Over the past two decades automated Sanger sequencing

Transcription

1 On the Use of ESI-QqTOF-MS/MS for the Comparative OnSequencing the Use of of ESI-QqTOF-MS/MS Nucleic Acids for the Comparative Sequencing of Nucleic Acids Herbert Oberacher, Florian Pitterl Institute of Legal Medicine, Innsbruck Medical University, Innsbruck, Austria Received 8 September 2008; revised 19 January 2009; accepted 19 January 2009 Published online 2 February 2009 in Wiley InterScience ( DOI /bip ABSTRACT: The usability of a quadrupole quadrupole time-offlight (QqTOF) instrument for the tandem mass spectrometric sequencing of oligodeoxynuleotides was investigated. The sample set consisted of 21 synthetic oligodeoxynucleotides ranging in length from 5 to 42 nucleotides. The sequences were randomly selected. For the majority of tested oligonucleotides, two or three different charge states were selected as precursor ions. Each precursor ion was fragmented applying several different collision voltages. Overall 282 fragment ion mass spectra were acquired. Computer-aided interpretation of fragment ion mass spectra was accomplished with a recently introduced comparative sequencing algorithm (COMPAS). The applied version of COMPAS was specifically optimized for the interpretation of information-rich spectra obtained on the QqTOF. Sequences of oligodeoxynucleotides as large as 26-mers were correctly verified in >94% of cases (182 of 192 spectra acquired). Fragment ion mass spectra of larger oligonucleotides were not specific enough for sequencing. Because of the occurrence of extensive internal fragmentation causing low sequence coverage paired with a high probability of assigning fragment ions to wrong sequences, tandem mass spectra obtained from oligonucleotides consisting of 30 and more nucleotides Correspondence to: Herbert Oberacher; herbert.oberacher@i-med.ac.at Contract grant sponsor: Austrian Research Promotion Agency (FFG), Österreichisches Sicherheitsforschungs-Förderprogramm KIRAS - eine Initiative des Bundesministeriums für Verkehr, Innovation, Technologie (BMVIT) Contract grant number: VC 2009 Wiley Periodicals, Inc. could not be used for sequence verification neither manually nor with COMPAS. # 2009 Wiley Periodicals, Inc. 91: , Keywords: electrospray ionization; tandem mass spectrometry; comparative sequencing; oligodeoxynucleotide; time-of-flight This article was originally published online as an accepted preprint. The Published Online date corresponds to the preprint version. You can request a copy of the preprint by ing the editorial office at biopolymers@wiley.com INTRODUCTION Over the past two decades automated Sanger sequencing has served as the most widely used analytical tool for DNA sequencing. 1 Sanger sequencing offers high quality data and long read lengths. Thus, it can be used to establish the identity of both known and unknown sequence specific nucleotide variations and has been proclaimed the gold standard against which other technologies must be judged. Applications span numerous research interests, including sequence variation studies, comparative genomics, forensics, and diagnostics. Sanger sequencing, however, is rather time-consuming, laborious, and expensive. Thus, several alternative sequencing strategies have been introduced in recent years. 2 Among those techniques mass spectrometry (MS) represents a promising method especially suited for the characterization of small or modified nucleic acids, which are hardly amenable to standard sequencing methods. The widespread use of mass spectrometric methods for nucleic acids research was triggered by the invention of the soft ionization techniques electrospray ionization (ESI) and matrixassisted laser desorption/ionization (MALDI). 3 8 Applications of mass spectrometric methods include quality control of synthetic oligonucleotides, characterization of DNA adduct forma- Volume 91 / Number 6 401

2 402 Oberacher and Pitterl tion, typing of genetic markers, study of DNA methylation, and DNA sequencing. Mass spectrometric sequencing of nucleic acids often relies on the reconstruction of sequences from measuring molecular mass differences between members of oligonucleotide ladders synthesized via molecular biological methods such as degradation, cleavage, or chain termination synthesis. Despite considerable success in obtaining sequence information of oligonucleotides consisting of [100 nucleotides (nts), the combination of various enzymatic steps with mass spectrometric measurements usually renders enzymatic assays too costly and time consuming for large-scale application. Sequencing based on tandem mass spectrometric analysis of nucleic acid molecules represents an alternative approach to enzymatic techniques. The term tandem mass spectrometry (MS/MS) summarizes mass spectrometric methods concerned with the selection of a particular ion (5 precursor ion) and its activation to generate characteristic secondary fragment ions The principles of MS/MS of nucleic acids have been reviewed recently. 