Webinar Highlights: Advanced Techniques in Peptide Isolation

By October 10, 2016

Hello, I’m Andy Aubin, Principal Consulting Chemist at Waters.  Jo-Ann Jablonski, Principal Scientist at Waters, and I recently conducted a webinar that addressed some techniques to make peptide purifications more efficient.

We discussed three techniques:

  1. Focusing the gradient
  2. At-column dilution
  3. Temperature control

These techniques are easy to implement, and are useful for improving the isolation of synthetic peptides at the lab scale, typically from a few milligrams to a few hundred milligrams. Tight timelines in research and development groups require processes that work well and accomplish the goal.  In this case, getting pure peptide in the fraction tube, quickly, effectively, and confidently was the goal.  The techniques we introduced are the ones we use all the time in our lab, as well as promote to other scientists who need pure peptide for their experimental studies.

If you missed the webinar or you would like to revisit any of the subject matter, you can view the slides or watch the webinar on demand.

We had many great questions! Here are some of the highlights from our Q&A:

How we can purify the negatively-charged peptides that contain cysteines considering that we are using mobile phase with ammonia?

There are two approaches that may be pursued. First, the purification and sample handling can be performed at pH 2.0. This will keep most of the material in the reduced state.

Second, the purification can be done before the removal of the protecting groups. In that way, the exposure of the free sulfhydryls to oxygen will be minimized.

What Waters columns are best for separation of large peptides (>20,000) according to molecular size?

For analytical scale size separations, samples less than a few hundred µg in volumes less than 40-50 µL, a 4.6 x 150 mm column packed with 1.7µm particles will be best. A 300mm column length can be used for better resolution. This material is the ACQUITY UPLC Protein BEH SEC family of columns available in 125 Å, 200Å, and 450Å pore sizes. For peptides greater than 20,000 molecular weight, I would suggest the 125Å material up to 40,000 molecular weight and 200Å up to about 200,000molecular weight.

It should be noted that these packing materials have a larger linear molecular weight range, but I suggest using the lower half of that range for best resolution. If larger samples are required, the XBridge Protein BEH SEC family of columns can be chosen. These provide 3.5µm particles in diameters to 7.8mm for about a three-fold increase in sample load. They are available in the same three pore sizes, and the selectivity is the same. The peaks will be somewhat wider.

For all SEC experiments with peptides, these molecules will usually behave as though they are somewhat larger because their three-dimensional structure is not as tightly well-ordered as the proteins used to establish the molecular weight ranges.

What effect does cyclisation of a linear peptide have on separation?

We don’t have as much experience with cyclized peptides. They are usually more hydrophobic than the corresponding linear peptides. And they have a smaller radius or apparent molecular weight in size exclusion.

If we have epimers, is it possible to separate them by increasing the temperature in the column or the temperature of the solvents?

We don’t have as much experience with epimers. There are, however, many examples of the sensitivity of three-dimensional configuration to temperature. I expect that elevated temperature will usually be helpful for these separations.

What if cysteines form disulfides in alkaline condition?

Cysteine monomers or side chains will tend to form disulfides in the presence of oxygen at modestly high pH. We must also remember that the thiol side chain is a very reactive reducing agent. It will react with any reducible functionality so you may observe more species than the homodimer.

Can you give an example of segmented gradient?

Please see the examples below.


Linear Gradient


Segmented Gradient


Focused Gradient

What can be the reason for low yield purification?

There are many reasons one might suffer from a low yield when performing purifications. Some of the more common ones are: Improper fraction collection settings (wrong threshold level, wrong collection wavelength, wrong delay time).

Mechanical Issues – (poorly performing injector, pump, or fraction collector).

Sample issues (poor solubility, sample degradation, peptides that elute as multiple resolved peaks because of relatively stable conformational isomers where only one is collected, non-specific binding to the column or some component of the fluidics).

