Label-assisted laser desorption/ionization mass spectrometry (LA-LDI-MS): an emerging technique for rapid detection of ubiquitous cis-1,2-diol functionality

Abstract

The Label-Assisted Laser Desorption/Ionization (LA-LDI) technique has recently been applied to the detection of small molecules through a Time of Flight (TOF) mass spectrometric measurement. By excluding the external matrix, the mass spectrum becomes a lot cleaner being free of matrix related peaks and noise. In this paper we report a Label-Assisted Laser Desorption/Ionization Mass Spectrometry (LA-LDI-MS) based method for detection of various cis-1,2-diols, including the ones ubiquitous in biological samples.

---

Introduction

Identification of biologically important small molecules is an important goal of bioanalytical chemistry¹ having vast applications in medical diagnostics, microbiology, pharmacology² and environmental sciences. Traditionally, techniques like UV-Vis Spectroscopy,³ Fluorescence Spectroscopy,⁴ Liquid Chromatography-Mass Spectrometry⁵ and High Performance Liquid Chromatography⁶ have been used for this purpose. Other methods include surface plasmon resonance based optical sensors⁷ and microfluidics based sensors.⁸ Mass Spectrometry can be a useful method for profiling a mixture of molecules. Recently in 2013, a novel method for reaction monitoring has been established by Kozmin et al.⁹ using Label-Assisted Laser Desorption/Ionization Time of Flight Mass Spectrometry (LA-LDI-TOF-MS). In this technique, detection of small molecules was carried out using Time of Flight Mass Spectrometer – an instrument which is typically used for detection of macromolecules by MALDI.¹⁰ The most important merit of this method is its high selectivity. Encouraged from this work we envisioned that a pyrene based boronic acid could be useful for the detection of various cis-1,2-diols. The selectivity of the method lies in its working principle. It is only the polyaromatic associated species which will undergo desorption/ionisation and show their presence in the mass spectrum. That is why, cis-1,2-diols, which form complexes with the pyrene based boronic acid should be detected selectively in LA-LDI MS. Other unreacted diols or unreacted contaminants will not interfere in the mass spectrum. This selectivity issue is the main advantage of this method over other available methods in the literature. Although, soft ionization based methods like Desorption Electrospray Ionization MS has been described for small molecules by Cooks et al.,¹¹ LA-LDI-MS offers the advantage of giving highly intense molecular ion peaks, also being very selective at the same time. Recently¹² we have reported LA-LDI-MS method for selective detection of Zn²⁺. In the present work, we have applied the same method for detection of biologically important cis-1,2-diols especially catecholamine neurotransmitters dopamine and epinephrine thus demonstrating the scope of this technique. 1,2-Diol functionality is important because of its presence in various molecules of biological relevance. Abnormal levels of dopamine and its metabolites in cerebrospinal fluid (CSF) are implicated in several neurodegenerative diseases, most notably Parkinson's disease.¹³ Also, other biologically important molecules like ascorbic acid, citric acid and simple sugars were studied.¹⁴

Commercially available MALDI-TOF-MS instruments are equipped with UV lasers. On laser irradiation, the pyrene label readily generates radical cation, which is detectable in MS. This happens due to high molar absorptivity of pyrene or other polyaromatic compounds. Based on the aforementioned facts, we selected pyrene-1-boronic acid (PBA) as a selective probe for identification of the cis-1,2-diol molecules (Fig. 1).

Fig. 1 Pyrene boronic acid and expected complexation behavior with cis-1,2-diols.

Results and discussion

Commercially available PBA was directly used as a probe for cis-1,2-diols without further modification. Its solution was prepared in phosphate buffer (0.1 M, pH 7): THF (1 : 1). The mass spectrum recorded in positive mode showed an intense molecular ion peak at m/z 246. Under the given conditions, it also gives an intense fragmentation peak at m/z 218 corresponding to pyren-1-ol, presumably generated by aerial oxidation during spotting (Fig. 2).¹⁵

Fig. 2 LDI mass spectrum of PBA (+ve mode).

