Method transfer - lipophilic, apolar analyte; RP18; MeOH ...

[bisnettrj2]

1. A published method is known to work to separate lipids in a relatively uniform matrix (of similar apolarities) – RP18 and methanol:water 90:10. They did not specify the brand nor particular packing material of RP18 column. How do we modify and optimize this method – for resolution/selectivity and run-time? (The base peak of the analyte, a lipid, was > 45 minutes.) Why does one poster say that the lipophilicity (hydophobicity) of the column, compared with that of the analyte, is irrelevant?

Buy any C18 to test out the method using the parameters provided in the paper you read. The particular features of a column are perhaps not irrelevant, but they may be secondary to your needs for the analysis. All C18s are similar, in that they all have the same bound ligand. They can, however, differ in a lot of ways, and Uwe Neue is the guy to ask about columns (search the forum for posts by him in regard to C18s, you should find many - or read his book). You may not need to retain a purified analyte as long as 45 minutes - hence you might use a C8, a C4, or a phenyl or cyano column in your analysis. To modify the method you found, the simplest way is to adjust the retention time of the observed peak of interest with a combination of temperature and proportion of organic solvent:water. For multiple analytes, typically a gradient of low organic:water to high organic:water is run to elute each analyte separately, if an isocratic condition cannot be found that separates the peaks effectively.

1. Buy my book.
2. Buy my column.
3. Hire my lab.
4. Just do what the “expertsâ€

[ksharp]

Hi Ben,

I understand that you just want some answers to your very specific questions. I have a little time to spare so I'll give it my best.

How do we modify and optimize this method – for resolution/selectivity and run-time?

I believe that all the other method information was provided, so start by buying a column. If you are hung up on absolute hydrophobicity then get a column with a high carbon load. The Kromasil C18 has a high absolute hydrophobicity based on NIST SRM 870.

Believe the good people in this forum when they tell you that it doesn't really matter. If the exact C18 column was so important the authors probably would have specified it. The optimization step is where you independently probe variables like column compartment temperature, %organic, pH etc. to measure their impact on retention and selectivity. In general you're trying to get the best separation in the shortest amount of time (which of course = money). If the paper says do the analysis at one temperature, see what happens if you raise it 10°C.

The fact is, one column might give you the best results at 90% methanol, and another might give you equivalent results at 92% methanol. You won't know until you buy your column and start using it. You might get the best results with an acetonitrile gradient at 25°C. Every C18 column will probably give you slightly different results, but they're not going to be so radically different that you can't tweak one of your variables to achieve the desired separation.

What do you think of the Mac-Mod review of C18 columns?

It's both thorough and interesting, and could be a useful shopping companion in certain situations.

Does it make sense to choose a column, based on this review and the affinity of the analyte for the column?

When choosing a column, rather than scope out the most hydrophobic column on the market, I would tend to identify a specific brand that has proven reliable and consistent and then have them send me a column. Then I work out the details.

But if you are dead-set on one of the 3 low-polarity, high-hydrophobicity columns as certain as you appear to be, that's a unique approach but I don't think it will particularly hurt. You will have a column and that's the important thing. The details, like mobile phase composition and temperature, will be worked out during optimization experiments. It's doubtful that the column you choose won't work. However, it's also doubtful that the column you didn't choose won't work.

If not, why not? If this review is nonsense, biased, incomplete, or faulty, in your opinion, then please tell me why and how.

The review is fine and the data is conceivably useful, but like I said, for most of us it is not something we would normally consult when selecting a column. The selectivity and retention depends on other variables than column characteristics.

Similarly, does it make sense to choose the solvent based on this affinity also? If not, why not?

Keep it simple, Ben. Start with one thing: your column. Then figure out variables like mobile phase composition. The right chromatographic conditions aren't something that you choose in advance. The right conditions are the ones that give you separation. That is something you work out during development and optimization. It might be 90:10 MeOH/H2O, or it might be 90:10 MeOH:Hex. You won't know until you set up your instrument and start running samples.

How do we pre-validate the method..."

If the accredited lab is doing the validation, then they probably have templates that they work from. You provide them with a method that works and they perform the validation protocol. If you want to perform a partial validation for your own confidence, look at parameters like interference from the matrix, linearity of analyte response, analyte recovery, any sample extraction procedures, repeatability, limits of detection. It's all useful data that can help the accredited lab out, but you have to develop and optimize your method before you validate it.

