-specific binding (NSB) with histidine-tagged protein and Ni-NTA sensors but with Biotinylated protein and streptavidin sensros, the NBS was greatly reduced. It would help if you note the buffer you used. We used 20 mM MOPS-Tris 50 mM KCl, pH 8 buffer with 0.5 mg/ml BSA (final) and we saw that the NSB was minimum. You have the right approach when you say you added PEG, although we didn't observe any changes with MOPS. Regarding Tween-20, we have observed that 0.005% final concentration of Tween-20 is the most suitable; with 0.05% final conc actually increasing the NSB. With DTT, NSB gets worse. TCEP was the more suitable reducing agent for us. I admit, we have seen some problems with peptides when other labs have used our machine; not with protein.

Reading your problem, my opinion would be to try ordering peptides from a different company. We have seen 10 fold difference in binding of same biotinylated peptides ordered from different companies. Your options 1,3 and 4 also sound reasonable. In the meanwhile, you can check this video out which we recently made for explaining the working of Octet http://www.jove.com/video/51383/bio-layer-interferometry-for-measuring-kinetics-protein-protein

You can see from our paper that we were able to successfully use the system and counter the NSB here- http://www.ncbi.nlm.nih.gov/pubmed/23400782

I know my answer is not the most complete one but we can talk more on this. Hope I was able to help.

Roger, you may also find useful info from agarose beads type of systems for pull-down assays. We see big difference in nonspecific binding to streptavidin versus avidin agarose beads with basic (pI ~10.5) protein. There is a short description of streptavidin properties in

http://www.lifetechnologies.com/us/en/home/references/molecular-probes-the-handbook/antibodies-avidins-lectins-and-related-products/avidin-streptavidin-neutravidin-and-captavidin-biotin-binding-proteins-and-affinity-matrices.html

Roger, I have seen other cases where blocking with biotin or biocytin did not help. Our Octet work has mainly used non-glycosylated protein (see Naman Shah's response), but I have experienced such problems with several colleagues who have done peptide/protein interactions on our Octet system, and not all used glycoproteins. The nonspecific binding signal varied between protein preps, so my suspicion is that the protein may contain some denatured fraction that sticks tightly to the biosensors surface, regardless of the streptavidin. Higher concentrations of BSA or other blocking protein might help with this, but is also worth testing the homogeneity/stability of the protein. If Sergey's comment above is right, it may be a distinct problem for glycoproteins interacting with streptavidin. Best of luck.

In my former lab we have been using the Octet system, but especially for protein-peptide interaction, we observed unspecific binding as well even in the presence of several mg/ml BSA.

If you would like to test your interaction in a completely immobilization-free set up, I would recommend using MicroScale Thermophoresis (MST)! This technique allows you to measure your interaction free in solution, without any coupling to surfaces!

Just watch this 3min video to explore how it works:

http://www.nanotemper-technologies.com/news-events/news/article/watch-a-3-minutes-movie-to-explore-microscale-thermophoresis-mst/

You could also join one of our MST Meetings where you can measure your own samples! In addition, there will be presentations by NanoTemper as well as by guest speakers.

http://www.nanotemper-technologies.com/news-events/events/article/10th-international-mst-meeting/

Please also feel free to contact me ([email protected]) for any questions!

Many thanks to all of you for your suggestions. Just some more specifics for those that have asked: I have been using the SA (as opposed to SSA) sensors. I would be interested to hear if either type has a benefit wrt non-specific binding. My buffer was standard PBS, pH 7.4, 0.1% BSA, 0.1% PEG 8K (the BSA and PEG were the recommendation of a colleague who had found they worked well with his particular proteins). I have now tried also Biocytin and Biotin-BSA blocks with little change, unfortunately.

I will certainly take on board your suggestions as to other things I can try:

1. Try changing buffers / salt concentration / detergent conc.

2. Trying casein as a block (+ biocytin).

3. Perhaps increasing the BSA.

4. Trying a lower level of Tween-20.

Of course, it may be that my protein is exhibiting some heterogeneity/unfolding/aggregation, although it seems quite well behaved in terms of gel filtration and thermal melting behaviour.

@Heidi Thank you for your advice about MST. I believe our neighbours at the LMB in Cambridge have one of these instruments, so I may ask them about trying it out. I may have to go to a completely solution phase assay, this protein seems sticky.

I believe SSA are made from different material - SA (and all other sensors) are white/transparent while SSA are yellow. May be your protein binds to the sensor surface, not streptavidin ? Also try higher salt, say 500 mM NaCl.

