Any experts on lactose impurities?

[Mattias]

Hi,

I have found a quite large peak in the lactose monohydrate that we use for our products. It seems to vary between batches, but is always there.

In reversed-phase, it elutes in the front and therefore it has always been ignored. In HILIC, I could get good retention of this peak and it was found to have two UVmax (215 and 277 nm). I have tried LC/MS (ESI) but it does not ionise in either positive or negative mode. In SEC it elutes with all the small molecules, so it seems to be fairly small.

We do not have NMR so I feel a bit stuck here. It is not 5-HMF (https://en.wikipedia.org/wiki/Hydroxymethylfurfural, since I see this peak as well but it is very small.

Any suggestions?

[Consumer Products Guy]

I had thought I was lactose intolerant.

Then I figured out that I was simply just intolerant...

[Blazer]

Not an expert, but I'll take on the low hanging fruit. Lactose is a disaccharide made up of glucose and galactose. Is your impurity possibly one of these?

[tom jupille]

Sugars wouldn't fit with an absorbance band at 277. That implies either an aromatic moiety or at least a set of conjugated double bonds.

[Mattias]

The absorbance band at 277 nm is quite strong too, it is just as strong as the one at 215 nm. Doesn't really fit with an aromatic ring, they usually absorb around 230 nm as well?

It is quite a big peak. When I prepare a 20 mg/ml solution of lactose (only) the peak height of the impurity is 1.2 AU (at 277 nm). But I have no clue about the true amount.

Since I have good separation of this peak now, I can at least start to trend the size of the peak against the properties of the product. But of course it would be nice to know the identity. I cannot be the only one seeing this peak? The lactose is from Merck BTW.

[bunnahabhain]

You could collect the peak, evaporate the solvent and let a friend of yours who is expert in GC/MS play with it...

[krickos]

Hi

not an expert but there were some studies on varoius lactose grades including pharma grades in mid 2000s. Lactose phosfate (bonded not salt) was found in the 0,01-0,04% range. CE was used.

[JMB]

Aquick Google search suggests,
Riboflavin
Proteins
Lactose phosphate
Lactic Acid.

Check Uv spectrum of Riboflavin.....

[Mattias]

Thanks for your suggestions!

The UV-spectrum of riboflavin looks very promising, but I have a feeling that it should have quite good retention in reversed-phase? I cannot get any retention of the unknown on a C18 column, even when using 0.1% formic acid in water as the mobile phase.

Facts:
No retention in reversed-phase
Elutes slightly before 5-HMF (Mw: 126) in size-exclusion, but after a Mw 1000 marker
UV spectrum shows strong absorption at 277 nm
No signal in ESI-MS (positive or negative)

I performed fraction collection of the impurity in SEC, concentrated it with the RotoVac and performed infusion on the MS. Then I got a clean mass spectrum showing two masses only in negative mode, 98.9 and 101.9 (about 1/3) of the intensity. The elemental/isotopic prediction for this compund was ClO4
This would be perchlorate, but that has no UV-absorbance at all. Apparantly perchlorate elutes at the same time as my unknown. I should do fraction collection in the HILIC system instead where I have better separation. But there is this Christmas thing now.

[LCFan]

Furfural?

[JMB]

Mattias,

[1] ClO4 would give m/z 98.9 and 100.9 in -ve ESI; presume that your '101.9' is a typo ?

If not a typo, then m/z 98.9 and 101.9 are not Cl isotopes of course, and are probably not related.

[2] From SEC, 126 < MW < 1000. Are the elution times on SEC from a single run, that is are the MW markers added to your sample ?? Can you run a MW 500 marker ?

[3] Is the SEC solvent 100% water ?

[4] Can you post your UV spectrum, and the range that was scanned ?

Regards,

JMB

[Mattias]

UV spectrum (200-300 nm)

[Mattias]

Mass spectrum (of the collected fraction in SEC, the impurity in UV has no mass response)

[JMB]

Hmmm........

[1] I agree that m/z 99, 101 is very likely due to ClO4-; there appear to be cluster ions at m/z 199 [M+ClO4]-, M =HClO4;
m/z 216 [M+ClO4]-, M= NH4ClO4; m/z 221 [M+ClO4]-, M = NaClO4.

[2] I would like to see the UV spectrum scanned out to ~800 nm, to check on possibility of riboflavin-derived impurity.

Regards,

JMB

[JMB]

Mattias,

I did a quick Google for lambda 277 nm, and a drug "pentoxifylline" (not a veterinary drug) came up.

This has a 2,6-dione-purine type structure; riboflavin (RF) has a 2,6-dione-pteridine type structure.

This suggests that you possibly have a biosynthetic precursor to RF or degradation product of RF ??

I note that HClO4 is used to precipitate proteins from solution, which suggests the source of m/z 99/101.

JMB