[the.khemist.ds]

I'm going to preface my statement with the admission that I am neither a NMR spectroscopist or an expert in the specifics of peptides. But... I think a lot of people's assessment of the NMR's has been somewhat 'unfair' it's easy to be critical of someone else's work.

I suspect that the NMR spectra are generally fine (perhaps apart from the missing aldehyde peak). Some factors that I think are worth considering are:

Rotamer effects
Tumbling/averging effects - by which I mean NMR relies on the concept that the molecules tumble freely - as only the N-terminus of the peptide is protected, I suspect that it is likely that there aggregates forming in the tube - in which case these large aggregates would tumble more slowly - meaning that the averaging effect is probably not working properly.

I'll concede that those two statements are somewhat speculative, and I'd be happy to hear the opinions of others if they disagree.

In the end as others have stated I think that a proper assessment should be based on HPLC-MS  (if you really doubt the purity, perform your own analysis on their product - if it doesn't match their stated purity levels you would be in your rights to request a replacement of the sample with one of the quality that they promised to supply or a refund)

[molecule_787]

According to that NMR which lab B sent, this compound is far from being >98% pure. They should really provide a much better NMR. This one is a joke.

Below you will find the HPLC reports from both LAB A and LAB B.  What are your thoughts on it?  I read the results/purity as: LAB B (97.87%) and LAB A (91.46%), correct?

@ the.khemist.ds, arit and anyone else who would like to chime in,
please feel free to also further/add your opinion on the HPLC reports and overall purity.

*I have whited out some of the sensitive information regarding lab names, files et. for privacy reasons.


LAB A (CHINA) HPLC REPORT (91.46%)



LAB B (USA) HPLC REPORT (97.87%)

[the.khemist.ds]

My response would be: Looking at the two chromatographs, the approach to peak area measurement is not the same (ie you are not comparing the same data) the main point is that lab A have treated the shoulder at RT 7.887 as a separate peak. Looking at lab B's report, you can see a similar shoulder which is not treated as a separate peak.
Of course any assessment is based on the principle that the chromatographic conditions are suitable to separate impurities from the desired peptide.

In the end, you have to either make the peptide yourself or assume that the supplier can provide the product at the specification they have stated - this isn't saying that you shouldn't perform your own QC when you receive the sample (remember that whatever data they provide is only true about the sample they measured at that time and it is possible that the sample you receive may be of different quality for any number of reasons)

If you are interested in other possible suppliers check out ChemSpider (I think the correct record is: http://www.chemspider.com/Chemical-Structure.8136267.html)

[molecule_787]

My response would be: Looking at the two chromatographs, the approach to peak area measurement is not the same (ie you are not comparing the same data) the main point is that lab A have treated the shoulder at RT 7.887 as a separate peak. Looking at lab B's report, you can see a similar shoulder which is not treated as a separate peak.
Of course any assessment is based on the principle that the chromatographic conditions are suitable to separate impurities from the desired peptide.

In the end, you have to either make the peptide yourself or assume that the supplier can provide the product at the specification they have stated - this isn't saying that you shouldn't perform your own QC when you receive the sample (remember that whatever data they provide is only true about the sample they measured at that time and it is possible that the sample you receive may be of different quality for any number of reasons)

If you are interested in other possible suppliers check out ChemSpider (I think the correct record is: http://www.chemspider.com/Chemical-Structure.8136267.html)

Thanks for your response.  It is my understanding that the sample reports of the HPLC and results will always be variable meaning it is an average taken at a certain point in time as the structure itself will never be completely uniform or the results consistently exact et.  However my own or third party QC to verify structure and purity should be within the stated or same parameters.

Basically, what you are saying is that the HPLC reports do reflect the stated percentages in regards to purity and the separation of impurities (albeit by different methods/means) but what it comes down to is whether or not I can lend credence to the supplier and that they will supply the peptide at the stated purity from these reports, correct?

LAB A has stated that they are going to bring the purity level up to 98%, they just wanted me to verify that the compound/structure (HNMR report) is satisfactory before they proceed in furthering the purification and yield et.  

