Anal.Chem. 2005, 77, 673-680
`
`Analysis of Biological and Synthetic Ribonucleic
`Acids by Liquid Chromatography-Mass
`Spectrometry Using Monolithic Capillary Columns
`
`Georg Ho1 lzl,† Herbert Oberacher,‡ Stefan Pitsch,§ Alfred Stutz,§ and Christian G. Huber*,§,|
`
`InstituteofAnalyticalChemistryandRadiochemistry,Leopold-Franzens-University,Innrain52a,andInstitueofLegal
`Medicine,InnsbruckMedicalUniversity,Mu¨llerstrasse44,A-6020Innsbruck,Austria,InstituteofOrganicChemistry,
`EcolePolytechniqueFederaledeLausanne,Ecublens,1015Lausanne,Switzerland,andInstrumentalAnalysisand
`Bioanalysis,DepartmentofChemistry,SaarlandUniversity,Building9.2,66123Saarbru¨cken,Germany
`
`Ion-pair reversed-phase high-performance liquid chro-
`matography (IP-RP-HPLC) has been evaluated as a method
`for the fractionation and desalting of ribonucleic acids
`prior to their characterization by electrospray ionization
`mass spectrometry. Monolithic, poly(styrene-divinylben-
`zene)-based capillary columns allowed the rapid and
`highly efficient fractionation of both synthetic and biologi-
`cal ribonucleic acids. The common problem of gas-phase
`cation adduction that is particularly prevalent in the mass
`spectrometric analysis of ribonucleic acids was tackled
`through a combination of chromatographic purification
`and the addition of ethylenediaminetetraacetic acid to the
`sample at a concentration of 25 mmol/L shortly before
`on-line analysis. For RNA molecules ranging in size from
`10 to 120 nucleotides, the mass accuracies were typically
`better than 0.02%, which allowed the characterization and
`identification of failure sequences and byproducts with
`high confidence. Following injection of a 500 nL sample
`onto a 60 (cid:2) 0.2 mm column, the limit of detection for a
`120-nucleotide ribosomal RNA transcript from Escheri-
`chia coli was in the 50-80 fmol range. The method was
`applied to the analysis of synthetic oligoribonucleotides,
`transfer RNAs, and ribosomal RNA. Finally, sequence
`information was derived for low picomole amounts of a
`32-mer RNA upon chromatographic purification and
`tandem mass spectrometric investigation in an ion trap
`mass spectrometer. Complete series of fragment ions of
`the c- and y-types could be assigned in the tandem mass
`spectrum. In conclusion, IP-RP-HPLC using monolithic
`capillary columns represents a very useful tool for the
`structural investigation and quantitative determination of
`RNAs of synthetic and biological origin.
`
`RNAs play an essential role in a variety of biochemical
`processes, in which they serve as a temporary copy of genes that
`
`* To whom correspondence should be addressed at Saarland University.
`Phone: +49
`302-3433. Fax: +49
`(0)681
`(0)681
`302-2963. E-mail:
`Christian.Huber@mx.uni-saarland.de.
`† Leopold-Franzens-University.
`‡ Innsbruck Medical University.
`§ Ecole Polytechnique Federale de Lausanne.
`| Saarland University.
`
`10.1021/ac0487395 CCC: $30.25 © 2005 American Chemical Society
`Published on Web 12/17/2004
`
`is used as a template for protein synthesis, function as adaptor
`molecules for translation of the genetic code, and catalyze the
`biochemical synthesis of proteins. In contrast to double-stranded,
`genomic DNA, RNA is a single-stranded polynucleotide that
`spontaneously folds into a variety of secondary and tertiary
`structures such as hairpins, bulges, pseudoknots, and internal
`loops.1 These structural elements can either constitute binding
`sites for regulatory proteins or directly mediate a biological
`process.2 In addition to these classical roles in cell biology, RNA
`is emerging as a key regulator of gene expression3 and has been
`shown to be capable of catalyzing a range of chemical reactions
`both in vivo and in vitro.4 Because of these multiple functions of
`RNA in many biochemical contexts, powerful analytical tools are
`required to determine their structure, sequence, purity, concentra-
`tion, and spatial distribution. Moreover, the rapidly growing
`demand for chemically synthesized RNA necessitates the elabora-
`tion of suitable protocols for high-throughput quality assurance.