The ability of PBA to complex with cis-1,2-diols and subsequent detection of the complex by LDI-MS was then proved. As an initial attempt, catechol was chosen as the representative cis-1,2-diol because of its well-known complexation behavior with boronic acid. A solution of PBA (2.5 mM) in phosphate buffer (pH 7.0)-THF mixture (1 : 1) was incubated (1 h) with a solution of catechol in THF (1 mM). LDI-MS recorded on the mixture showed a strong peak at m/z 320 (Fig. 3) corresponding to the boronate complex A (in Fig. 4).

Fig. 3 LDI mass spectrum (+ve mode) of PBA–catechol complex (molecular ion peak at m/z = 320).

Fig. 4 ‘Capture’ complexes of catechol and related molecules.

The experiment was repeated with 2,3-dihydroxynaphthalene and the peak at m/z 370 was found in LDI-MS (Table 1, see ESI†) corresponding to the expected complex B (for spectrum see ESI†).

In order to extend the scope of the method, several other compounds with cis-1,2-diol functionality were screened with PBA. These include simple diols like ethylene glycol and tartaric acid to biologically relevant molecules like dopamine, epinephrine, L-DOPA and ascorbic acid. LDI mass spectra were recorded in both positive and negative ion modes (Table 2, see ESI†). An interesting observation was that while ethylene glycol (see ESI† for spectrum), dopamine and epinephrine gave molecular ion peaks for the respective boronate complexes in positive ion mode, the corresponding complexes E and F of tartaric and ascorbic acid respectively (Fig. 5), could be detected in the negative ion mode. One other point to note is the appearance of peaks corresponding to free analytes in the mass spectra recorded in the negative ion mode. For example, for ascorbic acid, the peak at m/z 175 and for L-DOPA, the peak at m/z 196 corresponding to respective molecular ions could be observed. We believed that the unreacted PBA acted as a matrix which in turn, helped the desorption cum ionization of the analytes. The various molecular ions from different analytes along with the proposed structures of the boronate complexes are shown in Fig. 4–6.

Fig. 5 ‘Capture’ complexes of tartaric acid, ascorbic acid and citric acid.

Fig. 6 Structures of various possible species derived from PBA observed in different mass spectra. Sometimes the deprotonated form of the same species arrived in negative mode of the spectra.

As representative examples, LDI mass spectra of boronate complexes of dopamine, epinephrine, L-DOPA and ascorbic acid are shown in Fig. 7–10. Besides the molecular ion peak at m/z 363 (C), dopamine showed fragmentation peaks at m/z 347 (M) and m/z 333 (N), possibly due to loss of NH₂˙ and CH₂=NH₂⁺˙ respectively (Scheme 1).

Fig. 7 LDI mass spectrum (+ve mode) of PBA–dopamine complex.

Fig. 8 LDI mass spectrum (+ve mode) of PBA–epinephrine complex.

Fig. 9 LDI mass spectrum (−ve mode) of PBA–L-DOPA complex.

Fig. 10 LDI mass spectrum (−ve mode) of PBA–ascorbate complex.

Scheme 1 Molecular ion and fragment species from dopamine.

Mass spectrum of epinephrine did not show the molecular ion peak at m/z 393. Instead it gave a prominent fragmentation peak at m/z 376 (shown in inset in Fig. 8), due to the loss of benzylic –OH. There are minor peaks at m/z 474 and m/z 684, which can be attributed to formation of dimer and trimer respectively due to self-condensation of PBA (structures I and J, in Fig. 6). For ascorbic acid, the MS spectrum, recorded on a negative mode, showed a peak at m/z 385 assigned for the complex F. It also showed additional peaks at m/z 613 which is attributed to the bis-complex G. We have also recorded the complexation behaviour of citric acid, an α-hydroxy acid, with PBA, which showed the molecular ion peak at m/z 401 corresponding to the complex H (for spectrum, see ESI†).