12 Oligonucleotides consist of a limited number of different nucleotides and the bonds between these building blocks are well known that are preferably broken in MS/MS experiments. 13,14 In the majority of cases, collisioninduced dissociation (CID) experiments are used to obtain structural information from multiply charged nucleic acid ions. Other techniques applied for oligonucleotide fragmentation with Fourier transform mass spectrometry include infrared multiphotone dissociation, electron capture dissociation, and blackbody infrared radiative dissociation. 15 For oligonucleotides, CID typically produces a n -B n - and w n -type fragment ions whereas for oligoribonucleotides c n - and y n -type ions dominate. Different types of mass spectrometers can be applied for activation and scanning of fragment ions. These can be classified either as tandem-inspace"- or as tandem-in-time"-instruments. Tandem-intime MS, as implemented on an ion-trap, is the process whereby precursor ions are created, stored in a trapped ion cell, and then sequentially fragmented to form product ions by translational excitation and subsequent collision with a background gas. Tandem-in-time MS has been extensively used to study the fragmentation mechanisms of oligonucleotides, 12 and was successfully applied for the sequence verification of oligonucleotides consisting of [50 nts Tandemin-space experiments are typically performed on instruments consisting of three elements. In the first segment, which is most often a quadrupole filter, precursor ions are selected for fragmentation, which is accomplished in the subsequent part of the instrument (5 gas-filled collision cell). In the third section of the instrument which can either be another quadrupole, a linear ion trap, or a time-of-flight mass analyzer, the fragment ions are scanned. Several groups have studied the fragmentation behavior of oligonucleotides in tandemin-space MS/MS In comparison with tandem-in-time MS the beam-type collision activation allows for multiple competitive dissociation reactions to be observed, which tend to give more extensive structural information for short oligonucleotides. 12 However, whether this type of fragmentation can be used for the sequence verification of larger oligonucleotides remained questionable. A distinct advantage of tandem mass spectral sequencing is its inherent speed of data generation. Fragment ion mass spectra can be obtained very rapidly in a time frame of several seconds on fragmentation and subsequent mass analysis of the fragments directly in a mass spectrometer. Data interpretation, however, still represents a bottleneck. The complexity of fragment ion mass spectrum interpretation increases with the length of the precursor ion rendering manual interpretation of MS/MS spectra a difficult and time-consuming task. Although computational data interpretation routines are available, MS/MS has hardly been used for sequence elucidation of oligonucleotides larger than 30 nts. 16 We have recently introduced a comparative sequencing algorithm (COMPAS) for the computer-aided interpretation of fragment ion mass spectra generated by ESI-MS/ MS. 18,28,29 By applying the algorithm for the interpretation of data obtained from CID in a quadrupole ion trap instrument, we were able to successfully verify the sequences of oligodeoxynucleotides as large as 80-mers. 17,18 Inthisreport,wehaveevaluatedtheusabilityoftandemin-space MS on a quadrupole quadrupole time-of-flight (QqTOF) mass spectrometer for the generation of sequencespecific fragment ion mass spectra. Experiments with synthetic oligonucleotides consisting of 5 42 nts were conducted to study the extent of nucleic acid fragmentation by CID in a linear multipole collision cell. Each oligonucleotide was fragmented using several different collision energy settings. An updated version of COMPAS, which had been specifically optimized for the interpretation of information-rich spectra obtained from tandemin-space experiments, was used for sequence verification. The impact of oligonucleotide length, collision voltage, and charge state of the precursor ion on the principle usability of tandemin-space MS for sequence verification was studied. RESULTS AND DISCUSSION Evaluation of the Usability of Fragment Ion Mass Spectra Acquired on the QqTOF for Comparative Sequencing We have recently shown that the COMPAS is a valuable tool for the interpretation of data obtained from CID of nucleic