`Both electrospray ionization mass spectrometry (ESI-MS)5 and
`matrix-assisted laser desorption/ionization mass spectrometry
`(MALDI-MS)6 have been successfully applied as analytical meth-
`ods to the characterization of in vivo RNA transcripts,7 transfer
`RNA,8,9 synthetic RNA,10 ribosomal RNA subunits,11,12 and RNA-
`RNA complexes.13
`One characteristic of RNAs is their high affinity for metal ions,
`particularly bivalent cations that stabilize proper secondary and
`
`(1) Chastain, M.; Tinoco, I. Jr. Prog. Nucleic Acid Res. Mol. Biol. 1991, 41,
`131-77.
`(2) Toulme´, J. J.; Le Tine´vez, R.; Brossalina, E. Biochimie 1996, 78, 663-73.
`(3) Filipowicz, W. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 14035-37.
`(4) Latham, J. A.; Cech, T. R. Science 1989, 245, 276.
`(5) Yamashita, M.; Fenn, J. B. J. Chem. Phys. 1984, 88, 4451-59.
`(6) Karas, M.; Bachmann, D.; Bahr, U.; Hillenkamp, F. Int. J. Mass Spectrom.
`Ion Processes 1987, 78, 53-68.
`(7) Kirpekar, F.; Nordhoff, E.; Kristiansen, K.; Roepstorff, P.; Lezius, A.; Hahner,
`S.; Karas, M.; Hillenkamp, F. Nucleic Acids Res. 1994, 22, 3866-70.
`(8) Limbach, P. A.; Crain, P. F.; McCloskey, J. A. J. Am. Soc. Mass Spectrom.
`1995, 6, 27-39.
`(9) Little, D. P.; Thannhauser, T. W.; McLafferty, F. W. Proc. Natl. Acad. Sci.
`U.S.A. 1995, 95, 2318-22.
`(10) Hahner, S.; Lu¨demann, H.-C.; Kirpekar, F.; Nordhoff, E.; Roepstorff, P.; Galla,
`H.-J.; Hillenkamp, F. Nucleic Acids Res. 1997, 25, 1957-64.
`(11) Kirpekar, F.; Krogh, T. N. Rapid Commun. Mass Spectrom. 2001, 15, 8-14.
`(12) Kowalak, J. A.; Pomerantz, S. C.; Crain, P. F.; McCloskey, J. A. Nucleic Acids
`Res. 1993, 21, 4577-85.
`(13) Hoyne, P. R.; Benson, L. M.; Veenstra, T. D.; Maher III, L. J.; Naylor, S.
`Rapid Commun. Mass Spectrom. 2001, 15, 1539-47.
`
`AnalyticalChemistry,Vol.77,No.2,January15,2005 673
`
`1
`
`MTX1044
`
`

`

`tertiary structures.14 This represents a significant problem in ESI-
`MS of RNAs, as the substitution of some of the protons in the
`sugar-phosphate backbone by metal cations results in the
`formation of a multitude of adducted species. Sensitive and
`accurate mass measurements of RNAs are severely hampered by
`such cation adduction, because the numerous signals for the
`higher charge states of adducted species merge into one broad
`peak shifted to higher mass relative to that of the fully protonated
`species. Effective removal of cations is, therefore, essential to
`obtain mass spectra of high quality from which the correct
`molecular masses of RNAs can be deduced. As a consequence,
`several purification and desalting protocols such as multiple
`ethanol precipitation,8 addition of organic bases, 8,15 chelating
`agents,8 or cation-exchange beads,13 solid-phase extraction,16
`affinity adsorption,10 denaturing polyacrylamide gel electrophore-
`sis,10,13 and liquid chromatography,17-19 as well as various com-
`binations of the aforementioned methods, have been devised for
`the removal of cations and other impurities.
`The prime prerequisites that have to be met for the applicability
`of purification protocols to routine analysis of RNAs are high
`desalting efficiency, low sample requirement, high analyte recov-
`ery, and potential for full automation. In this regard, we have
`employed miniaturized ion-pair reversed-phase high-performance
`liquid chromatography (IP-RP-HPLC) using monolithic, poly-
`(styrene-divinylbenzene) (PS-DVB)-based capillary columns for
`the purification of oligodeoxynucleotides and double-stranded
`DNA prior to investigation by ESI-MS.20 Nevertheless, while IP-
`RP-HPLC has been shown to efficiently remove cationic adducts
`from DNA, the affinity especially of Mg2+ ions for RNA is still
`strong enough to endure the chromatographic process. And
`although Taniguchi and Hayashi have shown that HPLC of RNAs
`in 150 (cid:2) 0.3 mm columns packed with PS-DVB perfusion
`particles and a gradient of acetonitrile in 0.3 mmol/L tributylam-
`monium acetate (pH 5.5) enabled the successful characterization
`of prepurified transfer RNAs,17 we found that the desalting
`efficiency of this approach was not sufficient for ESI-MS analysis
`of unpurified transfer RNAs or raw products of RNA solid-phase
`synthesis. In this paper, we therefore report on an attempt to
`combine on-line liquid chromatographic purification under dena-
`turing conditions17,21 and addition of chelating agents8 to efficiently
`desalt RNAs for ESI-MS analysis. This combination is shown to
`be very effective, permitting highly accurate mass measurements
`with small amounts of both synthetic and biological RNAs. Finally,
`IP-RP-HPLC in combination with tandem mass spectrometry is
`utilized to derive sequence information about the purified RNAs.
