Protein Characterization - Science topic

I think Zeyaul is on the right track. Isn't CD more of a qualitative technique? In other words, you can't tell the amount of beta sheets in your sample only the percentage. So, it might be what you are observing is your relatively less heat stable helices unfolding and only beta sheets remaining.

Good luck!

Hi Yuhong, did you later find out the reasons for the problem? I used the same OPA reagents (Church et al., 1983) as yours but used 400 ul sample + 3 ml OPA reagent for the measurement. I also noticed the absorbance at 340 nm did not increase for the samples with a longer hydrolysis time while SDS-PAGE gel results showed the proteins of these samples had more hydrolysis.

Thanks.

I do have a similar question Though I tried Hitchcock equation with what seems to be a better result, I still had -ve binding constant. Have been able to solve it? Harpreet Singh

Thanks

If you are getting the desired band in an uninduced sample also, then that might be due to leaky expression. As many articles suggest to add 1% glucose. I tried but didn't work in my case, if you want, give it a shot.

The molecular weights of the bands, either on a reducing gel or a non-reducing gel, will be enough information to distinguish the two proteins. On a reducing gel, immunoglobulin G will have bands of 25 and 50 kDa (150 kDa on a non-reducing gel). Insulin will have 2 very small bands that will run near or at the dye front on a reducing gel, depending on the type of gel. The total mass is only 5.8 kDa.

About the Deborah Number ― Referring to the Bible, Prophetess Deborah sang: "The mountains flowed before the Lord", celebrating victory over the Philistines. It is now scientifically established that mountains indeed "flow", considering the geological time scale. There are two different perspectives of time to be generally considered when a transient relaxation phenomena is observed; one that is intrinsic to the phenomena, another that relates to the experimental time scale. Born as a rheological concept, the dimensionless ratio between the 'time of relaxation' and the 'time of observation' was named Deborah number (De). References: M. Reiner, The Deborah Number, Physics Today, January 1964, p. 62; https://physicstoday.scitation.org/doi/10.1063/1.3051374?journalCode=pto The following reference discusses the concept of Deborah number with regards to the glass transition:

Nah, termini are almost always disordered, the rest seems fine. Also, I believe the charge would explain only why it doesn't run as it supposed to (17 kDa), not why it runs differently on 2 different gels. Other reasons why it doesn't run at 17 could be some PTM, type of buffer used, whether protein is in detergent, generally shape of a protein in case it doesn't get totally denatured and so on. Anyway, I've seen different markers giving different masses before so I wouldn't be surprised if that was the case. Unless your lab used both of these markers for a long time and they were always fine.

Fixation with formaldehyde will affect your BCA quantification and you will end up getting skewed values that will affect your SDS-PAGE outcomes.

Negar Saeedi are you sure that Ion Mobility is the technique that you are looking for?

regards,

GB

I think for drawing of protein 3-D structure, softwares such as SPDBV and PyMOL are proper.

So, what is the issue to run in negative mode? The negative mode will give you better signal/noise ratio; therefore, you will get better sensitivity. If you need to monitor other peaks in positive mode, you can make a separate acquisition window for pos and neg mode.

First thing is to check is if anything in the sample buffer is absorbing strongly either in this region or at shorter wavelengths.

If the peak is specific to the protein, a positive peak at 230 nm in protein CD spectra may be an indication of a strong pi/cation interaction, usually between a charged amino acid like lysine an and an aromatic amino acid like tryptophan or a metal and an aromatic amino acid.

Dear Nathaniel Ninyio ,

Circular dichroism is relatively good in comparing (relative) results where for example your peptide is studied at different temperatures, with or without anionic phospholipids and so on. This allows you to say “upon increase of temperature one sees a decrease in helical content”.

See for example:

Unfortunately the deconvolution of CD spectra varies a lot and depends on the method used and some tent to over-estimate the alpha helical content. So an absolute answer based on one method is not THE answer of the analysis of your spectrum. The best you can do is comparing the results of different CD analysis since numerous programs are available (often your CD equipment includes pretty good software as well).:

Dichrocalc:

Dichroweb:

BestSel:

K2D3:

Dicroprot:

There are many more different programs to deconvolute your CD signal in order to extract secondary structure contents (CDPro, CDNN etc.).

For a nice introduction see:

PS. To get some ‘grip’ on what to expect in terms of secondary structural elements you might consider to do some secondary structure prediction, like:

SOPMA

or

JPred

Hope this answers your question somewhat. Good luck!