3 Use of ESI-QqTOF-MS/MS for the Sequencing of Nucleic Acids 403 Table I Summary of Oligodeoxynucleotides Analyzed in this Study Sequence Length (nt) Charge state: Collision Voltage Range (V) 5-mer_ ACGTA : mer_ CCGAT : mer_ TTAGC : mer_ CGTATTAGCC : mer_ ATTTGTACGT : mer_ GCTCGGAATC : mer_ CGTATTAGCCACGTA : : mer_ ATAGCAGTCCGATTC : : mer_ ACGCATTACGGCGGT : : mer_ GCACCCATTACCCGAATAAA : : : mer_ TGCACTCCAGCCTGGGCAAC : : mer 5 0 -TTGGTGCACCCATTACCCGAAT : : mer 5 0 -TCCAGAGACAGACTAATAGGAGGT : : mer_ GGTAGATAGACTGGATAGATAGACGA : mer_ GTCTTACAATAACAGTTGCTACTATT : mer_ CCCTAGTGGATGATAAGAATAATCAGTATG : : mer_ GGACAGATGATAAATACATAGGATGGATGG : mer 5 0 -GTGCTGAACTAACCAACGCTCCGAAACGACTGAA : mer 5 0 -TTCAGTCGTTTCGGAGCGTTGGTTAGTTCAGCACTTC : mer 5 0 -GCCCAGTGTAGAGCTATGTTAGCATTTAGGTTTTAAGTTAA mer 5 0 -TAAAACCTAAATGCTAACATAGCTCTACACTGGGCTCTAGAG : acids in tandem-in-time experiments. We were able to successfully verify the sequences of oligodeoxynucleotides as large as 80-mers. 17,18 In this report, the usability of spectra generated by tandem-in-space MS on a QqTOF instrument for sequence verification of nucleic acids was evaluated. The QqTOF was operated in a way that parent ion selection in the first Q, parent ion dissociation in the second q, and product ion mass analysis in TOF occur sequentially in space as ions traverse the instrument. The sample set consisted of 21 synthetic oligodeoxynucleotides ranging in length from 5 to 42 nts (Table I). The sequences were randomly selected. For the majority of tested oligonucleotides two or three different charge states were selected as precursor ions. Each precursor ion was fragmented applying several different collision voltages. The most intense fragment ions detected belonged to the a n -B n - and w n -type ions (Figure 1a). Formation of these species includes loss of a base and subsequent cleavage of the 3 0 C O of the sugar from which the base has been lost. As expected, positions containing thymine exhibited comparably lower signal intensities for the corresponding fragment ions. Overall 282 fragment ion mass spectra were acquired and used as input for automated sequence verification with the COMPAS algorithm. The principles of the applied algorithm are described in detail elsewhere. 17,18,28 A summary of the comparative sequencing results obtained from untreated spectra is given in Figure 2a. Only for 35% of all spectra, the correct sequence was retrieved as sequencing result. The highest number of correct results was obtained for the

4 404 Oberacher and Pitterl FIGURE 1 Fragment ion mass spectrum of 10-mer_1. (a) Full spectrum obtained from the triply deprotonated oligodeoxynucleotide using a collision voltage of 35 V. (b) Expansion of the spectrum close to the precursor ion. 22-mer (93.8%). COMPAS completely failed, however, to verify the sequences of the 5-mers and of all oligonucleotides consisting of 30 and more nucleotides. Obviously, in comparison with data obtained from tandem-in-time MS on an ion trap instrument the spectra measured by tandem-inspace MS on a QqTOF seem to carry a higher load of unspecific signals hampering sequence verification with COMPAS. To reduce the complexity of the spectra and to improve the performance of the algorithm, the tandem mass spectral data was treated. In a first filtering step heavier isotopic peaks were removed. The resolution of the mass analyzer reached a value of 9000 according to the full width at half-maximum definition. Thus, the isotopic patterns of multiply charged fragment ions were resolved (Figure 1b). Generally, the number of isotope peaks observed and their relative signal intensities depend on the chemical formula of the fragment and the natural isotopic composition of its constituent elements. For oligonucleotides and their fragment ions, the monoisotopic peak mass is the sum of the masses of the lightest isotopes from each element in the molecule and represents the measured FIGURE 2 Oligodeoxynucleotide length dependence of correct comparative sequencing results. (a) Untreated spectra were used as input for the sequencing algorithm. (b) Sequencing results obtained after removing isotopic peaks. (c) Sequencing results obtained after removing isotopic peaks and signals with intensities smaller than 0.5% of the intensity of the most intense fragment ion.