`
`MATERIALS AND METHODS
`Chemicals and RNA Samples. Acetonitrile (HPLC gradient
`grade) and disodium ethylendiaminetetraacetate dihydrate (EDTA;
`
`(14) Bukhman, Y. V.; Draper, D. E. J. Mol. Biol. 1997, 273, 1020-31.
`(15) Simmons, T. A.; Limbach, P. A. Rapid Commun. Mass Spectrom. 1997, 11,
`567-72.
`(16) Berhane, B. T.; Limbach, P. A. J. Mass Spectrom. 2003, 38, 872-78.
`(17) Taniguchi, H.; Hayashi, N. Nucleic Acids Res. 1998, 26, 1481-86.
`(18) Patteson, K. G.; Rodicio, L. P.; Limbach, P. A. Nucleic Acids Res. 2001, 29,
`e49.
`(19) Schu¨rch, S.; Bernal-Me´ndez, E.; Leumann, C. J. J. Am. Soc. Mass Spectrom.
`2002, 13, 936-45.
`(20) Huber, C. G.; Oberacher, H. Mass Spectrom. Rev. 2001, 20, 310-43.
`(21) Premstaller, A.; Oberacher, H.; Huber, C. G. Anal. Chem. 2000, 72, 4386-
`93.
`
`674 AnalyticalChemistry,Vol.77,No.2,January15,2005
`
`analytical reagent grade) were purchased from Sigma-Aldrich (St.
`Louis, MO), and butyldimethylamine (analytical reagent grade)
`was purchased from Fluka (Buchs, Switzerland). A 0.50 M stock
`solution of butyldimethylammonium bicarbonate (BDMAB) was
`prepared by passing carbon dioxide gas (AGA, Vienna, Austria)
`through a 0.50 M aqueous solution of butyldimethylamine at 5
`(cid:176)C until pH 8.4 was reached. For preparation of all aqueous
`solutions, high-purity water (Epure, Barnstead Co., Newton, MA)
`was used.
`5S ribosomal RNA from a 50S ribosomal subunit from Escheri-
`chia coli strain MRE600 was obtained from Sigma-Aldrich.
`Synthetic RNAs (32-mer, GGC GUU UUC GCC UUC GGG CGA
`UUU UUA UCG CU, Mr ) 10124.9; 55-mer, AGC GCC GAU GGU
`AGU GUG GGG UCU CCC CAU GCG AGA GUA GGG AAC UGC
`CAG GCA U, Mr ) 17837.9) were prepared by solid-phase
`synthesis22 and used either without purification or with purification
`by means of anion-exchange HPLC (13-mer, 32-mer, 55-mer) or
`reversed-phase HPLC (21-mer).
`The chemical synthesis and purification of aminoacylated
`tRNAs (protected E. coli Leu-tRNAGly-1), prepared by chemical
`ligation, GGG AGA AUA GCU CAG UUG GUA GAG CAC-(2¢-O-
`NPEOM)-GAC CUU CUA AAG GUC GGG GUC GCG AGU UCG
`AGU CUC GUU UC-(2¢ -O-NPEOM)-U CCC UCC A-{(3¢ -O-
`NPEOM)-(2¢-O-[(N-NPEOC)-Leu])}, Mr ) 25276.6, and the pro-
`tected E. coli Leu-tRNAGly-2, prepared by enzymatic ligation, pGGG
`AGA AUA GCU CAG UUG GUA GAG CAC GAC CUU CUA AAG
`GUC GGG GUC GCG AGU UCG AGU CUC GUU UCU CCC UCC
`A-{(3¢ -O-NPEOM)-(2¢ -O-[(N-NPEOC)-Leu])}, Mr ) 24997.2;
`NPEOM ) [1-(2-nitrophenyl)ethoxy]methyl; NPEOC ) [1-(2-
`nitrophenyl)ethoxy]carbonyl) have been described in detail in refs
`23-25.