Micro BCA is more sensitive than standard BCA, but the catch is that micro BCA is inhibited by the most common buffers such as RIPA or TBS, and many more chemicals. It is recommended a 1:10 dilution in ultrapure water before doing measurements with micro BCA. RIPA or TBS do not inhibit standard BCA at all, so I would go for standard BCA unless I do not have much exosome sample, which happens often.

Short peptides, isolated proteins and soft tissues. Using a piece of laboratory towel.

If there is a buffer that has been commonly used by other researchers for this enzyme, you can start with that.

If not, the buffer should be optimized for the particular enzyme you are using. The optimization can include choosing the identity and concentration of buffering compound, the pH, the identity and concentration of salt, the Mg²⁺ concentration (ATPases require Mg²⁺ for activity), the reducing agent identity and concentration if needed, the detergent identity and concentration if needed, and a stabilizing excipient identity and concentration if needed.

Prepare concentrated stock solutions of all the ingredients you plan to test, and prepare buffer solutions by mixing them together and diluting with Milli-Q water to get the desired final concentrations. This gives you the flexibility to try many different combinations.

You also need to consider the substrates (one of which is ATP), and their concentrations.

Hello Shobana, i do not see where you have the difficulty because the classical methods of protein separation and characterization have not changed, at least not to my knowledge. Protein characterization in recent times involves five protocols; protein quantification (involves Absorption at 280 nm, Lowry assay, Bradford assay and Mass Spectrometry - a technique by which you can determine the mass of a protein with remarkable precision. It can detect any important modification (post translational modification) or variation in a protein structure), Electrophoresis (migration of a protein in a gel as a function of size, shape and charge), dynamic light scattering (a technique used to determine the size distribution profile of small particles in solution) and circular dichroism (a variant of absorption spectroscopy which measures the difference in absorption of left and right polarized light (in the ultraviolet) by a medium. For protocols visit the links below: https://www.uq.edu.au/pef/content/characterisation-techniques-0.

Hello all,

do you have any good experience with a pulse sequence to reduce T1 noise on HSQC without any application of gradients? (Bruker systems)

Thx...

I have been looking at the translocon and use 4-20% gradient gels. I use a normal blotting protocol but fix the blot in 5% acetic acid before blocking. This works well for the beta subunit.

For simple protein enzymes like insulin you can use bacterial expression systems to create proteins.

For any protein that requires complex folding, and post translational modifications, you would need eukaryote expression systems eg. CHO, yeast etc, that can express these complex proteins, that require chaperones, correct tRNAs, and post translational modifications that a bacterial system can't provide.

I think that is very important a refined bioinformatic study of hypothetical proteins for planning a very detailed set of wet-laboratory experiments

Nicolas, grab it from here: http://www.kazusa.or.jp/codon/

Hi, 

I am also facing a similar problem. If the problem is not solved yet with you, the reason for lateral spread of protein leading to fusing of bands could be that the Ionic strength of sample lower than that of gel (if other obvious possibilities are obviated) (This is something I gathered from this site: http://www.bio-rad.com/en-us/applications-technologies/performing-protein-electrophoresis) .

Let me know if this works.

The question is whether the measurement of 109 and 100 Da made on separate strips is a statistically significant difference or is within measurement error of being the same mass. One way to address this question is to make the measurements several times, calculate averages and standard deviations, and perform a t-test to see if the difference is significant. Another approach is to combine both antibodies in one blot and see whether you get one band or two, compared with the individual antibodies. This should be done with a light loading of antigen to maximize the chance of seeing two separate bands. Also, you should optimize the separation, if there are two bands, by choosing the most appropriate acrylamide percentage.

I agree with Jaime. Your final concentration of Sypro orange is too high. We normally just use 0.5x - 1x final concentration and this gives us a low baseline signal.

It sounds like you are purifying a recombinant protein in a cell-line that has chaperones to help with folding. If your fusion protein and chaperone are tightly associated, no purification technique will work unless you dissociate the complex (they will co-purify). Try adding ATP (5mM) to your column wash buffer to release the chaperonin:fusion protein complex before eluting your fusion protein from the amylose resin. In the past, I have encountered some recombinant proteins that are not released from chaperones even in the presence of ATP and needed to add denatured E coli lysate plus ATP to break apart the complex.

I think this is more bigger to consider a shoulder. but I'm not sure. i  will send you .Cary Eclipse fluorimeter.