5 Use of ESI-QqTOF-MS/MS for the Sequencing of Nucleic Acids 405 variable which is most useful for sequencing. All other isotopic peaks just increase spectral complexity and could be misinterpreted. Thus, we tried to remove these species by filtering the tandem mass spectral data. First, all signals were sorted in increasing m/z order. Next, all signals were deleted that were not [1.1 mass units heavier and less intense than the previous entry in the peak list. The applied strategy has some weak points. Incomplete removing of isotopic peaks might be observed for signals corresponding to large and highly charged fragment ions as for these species the monoisotopic peak might not represent the most intense signal of the isotopic distribution. Furthermore, very extensive removing of isotopic peaks might be observed in the case of overlapping isotopic distributions. Nevertheless, the applied strategy is simple, can be easily automated, and improved the performance of COMPAS significantly. After processing the data, 65% of all sequencing results were correct (Figure 2b). For the oligodeoxynucleotides ranging in length from 5 to 26 nts, overall correct results were obtained in [80% of cases. For the larger oligonucleotides, the sequencing strategy still failed to verify the correct sequence in the vast majority of cases. To gain a further improvement of the sequencing outcome, an intensity cut-off was introduced to suppress matching of unspecific signals. As signal intensities largely depend on the applied collision voltage, a dynamic threshold was used to remove putative noise within sample spectra. The intensity threshold was set by multiplying the intensity of the most intense fragment ion by a user-defined factor. Different settings were tested. Best performance was obtained by applying Therefore, only those signals with intensities larger than 0.5% of the base fragment ion intensity were considered as compound-specific and were allowed to match. The implementation of the intensity cut-off had a positive impact on sequencing efficiency (Figure 2c). Correct sequencing results were retrieved in 66% of cases. For [94% of all 192 spectra collected from the oligodeoxynucleotides ranging in length from 5 to 26 nts, the correct sequence was retrieved as sequencing result. In the vast majority of cases, those spectra which failed to qualify were acquired at a collision voltage of either 15 or 50 V exhibiting rather low or high extensive fragmentation. For larger oligonucleotides even the use of the intensity cut-off" did not improve sequencing efficiency significantly. In Figure 3, a representative sequence correlation diagram of fitness value (FS) versus position obtained from the verification of the 26-mer_2 sequence (z 5 5, collision voltage 5 35 V) is shown. Because of the change of the equation to calculate FS in comparison with the published version, 28 the highest FS value is retrieved from the best matching sequence. All fitness values were divided by the fitness value of the tested reference sequence to improve the comparability of sequence correlation FIGURE 3 Representative sequence correlation diagram of fitness versus position for 26-mer_2 (z 5 5, collision voltage 5 35 V). All fitness values were divided by the fitness value of the tested reference sequence. The height of a triangle represents the FS-to-FS (tested sequence) ratio of the corresponding sequence matched to the sample spectrum. Different colors are used to code the kind of nucleotide incorporated at a certain position of the tested sequence during comparative sequencing (A: green; C: blue; G: black; T: red). diagrams. Triangles were used to represent the different FS-to- FS (tested sequence) ratios of sequences obtained from sequential variation of the tested sequence. Moreover, different colors were used to code the kind of nucleotide incorporated at a certain position. Because of these changes correlation diagrams resembled to a large extent sequencing electropherograms known from Sanger sequencing. Influence of the Collision Energy on the Sequence Verification of Oligodeoxynucleotides Consisting