? print your protein onto cellulose membrane (BSA as control), block by glucose, incubate with starch in PBS, wash, then develop with Iodide?

Thanks I will try. I am performing western blots so this would be easy to do but do you have some citation on the development with iodide? 

Iodide meets starch shows dark blackish blue colour. When you do western like blot, be ware that a denatured protein may lose its ability to bind starch.

You could try an affinity chromatography: Immobilize your purified protein (say 1..5mg) to 1ml of activated sepharose (there are nice columns available) and then pass a diluted starch solution through it. then elute with high salt and/or at pH 2..3, neutralize the fractions and test them with KI/I2 for the presence of starch.

Alternatively, do all steps with loose material on a rotator and probe the resin directly for bound starch. (Assuming/expecting/hoping that sepharose doesn't interact with iodine).

There are also ready-made amylose columns usually used to bind MBP-tagged proteins (e.g. MBPTrap columns from GE). You could try binding your protein to such a column and then elute with 10 mM maltose. If you get no elution, try with a pulse of SDS, guanidine hydrochloride, sodium hydroxide or similar.

2 methods, 1 analytical and another screening same technique (group separation/GF). Both are gentle and should not disturb binding.

1: Set-up your protein binding with starch then desalt in to the same buffer without starch, using PD G-25 minitrap columns (GE Healthcare). Then you can perform the iodine test on the solution. This will separate the protein and bound starch from the buffer containing starch. Also assuming you use something like Sigma ACS MW 325 starch.

This method is also good for determining optimal conditions for binding. Assuming your expressed protein is >5kD.

2: Set-up your protein binding with starch, then apply to a Gel Filtration column.

The starch will increase the hydrodynamic volume of the protein and should give a marked shift in the apparent MW. 

This method is more analytical and you should see more protein "shifted" in conditions that are optimal for binding. You can use the iodine test on fractions too.

Dear colleagues I have not checked the recent answers to my problem on starch binding for more days and now I see that you give me several possibilities to try. Many thanks for that and I will consider them and try the most suitable for my protein. If I will meet some problems than I will ask you for more explanation. Have a nice day. 

Dear Stanley Ng,

your suggestion is very interesting for me but I have no experience with starch types. I have looked online in Sigma-Aldrich catalog and found "soluble starch" with MW 342.3 Is this what you suggested or please specify what you mean with Sigma ACS MW 325 starch? My protein is over 50kDa.

The one you have found is fine. You are looking for a MW less than 2.5 kD, which is the exclusion limit of the columns. So your protein will pass through and come out in the void volume and the starch will be in the retained buffer.

I would sub-load, i.e. use less sample volume, than they suggest to maximise the sample separation (see attached picture for clarity). Collecting fractions of 0.1 ml.

Analyse a portion of it, say 10 ul

You can probably use a Nanodrop in scan mode to pick up your  A280 and KI/I soln response at 610nm. I'm sure there are lots of starch KI assays out there that you can try to optimise from.

https://www.gelifesciences.com/gehcls_images/GELS/Related%20Content/Files/1314774443672/litdoc28922529AB_20110831095410.pdf

Dear Stanley Ng, thank you for your advice. I was using PD10 columns for desalting my protein can I use it also for this purpose or you recommend me to change for  PD G-25? But what suprises me more: such "soluble starch" is in principle a disaccharid unit (otherwise it would probably not be soluble) and this is sufficient to verify the binding on a protein domain? 

PD10 will be  fine, it should be packed with Sephadex G-15 which has an exclusion limit of 5kD.

This experiment will answer your question regarding the type of starch recognised.

If your protein only recognises insoluble starch then that will create very different challenges on how you will be able to determine binding or not.

I would expect even in insoluble starch for their to be carbohydrate branches that do enter the solution.

