Article
Reference
Rapid log P determination of natural products in crude plant extracts from UHPLC-TOF-MS profiling data - an additional parameter for
dereplication and bioavailability
EUGSTER, Philippe, et al.
EUGSTER, Philippe, et al. Rapid log P determination of natural products in crude plant extracts from UHPLC-TOF-MS profiling data - an additional parameter for dereplication and
bioavailability. Planta Medica, 2009, vol. 75, no. 9, p. 913-914
DOI : 10.1055/s-0029-1234363
Available at:
http://archive-ouverte.unige.ch/unige:9930
Disclaimer: layout of this document may differ from the published version.
1 / 1
• The method previously described for small and simple reference compounds was not directly applicable to the log P determination of more complex natural compounds.
• The deviation between measured and expected log kw can be explained by intermolecular interactions .
Therefore the model has been adapted and a new equation has been proposed including a linear combination of structural parameters.
References
57th International Congress & Annual Meeting of the Society for Medicinal Plant Research and Natural product Research
Rapid log P determination of natural products in crude plant extracts from UHPLC-TOF-MS profiling data –
an additional parameter for dereplication and bioavailability
Philippe Eugster, Sophie Martel, Davy Guillarme, Pierre-Alain Carrupt, Jean-Luc Wolfender
School of Pharmaceutical Sciences, University of Geneva, University of Lausanne, Quai E-Ansermet 30, CH-1211 Geneva 4, Switzerland
• A test set of 42 NP were simultaneously injected in UHPLC system equipped with Aquity UPLC BEH shield RP18 100x2.1mm, 1.7mm stationary phase with 3 different mobile phase pH.
Ionization
• Because retention factor of the neutral form is higher than log kw of the corresponding ionized form, some ionization information can be directly extracted from the retention behavior measured at different pH.
Results and discussion
Conclusions & Perspectives
[1] J.-L. Wolfender, E. F. Queiroz, K. Hostettmann. Expert Opinion on Drug Discovery 1, 237-260 (2006);
[2] Y. Henchoz, D. Guillarme, S. Martel, S. Rudaz, J.-L. Veuthey, P.-A. Carrupt. Analytical and Bioanalytical Chemistry, 394 (7), 1919- (2009);
[3] A. Guillot, Y. Henchoz, C. Moccand, D. Guillarme, J.-L. Veuthey, P.-A. Carrupt, S. Martel. Chemistry & Biodiversity. In Press. (2009)
[4] E. Grata, J. Boccard, D. Guillarme, G. Glauser, P.-A. Carrupt, E. E. Farmer, J.-L. Wolfender, S. Rudaz. Journal of Chromatography B, 871, 261-270 (2008).
• Liquid chromatography is a fast and low sample consuming technique fully used in log P determination. The method is based on the relationship existing between retention factors and log P using specific chromatographic conditions [2].
• The development of column packed with sub-2μm particles working at high pressure (Ultra High Pressure Liquid Chromatography) also allows higher throughput.
• NP’s retention factors extracted from UHPLC-MS metabolite profiling data could then provide log P of compounds of interest prior isolation.
Introduction
• In phytochemical analysis, HPLC metabolite profiling methods provide a large amount of data on the composition of a given crude plant extracts for both dereplication or rapid on-line structure determination of given natural products (NP’s) [1].
• Many physicochemical properties could be extracted from HPLC data, such as lipophilicity.
• Lipophilicity (described by log P) is a key-parameter involved in pharmacokinetic (absorption, distribution, metabolism, elimination and toxicity) and pharmacodynamic processes (ligand-target interactions) and has to be determined as early as possible.
O
O
O O
O
O
log kw
log P
log P log k
NP’s mixture, or crude plant extract
Retention time in 2 gradients (slope ration 3:1) UHPLC-MS analysis
log P determination using a calibration curve Profiling data
Online stucture elucidation Dereplication
• This method provides information in ionization state variation with mobile phase pH. Furthermore, the direction of the variation is directly related to the ionizable function type.
Relation log kw – log P
• In a previous work, a calibration curve was established on a shorter Aquity UPLC BEH shield RP18 (30 mm) using a test set of 38 compounds injected individually [2, 3].