of Equal or <26 Nucleotides For the majority of tested oligonucleotides ranging in length from 5 to 26 nts two or even three different charge states were selected as precursor ions, and each precursor ion was fragmented applying several different collision voltages (Table I). The applied collision energy had a severe impact on the coverage of the sequence with fragment ions. As depicted in Figure 4a, the collision energy best suited for getting utmost sequence coverage (black dots) increased linearly with oligonucleotide length. The most comprehensive data set was collected from 20-mer_1. Triply, quadruply, and quintuply deprotonated oligonucleotide ions were selected as precursor ions. For each precursor ion charge state fragment ion mass spectra were collected at several different collision voltage values. The number of noncovered positions were used to qualify the combined effect of the collision voltage settings and the precursor ion charge state on the comprehensiveness of the sequence information obtained from CID of 20-mer_1. The results are depicted in Figure 4b. A collision energy value of about 160 ev was necessary to obtain utmost sequence coverage. The maximum obtainable sequence coverage was 62.2% for the triply deprotonated species. A value of 86.5% was obtained for the precursor ions

6 406 Oberacher and Pitterl V for the fragmentation of precursor ions carrying one or more charged nucleotides per 7 nts. FIGURE 4 Impact of the collision energy on the coverage of oligonucleotide sequences with fragment ions. (a) Dependence of the collision energy suited to obtain utmost sequence coverage (black dots) as well as to receive correct sequencing results (gray shadowed area) on oligonucleotide length. (b) Effect of the collision energy on the coverage of the 20-mer_1 sequence with fragment ions. carrying more charges. Under these circumstances, only the fragment ions resulting from fragmentation of the phosphate bonds at the 3 0 -side next to thymines were missing. At collision energies below the optimum value, a decreased amount of fragment ions were observed increasing the number of noncovered positions. Likewise, collision energies higher than the optimum value had a negative effect on sequence coverage as extended decomposition of larger fragment ions into smaller fragments was induced. Nevertheless, for all 20- mer_1 spectra but one (z 5 3, collision voltage 5 30 V) the correct sequence was retrieved as COMPAS result. For all oligonucleotides consisting of equal or \26 nts tested, correct sequencing results were retrieved in [94% of cases (182 of 192 spectra acquired). Obviously, a wide range of experimental conditions (gray shadowed area in Figure 4a) were found to be suitable for data acquisition. Sequence verification was even possible with spectra exhibiting sequence coverage as low as 20%, which clearly indicates a considerable tolerance of COMPAS for missing fragment ions. To obtain a sufficient amount of sequence-specific fragment ions, however, we recommend the use of collision voltages ranging from 20 to 45 CID of Oligonucleotides Consisting of 30 and More Nucleotides and Its Impact on Comparative Sequencing Problems were met using MS/MS for the sequencing of oligonucleotides consisting of 30 and more nucleotides. A decrease of the number of sequence specific a n -B n - and w n - type ions in combination with an increase of the number of unspecific ions was considered as putative cause for the observed difficulties to verify the sequences of larger species. To prove this hypothesis, the sequence coverage as well as the number of observed fragment ions were determined in fragment ion mass spectra of the mer oligonucleotides obtained from fragmentation of precursor ions containing on average one negative charge per 5 to 7 nts at a collision voltage of 35 V. Evaluation of the data revealed that the number of noncovered positions as well as the number of detected fragment ions increased with increasing oligonucleotide length (see Figure 5). A closer look at the fragment ion mass spectra of the larger oligonucleotides revealed that small a n - B n - and w n -type ions as well as rather small internal fragment ions were the most intense fragment ions observed (Figure 6a). Internal fragments originate from two subsequent fragmentation steps. 