You can check the attached paper.  You have to optimize ATP and MgCl2 concentration.

You can model your proteins (WT and truncated proteins) using servers like ITASSR, HHPRED (MODELLER) etc. and then you may get some clues. Zymography is an assay in solid phase but spectroscopic assay is in liquid phase (could be different mobility of catalytic loops!!) 

ddutta@ualberta.ca is my id.

I will be more then happy to help.

Hello Heide,

Thanks.

No I don't have ant preference for GST tagged protein. I just wanted to gain the information. Yes, now I have got the idea about cleavable GST.

Citrate is not only a good buffer in the slightly acidic range (pH 2-6) and a chelating agent for bivalent metals, it is also a cosmotrope (right side of the Hofmeister series of anions) that stabilises protein structure.

These articles may be helpful:

Hi Riddhi,

The responses above are elaborating on how to indentify the interactions of this gene. To identify the function, you gotta knock-out/off this gene and do a mutant analysis.

Thanks,

Aniket

Hi Avni, it could be that your  protein is dimerizing in the gel because mercaptoethanol does not travel into the gel along with the proteins because it is uncharged. The same goes for DTT.

If this is the case, then treating the sample with iodoacetamide (~5x concentration over the 2-ME) should explain if this is the problem.

Iodoacetamide will bind all free -SH groups and prevent any S-S linkage.  

many thanks james.....your suggestions and explination is really helpfull...thanks again

Hi Deepika 

Your protein is 12 KDa with no-disulfide bonds.

Arginine avoid dis-aggregating in refolding process but concentration depends on our hydrophobic amino acid.  The 0.02M-0.3M Arginine is preferred for more native structure of protein (depend on refolding protein concentration).

Before that you have to denature the protein with correct concentration of Urea/GuHcl and BME/DTT.

If i  know more about you protein, suggest better way. 

Hi Waheed, it all depends on the antibody. Some monoclonals are rock solid and some monoclonals loose activity when diluted into water.

If you have a reliable ELISA assay for your antibody then you should use it to determine any loss of activity.  

Aromatic side chains do not actually have an intrinsic CD signal, because they do not possess any center of asimmetry. However, they may show a CD spectrum if they are embedded in an asimmetric environment, such as that typical of the tertiary structure of the peptide backbone. Thus, it seems that your peptide has some secondary, but no tertiary structure, into which it could reorganize as a consequence of conjugation to dermatan sulfate. I mean that the entire complex peptide-dermatan sulfate acquires a tertiary structural arrangement, to which the 235 signal could be ascribed. Take a look at the near UV CD spectrum up to 300 nm, which should not be flat, but show novel absorption bands, even if they are less intense than the far UV ones.

Hi Daniel,

your crystallization solution might actually be pretty close to being cryoprotective. Both glycerol and MDP are cryoprotectans. You can test this by freezing a drop of the solution (without a crystal) in a loop and check if it remains clear, which suggests its cryoprotected or whether it's turns hazy. If it turns hazy, I would make a solution as closely matching your mother liquor as possible and slowly increase glycerol or MDP (usually by 5% steps at a time). Glycerol usually is croprotective around 15% but with MPD in your solution you might need less. Generally it's hard to predict what works best since it is different for different crystal systems, loop size used and procedures. A good rule of thumb is to change the condition as little and as slowly (stepwise) as possible.

Good luck.

Bernhard

I can recommend a new fluorescent method very sensitive that can help you, here is the information

Good Luck

Hi Arnab

            I will suggest you to run SDS page in non reducing and denaturing condition i.e in the loading dye do not put any reducing agent. If you protein forms disulphide linkage, you can clearly see a band near dimer of your protein of interest.  I have done this and was able to see clearly a dimer. The control for it will be a protein sample denatured with normal SDS PAGE dye i.e reducing and denaturing dye. Hope this will help you. 

Thank you so much

Not enough information provided to answer your question.

Mass spec's are NOT universal detectors. They do not and can not "see" everything.

For an LC/MS system to detect a compound, it must ionize well under the conditions used. This relies on many things, including, the concentration of the sample present, the experience and skill of the operator (This has the largest effect of all!), the type of MS system used, the mode the system is operated in (pos/neg) and of course, the many settings and parameters used to detect it (all of which can result in no signal, noise, signal or just about anything else).

For now, perhaps forget about the MS aspect and ask yourself if you can detect the sample using a good quality HPLC method with UV detection? Alternatively, if enough sample is available, try flow injection with the MS set up in a wide scan (look at the TIC) to see if anything is there and then a narrow SIM mode to look for fragments or clusters of the protein at higher energies.