• A mixture 8 of 38 of reference compounds was injected in UHPLC system with the 100 mm stationary phase and log kw obtained were perfectly superimposed with calibration curve. No significant effect of column’s length and simultaneous injection was observed.
• Only the 29 NP existing under their neutral form at least at one of the 3 pH and the corresponding log kw were kept.
• log kw measured were generally overestimated and the differences between experimental log kw and expected log k depend on compound’s structure.
• 42 NP’s were selected by cluster analysis based on 4 molecular descriptors : H-bond donor or acceptor properties (α, β), polarizability (π*), and Mc Gowan volume from a database of 298 well-known NP’s.
• According to previous works [2, 3], the 42 NP and 8 reference compounds (see below) were tested on Acquity UPLC BEH Shield RP18 (100x2.1 mm, 1.7 μm) stationary phase.
• Retention times were measured using 2 generic gradients differing only in gradient time (3:1 ratio, i.e. 33 and 11 min).
Materials and Methods
• log kw was extrapolated by a chromatographic modeling software (Osiris, Datalys, Grenoble France).
• log kw obtained were compared to experimental log Poct from literature.
• 3 buffer solutions were prepared : formic acid/ammonium acetate (pH 2.5), acetic acid/ammonium acetate (pH 5.0 and pH 10.0)
• Detection : a Time-of-Flight mass spectrometer equipped with an electrospray interface was used in both positive and negative modes [4].
• These observations suggest that the chromatographic behavior of complex natural compounds in our conditions can be attributed to a complex influence of several inter-molecular interactions between the natural solutes and the chromatographic system.
• Multilinear analyses (MLR) have to be used to identify the combination of structural parameters responsible of the peculiar behavior of NP’s.
However, the very high correlation (80 – 94 %) between the four solvatochromic parameters for the studied compounds forbid the simultaneous usage of these parameters in a single MLR equation.
• A principal component analysis confirmed these high correlations since a single component PC1 described 88 % of the solvatochromic variation in the chemical space of the 81 reference and natural compounds explored . PC1, composed by molecular mass (24.9 %), α (23.1 %), β (26.2 %) and π∗ (25.8 %), can thus be used in a regression analysis.
QSPR analysis
• Preliminary graphical analyses demonstrated that the deviation of lipophilicity prediction was not linearly related with a single structural parameter such as molecular volume (or exact mass), polarizability (π*), the H-bond donor capacity (α) or the H-bond acceptor capacity (β).
• Equation demonstrates that PC1, the linear combination of the four solvatochromic parameters, allows to predict log P from the UHPLC measurements for the reference and the natural compounds.
log P=1.26 ( 0.07) log kw -1.08 ( 0.37) PC1-1.31( 0.04) n=81; r2=0.92; q2=0.90; F=470
• 3 outliers have been identified ; the reason of their deviation has to be investigated.
• A rapid NP’s structural parameters determination has to be developed in order to directly apply the new method on non isolated compounds with unknown structure.
• Therefore the new model will be applied to complex matrices such as crude plant extracts and the determined physicochemical properties would be of great value in the identification of new NPs of interest.
0 1 2 3 4 5 6
log kw
log kw at different pH
pH 2.5 pH 5.0 pH 10.0
No change: neutral form at pH 2.5, pH 5.0 and pH 10.0
apigenin pKa= 8.54
acidic aescin
pKa= 3.10 acidic
caffein pKa= 0.60
basic
codein pKa= 8.22
basic
O O
O
O
O
O O
O
OH
OH OH
H
HO O
O HO O H
OH HO HO HO
OH OH HO HO
H
N
N N
N O
O
Me
O^Me O
N
OH H H
O
O OH HO
OH
Neutral form at pH10.0
Ionized at pH 2.5 and pH 5.0 Ionized at pH 10.0
Ionized at pH 10.0 Neutral form at pH 2.5
and pH 5.0 Partially ionized at
pH 2.5 and pH 5.0
-2 0 2 4 6 8
0 2 4 6 8
NP's test set Reference
Calibration
Calibration curve
log Poct
logk w
0 5 10
-2 0 2 4 6 8 10
NP's test set Calibration
predicted log P
experimental log P
View publication stats View publication stats