16,22,23 As a result of multiple collisions of the precursor ions with gas molecules internal fragments are known to appear with an extraordinary high-signal intensity on fragmentation within linear multipole collision cells. 22 The QqTOF applied in this study used such a gas-filled cell for CID. The major class of internal fragment ions that were identified in the acquired mass spectra belonged to the (a n - B n )-a m -type ion series (n [ m). These species are formed either by subsequent w cleavage of a n -B n -ions or by a-b cleavage of w n -ions. Interestingly, internal fragments containing thymine exhibited highest signal intensities. Somehow, thymine residues seem to have a stabilizing effect on internal fragments preventing an even more extensive fragmentation. Internal fragments containing adenine on the other hand seemed to represent the most instable ones. The (pn i pf) 2 ions (with N i for the deoxynucleoside at position i, p for phosphate, and f for furan) were used to evaluate the length dependence of the internal fragment ion formation. As can be deduced from Figure 6b, the relative signal intensities of these fragment ions was below 10% within the fragment ion spectra of the species consisting of 26 or less nucleotides, but increased significantly for all larger oligonucleotides. This observation suggests that the occurrence of extensive internal fragmentation renders mass spectrometric sequencing of oli-

7 Use of ESI-QqTOF-MS/MS for the Sequencing of Nucleic Acids 407 CONCLUSIONS Tandem mass spectrometric sequencing of oligonucleotides using ESI-QqTOF-MS/MS for the generation of fragment ion mass spectra represents a valuable tool for the characterization of oligonucleotides smaller than nts. Because of inherent high-mass accuracy and resolution of the QqTOF sequences of short oligonucleotides can be verified with high accuracy. A large variety of experimental conditions can be used to generate sequence-specific information. Applications may include quality control of synthetic oligonucleotides, characterization of DNA adduct formation, and metabolic profiling of nucleic acids used as pharmaceuticals. However, because of extensive secondary fragmentation within the linear multipole collision cell, the QqTOF seems to be an improper platform for the characterization of larger nucleic acids. Thus, for the characterization of genetic markers, the more gentle fragmentation processes occurring within an ion trap seems to yield more sequence-specific information rendering this platform especially in combination with a highperformance instrument like an orbitrap or a Fourier-transform ion cyclotron resonance mass analyzers more suitable for nucleic acids sequencing. FIGURE 5 Oligodeoxynucleotide length dependence of (a) the sequence coverage and of (b) the observed fragment ions. All spectra obtained from the fragmentation of precursor ions containing on average one negative charge per 5 to 7 nts at a collision voltage of 35 V were considered. gonucleotides larger than nts a difficult to accomplished task on a QqTOF instrument. We want to point out that the performance of COMPAS does not represent the limiting factor because in previous studies, we could demonstrate that COMPAS can be used to verify the sequences of oligonucleotides as large as 80-mers by interpretation of fragment ion mass spectra generated on a quadrupole ion trap instrument. 17,18 Obviously, ion trap collision activation, which samples dissociation reactions at lower rates and therefore, lower energy than the beam-type collision activation, 30 seems to be the more appropriate platform for sequencing larger oligonucleotides. FIGURE 6 (a) Fragment ion mass spectrum of 30-mer_2. The five-times deprotonated oligodeoxynucleotide was fragmented using a collision voltage of 35 V. (b) Oligodeoxynucleotide length dependence of the relative signal intensities of internal fragments of the type (pn i pf) 2.