Having been successful with the LDI-MS based detection of small molecules including catecholamine neurotransmitters, we turned our attention to sugars. Detection of simple sugars and their derivatives is an important part of analytical biochemistry, having applications in clinical diagnosis, elucidation of biological pathways and reaction mechanisms. Two representative examples, namely D-glucose and D-ribose were chosen. Although stereochemically pure sugars were used for analyses, they have complex solution chemistry, especially in alkaline medium and exist as equilibrium of various cyclic and acyclic forms. So, the exact nature of the captured complexes was not determined. Mass spectra were recorded in both positive and negative ion modes. For ribose, ions were detected in positive mode only at m/z 360 (100% intensity) and m/z 570 (∼20% intensity) – corresponding to complexation with one and two molecules of PBA respectively (for spectrum see ESI†). However, for glucose, the situation was more complex. In the positive ion mode, the spectrum showed a peak at m/z 600 while in the negative ion mode, a peak appeared at m/z 599. With time, the intensity of peak at m/z 600 decreased while that at m/z 599 increased.¹⁶ The latter can arise through loss of one proton from the doubly-complexed glucose molecule. However, the loss of a proton from the doubly complexed species appears to be intriguing since there is apparently no significantly free ionisable group in that species. This can be explained based on a similar observation reported by Norrild et al.,¹⁷ in which, all five –OH groups of α-glucofuranose are reported to be involved in complexation with overall loss of a proton. We believe that the peak at m/z 600 arises due to bis-complexation with α-glucopyranose (structure “O”, Scheme 2), which in aqueous medium slowly converts to the thermodynamically more stable α-glucofuranose complex with m/z 599 via mutarotation followed by coordination of the C₃–OH to the boron centre and concomitant loss of a proton to generate structure “P”. The latter assumes a negative charge and the species shows a peak at m/z 599 recorded in the negative ion mode.

Scheme 2 Structures of two complexes of glucose.

This was further verified by recording the LA-LDI-MS of aliquots taken out from the incubating mixture at different time points in both positive and negative ion modes (Fig. 11). There was a gradual decrease of peak intensity at m/z 600 (O) with concomitant increase of the peak intensity at m/z 599 (P) thus supporting the glucopyranose to glucofuranose conversion.

Fig. 11 Time dependent variation of intensities of peaks at m/z 600 (+ve mode) (left panel) and m/z 599 (−ve mode) (right panel); time of data recording: 30 min (red), 4 h (pink), 8 h (violet), 12 h (blue), 24 h (black).

We have also checked the possible interference by glucose on the detection of dopamine whose complex with PBA showed up only in the positive ion mode. Solutions of dopamine of different strengths (100 μM, 50 μM, 10 μM) were kept incubated with PBA (2.5 mM) in presence of large excess of glucose (10 mM). Peaks for the dopamine–PBA complex showed up in the positive ion mode LDI-MS even when glucose is present in 1000-fold excess by molar ratio. Thus presence of glucose has no interference with the detection of dopamine. The peaks for the glucose–PBA complex at m/z 599 were obtained in negative ion mode (all relevant spectra are included in ESI†).

Sensitivity of detection of dopamine was also determined thereafter using similar reaction conditions. In presence of 2.5 mM pyrene-1-boronic acid, it was found that dopamine is detectable in the concentration as low as 5 μM, which amounts to 5 picomoles per 1 μL spotting (Fig. 12).¹⁸

Fig. 12 Concentration dependent variation of intensities of peaks for dopamine at m/z 363, m/z 347 and m/z 334 (+ve mode).

One more point that needs to be mentioned here. Apart from selectivity, the LA-LDI MS has advantage over MALDI-MS for the detection of small molecules in terms of quality of the spectra. This can be demonstrated by comparing the spectra of the same analyte, dopamine, using these two techniques separately. The LA-LDI mass spectrum (Fig. 13(i)) clearly shows the peaks for dopamine (complexed with PBA) at m/z 363, 347 and 333 along with pyren-1-ol (m/z 218) originating from PBA. The MALDI mass spectrum (Fig. 13(ii) and (iii)) shows the peaks from dopamine at m/z 154 (dopamine + H⁺), 137 (from α-cleavage of NH₂) and 123 (from benzylic cleavage); however, appearance of several other peaks arising from the matrix make the assignment difficult. The analysis may become even more uncertain if masses of the fragments arising from the matrix have same or very close m/z values to those of the analyte. Between MALDI and LA-LDI, MALDI is more sensitive as indicated by the absolute intensity for same concentration of dopamine. This is understandable as the matrix is used in large excess. However, LA-LDI has the following advantages: it gives much cleaner spectrum; the spectrum mainly shows the peak for the complex while in MALDI, the peaks from the analyte and from the matrix may overlap as can be seen in case of dopamine (Fig. 13). We believe that the sensitivity of LA-LDI can be increased further by using a label with larger polyaromatic network. We are currently checking other polyaromatic labels. We have already synthesized phenyl ethynyl pyrene boronic acid (PEBA). Initial studies showed that the spectra are much cleaner and ions due to dimer or trimer are much less (Fig. 14 and 15).