OK I will try with that what we have here (I have even found a soluble starch in our department) and aftewards I will evaluate the result... Have a nice day

Dear Stanley Ng, I still Need your advice. I am collecting everything I Need for the test of binding of soluble starch on my Protein. I searched for Somogyi starch-iodine colour method that was published in various modifications in 1970´s so the full text is hardly available for me. Anyway, the colorimteric detection at 610nm is for the starch-iodine complex. I have found KI in our store but as I have understood I need also iodine (I2) to form I3- that slides into starch coil to give a blue-black color. But I do not have iodine here, maybe it is problematic as it is irritable? But otherwise there are also some commercial kits like that from Abcam ab83393 where the color detection is at 570nm because of oxidized Glucose (from starch) .

Please give me advice: shall I use KI/I2 according to Somogyi or can I use the above mentioned kit. To summarize: I try to mix my protein with soluble starch, incubate for some time, then I separate my protein from unbound soluble starch by PD-10 column and in the protein fraction I will detect the starch bound to my protein? 

Please help    

If you can't get iodine the "normal" way, buy some povidone-iodine ointment from the next drugstore.

Wolfgang thanks, yes this is a possibility to get a tincture of iodine from a drug store but I wait also for the answer of Stanley Ng...

Both iodine and potassium iodide are cheap and someone should have some kicking around. You may need to ask some chemists/biochemists nicely for some.

I would say that it's no more harmful than many chemicals used normally in the laboratory and any risk is small compared to handling conc. acids.

Your summary covers everything.

How much protein do you have to play with? This will determine your theoretical concentration of starch.

As for the abcam kit if you have it available you can try, but i would add some controls.

So run a column with protein only for a no substrate control.

Without protein and starch

With starch

This will allow you to establish what is going on if you get some cross reaction from something.

You will need to start at the right point in the procedure given as you are starting with soluble starch.

Dear Stanley many thanks so I will just get some iodine and then I will start.

I hope you will just try the starch assay before trying anything with valuable protein!

Dear Stanley Ng et al., thank you so much for your advices. The modified method of Somogyi works at my lab-bench and I can recommend it for routine uses. I do not need any commercial assay, the (cheap) method with KI/I2 works perfectly (I got I2 from my colleagues here, did not need to visit drug store for this purpose :-) ). Now I even can quantify the amount of bound soluble starch as I constructed a calibration curve of 2nd order at 610.0 nm. With a PD-10 column I can separate the protein from unbound starch and afterwards check the fractions. So thanks and I wish you success in your experiments ...

Excellent, good luck with your research!

Dear Stanley Ng,

how are you? I am writing to you after a longer time. I have presented my results in a seminar in our department today. There were some questions about the specificity of the KI/I2 starch binding method, mainly if it specifically binds only to soluble starch or if it can have some background also with other sugars? I only have an old citation on this photometric method: Jacobsen and Hensten-Pettersen Caries Res. 1970, 4:193-199 (DOI: 10.1159/000259641) but I have no access to the full text (not available on Research Gate although I requested it 2 months ago). Can you please help me with  the specificity of this KI/I2 method for starch quantification? I even have separated my protein from unbound starch with a Superdex200 column to be pretty sure that there is no unbound starch remaining in the fraction of protein that I use for KI/I3 method. Can this reagent also bind on glycosyl moiety of N-glycosylated Proteins? This is also an important aspect for me at the moment. Please help...

Your recombinant protein is expressed in what system, e.coli or mammalian?

I express my protein in the yeast Pichia pastoris and it is surely N-glycosylated...

You're test includes a blank and a no starch added control?

If you have then your results will show any background in the no starch control and your final result can be corrected for any background.

However, i would not expect any as the only carbohydrates that would give any response would have to be amylose like. However, n-glycosylation is linear/branched and the mechanism of action for iodine to work is for the tri-iodide ions  to interact with the helical amylose, hence why the test is specific for starch/amylose and no other carbohydrates respond (because only starch and amylose are helical).

Dear Stanley thanks for the explanation. Yes, I have included controls. I also think that the N-glycosylation cannot contribute to the specific tri-iodide reaction with starch amylose but those were the questions of colleagues that are also not familiar with this assay. Wish you a nice evening...

If you get a chance, I would appreciate it if you hit the recommend on the answers which you found  useful. Best of luck in your research.