As a final note, I know you've got specific questions and want specific answers, and I sense some frustration. But the experts who frequent this forum have enough experience to tell when people have got things muddled. You've focused on a very specific list of candidate columns (not very important) and you're making mobile phase decisions way too early.

There's thousands of different ways to get from New York to LA, and in the same way there's a thousands ways to effect a separation, it's all how you balance the variables. This is the equivalent of choosing to drive exclusively on local roads making nothing but left turns, but also making the trip in a cement mixer.

KS

[mbicking]

If anyone still cares, I can give additional thoughts on the Mac-Mod Guide to C18 columns. It is just that, as mentioned earlier; a comparison of some C18 columns. I use it often as an example of how to get started with column selection.

The main conclusion is that not all C18 columns are created equal. They come in many flavors, and you can group them into nine similar categories, depending on they behave toward some specific test probes. However, these differences are most apparent when separating compounds that are similar to the test probes - namely, more polar compounds. If a column in one category doesn't work, for whatever reason, others in that same category will probably not work either. You should then choose a column from a different category.

However, when you are separating less polar compounds (lots of carbons), some of these differences between columns are reduced, and their behavior becomes more alike. That does not mean identical, but more similar. This is the origin of the "any C18 will do" remark. It is a bit of a simplification, but the point was that the "flavor" of the column is not as important for your particular separation problem (from what we know of it).

There are some methods that use only organic solvents with a C18 column (no water). It historically has been known as non-aqueous reversed phase (NARP) and was used for very hydrophobic molecules (retinols, for example). It's not as common now because we have more choices than C18 for stationary phase, and the very non-polar mobile phase is no longer needed.

The other important point is that LC is primarily an empirical science. You have been given lots of "advice" here from "experts" with the understanding that we can only get you started; you have to figure out the final details. If you have a method reference, use it as a starting point, with whatever columns you have, and see if it works. We can't give you all the answers up front because we don't know them; we need more information and data - that is, results from your first experiments.

[benrock]

Sincere thanks for your clear, specific answers to my questions, particularly about the Mac-Mod C18 column review, column selection, solvent, and validation of a method, including your constructive criticism.

The principal issue is that you don't even have a characterised extract, but you want to freely mine other peoples' IP as though their time has no value, and also ignore their advice. Your research appears to be minimal, your assumptions about participants abusive and woeful. You attempt to justify your behaviour by making a virtue of willful ignorance... My tolerance is somewhat less than others, because you have consistently failed to respect sensible advice.

In fact, we do have a characterized extract, so you do not know what you are talking about. I am looking for a lab, probably two, to hire, so this discussion is about the specifications to the lab on this forum. My research goes back almost two years into the molecules, so again you do not know what you are talking about. The LC details are just one piece of a puzzle, with many clues (starting points for the analysis) already published, that include UV, FTIR, MS/2, NMR, SFC, TLC, and SFE techniques. My tolerance is minimal, if you appear not to know what you are talking about and still strut around as if you were a guru or an imam. I simply am asking basic questions to understand the analysis and to expedite the analysis by contract laboratories based on knowledge of the analyte (specific lipid) and the matrix (relatively uniform apolarities). Again, and forgive me for repeating myself, but I am also legally liable for the results, since even the accredited lab for the validation will be a subcontractor in the IND/IMPD.

As some of you point out, the Mac-Mod review is not complete and only applies to C18 columns, but seems an objective start for our application to optimize and validate a published method based on known characteristics of the analyte, a specific lipid, and affinity with the column.

I am interested in hiring an independent or university lab to repeat and optimize the published method, and also later an accredited lab to validate the repeated, optimized method. I have no preconceived ideas. I believe I have an open mind about this, despite or perhaps because I have no training nor practical experience with LC analysis. I prefer an evaluation of columns that is independent from the instrument and column suppliers. Also, I simply would like to understand enough to specify to the lab how the method is to be developed (repeated) and what is to be varied if it fails, such as the solvent mix, the column & auto-sampler temperature, and the flow. I also believe I am keeping an open mind about this, as shown by my tolerance of the personal insults in this thread after I admitted in the first posting that I just have book learning from reading the published papers, attending conferences, meeting chemists, and asking questions, also with a sceptical frame of mind, even of the so-called experts. Naturally, I am interested in doing this as quickly as possible and at the lowest possible cost. For me, the principal issue is becoming: do you have knowledge and experience with lipids & natural products? Your advice is then much more welcome and your insults and nonsense more tolerable.