User issues (injecting from wrong vial. re-analyzing wrong collection tube, mis-calculating theoretical yield, erroneous estimates of the amount of mass in the starting material, failed synthesis, a gradient that is so steep that the peptide can precipitate in high organic before completely eluting).

Using the at column dilution loading technique, is there still a risk of precipitation of the crude in the tee when the high organic sample meets the high aqueous eluent?

There is still a risk, but it is significantly reduced. The mixing tee is placed as close to the inlet of the column as possible. In this way, the time between mixing and entry to the column is minimized. In Waters systems, this time is much less than one second. Any precipitation or aggregation that occurs, therefore, forms as a thin film on the large surface of the packing material, reducing the chance for plugging the column and helping to ensure that that material re-dissolves easily for elution from the column.

At-Column Dilution Application Notes 71500078010rA

What is the largest scale purification you have conducted using temp control purification?

The work done here was with a 19 X 150 mm column.

Effective Use of Temperature Control in Compound Isolation 

Does the temperature variation for a repeated use affect the column life?

No, modern columns are designed to work at elevated temperatures. Check the care and use guide for temperature operating limits.

Are there any downsides to a 100% DMF injection instead of DMSO?

They are equally problematic.

How did you maintain the temperature of the water bath during the duration of the chromatographic run? It looks like it is just in a bucket of water.

It was a heated water bath set to 60 C.

What would you recommend as a gradient slope for an analytical scout run for a crude peptide mixture?

For a crude peptide mixture, I typically would use 0-50% or 0-60% at 1.5%/column volume. I often do include a “regeneration” quick gradient to 90% at the end of a scouting run.  Very few peptides require more than 50% acetonitrile to elute.  For materials that hydrophobic, they are unlikely to give good peak shape and resolution without elevating temperature and/or adding propanol.

How much does the counter ions matter for separation like TFA vs Formate in Solvent systems?

Formic acid lowers the pH of the mobile phase, thereby protonating the acidic side chains on the peptide and removing some of the negative charge. The reduced number of charged peptides increases the attraction to the hydrophobic surface of the column, which promotes better separation.  Although some silanols on the surface of the column may also be protonated, free silanols will still be available for binding with the peptide.  These peptide-silanol interactions will appear as tailing peaks in the chromatogram.

Trifluoroacetic acid, the most commonly used modifier, differs in the mechanism of interaction with the column and sample. The low pH protonates acidic peptide side chains, and ion pairing neutralizes basic side chains, inhibiting their binding to free silanols which cause peak tailing.  The reduced charge on the peptide increases its attraction to the hydrophobic column surface.

The TFA ion-pairing reagent also affects the way the peptide interacts with the bonded phase surface. In addition to neutralizing the basic side chains, it increases the apparent hydrophobicity of the peptide.  This is a useful phenomenon because other perfluorinated acids can be substituted for TFA, e.g., perfluoropropionic acid or heptafluorobutyric acid, to give altered selectivity that may improve the purification.

What about the use of isocratic methods for purification?

Isocratic methods can be used for purification. They are very efficient and can provide very high throughput. One must be confident that the isocratic method has eluted all compounds from the column.  If it has not, these late eluting peak may interfere with subsequent separations and collections.

In addition, isocratic separations of peptides can be delicate in the sense that they only give reasonable results over a very narrow range of organic concentration. This is because the slope of k’ versus solvent strength is much steeper for peptides than for small molecules. There is a relatively small window where peptides elute with reasonable k’.

Are you aware of temp control being employed for a GMP purification?

I am not personally aware of this, but there is no reason that temperature control cannot be implemented in a GMP environment.

Can you talk about alkaline mobile phases?