Fig. 13 LDI MS (+ve mode) of dopamine (50 μM) taken in presence of (i) PBA (ii) external matrix CHCA and (iii) external matrix + PBA (the yellow shades show the relevant peaks (matrix : analyte = 4000 : 1 (molar ratio).

Fig. 14 LA-LDI (+ve mode) spectrum of the synthesized probe.

Fig. 15 LA-LDI (+ve mode) spectrum of the synthesized probe and dopamine complex (intensity is much higher, and fragmentation is much less).

Since a biological fluid will contain all the catecholamine neurotransmitters as well as other small analytes, we were curious to know whether the method can detect them all in a mixture. Thus we have recorded LA-LDI mass spectrum on a mixture of dopamine, epinephrine, L-DOPA, glycine and L-alanine dissolved in a solution of THF and buffer (1 : 1, pH 8). To this solution PBA was added keeping its concentration 2.5 mM and the final concentration of each neurotransmitter was 30 μM. The solution was incubated for 1 h and 1 μL was spotted for LDI MS study. In the positive ion mode we have found peaks corresponding to dopamine (at m/z 363, 347 and 334) and epinephrine (at m/z 376). In the negative ion mode we have found peak corresponding to L-DOPA (at m/z 406) (Fig. 16).

Fig. 16 LDI MS of a mixture dopamine, epinephrine, L-DOPA, glycine and L-alanine incubated with PBA and recorded on (i) positive ion mode (ii) and negative ion mode.

Conclusion

Thus we have developed a new label-assisted LDI mass spectrometry based technique for the detection of small molecules with cis-1,2-diol functionality via selective capture.¹⁹ The method has used the commercially available pyrene boronic acid acting as a dual role of a label as well as a ligand. The method has been demonstrated to be successful in detecting biologically important neurotransmitters catecholamines and also sugars. The method is free from cross interference. We believe that the technique will be useful in detecting such small molecules in biological specimen. Current efforts are aimed towards that direction.

Acknowledgements

DST is acknowledged for an SERC grant and for JC Bose National Fellowship to AB which supported this research. PSA is grateful to CSIR, Govt. of India for a Senior Research Fellowship. ANB thanks DST, Govt. of India for INSPIRE Undergraduate Scholarship.