`High-Performance Liquid Chromatography Interfaced
`with Electrospray Ionization Mass Spectrometry. Monolithic
`capillary columns (60 (cid:2) 0.20 mm i.d.) were prepared according
`to the published protocol.21 The HPLC system consisted of a low-
`pressure gradient micropump (model Rheos 2000, Flux Instru-
`ments, Basel, Switzerland) controlled by a personal computer, a
`vacuum degasser (Knauer, Berlin, Germany), a column thermostat
`made from 3.3 mm o.d. copper tubing which was heated by means
`of a circulating water bath (model K 20 KP, Lauda, Lauda-
`Ko¨nigshofen, Germany), and a microinjector (model C4-1004,
`Valco Instruments Co. Inc., Houston, TX) with a 500 nL internal
`sample loop.
`ESI-MS was performed with a quadrupole ion trap mass
`spectrometer (LCQ, Thermo Finnigan, San Jose, CA) equipped
`with a modified nanospray ion source. The spray capillary (fused
`silica, 90 (cid:237)m o.d., 20 (cid:237)m i.d.; Optronis, Kehl, Germany) was
`directly connected to the capillary column by means of a T-piece
`(Upchurch Scientific, Oak Harbor, WA). A syringe pump equipped
`with a 250 (cid:237)L glass syringe (Unimetrics, Shorewood, IL) and the
`T-piece were used for adding a makeup flow of 3.0 (cid:237)L/min of
`acetonitrile. For analysis with pneumatically assisted ESI, an
`electrospray voltage of 5.0 kV and a nitrogen sheath gas flow of
`
`(22) Pitsch, S.; Weiss, P. A.; Jenny, L.; Stutz, A.; Wu, X. Helv. Chim. Acta 2001,
`84, 3773-95.
`(23) Stutz, A. Methoden zur Ligation von Oligoribonukleotiden und ihre Anwen-
`dung bei der Totalsynthese aminoacylierter tRNAs. Dissertation, Swiss
`Federal Institute of Technology, Zu¨rich, 2003.
`(24) Stutz, A.; Pitsch. S. Helv. Chim. Acta 2000, 83, 2477-503.
`(25) Pitsch, S. Chimia 2001, 55, 60-63.
`
`2
`
`

`

`100 arbitrary units were employed. The temperature of the heated
`capillary was set to 250 (cid:176)C. Total ion chromatograms and mass
`spectra were recorded on a personal computer with the LCQ
`Navigator software version 1.2 (Thermo Finnigan) with the
`biomass deconvolution module. Mass calibration and coarse
`tuning were performed in the positive ion mode by direct infusion
`of a solution of caffeine (Sigma, St. Louis, MO), methionylargin-
`ylphenylalanylalanine (Finnigan), and Ultramark 1621 (Finnigan).
`Fine-tuning for ESI-MS of oligonucleotides in the negative ion
`mode was performed by infusion of 3.0 (cid:237)L/min of a 20 pmol/(cid:237)L
`solution of (dT)24 in 25 mmol/L aqueous BDMAB containing 10%
`acetonitrile (v/v).
`
`RESULTS AND DISCUSSION
`Effect of a Chelating Agent on Desalting Efficiency. RNAs
`feature four potential sites for complexing metal ions, namely, the
`strongly acidic phosphodiester groups, the ribose hydroxyls, the
`nitrogens of the nucleobases, and the exocyclic keto groups of
`the nucleobases, the phosphate groups and the hydroxyl groups
`of which are the preferred binding sites.26 As a consequence,
`RNAs have high affinity for cations both in aqueous solution and
`in the gas phase. Upon transfer of the RNA molecules from the
`liquid phase to the gas phase during electrospray ionization,
`cations may or may not remain associated with the RNAs.
`Different molecular species result with various degrees and types
`of cationization that differ in their molecular masses and show,
`depending on their charge states and the resolution of the
`employed mass analyzer, multiple or broadened signals in the
`mass spectrum. In the ion-pair reversed-phase chromatographic
`purification of RNAs using BDMAB as a mobile phase additive,
`the suppression of cationic adducts is based upon the exchange
`of metal cations with butyldimethylammonium ions from the
`mobile phase during the chromatographic process. In contrast to
`metal ions, butyldimethylammonium ions may dissociate into
`volatile butyldimethylamine and a proton that is left with the
`analyte molecules in the course of desolvation of analytes through
`the electrospray process, resulting in the reduction of metal cation-
`adducted species.