Extremes of pH change the selectivity of separations. Acidic peptides retains better on the reversed phase packing at low pH because it is completely unionized and strongly interacts with the hydrophobic column surface.  At high pH, on the other hand, the peptide moves to an earlier retention time due to a greater number of negative charges, which prevent it from adhering as strongly to the column packing.  Basic peptides exhibit the opposite behavior at the extremes of pH, as expected.  At high pH, these peptides are completely unionized and its strong interaction with the column requires a higher percentage of organic mobile phase to elute it from the column.  The change in selectivity, with improved resolution at pH 10, makes the higher pH more attractive for the end goal of isolating peptide with increased purity. Before using higher pH, consult the care and use guide of your column for the usable pH range of that particular column.

If we want to purify less than 5mg of peptide sample, can it be done by using higher sample volume (about 100ul), collecting fractions from analytical HPLC run and concentrating it?

Yes, this is a perfectly reasonable thing to do. It is also very practical because of the steep slope of k’ versus solvent strength. The peptide can be loaded in quite a large volume at low solvent strength without any compromise to the separation.

What would you suggest for the purification of very polar peptides that elutes with the solvent front on a C18 column?

Hydrophilic-interaction chromatography has been practiced for a long time, but the term HILIC has only recently come into common usage. Analytes are very polar compounds such as polar metabolites, carbohydrates or peptides.

Caution must be used with HILIC to ensure that the peptide does not precipitate in the high acetonitrile concentration used as the initial solvent in HILIC.

You could also consider the following alterative as well:

Ion-exchange chromatography for very polar peptides.

The use of longer chain perfluorinated acids for very polar peptides. PFPA and HFBA will give longer retention than TFA.

Use of an alkyl amine at higher pH to give increased retention of acidic peptides.

Do you have any experience with HILIC mode for peptide purification?

What about the pore size of the stationary phase for peptide purifications.

130Å or 300Å Columns are typically used.

For very hydrophilic peptides, would the at-column dilution necessary, or help increase loading?

For peptides that are dissolved in highly aqueous conditions, using reversed phase separations, the volume loading capacity is virtually unlimited. Capacity is limited by solubility during elution. In these cases ACD would offer little benefit. The peptides, however, may be dissolved in much higher organic, depending on the preceding steps.  Then, of course, ACD would be helpful.  ACD can also be used primarily to adjust the sample pH or to add TFA for increased retention.

Any practical way to purify amyloid?

I did see some example in the literature of amyloids being separated and purified by HPLC.

Can we dissolve the peptide in pure DMSO and inject directly without at column dilution?

Yes but too large of an injection volume will distort your peak shapes.

Why do you use isopropanol instead of methanol? Is it specific to peptide purifications?

The main reason for using isopropanol is the increased solubility of hydrophobic peptides. The longer chain alcohol is a stronger solvent than methanol or ethanol. So it is more effective in both solubilizing peptides, unfolding them, dis-aggregating them, and eluting them.

Although methanol, ethanol, or isopropanol may be considered for fulfilling the role of strong solvent, acetonitrile is usually the solvent of choice because of its low viscosity, slightly higher solvent strength, and its usability in the low wavelength UV detection range (185-205 nm). Acetonitrile usually provides the best peak shape, is easy to evaporate, and reduces the number of side reactions that can proceed during isolation and workup.  Good UV transparency gives high signal to noise for better peak detection.  Some users, however, may object to the use of acetonitrile if they want to use the peptide in biological testing.  Substituting an alcohol, such as ethanol, reduces the residual toxicity that may be associated with acetonitrile.  In addition to concerns about biocompatibility, propanol or acetonitrile:propanol mixtures often increase the solubility of large or hydrophobic peptides and break up aggregates that sometimes form in solution.

If after preparative purification we are getting compound in salt form e.g. ammonium salt and chemical purification to remove ammonium salt is not possible, can we do it using preparative column?

This is not usually possible with a reversed phase column.

What are the ways to clean the column and how we can clean the system from impurities that elute with organic phase?

There are many options for column cleaning. Please consult the care and use guide of your column for specific details. The following is an excerpt from the Waters XBridge Care and Use Guide.