References and notes

1. (a) M. McKeague and M. C. DeRosa, J. Nucleic Acids, 2012, 2012, 1; (b) J. Park, D. Bang, K. Jang, E. Kim, S. Haam and S. Na, Nat. Commun., 2014, 5, 1.
2. (a) H. Glauner, I. R. Ruttekolk, K. Hansen, B. Steemers, Y. D. Chung, F. Becker, S. Hannus and R. Brock, Br. J. Pharmacol., 2010, 160, 958; (b) Z. E. Perlman, M. D. Slack, Y. Feng, T. J. Mitchison, L. F. Wu and S. J. Altschuler, Science, 2004, 306, 1194; (c) P. Lang, K. Yeow, A. Nichols and A. Scheer, Nat. Rev. Drug Discovery, 2006, 5, 343.
3. (a) A. Manz, N. Graber and N. M. Widmer, Sens. Actuators, B, 1990, 1, 244; (b) D. J. Harrison, A. Manz, Z. Fan, H. Ludi and M. H. Widmer, Anal. Chem., 1992, 64, 1926; (c) S. Hiki, K. Mawatari, A. Hibara, M. Tokeshi and T. Kitamori, Anal. Chem., 2006, 78, 2859.
4. (a) Z. Dai and J. W. Canary, New J. Chem., 2007, 31, 1708; (b) J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, New York, 2006; (c) B. Valeur, Molecular Fluorescence Principles and Applications, Wiley-VCH, New York, 2002; (d) Y. Choi, Y. Park, T. Kang and L. P. Lee, Nat. Nanotechnol., 2009, 4, 742.
5. (a) J. J. Pitt, Clin. Biochem. Rev., 2009, 30, 19; (b) C. Lauridsen, S. W. Leonard, D. A. Griffin, D. C. Liebler, T. D. McClure and M. G. Traber, Anal. Biochem., 2001, 289, 89.
6. G. J. Van Berkel and V. Kertesz, Rapid Commun. Mass Spectrom., 2013, 27, 1329.
7. (a) K. V. Kong, Z. Lam, W. K. O. Lau, W. K. Leong and M. Olivo, J. Am. Chem. Soc., 2013, 135, 18028; (b) S. K. Srivastava, R. Verma and B. D. Gupta, Sensors and Actuators B., 2011, 153, 194.
8. H. F. Li and J. M. Lin, Anal. Bioanal. Chem., 2009, 393, 555.
9. J. R. Cabrera-Pardo, D. I. Chai, S. Liu, M. Mrksich and S. A. Kozmin, Nat. Chem., 2013, 5, 423.
10. (a) K. Dreisewerd, Chem Rev., 2003, 103, 395; (b) D. C. Muddiman, R. Bakhtiar, S. A. Hofstadler and R. D. Smith, J. Chem. Edu., 1997, 74, 1288.
11. H. Chen, I. Cotte-Rodríguez and R. Graham Cooks, Chem. Commun., 2006, 597.
12. P. S. Addy, S. B. Roy, S. M. Mandal and A. Basak, RSC Adv., 2014, 4, 23314.
13. (a) K. Hyland, Future Neurol., 2006, 1, 593; (b) T. Suominen, P. Uutela, R. A. Ketola, J. Bergquist, L. Hillered, M. Finel, H. Zhang, A. Laakso and R. Kostiainen, PloS. One, 2013, 8, 1; (c) S. Hisahara and S. Shimohama, Int. J. Med. Chem., 2011, 2011, 1.
14. (a) O. E. Owen, S. C. Kalhan and R. W. Hanson, J. Bio. Chem., 2002, 277, 30409; (b) K. Iqbal, A. Khan and M. M. A. K. Khattak, Pak. J. Nutri., 2004, 3, 5.
15. (a) A. N. Cammidge, V. H. Goddard, C. P. Schubert, H. Gopee, D. L. Hughes and D. Gonzalez-Lucas, Org. Lett., 2011, 15, 6034; (b) Y.-Q. Zou1, J.-R. Chen, X.-P. Liu, L.-Q. Lu1, R. L. Davis, K. A. Jørgensen and W.-J. Xiao, Angew. Chem., Int. Ed., 2012, 51, 784.
16. After 24 h, the intensities of peaks at m/z 600 and at m/z 599 are ∼5% and ∼16% respectively. This observation indicated a slow conversion of complex “O” to complex “P”.
17. M. Bielecki, H. Eggert and J. C. Norrild, J. Chem. Soc., Perkin Trans. 1, 1999, 2, 449.
18. All the experiments were carried out using Voyager-DE PRO instrument. Linear mode; UV laser: N2 LASER, 337 nm wavelength; LASER Intensity 3200; LASER Rep Rate 20.0 Hz. Dopamine and epinephrine were procured in injectable form as dopamine hydrochloride and epinephrine bitartrate respectively. Solutions containing 0.03 mM of dopamine and epinephrine were incubated with 1 mM of pyrene-1-boronic acid in PBS (0.1 M, pH 7) + THF (1 : 1) at room temperature for 1 h. The mass spectra were recorded in positive ion mode.
19. Detection of large macromolecules like proteins via selective capture by photo-cross linking followed by MALDI MS is a currently active area: (a) H. Köster, D. P. Little, P. Luan, R. Muller, S. M. Siddiqi, S. Marappan and P. Yip, Assay Drug Dev. Technol., 2007, 5, 381; (b) P. S. Addy, B. Saha, N. D. P. Singh, A. K. Das, J. T. Bush, C. Lejeune, C. J. Schofield and A. Basak, Chem. Commun., 2013, 49, 1930; (c) J. T. Bush, L. J. Walport, J. F. McGouran, I. K. H. Leung, G. Berridge, S. S. van Berkel, A. Basak, B. M. Kessler and C. J. Schofield, Chem. Sci., 2013, 4, 4115.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra07499h

---