8 408 Oberacher and Pitterl MATERIALS AND METHODS Chemicals and Oligodeoxynucleotides Acetonitrile (HPLC gradient-grade) was obtained from Sigma- Aldrich (St. Louis, MO). A stock solution (1.0M) of cyclohexyldimethylammonium acetate (CycHDMAA) was prepared by titration of a cyclohexyldimethylamine solution (Fluka, Buchs, Switzerland) with acetic acid (Fluka) at 58C until ph 8.3 was reached. For preparation of all solutions, HPLC-grade water (Merck, Darmstadt, Germany) was used. Synthetic oligodeoxynucleotides were obtained from Sigma-Aldrich. The sequences were randomly selected and are summarized in Table I. Mass Spectrometric Measurements ESI-MS and -MS/MS was performed on a QSTAR XL mass spectrometer (Applied Biosystems, Foster City, CA). A modified TurboIonSpray source was used for all experiments. The modifications included the replacements of the Peek tubing transfer line and of the stainless steel sprayer capillary by fused silica capillaries (transfer line: 375 lm o.d., 20 lm i.d., sprayer capillary: 90 lm o.d., 20 lm i.d., Polymicro Technologies, Phoenix, AZ). 31,32 Oligodeoxynucleotides were analyzed in the negative ion mode by continuous infusion of 5 20 pmol/ll solutions in 25 mm aqueous CycHDMAA containing 50% acetonitrile (v/v) at a flow rate of 2.0 ll/min. Cations present in the sample solutions were removed by on-line cationexchange using a 20 mm mm i.d. cation-exchange microcolumn packed with lm Dowex 50 WX8 particles (Serva, Heidelberg, Germany). 33 The spray voltage was set to 3.9 kv. Gas flows of arbitrary units for the nebulizer gas and 40 arbitrary units for the turbo gas were employed. The temperature of the turbo gas was adjusted to 2008C. For MS/MS, the Q1 resolution was set to unit resolution. The collision gas (N 2 ) flow was set to five arbitrary units. The collision energy, which is the translational energy of ions entering the collision cell, was changed by varying the voltage applied to the collision cell (Table I). The accumulation time was set to 1 s. Mass spectra collected over a period of min were averaged. Mass spectra were recorded on a personal computer with the Analyst QS software (service pack 8 and Bioanalyst extension, Applied Biosystems). Computer-Aided Data Interpretation Measured MS/MS spectra were exported from the Analyst software as txt-files. Each file contained a list of the centroided fragment ion m/z-values and the corresponding relative signal intensities. To improve sequencing efficiency, files were processed. Data treatment was accomplished fully automatically using a program written in ActivePerl (Active State Corporation, Vancouver, BC, Canada) and started with the deletion of all signals detected within a range of 65 mass units around the m/z of the precursor ion. Furthermore, all signals having either a signal intensity smaller than 0.5% of the intensity of the most intense fragment ion or which were identified as isotopic peaks were deleted. Those signals were called isotopic peaks that were not [1.1 mass units heavier than a larger peak. The principles of the applied comparative sequencing algorithm (COMPAS) are described in detail elsewhere. 17,18,28 Briefly, COMPAS involves the comparison of a measured MS/MS spectrum to a set of fragment ion m/z values predicted from a reference sequence. The closeness of matching between the measured spectrum and the predicted set of ions is characterized by FS. FS determination starts with the generation of a list of monoisotopic m/z values representing all theoretically possible a n -B n - and w n -ions for the given reference sequence. Then, the predicted m/z values and those obtained from the experimental spectrum are compared and FS is calculated. The maximum tolerable mass deviation for matching predicted and calculated fragment ions was fixed at 0.1. FS takes into account the difference (D) between measured and predicted m/z values, the relative intensity I% of the fragment ions, the number K of fragments assigned, and the number M of nucleotide positions not covered by fragment ions in the experimental spectrum. The