c.constantinidou at bham.ac.uk (C. Constantinidou) wrote:
>I have a problem deglycosylating a bacterial protein. The protein is
>glycosylated according to results obtained from the Oxford GlycoSystems
>Glycotrack Carbohydrate Detection kit (periodate oxidation/biotin
>hydrazide method). I have tried both chemical and enzymatic methods to
>deglycosylate partially purified protein. The chemical method used was
>the GlycoFree degylcosylation kit utilizing TFMS, and the enzymatic >method was the GlycoShift de-N-glycosylation kit utilizing both
>peptide-N-glycosidase F and a mixture of Endo-F/peptide-N-glycosidase F.
>Both kits were purchased from Oxford GlycoSystems. The control proteins
>supplied with the kits have always shifted molecular weight (on
>SDS-PAGE) after treatment, as expected. The protein of interest has >failed to do so.
If your protein is bacterial, it's probably unlikely for N-glycosidase F
to work unless the structure really has a GlcNAc-GlcNAc-Asn core. There
is a report of a N-glycan from Streptococcus sanguis which appears to be
cleavable with N-glycanase (ie. peptide N-glycosidase F) - see Erickson &
Herzberg J. Biol. Chem. 268:23780 - but other 'wierd' structures have
been found - on Endo F2 and Endo F3 from Flavobacterium, there are
O-links with mannose, rhamnose and other sugars (Reinhold et al J. Biol.
Chem. 270:13197 and Plummer et al J. Biol. Chem. 270:13192) - in fact
even in eukaryotes there's a very wide range of O-linked glycosylation
types, ranging from mucins with GalNAc to fungi with mannose and some
unusual Xyl-Glc and Fuc modifications in some mammalian proteins.
Of course, can one rule out that the original glycan detection kit gave a
false positive. Perhaps you should try other means of finding that your
protein is really glycosylated - if it contains GalNAc or GlcNAc, an
amino acid analysis would pick these up. Additionally, it's worth bearing
in mind that even if a protein is N-glycosylated with a GlcNAc-GlcNAc, it
doesn't mean that N-glycanase will work - plant and insect glycoproteins
with a alpha-1,3-linked fucose on the reducing terminal GlcNAc will not
respond to the enzyme - only PNGase A from almonds will remove such a
structure - and then only from glycopeptides - not the whole protein.
> Unfortunately I had only two attempts using the chemical kit before
>the VERY expensive reagents run out. So I can't speculate on what went
>wrong. With the enzymatic method I think the problem may be failure to
>sufficiently denature the protein before treating it with the
>glycosidases. Currently I am denaturing by boiling the sample with SDS
It's probably far cheaper to go to follow the TFMS procedure described by
Edge et al (Anal. Biochem. 118:131 (year 1981)) - but they say the
procedure's efficiency varies depending on the type of glycosylation. And
beware that degradation of the protein can't be ruled out.
>One more thing. The controls work in the presence of my protein
>preparation so the presence of inhibitors in the preparation is not
>>Any suggestions on what to try next would be appreciated.
Can you metabolically label this stuff - mannose, glucose or whatever? Or
use hydrazine (great care and absolute dryness required) and then label
any products either with boro[3-H]hydride or a fluorescent label and
analyse the size?
If you feel inclined to use lectins - they're not a cure-all, since they
do exhibit complicated specificities.
Iain Wilson Institut fuer Chemie
Tel: 43-1-47654-6065 Universitaet fuer Bodenkultur
Fax: 43-1-310-5176 Gregor-Mendel-Strasse 33
E-mail: wilson at edv1.boku.ac.at A-1180, WIEN, Austria