As stated, I have sent them an email stating that the HNMR report is unsatisfactory due to the missing Aldehyde signal and the ppm being cut off on the report at the critical region et.  They have responded by forwarding the information onto their engineers and will forward a revised document.  So I will likely post that for further feedback once it arrives.

In my humble opinion, I don't think LAB B would go to the trouble of fabricating an HPLC when they have produced (what appears to be) supportive data (HNMR and Mass Spec) that is satisfactory in terms of verifying the compounds identity et.  The quality/resolution of the report is an issue in itself but the data, from what others have stated, seems to be accurate with the desired compound.  Would you agree with this statement?

I have in fact contacted a 3rd supplier and received an initial quote with a guarantee of purity and identity of the peptide with supportive data of a HPLC and Mass Spec.  They will not however, produce and provide tests until I commit to an order, at which point, they will provide a written guarantee in regards to the purity and identity by internal test reports et.  Obviously if things were to not mesh, the guarantee would be void and funds reimbursed (as you outlined).

Essentially, it would seem that LAB B has the correct or desired structure and purity, so purchase from this lab and third party verification is an option with this supplier.  I will however, perhaps wait on LAB A to correct and forward the HNMR so it includes the missing information.  Still, some of the other experts in this thread have stated that LAB A seems to be slightly off not only due to the missing Aldehyde signal but also on the subsequent structure itself.

Thank you again for your time and comments.

[molecule_787]

My other question is, can the identity and purity of this peptide be properly verified by HPLC and Mass Spec alone?  Is including a HNMR vital in determining the identity of the compound?

[discodermolide]

My other question is, can the identity and purity of this peptide be properly verified by HPLC and Mass Spec alone?  Is including a HNMR vital in determining the identity of the compound?

If you run the HLPC against a known standard compound you can measure the content in your sample, otherwise it's just area%.
The NMR is not vital but it does give a lot of structural information, for example, it will indicate if the compound has epimerised, i.e. a mixture of diastereoisomers. A 2D spectrum would also be nice to have as well as a 13C.

[molecule_787]

Any other thoughts on the HPLC reports and stated purity, specifically LAB B's? 

[Honclbrif]

The LC's look pretty good, but I don't know the method of detection. Probably UV. Could hide something depending on the detection wavelength, but I prefer to believe in the intrinsic goodness of Humanity (feel free to laugh here). Your MS has a big fat 619.5 peak which I'm assuming is your [m+H]+, and a 641.5 which is probably [m+Na]+. I've no idea what the 673.7 peak is, but its about 54 units too heavy. Works out to the mass of 3 H2Os. Might be a coincidence, might not. When I do MS of peptides I don't see H2O bound, but I've got fairly harsh ionization conditions.

My major concern in all this is that, as others have pointed out, the NMRs are a little wacky for a couple of reasons:
A. One lab cut off before the critical aldehyde region. Makes it look like they either don't know what they're doing or they have something to hide. Neither smells good.

B. Both specs' integrations are all over the place. In the first one they picked a random peak to be A = 1, in the second (seriously, who phased that thing?) they picked the aldehyde to be A = 1, but the phenyl group integrates to 12.5. They're both soaking wet so the aldehyde may have hydrated making it look like the aldehyde peak is smaller than it really is. I haven't done many NMRs of aldehydes in wet DMSO, so I don't know if this happens readily under those conditions or not.

Overall: You've got a clean HPLC, and an MS that has the right mass as the major peak and those facts are difficult to ignore. If you can get an MSMS you'll know the sequence and that should eliminate pretty much all doubt. Based on the LC and MS evidence I'm inclined to believe its probably the right compound and there's some weirdness happening in the NMR for what could be many reasons. If you can afford to use it, I'd go ahead and try a positive control. If it fails, you know the stuff is bad, stop using it, make a really angry phone call to the supplier, and try to get your money back and get it from a better supplier. If you can't afford to use it in an experiment skip all that, get your money back, and get it from a better supplier.