*Most importantly of all, get some help locally with this (not the web) as a student would not be expected to do this on their own. It takes many years to learn how to analyze samples by HPLC and/or LC/MS and your focus should be on having someone help make sure you are pursuing a correct approach to solve the problem. Let them help you insure the instrument(s) and method(s) used are applicable to your sample and problem.

I agree with Omar and Bobbi, you are loosing your protein in the dialysis procedure. If you are trying to concentrate a protein solution ~below 10 ng/ml, you will loose most of it. Dialysis bags have a huge surface area that can bind proteins.  

If you can, lyophilize the solution. If you can't do that then use centricon fiters.   

Unless there is something unusual about your protein, you can use standard SDS-PAGE methodology. To determine molecular weight, using a single acrylamide concentration gel (not a gradient gel), and a set of commercial molecular weight markers run in the lane next to your protein. Plot a graph of Rf versus log molecular weight of the standards, and use the Rf of your protein and the standard curve to measure the molecular weight of the monomer. (Rf is the ratio of the distance from the top of the separating gel to the band divided by the distance from the top of the separating gel to the dye front.).

If you are interested in knowing the molecular weight of the oligomer, then you should not use SDS-PAGE. There is a procedure to determine the molecular weight by non-denaturing PAGE, which involves running the protein on gels with varying acrylamide concentrations. You could also use analytical ultracentrifugation, sucrose gradient ultracentrifugation with size standards, and size-exclusion chromatography coupled with multiangle light scattering, all of which require specialized equipment.

How far did you go in terms of urea and GdmCl concentrations? Are you sure that both transitions are completed? Does your protein unfold through intermediate state? For many globular proteins, the use of GdmCl allows decoupling native state <--> intermediate (molten globule) state and intermediate state <--> unfolded state transitions, whereas in urea these transitions typically overlap. Also, GdmCl-induced curves are typically noticeably steeper than urea-induced unfolding curves. To verify complete unfolding mechanism of your protein you should use more than one technique. For example, near- and far-UV CD are typically used to detect changes in tertiary and secondary structures, respectively.  

Hi Naga,

I am curious. How do you plan to deal with the interaction between the two proteins? In my study of association between an ordered protein and a disordered one, I used coarse-grained protein models and I did change the force parameters for each protein separately to match with the experimental measurement of the isolated proteins.

I couldn't find it either, but looking at the structure of quinpirole, I would not expect it to have any absorbance in the visible range. You can see this for yourself by the fact that it is a white powder. A solution of it should be colorless. Therefore it will have no absorbance at 540 nm, which is in the visible range.

I agree with the previous suggestions. Three main things could be occurring, perhaps more than one of these at the same time:

1. The peptides binding site was lost by cleavage

2. The peptides are self-aggregating

3. The peptides are not correctly folded and the binding site is hidden

Since you believe there is aggregation going on, I would suggest adding a small amount of a non-ionic detergent to prevent aggregation. After aggregation you would need a stronger chaotropic agent to disrupt the aggregates. Good luck.

It could be, it depends on the number of lysine residues on the surface of your proteins. I prefer to run reactions with glutaraldehyde and proteins at 4C for 2 h. At 37 C the reaction rate will be fast and since you cannot vigorously stir your protein sample you need to give glutaraldehyde some time to generate homogeneous reaction conditions. Add a reduction step using sodium borohydride. 

Right, I do understand that. Thank you for your comment.

But, through flow cytometer analysis, cell population that have bound antibody between control cells and the cell pulsed with peptide of interest are almost identical.

To elucidate, the question I have is should the peptide at least stabilize the MHC with small amount that we see a difference between control and peptide pulsed cells.

 

Has anyone else done T2 stabilization with peptide pulsing? A little insight would be really helpful.

Hi Ben, thank you.

Anyway, I am trying to monitor aggregation using ThT and I am having outrageous results. Final concentrations and conditions:

- Abeta 1-42 (pretreated with NH4OH and resuspended in NaOH (no sanitation, no vortexing): 56 µM

- ThT: 0.2 mM (in PBS)

- PBS 10 mM till 100 µL

- Greiner black flat bottom plates

Everything at 4ºC till plating and start measures (440 nm exc and 484 nm emission) each 5 min, 15 sec shaking (6 mm amplitude) before measurement, 37ºC. Equipment (tecan infinite 200) gain setting at 140.

