M. M. Lehmann et al.
conversion/elemental analysis-isotope ratio mass spectrometry
(TC/EA-IRMS ),[11–13] since precise and reliable methods for
compound-specific δ18O analysis of individual carbohydrates
are rare, due to the highly complex and sensitive nature of
the sugars.
as solvent and dimethyl sulfide as catalyst, causing
permethylation of all carbohydrates within 24 h,
and this derivatization reaction was above all also
easy-to-handle. Based on this protocol we have
established a new methylation method for individual
carbohydrates for GC/Pyr-IRMS, with unrivalled accuracy
and precision for δ18O measurements. The new method
was tested with (1) various commercial carbohydrates
and (2) standard mixtures of glucose, fructose, pinitol,
and sucrose, and it also was applied to (3) honey samples
from different countries and to (4) Pinus sylvestris leaf
samples.
During the last two decades technical and analytical
developments have opened the door for online δ18O analyses
of individual carbohydrates. Technically, this is based on
the coupling of a gas chromatograph to a pyrolysis reactor
that converts carbohydrates in helium gas under high
temperatures into CO, which can be subsequently measured
by isotope ratio mass spectrometry (GC/Pyr-IRMS). One
crucial step was the design of a functional reactor,[14,15]
consisting of an outer ceramic tube and an inner platinum
tube, partially filled with nickel wires, which was used
successfully in several studies.[3,16–18] From an analytical
point of view, the highly polar carbohydrates need to be
derivatized to make them amenable to analysis by
GC/Pyr-IRMS. The first application of δ18O analysis on
individual carbohydrates was performed by Zech and
Glaser,[19] who described a derivatization method for plant-
and soil-derived carbohydrates with methylboronic acid. This
method is recommended for the pentoses, arabinose and
xylose, and for the deoxysugars, fucose and rhamnose,
but it is not applicable to hexoses. A further study based
on trimethylsilylation as the derivatization method of
individual plant carbohydrates yielded good results
especially for sucrose, with precision ranging from 0.5 to
1.1‰.[20] By contrast, this method was not satisfactory for
the analysis of fructose, glucose, and pinitol. Thus, a new
derivatization method is required to improve the accuracy
and precision of δ18O measurements and to enable the
analysis of a wider range of individual carbohydrates. An
improved method for δ18O analysis of carbohydrates would
be especially helpful in plant sciences, e.g., for the
understanding of how photosynthetically produced
carbohydrates influence the δ18O values of cellulose, which
is often used as a climatic signal to reconstruct past
environmental conditions.[4]
EXPERIMENTAL
Chemicals
All the carbohydrates used as standard material in this study
were of analytical grade (HPLC ≥95–99%) and commercially
available from Sigma-Aldrich (Buchs, Switzerland), Fluka
(Buchs, Switzerland), Merck (Zug, Switzerland), and Gerbu
(Heidelberg, Germany). Silver oxide (Ag2O) was freshly
produced to ensure optimal basicity for derivatization. In brief,
15 g silver nitrate (88 mmol AgNO3, Sigma-Aldrich) and 3.5 g
sodium hydroxide (88 mmol NaOH, VWR, Zurich,
Switzerland), each dissolved in 100 mL hot deionized H2O
(80–90 °C), were mixed in a beaker. This yielded a brown
precipitate consisting of Ag2O that was filtered with a fritted
glass filter (pore size of 40–100 μm; Winzer, Wertheim,
Germany) under vacuum and washed with 100 mL deionized
H2O and 100 mL 99% ethanol. The remaining pellet was dried
in an oven for about 5 h at 60 °C, milled to a fine powder with
a pestle and mortar, and kept until further use in a desiccator
filled with silica gel in a light-proof vial to protect the light-
sensitive Ag2O.
Sample material and preparation
As a good starting point for method optimization we
considered the methylation of individual carbohydrates, which
has been used for decades for various GC applications, but
which has not so far been used for isotopic analysis by
Leaf material of Pinus sylvestris was harvested in a forest
close to the Paul-Scherrer-Institute in Villigen, Switzerland.
The leaf material was dried at 60 °C in an oven and
milled to powder with a steel ball-mill (MM2000, Retsch,
Haan, Germany) for further analysis. Essentially as
described by Streit et al.,[27] 100 mg plant material were
transferred to 2 mL reaction vials and boiled for 30 min
at 85 °C in a water bath. After centrifugation (2 min,
10,000 g), the supernatant with the water-soluble content
was transferred into a new vial and purified as described
by Rinne et al.[9] In brief, top to bottom ordered columns
for trapping of amino acids, organic acids, and polyphenols
(OnGuard II A, H, and P, Dionex, Thermo Fisher Scientific,
Bremen, Germany) were flushed with 30 mL deionized
H2O. Subsequently, the extracted water-soluble content
was added and the purified sugar solution eluted with
6 mL deionized H2O and stored in a freezer (–20 °C).
After freeze drying for 24 h, the remaining pellet was
dissolved in 1 mL H2O and transferred to a new reaction
vial. After a second freeze-drying step, the pellet weight
was determined for optimal scaling of the methylation
derivatization protocol. For commercially available
GC/Pyr-IRMS.
A big advantage is that methylation of
carbohydrates does not introduce additional oxygen,
making it especially suitable for δ18O measurements. The
methylation of carbohydrates was first established by
Purdie and Irvine[21] in 1903, using silver oxide (Ag2O) as
proton acceptor and methyl iodide (MeI) as methyl group
carrier. However, the reaction was slow and incomplete,
yielding products that were not permethylated (methylation
of all possible O atoms). Improvements to the methylation
reaction during the last century have increased the speed
and completeness of the reaction by changing the proton
acceptor or solvent or by adding different catalysts.[22–24]
Methylation of individual carbohydrates is now a versatile
state of the art technique for carbohydrate analysis.[25]
However, it does require advanced chemical skills and issues
such as cleavage of the methylated product during the
reaction or the formation of other undesirable by-products
can still appear. Recently, Hou et al.[26] significantly
improved the “Purdie reaction” by using acetonitrile
wileyonlinelibrary.com/journal/rcm
Copyright © 2015 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2016, 30, 221–229