If the stationary phase is too hydrophobic for the analyte i.e. you have a hard time getting it out of the column, then the most obvious action is to decrease the hydrophobicity of the stationary phase. I don’t know why you obviously have fallen in love with hexane, but if you meant you interest in green technology seriously, I don’t think hexane is the right solvent for you – solely on that ground.

If the stationary phase is too hydrophobic for the analyte, isn't it also logical and rational to experiment - to increase the apolarity of the solvent (in concept)? Why would this not work in practice, for example substituting 10% hexane for 10% water? And if hexane is not miscible with water, then what about mixing it with ethanol? And if this failed or killed the UV detection limit, then couldn't we use MS for the detection? (TLC & FTIR have also been used in the past to separate & detect these molecules. There are also reports of detection by ELSD.) By the way, I am not in love with hexane. I have no fixed ideas about this. I simply revised my thoughts about green chemistry for what seems to be a rational guess that hexane would reduce the elution time, compared with water, is this not rational? According to Sigma-Aldrich, as linked by Uwe, hexane is not miscible with methanol, but it is miscible with ethanol. Another poster noted a published method that was hexane:methanol 40:60. Why is it so repulsive to try this, 10:90, or something similar to start with and then to vary the solvent mix (hexane/ethanol/methanol), temperature, and flow?

Retention is determined by an equilibrium constant between the stat. phase and the mobile phase. The interaction between mobile phase and analyte is only one aspect of this. Normally, the interaction between stat. phase and analyte is considered the more important, but in borderline situation this might be different. For instance if the solubility of the analyte is extremely low in mobile phase this aspect might dominate, the analyte will just get stuck somewhere (probably not even make it to the column). Just a hunch: I suspect that your compound will do this if hexane is used. If your compound is soluble in hexane it will elute at the dead time with hexane as mobile phase.

Yeah, ok with the theory. Is this inconsistent or contradictory with the possible method of any C18 & hexane:methanol 10:90 and selection of both the stationary phase and the mobile phase for their affinity with the analyte, among a relatively uniform analyte set of apolarities? We are not concerned with what is "normal" or "the usual way" or the "way it has always been done", concerning acetonitrile, methanol, water, and THF. Also, we have noticed the price of acetonitrile go up recently and the dependence on the automobile industry, and we do not want to depend on this solvent. We believe we are keeping an open mind about this, starting with the published method, the lipid analyte (the extract that we have and whose characteristics we know), the independent column comparison, and the characteristics and affinities of the analyte, column, and solvent. As crude or naive or backward or ignorant or illiterate or unconventional as this may seem to you, can you please explain why this line of thinking and basic specifications for the method and first analysis (RP18 column and hexane:ethanol or methanol solvent) do not make sense to try? If this thinking is a return to the stone age, then would you please explain this clearly and specifically? On the contrary, some of you have even stated: go ahead, and this is what I am looking for a contract lab to do.

You want to re-create a method you found in one paper - to do what? Do you want to purify an extract for milligram-gram quantities of one analyte, or many? Or do you want to test for trace levels of one (or many) analytes in a complex matrix?

In fact, there are various published papers that describe the C18 & MeOH:H20 method and other methods. Yes, we want to purify a plant extract for mg quantities of one analyte, including making an internal standard, using an auto-sampler, fraction collector, software, first chromatogram of an acceptable method, and the retention time of a published analyte, a molecule with a long carbon-chain (35-39), three hydroxyl groups, and a terminal lactone moiety.

Please make a separate offer to purify 10-15 related molecules with similar MW's.

Assume that the plant extract would be done using SFE (supercritical fluid extraction), so the apolarity of the various molecules in the mix would be relatively uniform, to repeat, very apolar. (These molecules, annonaceous acetogenins, are known & published to be selectively cytotoxic. There is no standard, because the molecules have not been patented, or else the patents have expired.)

Further assume that the extract would also be tested for pesticides, fungus, and heavy metals by a commercial testing laboratory, using LC-MS and GC-MS.

You're like a guy at a scientific lecture on evolution with only the Bible as a textbook.

You better believe that you'll never find me anywhere near the Bible. My psychiatrist says to never, ever call me a sex maniac. She says that I simply have a natural, healthy, clean interest in frequent copulation with the opposite gender. Nor am I ever, ever to be called a sex addict. Again, my pleasure is simply natural, healthy, clean, and of course safe. Costs can be controlled if it is not available for free, although it is better to get it for free.

While I do not work in a lab that performs validation of new methods, there are people here who could probably describe to you the process.

Ok, what is the process of validation, other than a statistical exercise, with analysis of accuracy, precision, limit of detection, limit of quantification, linearity / range, selectivity, robustness, and ruggedness.