Fluorescence values don´t increase but even go down 5-10 min from assay start and stabilize. I repeated with 3 different Abeta aliquots and I see the same. Maybe my aliquots are a bit old... but in ELISA still worked fine.

Any suggestion? Thanks in advance!

Hi

Overdose heparin can compete with your protein and protein/enzyme can be titrated out. So its better to use synthetic poly (dl:dc) as suggested by Bhaskar Gouda.

Best of Luck

(1) Check your amino acid content for side-chains which absorb at that wavelength: tryptophan (W), tyrosine (Y) and cysteine (C) (or use something like http://web.expasy.org/protparam/ which will give you a predicted extinction coefficient)

(2) Make sure to remove the imidazole from your sample either via a buffer exchange column or dialysis. High concentrations of imidazole will skew any data as the overall absorbance will be too high and out of the range of detection.

(3) Test your nano-drop with a different protein, nano-drops can break down or might need a good clean as the sensor area is very small.

(4) If you have access to a spectrophotometer, try using it to detect your protein instead, as these instruments tend to be much more sensitive.

Thank you very much for you help Betty,

Is it possible that I could use a decellularized section of human skin, which would contain the collagen matrix as a foundation for developing my artificial skin construct?

Great question, its a typical problem for proteins with N-terminal and C-terminal processing, and considering the function of a protein it is relevant to all proteins that are modified by tagging with epitopes (FLAG, HA, c-myc...) or fluorescent proteins (GFP, YFP, RFP....). For a GPI-anchored protein I would seriously consider producing a recombinant secreted derivative in a eukaryotic expression system and raising specific antibodies that you can use to immunoprecipitate the native protein. However, I understand that this takes time and often people try epitope tagging first.

1) One approach would be to add the tag between the signal peptide and the mature portion of the protein, possibly with a glycine-serine linker after the epitope tag. I would also insert the first 3 aminoacids of the native portion of protein before the epitope, in order to avoid any changes to the signal peptide processing site. Also eukaryotic signal peptides are less sensitive towards the type of amino acids immediately after the cleavage site compared to prokaryotic signal peptides, it can happen that the epitope sequence is causing problems with signal peptide processing at the ER translocation pore. If the signal peptide is not cleaved, you end up with a membrane anchored protein that will probably not fold correctly. I would also try different tags to make sure you find one that works.

2) Alternatively, you could add the epitope tag before the processing site of the C-terminal transmembrane domain (the hydrophobic portion that is cleaved before the GPI-anchor takes over the function of membrane anchoring). But again, you have to make sure you don't upset the processing site, so I would insert it a few aminoacids further upstream and again I would try different epitope tags in parallel to find one that keeps the protein working normally.

In an case it is a tricky task and you have to approach it with trial and error. You are absolutely right to be worried about the removal of the epitopes if you simply insert at the N-terminus (before the signal peptide) or the C-terminus (after the hydrophobic transembrane domain, it is not going to work, don't even think of trying that. If you detect a protein, then it has not been processed properly and it is probably retained in the ER as a malfolded protein. If it is processed correctly, the tag is removed and you can't detect it. In all likelihood, both will happen, so you may see functionality in some bio-assay, and you may detect something with your anti epitope antibodies, but both methods report on different molecules so the results will be misleading. 

Crude protein or stock feeds (as mentioned by Dr.Milham) may contain aminoacids,amides or non- protein -N.So a general factor of 6.25 (assuming 16% N in the protein) is used to convert estimated N to protein.However if one knows the amino acid composition of protein and its N concentration accordingly factor will vary (as per the example for arginine given by Chandini).So one can derive and use a factor if one knows the amino acid composition and N concentration in individual aminoacids.The paper attached by Dr.Weller also supports this view.

Hi Prakash,

The cells are L2 mouse fibroblast cells. I have also tried supplementing my glucose free media with 1mM pyruvate and the cells also die. 

Jessica

I agree with  Martinelli and just want to add that you should try the top two programs for deciphering percent of alpha and beta structure (one  of which is mentioned above- Dichhroweb). Remember that the curves are a combination of the percentages of different structures and thus only after doing this analysis you can get the whole story.

Gustavo

Hi Martin, neutravidin can bind 4 biotins and if your protein has more than 1 biotin you are seeing crosslinking between the biotinylated protein and the neutravidin resin.

Thanks for the answer and protocol Omar. It will come in very useful!

