Electrospray ionization (ESI) is a dynamic process that, when coupled with mass spectrometry (MS), serves as an invaluable tool for analysis of biomolecules. Our group, as well as others, has observed that there is a bias in signal intensity for one strand of a PCR amplicon over the complementary strand in an ESI mass spectrum. In this report, we have investigated the contributions of hydrophobicity and free energy of solvation to relative signal intensities in ESI-MS spectra of nucleic acids. We developed approaches for predicting which strand of the PCR amplicon will be the most intense: one based on a rate equation for calculating ion flux using values from the literature for hydrophobicity and free energy of solvation and the other based on the percentage of the relatively hydrophilic guanines present in the strand. A trend in signal intensity for deoxyribonucleotide triphosphates, oligonucleotides, and PCR amplicons was observed that was consistent with our model. On the basis of the observation that increased hydrophobicity correlates with greater signal intensity, we selectively enhanced the signal intensity of a 20-mer with the addition of an alkyl chain to the 5' terminus, which subsequently improved the limit of detection to 1 nM, an improvement by 1 order of magnitude. This was extended to a 53-bp PCR amplicon by modifying one primer with the hydrophobic moiety, which resulted in a 16% increase in signal intensity. We capitalized on this result to determine allele frequencies from pooled DNA for single-nucleotide polymorphisms down to 1%.

Electrospray ionization mass spectrometry (ESI-MS) has been utilized to obtain accurate mass measurements of intact PCR products; however, single-stranded PCR products are necessary to detect sequence modifications such as base substitutions, additions or deletions. The locations of these modifications can subsequently be determined using additional stages of mass spectrometry. The recombinant enzyme lambda exonuclease selectively digests one strand of a DNA duplex from a 5' phosphorylated end leaving the complementary strand intact. Using this rapid enzymatic step, we were able to produce single-stranded PCR products by digestion of an intact PCR product derived from the Human Tyrosine Hydroxylase (HUMTHO1) gene, which contains a tetrameric repeating motif. The non-template directed 3' adenylation common when using Taq polymerase resulted in three distinct species (blunt-ended, mono-adenylated and di-adenylated), which added complexity to the spectrum of the double-stranded product. The data from the single-stranded products shows that one strand is preferentially adenylated over the other, which cannot be determined from the mass spectrum of the double-stranded PCR product alone. The ESI-FTICR (Fourier transform ion cyclotron resonance) mass spectra of the lambda exonuclease treated PCR products exhibited less than expected signal-to-noise (S/N) ratios. This is attributed to inaccurate concentration calculations due to remaining double-stranded PCR product amplified with unphosphorylated primers, and to matrix effects contributed by the lambda exonuclease reaction buffer. To further test this hypothesis, we investigated and determined the limit of detection to be 0.27 microM using standard curve statistics for single acquisitions of a synthetic 75-mer. The concentrations of the noncoding and coding strands produced by lambda exonuclease digestion were calculated to be 0.29 and 0.37 microM, respectively, taking into account the presence of double-stranded product. The products were electrosprayed from concentrations at the limit of detection requiring the averaging of 5-10 acquisitions to produce a sufficient S/N ratio, indicating that product concentration, base composition and matrix effects play a combined, significant role in detection of lambda exonuclease treated PCR products. Although additional work will be required to further exploit this strategy, lambda exonuclease clearly provides mass spectrometrists with a method to generate single-stranded PCR products.

The utility of electrospray ionization Fourier transform ion cyclotron resonance (ESI-FIICR) mass spectrometry as a new approach for genotyping short tandem repeats (STRs) is demonstrated. STRs are currently valued as a powerful source of genetic information with repeats that range in structure from simple to hypervariable. Two tetranucleotide STR loci were chosen to evaluate ESI-FTICR mass spectrometry as a tool for genotyping: HUM-TH01, a simple STR with nonconsensus alleles, and vWA, a compound STR with nonconsensus alleles. For HUM-TH01, the genotype (i.e., repeat number of each allele) was determined for each of 30 individuals using mass measurements of double-stranded amplicons. Low-intensity peaks observed in the spectra of amplicons derived from heterozygous individuals were identified by mass as heteroduplexes that had formed between nonhomologous strands. Mass measurement of the double-stranded vWA amplicon was not sufficient for determining whether the individual was homozygous for allele subtype 18 or 18' since the amplicons differ by only 0.99 Da. Therefore, single-stranded amplicons were generated by incorporating a phosphorylated primer, prepared using T4 polynucleotide kinase, into the PCR phase and subsequently digesting the bottom strand using lambda-exonuclease. Accurate mass measurements were obtained for the single-stranded amplicons using internal calibration and the addition of a correction factor to adjust for the natural variation of isotopic abundances, confirming that the individual is homozygous for allele 18. Our results clearly demonstrate that ESI-FTICR mass spectrometry is a powerful approach to characterize both simple and compound STRs beyond the capabilities of electrophoretic technologies.

Electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry (ESI-FTICRMS) has been used to determine the mass of a double-stranded 500 base-pair (bp) polymerase chain reaction (PCR) product with an average theoretical mass of the blunt-ended (i.e. unadenylated) species of 308 859.35 Da. The PCR product was generated from the linearized bacteriophage Lambda genome which is a double-stranded template. Utilization of ethanol precipitation in tandem with a rapid microdialysis step to purify and desalt the PCR product was crucial to obtain a precise mass measurement. The PCR product (0.8 pmol/µL) was electrosprayed from a solution containing 75% acetonitrile, 25 mM piperidine, and 25 mM imidazole and was infused at a rate of 200 nL/min. The average molecular mass and the corresponding precision were determined using the charge-states ranging from 172 to 235 net negative charges. The experimental mass and corresponding precision (reported as the 95% confidence interval of the mean) was 309 406 +/- 27 Da (87 ppm). The mass accuracy was compromised due to the fact that the PCR generates multiple products when using Taq polymerase due to the non-template directed 3'-adenylation. This results in a mixture of three PCR products with nearly identical mass (i.e. blunt-ended, mono-adenylated and di-adenylated) with unknown relative abundances that were not resolved in the spectrum. Thus, the experimental mass will be a weighted average of the three species which, under our experimental conditions, reflects a nearly equal concentration of the mono- and di-adenylated species. This report demonstrates that precise mass measurements of PCR products up to 309 kDa (500 bp) can be routinely obtained by ESI-FTICR requiring low femtomole amounts. Copyright 1999 John Wiley & Sons, Ltd.