If you are with a lab or have a lab that might still be open to collaboration, would you please reply.

Assume that we would supply a C18 column, among the various discussed, preferably with 3 micron particle size, 50 mm length x 3 mm i.d., if your instrument can withstand the pressure.

Further assume that, if required, we would also supply solvents, methanol, ethanol, and hexane, which are to be procured for HPLC grade quality.

Detection is to be preferred with UV, according to the published range 205 to 220, but at least tested with MS-2 to confirm MW equal 622 and precise molecular formula, with NMR also to be done to identify the chirality of the single key analyte by comparing to the scientific literature.

Preference with lipids & natural products preferred.

In fact, two methods are to be modified, optimized, and validated, that is a preparative method for an internal standard and an analytical method with detection by FTIR and UV, according to published studies.

If you wished, we could possibly consider that you of course be free to publish the results of any studies, for a doctoral thesis or even for a postdoc.

We are also working on new ways to develop and enhance the specificity of chiral synthesis & enantiomers, specifically of acetogenins. These methods are being patented, so I am not at liberty to say much more.

You can find the various published methods, using even a cursory search of Elsevier sciencedirect, Wiley, Springer, or PubMed database, using the keywords "annonaceous acetogenins", "selective cytotoxicity", "mechanism of action", "method", and so on.

If the accredited lab is doing the validation, then they probably have templates that they work from.

Ok, if you have an accredited lab and you are replying, then can you please send an outline of the templates or else offer specific advice how to control the cost of the validation. How many samples are required for the validation? What if we just vary the % mix of hexane:methanol/ethanol, the temperature, and the flow?

If you want to perform a partial validation for your own confidence, look at parameters like interference from the matrix, linearity of analyte response, analyte recovery, any sample extraction procedures, repeatability, limits of detection. It's all useful data that can help the accredited lab out, but you have to develop and optimize your method before you validate it... I'm not sure I've seen anyone solicit you to hire their lab in this thread, but many of us work for labs who do the work you require, so I can't see why a solicitation would be in any way a surprise to you.

Ok, assume that there will little interference with the matrix (due to SFE). Further assume that your detectors, UV, MS, and FTIR, can be calibrated, or else that two of them can be calibrated, also with comparison to published method. The analyte recovery will be .01 - 10%, and we would be interested in measuring the yield of the analyte and also the yeld of the set of acetogenins in the extract, beside the runtime, setup time, and cleanup time. The extract would be supplied to you. The only sample preparation required is for the internal standard, again using an auto-sampler, fraction collector, software, and known retention time of the key analyte from the chromatogram.

How much would it cost and approximately how long would it take to:

a. redevelop & optimize the published method (RP18, MeOH:H20 90:10), including an RP18 column to be supplied among these discussed, hexane:ethanol (or methanol) 10:90 to start with, and then varying at least the solvent mix, temperature, and the flow,

b. validate the method, using a lab accredited for GLP by the FDA or national health authority.

This is part of a project by various clinical & pre-clinical researchers who are breaking out and who have broken out of Novartis, GSK, and Pfizer to develop these chiral molecules, the synthesis of their chirality, and also economic aspects of the development, such as a patent. See the paper published in 2005 by Professor Steven Ley of the Department of Chemistry of the University of Cambridge about annonaceous acetogens in "Angewandte Chemie..."

Whether we actually get a patent or not for synthesizing the chirality, we still intend to go ahead with this project, but at the lowest possible cost, so if you are not offering lab services, then could you please advise specifically how the costs of development/optimization and validation could be minimize or controlledd?

If you cannot make an offer of these services, what else would you need to know to make an offer, other than trying it at your lab with your instrument and your analyst?

How can I communicate this request for quote formally and in writing? How can I find such a supplier of these preparative/analytical services?

As much as anything, I am trying to control the costs of the experts in a field where I have no training nor practical experience. I am not interested in reading a book nor attending a course that is not specific to my application, the details of which have been described in detail and which are available by searching public databases of scientific papers. And yes, I am sceptical of experts, since it was the experts who gave us thalidomide (anybody old enough to remember thalidomide?), Vioxx, and the Great Flu Vaccine Quackery of 2009 (GFVQ-2009).

Thanks for your time and attention.

Regards,

Ben

[tom jupille]

i think that the tone of this thread has has deteriorated to the point where it really doesn't match the level of civility expected here on the Forum. I normally try to keep a low profile, but in this case I've locked the thread and banned "benrock".