`On the basis of the efficacy of desalting by means of IP-RP-
`HPLC, which enabled the mass spectrometric investigation of low
`femtomole amounts of double-stranded DNA restriction fragments
`up to a size of 500 base pairs,21,27 we could expect that the method
`should perform equivalently in the analysis of RNAs. The analysis
`of 14 pmol of a synthetic 55-mer RNA by IP-RP-HPLC-ESI-MS is
`illustrated in Figure 1a. The sample was chromatographed at 70
`(cid:176)C with a gradient of 4 -56% acetonitrile in 25 mmol/L BDMAB
`in 10 min. The poor signal-to-noise ratio in the raw mass spectrum
`(Figure 1a) as well as the multiple broad peaks in the deconvo-
`luted mass spectrum (inset in Figure 1a) clearly proved that
`desalting was far from sufficient. Moreover, the lowest molecular
`mass of 17876 Da observed in the mass spectrum differed by 39
`Da from the theoretical mass of 17837.9 Da of the 55-mer and
`indicated the prevalence of a potassium adduct. Two more intense
`signals were observed in the deconvoluted mass spectrum, which
`
`(26) Saenger Wolfram. Metal Ion Binding to Nucleic Acids. In Principles of Nucleic
`Acid Structure; Cantor, C. R., Ed.; Springer-Verlag: New York, 1993; pp 201-
`219.
`(27) Oberacher, H.; Oefner, P. J.; Parson, W.; Huber, C. G. Angew. Chem., Int.
`Ed. 2001, 40, 3828-30.
`
`Figure 1. IP-RP-HPLC-ESI-MS of synthetic 55-mer RNA: column,
`PS-DVB monolith, 60 (cid:2) 20 mm i.d.; mobile phase, (A) 25 mmol/L
`BDMAB, pH 8.40, (B) 25 mmol/L BDMAB, pH 8.40, 80% acetonitrile;
`linear gradient, 5-70% B in 10.0 min;
`flow rate, 2.0 (cid:237)L/min;
`temperature, 70 (cid:176) C; sheath liquid, 3.0 (cid:237)L/min acetonitrile; sheath gas,
`N2, 100 units; scan range, 500-2000 u; electrospray voltage, 5.0 kV;
`sample, (a) 14 pmol dissolved in water, (b) 14 pmol dissolved in 25
`mmol/L EDTA solution.
`
`were offset by 59 and 78 Da from the fully protonated molecule,
`probably mixed sodium/potassium and bipotassium adducts. The
`adduct-free species was visible in the mass spectrum only as a
`very small signal. Clearly, additional measures had to be taken to
`improve mass spectral quality.
`The potential of removing cations from RNA by complexation
`with chelating agents such as EDTA, nitrilotriacetic acid, or trans-
`1,2-diaminocyclohexane-N,N,N¢ ,N¢ -tetraacetic acid has been dem-
`onstrated by Limbach et al.8 The agents were added at an
`approximately 3-fold molar excess to solutions of RNA, which had
`been prepurified by ethanol precipitation before their investigation
`by direct infusion ESI-MS. In this approach, however, the excess
`of complexing agent has to be kept as low as possible, because
`the agent is not separated from the analytes during direct infusion
`analysis, which may result in signal suppression due to competitive
`ionization. Chromatographic purification of RNAs may be advanta-
`geous in that complexed cations and excess chelating agents elute
`in the flow-through, while purified RNA is first retained on the
`column and subsequently eluted by an acetonitrile gradient.
`Such an experiment is demonstrated in Figure 1b, in which
`the mass spectrum obtained by IP-RP-HPLC-ESI-MS analysis
`upon addition of EDTA to a final concentration of 25 mmol/L to
`a 28(cid:237)mol/L sample solution (900-fold molar excess) is shown.
`Multiply charged ions with charge states from 11- to 22- were
`clearly distinguishable in the mass spectrum and deconvoluted
`into an intact molecular mass of 17837 Da, which deviated only
`
`AnalyticalChemistry,Vol.77,No.2,January15,2005 675
`
`3
`
`

`

`0.0050% from the theoretical molecular mass of the fully protonated
`species. Nevertheless, a monopotassium adduct at about 30%
`relative abundance was still observed in the deconvoluted mass
`spectrum. This result confirms earlier investigations, which
`showed that also monovalent cations, especially potassium, are
`very important in forming secondary and tertiary structures and
`can be found in tetraloops with a strong binding affinity.28 Variation
`of the amount of EDTA added to the sample solution revealed
`that a minimum concentration of 1 mmol/L was necessary to
`induce effective cation complexation. Furthermore, increasing the
`concentration to 25 mmol/L had no untoward effect on chroma-
`tography, while desalting efficiency was slightly increased. Hence,
`we recommend the addition of 20-25 mmol/L EDTA for efficient
`cation removal, especially in the case of samples containing high
`salt loads.