If your protein is "sticky", there might be problems with MST too. Protein will be binding to the capillaries wall and it ruins the experiment, I had this experience.

SSA sensors are super-streptavidin sensors. They have higher amount of streptavidin on the surface of the sensor as compared to streptavidin sensors. This allows for a higher binding signal and interactions on larger surface area of the sensor will be specific; thus reducing non specific binding. Keep in mind that SSA sensors are more expensive than SA. So if you can get a few sensors to try out first, it will be a better idea than buying the whole 96-sensor box.

@Oleg and @Roger:

You are right, molecules sometimes stick to capillary walls. The good thing about it: you will see it! You will even see if the molcules are sticking to tube walls in which you prepare your binding reaction! And then you can improve your assay conditions, either by changing capillary type or by adding some passivating agents or detergents (such as BSA, Tween-20, Pluronic F-127, etc.) to your assay buffer.

Actually, sticking is not unique to MST, but with MST you have such small volumes and low concentrations that you are able to detect it.

Please let me know if anything is unclear or if you have any further questions!

I believe ForteBio does not reveal what is the composition of the "biocompatible matrix" beneath the streptavidin layer. Years ago, using the streptavidin chips from Biacore, we found the dextran base problematic. We had to resort to hydrophobic sensor chips, coating them first with biotinylated SUV-vesicles and thereafter with Neutravidin.

Hi Roger,

Have you managed to sort out your issues with non-specific binding ?

You were probably experiencing non-specific binding to the sensor matrix rather than to SA which is the reason why the best negative control, in your case, would have been an irrelevant or scrambled peptide. This is not unusual with Octet. You mentioned most of the possible fixes, I would start by loading as much peptide as possible and add a blocking step with biocytin and BSA prior to binding your analyte. You will definitely need to optimize your buffer for your protein interest. You can play with FBS, BSA, Tween and even Triton to find the best conditions.

Best,

Vincent

Simple solution - practical but not a very scientific I admit- is to simply run your assay again with no changes to samples or protocols. The non-specific binding has been blocked now, hopefully. and what you will see in the repeat is the real binding event.

Yes these sensors are problematic. We spent a lot of time optimising. By all means use BSA and Tween, but also try extensive reblocking after the peptide binding to reach a new base line. That might just help. We no longer look to this system for interaction work as it takes far to much time to optimise. Good luck!

Hi all,

Thanks for the continued advice. A quick update: I ended up jumping ship to a different system rather than spend time trying to avoid the non-specific binding. My protein tends to like sticking to surfaces if it get's the chance. So, I ended up using a microscale thermophoresis machine in a neighbouring institute to detect binding between a fluorescently-labelled peptide and my protein.

Dear Roger, 

FYI, I've heard that the 'biocompatible matrix' of the SA sensors is likely carboxymethyldextran, like the CM5 chips for the Biacore system, so the streptavidin is immobilized to the sensor surface by reductive amination.  

We've also found that we need to use a similar, but non-reactive molecule on our sensor surface as a reference, since biocytin did not perform well enough and our protein seems to like binding to either the protein component of SA or the underlying surface.

Hope this helps!

His tips work beautifully

if your protein have his tag use the his tips

It seems the SA tips are very problematic. In my hands, I realized that the sensor interacts strongly with basic proteins for a reason unknown to me. My protein has a pI ~9, and both the biotinylated and the unbiotinylated versions interacted with the probe appreciably. We then tested binding with BSA (acidic) and lysozyme (basic) and realized it binds strongly to the latter. Using unbiotinylated protein in buffer with 0.5M NaCl and 0.5% triton-x 100 as the reference improved the Kd measurement for protein-biotin (10^-12 M)..

I hope this helps someone.

We have standardized Fc-FcRn interactions on SA sensors in past and we used biotinylated FcRn (both with Fc and IgG). There was no nonspecific binding observed for majority of sensors in a pack and they were good to be used in our work-flow. We got values very close to published for WT and set of Fc mutants.

Now after couple of years we again did same experiment on new batch of SA sensors and they have huge nonspecific binding (OD>1.0) to Fc and also sensors give signal when dipped in just PBST. It does come down to OD 0.2 - 0.4 (for Fc) after three cycles of regeneration with glycine pH 2.0 and we still got trend that we expected but we think this may also has something to do with batch to batch variation and also sensor to sensor variation in same pack.