“Flushing with a neat organic solvent, taking care not to precipitate buffers is usually sufficient to remove the contaminant.”

Also, some peptides and other impurities are not soluble in very high concentrations of organic. For that reason, repeated gradients may be more effective for column cleaning.

How to setup the analytical scale purification? Running through the column, UV detector, and then the elute collection device?

Any standard analytical scale HPLC system could be used as long as there is a way to collect fractions.

Injector, column, detector and then to the fraction collector is the typical flow path.

Can the peptide be injected in the prep/analytical column 95% TFA?

I have never done this. I would expect very low retention.  As you go higher in acid concentration in general, the acid begins to act like an organic solvent.  95% TFA is also evaporating very quickly so you would have to take precautions from those effects but there is nothing to prevent a user from injecting a sample dissolved in 95% TFA. I would also be concerned that column lifetime could be reduced.

The collection of the peak of interest is after the UV detection, so how would you solve the issue that there is a delay between the UV detection and the fraction collector?

Fraction collection occurs when the detector sees a peak that meets user defined collection triggering criteria. When the peak is present in the detector, it has not yet traveled to the fraction collector. Therefore, if a fraction collector is triggered too soon, it would be missed by the fraction collector. A system specific, fraction delay time must be determined beforehand in order to time fraction collection correctly. Calibration typically involves injecting a concentrated dye and comparing the time it takes to be seen by the detector, to the time it takes to reach the fraction collector. Fraction delay time calibration is usually a one-time process and input into the software to be re-calculated for any flow rate. Fraction delay time must be re-calibrated when changes are made to the system flow path.

Could you recommend any elute collection device for small scale analytical purification?

The Waters Fraction Manager (WFM-A) is an analytical fraction collector for UPLC systems that minimizes fraction loss and carryover to better manage low volume peaks and allows for efficient collection of small amounts of material for further assays.  Providing high purity and recovery, the WFM-A has been designed to confidently minimize sample loss and reduce carryover of your complex mixtures and precious sample.

Small Scale Peptide and Impurity Isolation Using the ACQUITY UPLC H-Class and Waters Fraction Manager-Analytical Systems 720005500EN

Aren't the methanol and stronger solvent damaging for C18 silica column?

Common organic solvents such as methanol, ethanol, IPA, acetonitrile, (and many others) will not damage a silica based C18 column.

Can you explain a bit more about loading study?

The goal of a loading study is to maximize the amount of material on column and still achieve an acceptable purity for your collection. A loading study is simply a series of injections that increases the mass load on a column. When the peak shape or critical pair resolution has reached an un-acceptable level, you have reached the maximum loading.  Loading studies can be performed at an analytical scale and then be scaled to prep (this is the preferred way) or performed at the preparative level.

How do you know when it is time to replace the column?

When the degradation of chromatography reaches a point where your separation goals cannot be met, it might be time to replace the column.

What is the average lifetime of a preparative column in terms of injections?

Many factors will shorten the life of the column. Of things commonly encountered in purification, I would take precautions about particles and extremes of pH.  The pressure shock associated with injecting samples dissolved in neat DMSO or DMF also reduces column life. An acceptable column lifetime is approximately 1000 injections, but I have seen columns last for multiple thousands of injections without any performance degradation. The only requirement is reasonable column care. Inclusion of a guard column is an easy way to protect (and increase the usable life) of a more expensive column.  An in line particle filter at the head of the column can be helpful even if you have filtered the samples.

How long would you suggest for a column to be cleaned before the next injection in the series to ensure the column is ready for the next injection.

Column cleaning has been discussed in a previous question, however, to ensure a column (and system) are ready for the next injection, both must be fully equilibrated to the starting gradient conditions. They typical re-equilibration recommendation is 3 times the system volume plus 5 times the column volume.

Special Thanks to Tom Wheat, Principal Scientist at Waters, for his input on the answers to these questions!

Watch the webinar

Categories: Technologies