following equation was used for calculating FS: K 3 FS ¼ 1 þ 1 P K 1003 D I% 3ð1 þ MÞ 0:25 The larger the value for FS, the closer the match between the measured and the predicted spectra. Finally, to find a sequence most closely matching the experimental spectrum, the tested sequence is sequentially permutated by incorporating all four possible nucleotides A, T, G, and C at each position in the sequence. The correct sequence is then identified by that reference sequence having the highest FS value. To generate a sequence correlation diagram, all fitness values were divided by the fitness value of the tested sequence to improve the comparability. Triangles were used to represent the different FS-to-FS (tested sequence) ratios of sequences obtained from sequential variation of the tested sequence. Moreover, different colors were used to code the kind of nucleotide incorporated at a certain position. Thus, correlation diagrams resembled to a large extent sequencing electropherograms known from Sanger sequencing. All calculations were performed on a personal computer under Windows XP TM operating system (1.7 GHz Pentium, 1.0 GB RAM). Automated comparative sequencing was performed with a program written in ActivePerl (Active State Corporation). A copy of the program for academic use is available on request from the corresponding author. All data were collected, evaluated, and prepared for final output in Microsoft Excel 2002 TM. Correlation diagrams were prepared from the Excel data using Origin 7.0 (OriginLab Corporations, Northampton, MA). REFERENCES 1. Strausberg, R. L.; Levy, S.; Rogers, Y. H. Drug Discov Today 2008, 13, Metzker, M. L. Genome Res 2005, 15, Limbach, P. A. Mass Spectrom Rev 1996, 15, Nordhoff, E.; Kirpekar, F.; Roepstorff, P. Mass Spectrom Rev 1996, 15, Gut, I. G. Hum Mutat 2004, 23, Gross, J.; Hillenkamp, F. Encyclopedia of Analytical Chemistry; Meyers, R. A., Ed.; Wiley: Chichester, Huber, C. G.; Oberacher, H. Mass Spectrom Rev 2001, 20, Oberacher, H. Anal Bioanal Chem 2008, 391, McLuckey, S. A. J Am Soc Mass Spectrom 1992, 3, Shukla, A. K.; Futrell, J. H. J Mass Spectrom 2000, 35,

9 Use of ESI-QqTOF-MS/MS for the Sequencing of Nucleic Acids Gabelica, V.; De Pauw, E. Mass Spectrom Rev 2005, 24, Wu, J.; McLuckey, S. A. Int J Mass Spectrom 2004, 237, McLuckey, S. A.; Berkel, G. J.; Glish, G. L. J Am Soc Mass Spectrom 1992, 3, McLuckey, S. A.; Habibi-Goudarzi, S. J Am Chem Soc 1993, 115, Hofstadler, S. A.; Sannes-Lowery, K. A.; Hannis, J. C. Mass Spectrom Rev 2005, 24, Little, D. P.; Thannhauser, T. W.; McLafferty, F. W. Proc Natl Acad Sci USA 1995, 95, Oberacher, H.; Oefner, P. J.; Hölzl, G.; Premstaller, A.; Huber, C. G. Nucl Acids Res 2002, 30, e Oberacher, H.; Parson, W.; Oefner, P. J.; Mayr, B. M.; Huber, C. G. J Am Soc Mass Spectrom 2004, 15, Barry, J. P.; Vouros, P.; Van Schepdael, A.; Lay, S.-J. J Mass Spectrom 1995, 30, Boschenok, J.; Sheil, M. M. Rapid Commun Mass Spectrom 1996, 10, Crain, P. F.; Gregson, J. M.; McClosckey, J. A.; Nelson, C. C.; Peltier, J. M.; Phillips, D. R.; Pomerantz, S. C.; Reddy, D. M. In Mass Spectrometry in the Biological Science; Burlingame, A. L.; Carr, S. A., Eds.; Humana Press: Totowa, NJ, 1996; p Lotz, R.; Gerster, M.; Bayer, E. Rapid Commun Mass Spectrom 1998, 12, Premstaller, A.; Huber, C. G. Rapid Commun Mass Spectrom 2001, 15, Ni, J.; Chan, K. Rapid Commun Mass Spectrom 2001, 15, Rozenski, J.; McCloskey, J. A. J Am Soc Mass Spectrom 2002, 13, Muhammad, W. T.; Tabb, D. L.; Fox, K. F.; Fox, A. Rapid Commun Mass Spectrom 2003, 17, Kellersberger, K. A.; Yu, E.; Kruppa, G. H.; Young, M. M.; Fabris, D. Anal Chem 2004, 76, Oberacher, H.; Wellenzohn, B.; Huber, C. G. Anal Chem 2002, 74, Oberacher, H.; Mayr, B. M.; Huber, C. G. J Am Soc Mass Spectrom 2004, 15, Huang, T.; Liu, J.; Liang, X.; Hodges, B. D. M.; McLuckey, S. A. Anal Chem 2008, 80, Oberacher, H.; Niederstätter, H.; Parson, W. J Mass Spectrom 2005, 40, Oberacher, H.; Niederstätter, H.; Casetta, B.; Parson, W. Anal Chem 2005, 77, Huber, C. G.; Buchmeiser, M. R. Anal Chem 1998, 70, Reviewing Editor: Dan Fabris