750
D. H. Kim et al
the KA for boric acid (0.5 mM) complexation with NADC
(0.1 mM) was 1840 š 110 l molꢀ1 at pH 10.3. Since these two
KA values are in reasonable agreement, we believe that the
ionization efficiencies between the borate–nucleotide com-
plex and free nucleotide are similar. Therefore, ESI-MS is a
valid method for determining the KA for borate complexation
with nucleotides.
The KA values for different nucleotides were calculated
from the peak areas of the signals for complexed and
free nucleotides. The total concentration of the nucleotide
ꢀNuctotalꢁ is constant and is equivalent to the sum of all
the nucleotide-associated peak areas. The sodium-adducted
nucleotides also complex with boric acid at elevated
concentrations (data not shown). The concentration of the
free boric acid is then derived. Using this method, the
concentration of all the species at equilibrium is determined.
The KA is then calculated using Eqn (3). This method assumes
that the ESI-MS peak areas are proportional to concentration,
a stipulation that has been repeatedly verified both here
(unpublished data) and elsewhere28 and, as already stated,
we believe that the efficiencies of ionization of the free
nucleotides and the borate–nucleotide complexes are not
significantly different from one another.
We also investigated the stability of other borate–nucleo-
side complexes including those with adenosine, cytidine,
guanosine and uridine. The complex stability with these
nucleosides fits into the trend described above. Since these
nucleosides are not phosphorylated, they have higher KA
values than the monophosphate nucleotides. We believe
that only qualitative comparisons can be made with these
nucleosides because there might be a difference in ionization
efficiency between negatively charged complexed nucleo-
side and uncharged free nucleosides. Unlike mono-, di- and
triphosphate nucleotides, these non-phosphorylated nucle-
osides will have neutral charge even at pH 10.3. With this
caveat in mind, the KA for adenosine, cytidine, guanosine and
uridine analyzed by ESI-MS were 1150 š 110, 1900 š 170,
800 š 100 and 970 š 130 l molꢀ1, respectively.
results provide a more accurate measure of the complexes
formed between borate and nucleotides.
Borate complexation with nucleotides was also measured
at pH 7.4 in ammonium bicarbonate buffer. At 500 µM
boric acid and 100 µM nucleotide concentrations, of all the
nucleotides tested only the boric acid–NADC complex was
observed. The KA of this complexation was 46 š 7 l molꢀ1 and
indicates that borate complexation with NADC is relevant at
physiological pH. From this it is reasonable to suggest that
some of the effects of borate can be contributed to the in vivo
formation of an NADC –borate complex.
The inverse relationship between complex stability and
length of phosphate groups can be explained by electrostatic
and physical interactions of phosphate groups and borate
bound to the ribose group. First, at elevated pH, the depro-
tonated phosphate groups electrostatically repel borate ions
from complexing with the nucleotides. Second, our previous
experiment with borate–NADC complex demonstrated that
the phosphate backbone is sufficiently flexible to interact
with the ribose group. We showed that one of the MS/MS
fragments of NADC is a borate diester that connects two
ribose groups linked by two phosphate groups to form a
cyclic molecule.5 In the light of this, we propose that the
hydroxyls on the phosphate groups can physically interact
near the ribose ring to destabilize the borate binding. As
result, the complex stability between borate and nucleotides
is inversely proportional to the length of the phosphate
group. Therefore, the KA of borate complexes decreases in
the order mono-, di- and triphosphate nucleotides.
In order to use MS for quantitative purposes, the ion-
ization efficiencies of the species compared must be similar.
Unfortunately, because elevated boric acid concentrations
ꢀ>600 µMꢁ cause ion suppression of signals, we cannot shift
the equilibrium between two extreme conditions to examine
the ionization efficiencies of the borate–nucleotide com-
plex and the free nucleotide. Moreover, a standard for
borate–nucleotide complexes is unavailable, owing to the
reactivity of borate with hydroxyl groups in solvents includ-
ing water and alcohol. However, it seems reasonable to
assume that the ionization properties of the free nucleotides
and borate–nucleotide complex species are similar for the
following reasons. (i) At high pH, both complexed and free
nucleotide species are already ionized in solution owing to
deprotonation of the phosphate group on the nucleotides
ꢀpKa D 6–7ꢁ.26 Therefore, both the free nucleotide and
borate–nucleotide complex are ionized prior to electrospray.
(ii) The change in the mass due to borate complexation is
small compared with the mass of the free nucleotides (<10%
difference). (iii) The complex is covalently bound and there-
fore it is less susceptible to collision-induced dissociation in
the gas phase. (iv) MS analysis in water-ammonium hydrox-
ide at pH 10, instead of WAT solvents, also showed consistent
borate–nucleotide complex formation but with lower sen-
sitivity. The complexation of borate–nucleotide does not
depend on the organic solvents. (v) Johnson and Smith27
studied the kinetics of borate–NADC complex formation by
stopped-flow spectrophotometry and determined that the
KA for NADC (44 mM) and boric acid (2.5 mM) at pH 10.4
was 1500 š 400 l molꢀ1. Using ESI-MS, we determined that
In conclusion, we have examined the KA values of
borate complexation with 16 nucleotides at pH 10.3 and
7.4 using ESI-MS. We demonstrated that the stability of
borate–nucleotide complex is predicted by the charge and
phosphorylation state of the nucleotide and by the pH
of the solution. The cationic charge on the nicotinamide
group increases the KA as compared with the neutral
nicotinamide group. Also the length of the phosphate group
is inversely proportional to the KA of borate complexation
with nucleotides.
Acknowledgements
We thank Alex Dooley for assistance with ESI-MS and Beth Marbois
for continual guidance. This work was supported by UC Toxic
Substances Research and Teaching Program and US Borax.
REFERENCES
1. Ralston NVC, Hunt CD. Diadenosine phosphates and S-
adenosylmethionine: novel boron binding biomolecules
detected by capillary electrophoresis. Biochim. Biophys. Acta 2001;
1527: 20.
Copyright 2004 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2004; 39: 743–751