`The differences between RNA and DNA in the conditions
`required for efficient removal of cations probably rest within the
`divergence of the three-dimensional structures of the molecules.
`The high negative charge density of nucleic acids attracts a dense
`“cloud” of positive counterions which are usually delocalized and
`not associated with specific sites. These delocalized ions interact
`primarily through long-range Coulombic forces with the sugar-
`phosphate backbone and remain fully hydrated. The strength of
`interaction is determined by the charge of the ion and the
`magnitude of the electrostatic field. Double-helical DNA and
`regular single-stranded RNA are surrounded by such delocalized
`ions that can be rather easily replaced by other cations. This can
`change significantly in RNA with a globular structure containing
`pockets that are able to chelate the cations and displace some or
`all of its shell of tightly bound water molecules. This direct
`chelating of the ions makes it very difficult for mobile-phase
`additives to exchange the metal ions with protons.29
`Desalting of Ribosomal RNA. To evaluate the effectiveness
`of desalting for larger RNAs forming quite stable secondary
`structures, we utilized the 120-nucleotide ribosomal 5S RNA (5S
`rRNA) subunit as analyte for IP-RP-HPLC-ESI-MS analysis.
`Figure 2 illustrates the analysis of 2 pmol (75 ng) of 5s rRNA
`strain MRE600 from E. coli. The components were separated at
`70 (cid:176) C with a gradient of 4-56% acetonitrile in 25 mmol/L BDMAB
`in 10 min without and with the addition of EDTA to the sample
`solution prior to injection (Figure 2a,d). In the absence of EDTA,
`the spectrum revealed charge states from 21- to 51- with very
`poor signal-to-noise ratios. The two molecular masses of 38 915
`and 38 957 Da obtained from the deconvoluted mass spectrum
`were significantly higher with respect to the expected masses of
`the target compounds of 38 814.4 and 38 853.4 Da, which could
`be found only as a small peak and a shoulder in the mass spectrum
`(Figure 2c). This result demonstrates that cation adduction is also
`prevalent in biological RNA samples, necessitating additional
`desalting. Figure 2d depicts the chromatogram of the sample after
`ion-EDTA
`addition of EDTA. Due to the elution of metal
`complexes and excess EDTA, the intensity of the flow-through
`peak increased significantly, while that of the RNA peak remained
`practically constant. The effect of cation reduction can be clearly
`seen in the improved quality of the raw mass spectrum (Figure
`2e), from which the molecular masses of 38 812 and 38 852 Da
`
`(28) Basu, S.; Rambo, R. P.; Strauss-Soukup, J.; Cate, J. H.; Ferre´-D’Amare´, A.;
`Strobel, S. A.; Dounda, J. A. Nat. Struct. Biol. 1998, 5, 986-92.
`(29) Draper, D. E.; Misra, V. K. Nat. Struct. Biol. 1998, 5, 927-30.
`
`676 AnalyticalChemistry,Vol.77,No.2,January15,2005
`
`Figure 2. Analysis of 5S ribosomal RNA from E. coli strain
`MRE600: sample, (a) 2 pmol of 5S rRNA in water, (b) 2 pmol of 5S
`rRNA dissolved in 25 mmol/L EDTA solution. Other conditions as in
`Figure 1.
`
`were deduced with no signs of cation adduction (Figure 2f). The
`two detected masses correspond to transcripts of the two 5S
`ribosomal RNA genes that occur in E. coli and correspond
`excellently to the theoretical molecular masses.30
`In a previous report, HPLC was not recommended as a suitable
`purification method for RNA, mainly because of a large sample
`requirement, low analyte recovery, and possible degradation of
`RNAs in the presence of heavy metal ions.8 Nevertheless, using
`an HPLC pump with all wetted parts made of titanium or sapphire
`and a stainless steel injector which has been passivated with 1
`mol/L nitric acid, we observed neither degradation products nor
`any signs of sample loss, as evidenced by blank injections that
`showed no memory signals. In the next step we evaluated the
`limits of detection of the IP-RP-HPLC-ESI-MS method for the
`analysis of large RNA molecules by analyzing a sample of 160
`fmol of a 5S RNA sample containing 25 mmol/L EDTA. The
`
`(30) Specht, T.; Wolters, J.; Erdmann, V. A. Nucleic Acids Res. 1991, 19 (Suppl.),
`2189-91.