[molecule_787]

The LC's look pretty good, but I don't know the method of detection. Probably UV. Could hide something depending on the detection wavelength, but I prefer to believe in the intrinsic goodness of Humanity (feel free to laugh here). Your MS has a big fat 619.5 peak which I'm assuming is your [m+H]+, and a 641.5 which is probably [m+Na]+. I've no idea what the 673.7 peak is, but its about 54 units too heavy. Works out to the mass of 3 H2Os. Might be a coincidence, might not. When I do MS of peptides I don't see H2O bound, but I've got fairly harsh ionization conditions.

My major concern in all this is that, as others have pointed out, the NMRs are a little wacky for a couple of reasons:
A. One lab cut off before the critical aldehyde region. Makes it look like they either don't know what they're doing or they have something to hide. Neither smells good.

B. Both specs' integrations are all over the place. In the first one they picked a random peak to be A = 1, in the second (seriously, who phased that thing?) they picked the aldehyde to be A = 1, but the phenyl group integrates to 12.5. They're both soaking wet so the aldehyde may have hydrated making it look like the aldehyde peak is smaller than it really is. I haven't done many NMRs of aldehydes in wet DMSO, so I don't know if this happens readily under those conditions or not.

Overall: You've got a clean HPLC, and an MS that has the right mass as the major peak and those facts are difficult to ignore. If you can get an MSMS you'll know the sequence and that should eliminate pretty much all doubt. Based on the LC and MS evidence I'm inclined to believe its probably the right compound and there's some weirdness happening in the NMR for what could be many reasons. If you can afford to use it, I'd go ahead and try a positive control. If it fails, you know the stuff is bad, stop using it, make a really angry phone call to the supplier, and try to get your money back and get it from a better supplier. If you can't afford to use it in an experiment skip all that, get your money back, and get it from a better supplier.

Ok great, thanks for the insight.  I was thinking along the same lines due to the clean HPLC and MS from LAB B.  The NMR has some odd points but the general consensus was that overall it correlates with the desired structure.

I will await for LAB A to forward a revised NMR and also include a MS to make a final decision on that supplier. (I will likely post their results again for some feedback).

I have a third supplier lined up should the third party HPLC and MS testing on LAB B's peptide show any discrepancies et.

[molecule_787]

The LC's look pretty good, but I don't know the method of detection. Probably UV. Could hide something depending on the detection wavelength, but I prefer to believe in the intrinsic goodness of Humanity (feel free to laugh here). Your MS has a big fat 619.5 peak which I'm assuming is your [m+H]+, and a 641.5 which is probably [m+Na]+. I've no idea what the 673.7 peak is, but its about 54 units too heavy. Works out to the mass of 3 H2Os. Might be a coincidence, might not. When I do MS of peptides I don't see H2O bound, but I've got fairly harsh ionization conditions.

My major concern in all this is that, as others have pointed out, the NMRs are a little wacky for a couple of reasons:
A. One lab cut off before the critical aldehyde region. Makes it look like they either don't know what they're doing or they have something to hide. Neither smells good.

B. Both specs' integrations are all over the place. In the first one they picked a random peak to be A = 1, in the second (seriously, who phased that thing?) they picked the aldehyde to be A = 1, but the phenyl group integrates to 12.5. They're both soaking wet so the aldehyde may have hydrated making it look like the aldehyde peak is smaller than it really is. I haven't done many NMRs of aldehydes in wet DMSO, so I don't know if this happens readily under those conditions or not.

Overall: You've got a clean HPLC, and an MS that has the right mass as the major peak and those facts are difficult to ignore. If you can get an MSMS you'll know the sequence and that should eliminate pretty much all doubt. Based on the LC and MS evidence I'm inclined to believe its probably the right compound and there's some weirdness happening in the NMR for what could be many reasons. If you can afford to use it, I'd go ahead and try a positive control. If it fails, you know the stuff is bad, stop using it, make a really angry phone call to the supplier, and try to get your money back and get it from a better supplier. If you can't afford to use it in an experiment skip all that, get your money back, and get it from a better supplier.

