The LCGC Blog: Derivatization
Derivatization is a Powerful Strategy to Achieve Ultratrace Detection Limits for Quantitative Analysis Using HPLC–ESI-MS
Electrospray ionization (ESI) is the most commonly used source for ion generation when high performance liquid chromatography is coupled to mass spectrometry detection (HPLC–ESI-MS). High and moderately polar and ionic compounds of variable size and stability present in a flowing liquid stream of an aqueous polar organic solution can be reliably and efficiently converted into intact gas phase ions by ESI. This means that highly efficient liquid-phase separations can be coupled to sensitive and selective mass spectrometers for targeted quantitative analysis from complex mixtures. Parts-per-million detection limits are routine, parts-per-billion limits can be regularly achieved for certain analytes, and even parts-per-trillion concentration levels can be reliably detected if appropriate methods are devised.
Beyond enhancements that can be achieved by various chromatographic methods and the use of highly sensitive mass spectrometers, ESI efficiency is a key controller of detection limits. Ionization efficiency is highly variable for different analytes and is based squarely on a molecule’s physicochemical properties. To be readily converted into gas-phase ions, molecules of interest present in the electrospray droplet must be able to acquire a charge and reach the surface of the electrospray droplet . Since the inception of ESI, there have been several seminal reviews that have addressed the “practical” mechanism of ion generation (1–3). Although concepts such as ion evaporation or charged residues are quickly encountered when one seeks to study the mechanism of ESI, the term “practical” should point one directly to a model called the equilibrium partitioning model (EPM) (4). The EPM highlights that it is the concentration of ions at the surface of an ESI droplet that is directly proportional to the abundance of ions measured by the MS detector. Therefore, ionizable surface-active molecules are the most highly responsive species. While we will go into more detail into these concepts in future postings, it should not be surprising that a wide variety of chemical compounds can be synthetically modified, or derivatized, to give them these optimal properties (5).
When we started down the path of needing HPLC–ESI-MS for quantitative analysis in our laboratory in 2007, derivatization was an absolute necessity. Our charge was to develop a method for ultratrace (low parts-per-trillion) analysis of native estrogens from cerebrospinal fluid in conjunction with a preclinical trial on traumatic brain injury; we had a circa 2001 ion-trap mass spectrometer connected to a standard HPLC system to achieve this task. A triple-quadrupole instrument would have been much better for quantitative analysis than an ion-trap system, but we did not yet have one, so we turned to the use of dansyl chloride derivatization of the estrogens (estrone, estriol, 17α-estradiol, and 17β-estradiol) to reach the detection limits we needed. Dansylation is a derivatization technique used previously to enable fluorescent detection of phenols and amines. To these groups are appended a naphthyl amine (tertiary amine) group that is also ideally suited for enhancing ESI efficiency. For example, estradiol itself does not provide an ideal handle for ionization; perhaps its phenol group can be deprotonated at high pH to yield an anion for detection in negative ionization mode. However, when the phenol is dansylated, the tertiary amine is a readily cationizable group and together with the napthyl moiety confers substantial surface activity to vastly increase ionization efficiency.
We published a method for the simultaneous detection of the four native estrogens, dansylated, by HPLC–ESI-MS using the ion-trap instrument. (1) We achieved detection limits on the order of 20 ppt (parts-per-trillion), following a liquid–liquid extraction and reconstitution and dansyl derivatization before injection.
More recently, we have acquired new triple-quadrupole instruments. On an entry level triple-quadrupole instrument, we have been able to achieve parts-per-quadrillion detection limits for estrogens in the absence of matrix (unpublished results). Matrix effects still detract from these levels (into the low parts-per-trillion), but now we are able to start exploring on-line preconcentration, coupled with high efficiency fast HPLC separations, and higher sensitivity MS instruments.
With on-line preconcentration techniques using restricted access media (the topic of my previous post), the concept of bulk derivatization becomes possible (7). Bulk derivatization, where the derivatization reagent is applied directly to a biological fluid in which the analyte of interest is present, reduces the number of sample preparation steps to maximize recovery rates and achieve optimal detection limits. For this to work, a large concentration of derivatization reagent must be applied to the sample (along with all of the appropriate internal standards), so dilution must be considered, but if preconcentration can be achieved following this step, then the advantage can be retained. In our lab, we are working on a variety of bulk derivatization techniques to maximize HPLC–ESI-MS response. For us, the “extra step” disadvantage so often cited against derivatization is unfounded. The advantages far outweigh the limitations if two to three orders of magnitude of sensitivity can be reliably attained through derivatization.
(1) N.B. Cech and C.G. Enke, Mass Spectrom. Rev. 20, 362–387 (2001).
(2) R.B. Cole, J. Mass Spectrom. 35, 763–772 (2000).
(3) P. Kebarle and L. Tang, Anal. Chem. 65, 3654–3668 (1993).
(4) C.A. Enke, Anal. Chem. 69, 4885–4893 (1997).
(5) T. Santa, O.Y. Al-Dirbashi, and T. Fukushima, Drug Discov. Today 1, 108–118 (2007).
(6) H.P. Nguyen, L. Li, J.W. Gatson, D. Maass, J.W. Wigginton, J.W. Simpkins, and K.A. Schug, J. Pharm. Biomed. Anal. 54, 830–837 (2011).
(7) S.H. Yang, A.A. Morgan, H.P. Nguyen, H. Moore, B.J. Figard, and K.A. Schug, Environ. Toxicol. Chem., 30, 1243–1251 (2011).