Mass Spectroscopic Investigation of Nucleic Acid Degradation Products

The possibilities of structure elucidation of nucleic acid degradation products and of sequencing of oligonucleotides are discussed. There are three classes of natural products, the sequencing of which has been attempted by mass spectrometry, viz. oligosaccharides, peptides and nucleotides. While oligosaccharides offer some special problems due to structural variability, peptides and nucleotides are built according to the same basic principle: a backbone of recurrent structural units which is substituted at regular intervals by characteristic groups. In case fragmentation of 'the backbone occurs in a specific way (as at the amide bOnds in peptides) sequential information may be obtained from, a mass spectrum. There are, however, two further premises for the successful application of' mass spectrometry: it must be possible to distinguish the two ends of the strand from each other and it is necessary to obtain ions of the entire molecule or at least of parts of it still carrying sequential information. The backbone of nucleotides is an alternating sequence of sugar (ribose, 1, R = OH, and desoxyribose, !, R = H) and phosphoric acid residues which is substituted N-glycosidically by purine and pyrimidine bases (Fig 1) HOCH2 ?H H0C112_ ?ase]Base(.lor2H) Sugar JBase,3O HOR HOOH 1 The mass spectrometric behavior of pyrimidine and purine bases has been investigated in exteneo (Ref. 1), but is of minor importance in this context, since due to the stability of the aromatic nucleus, fragment formation in nucleosides and nucleotides takes place preferentially in the sugar/phosphate units. Identification of a base in bigger units has, therefore, to come from mass and elemental composition of pertinent fragments. That nucleosides (i.e. base ÷ sugar) are amenable to El mass spectrometry (with the possible exception of guanosine which shows severe pyrolytic degradation) was shown first by Biemann (2). They exhibit a rather straight forward fragmentation pattern (Ref's. 2, 3) (2), though the abundance of the typical fragment ions varies (Ref. 1). CI, upon suitable selection of the auxilliary gas, may enhance pseudo_M+ (M ÷ 1) and somewhat simplify the spectrum (Ref. 4). Fl (Ref. 5) and FD (Refs. 6, 7) spectra are dominated by M+ (occasionally EM H20]+) and show clearly the sugar and base fragments (Fig. 2). Mass spectra of nucleosides are especially important for the detection of unusual bases (Ref's. 1, 8-10) and for biogenetic studies by isotope labelling (Ref. 11). 160 HERBERT BUDZIKIEWICZ Fig. 1. Primary structure of ribonucleotides. Addition of a phosphate unit (i.e. formation of a mononucleotide) renders these compounds too involatile for direct El mass spectrometry. Various types of derivatives (TMS, methyl ethers, acetates, acetonates, phenyl boronic acid esters) have been suggested (Ref. 1) to enhance volatility. FD (Ref s. 6, 7, 12) allows the measurement of free nucleotides giving EM + H]+ and ions characteristic of the base and the sugar moiety (Fig. 3). Although not of much use for structural work — such information is extracted much more readily from the corresponding nucleosides which are also easier to handle they may serve as models to discuss problems encountered with higher units: Derivatisation. Trimethyl silylation is readily achieved (per—TMS derivatives allow even analysis by GC/MS; CI (i-butane) enhances the abundance of M4 (Ref. 13) ), but since each TMS group adds 72 amu to the molecular weight the limitation is obvious (Ref. 14). Methylation, depending upon the method used usually results in mixtures (Ref s. 7, 13). Differentiation between 3'and 5'—phosphates. FD spectra of the two isomers are within the range of variability (v. nfrcz) indistinguishable (Figs. 3c and 3d). At the best, it may be observed that loss of H20 occurs at lower temperatures for 3'—phosphates due to cyclophosphate formation (Ref .7). It has been reported by McCloskey (14) that a differentiation is possible with El spectra of per-TMS derivatives. Comparing the spectra of 3t and 5'-adenosine monophosphoric acid he found that m/e 169 (part of the sugar moiety) and EM P02(OTMS) 2J are both characteristically more abundant in the 5t-isomer. Examination of the other three pairs showed (Ref. 13) that higher abundance of mfe 169 for the 5'—isomer could be observed also for GMP and UMP, but not for CMP, and that EM P02 (OTMS) 2]+ was of about the same abundance for the GMP and CMP pairs and far more abundant for U—3' -MP. However one can rely on m/e 501 EM — base]+ which, in all four cases, is of much higher intensity for the 3'-isomer (Table 1) (Ref. 13). A differentiation between 2— and 3'—isomers seems not to be possible for any of the nucleoside phosphoric acids, owing to isomerisation (Ref s. 13, 14). Problems in FD work (Ref. 12). FD spectra of nucleotides change with the anode temperature. Starting from [M + HJ+ as sole ion in the spectrum (Fig. 3a) of A-3'—MP with increasing temperature fragments appear (m/e 136, [adenine + H]+ in Fig. 3b), increase in variety (loss of 1 or 2 H20 from EM + H]+ and from EA + H]4, m/e 250 and 232, EM adenine H20 + HJ+, m/e 195 and H4P044, m/e 99 in Fig. 3c). With higher temperature uncharacteristic ions appear. Hence a suitable temperature has to be selected for the 5' End