`
`4
`
`

`

`quality of the raw mass spectrum was adequate to calculate the
`two masses with relative mass deviations of 0.0060% and 0.011%,
`respectively. On the basis of signal-to-noise ratios in the mass
`spectrum of approximately 10:1 to 6:1, the detection limit for the
`5S RNA using this analytical system was estimated to be 50-80
`fmol. With analysis times of less than 5 min, including sample
`preparation, this example clearly demonstrates the potential of
`on-line combined separation and analysis using IP-RP-HPLC-ESI-
`MS compared to other commonly applied techniques that require
`both more time and more sample material.
`Application to the Quality Control of Synthetic Oligoribo-
`nucleotides. Synthetic oligoribonucleotides are considered as
`potential
`therapeutic agents for artificial regulation of gene
`expression.31 Recent experiments have shown that double-
`stranded RNAs of approximately 20-25 nucleotides are able to
`induce posttranscriptional gene silencing in animals and plants.
`It has been demonstrated that short-chain RNA duplexes specif-
`ically suppressed expression of endogenous and heterologous
`genes in different mammalian cell lines.32 The most prominent
`byproducts of chemical oligoribonucleotide synthesis are failure
`sequences resulting from incomplete chain elongation during
`solid-phase synthesis and partially protected sequences due to
`incomplete removal of protecting groups after complete assembly
`of the target sequence. As an example for the analysis of synthetic
`oligoribonucleotides, Figure 3a illustrates the reconstructed ion
`current chromatogram of a deprotected raw product of a synthetic
`21-mer. Upon application of a shallow gradient of 2-12% aceto-
`nitrile in 25 mmol/L BDMAB in 10 min, the analytical system
`was capable of separating and identifying sequences from 7- to
`20-mers together with the target product within 7 min. The results
`of the mass determinations and the identification of the different
`sequences, as well as the relative mass deviations, are summarized
`in Table 1. The mass spectrum extracted from the peak eluting
`at 3 min showed a predominant signal at m/z 1035.4 representing
`the 2- charge state of the eluting oligoribonucleotide. A molecular
`mass of 2072.8 Da was calculated and correlated with the mass
`of the 7-mer, which constitutes the smallest detectable failure
`sequence. The following peaks were identified as the 8-21-mers,
`only the 12- and 13-mers, the 17- and 18-mers, and the 19-, 20-,
`and 21-mers of which coeluted at least partially.
`The relatively high abundance of failure sequences, especially
`from 12- to 20-mers, indicates that the average coupling efficiencies
`during synthesis should be at least 98% to obtain the target in
`reasonable yields. Since the target sequence is generally the
`longest sequence generated by solid-phase synthesis, products
`eluting after the target sequence are most likely sequences from
`which removal of protecting groups was not exhaustive. A mass
`spectrum extracted from the peak eluting at 7 min yielded a
`molecular mass of 6707.2 Da, equivalent to a mass difference of
`54.8 Da relative to the mass of the target sequence. This mass
`difference is compatible with an additional cyanoethyl protecting
`group attached to the 21-mer target sequence (theoretical mass
`difference 54.1 Da) and indicated that
`the removal of
`the
`cyanoethyl groups by elimination was incomplete.
`The raw oligoribonucleotide obtained after solid-phase syn-
`thesis usually has to be purified by preparative gel electrophoresis
`(31) Tuschl, T. ChemBioChem 2001, 2, 239-45.
`(32) Elbashir, S. M.; Harborth, J.; Lendeckel, W.; Yalcin, A.; Weber, K.; Tuschl,
`T. Nature 411, 494-98.
`
`Figure 3. Quality control of a synthetic 21-mer oligoribonucleotide:
`mobile phase, (A) 25 mmol/L BDMAB, pH 8.40, (B) 25 mmol/L
`BDMAB, pH 8.40, 40% acetonitrile; linear gradient, 5-30% B in 10
`min; sample, (a) 250 ng of raw product, (b) 250 ng of the product
`purified by reversed-phase chromatography at 25 (cid:176)C dissolved in 25
`mmol/L EDTA solution. Other conditions as in Figure 1.