By the way, Have the coupling constants for the multiplets (doublets, triplets, etc.) been picked out in both these NMR's?

Thanks

[arit]

Unless they sent them over in a separate document, no. The spectrum from LAB A needs better S/N-ratio to start to estimate couplings and multiplicities.
Spectrum from LAB B on the other hand suffers from horrible shimming and phasing, and I wouldn't even try to calculate the couplings or guess multiplicities.
Both spectra would have gained a lot from a more homogeneous magnetic field, and maybe from a higher field spectrometer.

As to the integrals, LAB A hasn't even bothered to aim for quantitative spectra (D1 1s, NS 120). 1 second isn't probably even the average T1 for the compound,
and ideally you'd want it somewhere around 10*T1 for quantitative purposes. I'd assume spectrum from LAB B suffers from same problems, but can't really tell
since the parameters are cropped out.

[molecule_787]

Unless they sent them over in a separate document, no. The spectrum from LAB A needs better S/N-ratio to start to estimate couplings and multiplicities.
Spectrum from LAB B on the other hand suffers from horrible shimming and phasing, and I wouldn't even try to calculate the couplings or guess multiplicities.
Both spectra would have gained a lot from a more homogeneous magnetic field, and maybe from a higher field spectrometer.

As to the integrals, LAB A hasn't even bothered to aim for quantitative spectra (D1 1s, NS 120). 1 second isn't probably even the average T1 for the compound,
and ideally you'd want it somewhere around 10*T1 for quantitative purposes. I'd assume spectrum from LAB B suffers from same problems, but can't really tell
since the parameters are cropped out.

Ok, thanks again for the info arit.  By stating that LAB A has not even bothered to aim for quantitative spectra ie. 1 second is probably not even the average T1 for the compound, you are referring to a form of its relaxation rate in the NMR test, correct?

T1 relaxation are generally strongly dependent on the NMR frequency and so vary considerably with magnetic field strength.  Small amounts of paramagnetic substances in a sample speed up relaxation very much. By degassing, and thereby removing dissolved Oxygen, the T1/T2 of liquid samples easily go up to an order of 10 seconds.  Correct?  So in other words, LAB A has not even bothered to attempt to gather this data properly so it appears the average T1 stated (1 sec) could be incorrect because of this?

LAB B you cannot draw any conclusions on in regards to the relaxation rate or T1 for the compound because the parameters are cut off but you are assuming that they have not even bothered to aim for quantitative spectra (ie. T1 et.)?

[molecule_787]

Also, is it fair to say that these suppliers have used a simple 1-dimensional proton NMR spectrum? It wouldn't be a 2D spectrum (such as a COSY or NOESY)?

Thanks.

[arit]

Yes, I'm making a point about how in quantitative measurements you want the relaxation delay (D1) to be 10*T1. This follows
from the fact that at 10*T1, 99% of the spins have returned back to their equilibrium state, and you get the same amount of signal
when repeating the experiment.

When the relaxation delay is shorter, the spectrum can still be fine, just not reliably quantitative. The length of the delay doesn't of course
matter if the number of scans is 1.

To put it shortly, LAB A hasn't aimed for quantitative spectra. Either knowing what they were doing or not.

LAB B you cannot draw any conclusions on in regards to the relaxation rate or T1 for the compound because the parameters are cut off?

I can't tell whether the relaxation delay LAB B used is in the general vicinity of a "usual" T1 for a smallish organic molecule. The actual relaxation rate is a property of the sample, not experiment setup. Just the fact that the parameters are cut off raises a few questions in my mind, but I tend to worry too much
anyway. Since the parameters aren't there, you can't tell what relaxation delay, window function, line broadening etc. settings they used that might have an
effect on the spectrum. Actually, even calculating the coupling constants would be difficult, since you don't know the operating frequency of the spectrometer.

And yes, both suppliers have sent just plain 1D proton spectra. Which of course doesn't mean that they wouldn't have used an array of 2D experiments
in their internal QA process.

[molecule_787]

^Once again, thank you very much for the explanation arit.