I think you will find this paper useful:

My takeaway from it is that for larger proteins, the value of dn/dc falls within quite a narrow range, and depends on the amino acid composition rather than the protein shape.

According to this article, proteinase K contains 5 Cys residues and 2 disulfide bonds, so it is plausible that reduction of the protein could affect its activity.

Does the peptide have at least 3 primary amino groups (N-terminus and Lysine) to account for the 3:1 stoichiometry? If so, that is probably what you made.

I have never heard of self-reaction of the reagent being a problem. The dye has no amino groups to react with NHS ester.

By the way, you may find that having 3 FAMs/peptide is too many. The peptide may become very hydrophobic, and the dye molecules may quench each other's fluorescence.

Try to give a bigger value for omit command

Michael is right, RT data (from biopsy material?) not always reflect the changes in protein levels, especially if you are working with serum of patients. It depends on the production rate/secretion rate of your protein of interest (is it actively secreted or a sign of cell damage like the transaminases?), the stability in serum and the excretion/degradation.

Hi Snehasish! If you are certain that the peptides you see are of genuine interest (i.e. you repeated the experiment and always the same types of masses appear) than it is certainly worthwhile to try different pre-Maldi workups - The quality of MALDI-TOF spectra is strongly influenced by the purity of the sample (i.e. salts, contaminants which prevent proper co-crystallisation of the matrix etc..) and therefore pre-treatment of the samples can enhance the sensitivity significantly. Luckily for peptides many tools such as C18 ZipTips or mini SPE cardriges exist to remove contaminants. A 5 minute sample workup could increase the signal intensity sufficiently that you are able to obtain neat and meaningfull fragmentation spectra. Good luck!

I think some metalloproteins stain negative with silver stain which may be less visible than a positive stained control

Something must have changed without noticing it. We are dealing with coordination bonds which can only be perturbed by a low pH or the presence of a competing molecule. I would prepare new solutions with new reagents. 

I see. Thank you!

Hi,

maybe this publication could help:

Btw, you could also "stain" standard SDS-PAGE gels by soaking in TCE after running the gel, but then there is not much of an advantage left compared to other staining methods. The method also works with Trichlor-acetate (only after the run, because TCA can't be added to the gel directly before the run due to its charge).

I guess the imaging system has been optimized in terms of wavelength, exposure time and image processing so if you can afford it, you can save some time. Or spend some time at a regular imager and optimize yourself.

Hope that helps.

Thanks a lot i will follow all this

Dear Betul:

After studying your spectra - my advise to you: better change the method. In my view FTIR is incapale to solve your problem.

Kind Regards,                              Leonid

Thank you, I have not tried earlier. I will consider about that;

Dear Keith. Thanks for the detailed description. I will see these links and will use it for my studies

It is sometimes observed that ligand binding destabilizes proteins. One explanation is that the ligand preferentially binds to a non-native state of the protein. Here are 2 papers on the subject.

Hi Carlos,

I have got good experience with the CDPro program suite (http://sites.bmb.colostate.edu/sreeram/CDPro/), which is also on Kyle's list. It only runs on windows computers, but it has the advantage that no registration is needed.

One has to pay attention to convert the CD data to the correct units (per residue molar absorbance units (De M-1 cm-1)). Machine units in mdeg will give wrong results.

It might be worth to extend you CD measurements down to 185 or 180 nm, if possible, as this will increase the accuracy of the results significantly.

Regards, Andreas

Hi Heena,

I don't think peptide confirmation will influence its pI (maybe aggregation could..), but I do think that IEF can be stability-indicating: if you're peptide contains amino acids that can be modified upon storage (in formulation), such as asparagine/glutamine that can be deamidated, methionine/cysteine that can be oxidized, these modifications can alter the pI of the peptide.

In addition to the ones mentioned above, lysine glycation, arginine deimidation and/or N/C-terminal reactions, such as pyroglutamate formation can also influence the pI of a peptide (or protein). 

Hope this helps!

Thanks Manuele Martinelli and Farid Tejeda-Dominguez ! I will try as you said. Maybe my protein concentration is way too high in the previous experiments, I'll do some dilute this time. Your advices really helped me a lot.  Btw, the reason why we choosed to do that binding test using Co-IP was that we previously used the protein1-conjuagated column to pull-down that protein2 from insect tissue homogenetes and sent for a LC/MS-MS assay, so that we decided to verify their interaction under a more natural condition (in the solution),which we thinkwould be Co-IP.