`
`Table 1. Measured and Theoretical Masses of a
`Synthetic 19-mer and Its Failure and Partially
`Deprotected Sequences
`
`oligonucleotide
`
`7-mer
`8-mer
`9-mer
`10-mer
`11-mer
`12-mer
`13-mer
`14-mer
`15-mer
`16-mer
`17-mer
`18-mer
`19-mer
`20-mer
`21-mer
`21-mer + cyanoethyl
`
`molecular mass (Da)
`retention
`time (min) measured theoretical
`
`3.09
`3.51
`4.09
`4.44
`4.70
`4.91
`4.91
`5.09
`5.27
`5.43
`5.78
`5.78
`6.07
`6.07
`6.22
`7.03
`
`2072.8
`2418.8
`2751.5
`3078. 5
`3407.3
`3736.5
`4041.3
`4347.4
`5654.1
`5000.3
`5329.1
`5673.0
`6003. 0
`6307.9
`6654.0
`6708.0
`
`2074.3
`2419.5
`2748.7
`3078.0
`3407.2
`3736.4
`4010.6
`4347.7
`4653.9
`4999.1
`5328.3
`5673.6
`6002.8
`6309.0
`6653.2
`6707.2
`
`rel dev
`(%)
`-0.075
`-0.032
`0.100
`0.018
`0.003
`0.003
`-0.008
`-0.008
`0.005
`0.022
`0.015
`-0.010
`0.004
`-0.001
`0.013
`0.011
`
`or chromatography. In our example, the raw product was chro-
`matographed at room temperature on a 250 (cid:2) 4.6 mm column
`packed with an octadecyl stationary phase with a gradient of
`acetonitrile in 0.1 mol/L triethylammonium acetate. The peak of
`the main product was isolated and lyophilized. The success of
`preparative purification was checked by capillary IP-RP-HPLC-
`ESI-MS (Figure 3b). The chromatogram showed two dominant
`peaks in which four different oligoribonucleotides were identified
`on the basis of their molecular masses. Coeluting with the target
`
`AnalyticalChemistry,Vol.77,No.2,January15,2005 677
`
`5
`
`

`

`Figure 4.
`(a) Total synthesis of a 76-mer aminoacylated transfer
`RNA by ligation of a 27-mer, a 41-mer, and an aminoacylated 8-mer
`RNA. Only the coupling reaction between the 41-mer and the
`aminoacylated 8-mer is shown. PG ) protecting group, and EDC )
`N-(3-dimethylaminopropyl)-N¢ -ethylcarbodiimide. The template is a 2¢ -
`O-methylated RNA with a sequence complementary to that to be
`ligated. (b) Structure of the synthetic Leu-tRNAGly.
`
`sequence were the 19- and 20-mer sequences, which could
`obviously not be sufficiently separated by preparative reversed-
`phase HPLC. All other failure sequences have been removed quite
`efficiently, except the 12-mer that eluted as a clear and abundant
`peak at 4.9 min. This surprising finding could be explained by a
`strong interaction between the 12-mer and the longer sequences,
`which resulted in a shift of elution of the 12-mer toward longer
`retention time in the preparative separation, which was performed
`under nondenaturing conditions. Moreover, the small peak of the
`mono(cyanoethyl)-protected 21-mer was still present. To prepare
`RNA sequences of good quality, HPLC purifications, even of
`relatively short sequences, need to be carried out under denatur-
`ating conditions (e.g., anion-exchange HPLC at pH 11.5 and 25
`(cid:176)C or at pH 7 and 80 (cid:176)C), 31 thereby avoiding the formation of
`structures which result in peak-broadening and insufficient separa-
`tion of failure sequences and partially protected products.
`Verification of the Total Chemical Synthesis of Aminoacyl-
`tRNA. In the translation of the genetic code into an amino acid
`sequence, aminoacyl transfer RNAs serve as the carriers of the
`amino acids during the ribosomal biosynthesis of proteins. The
`total chemical synthesis of transfer RNAs facilitates the specific
`incorporation of alternative or modified amino acids, which
`constitutes a powerful tool for site-directed modification and
`mutagenesis. Automated solid-phase synthesis enables, in prin-
`ciple, the assembly of linear chains of polynucleotides up to chain
`lengths of 100 and more.22,33 Nevertheless, because of low total
`
`(33) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859-62.
`
`678 AnalyticalChemistry,Vol.77,No.2,January15,2005
`
`Figure 5. Analysis of synthetic E.coliLeu-tRNAGly obtained by (a)
`chemical (Leu-tRNAGly-1) and (b) enzymatic (Leu-tRNAGly-2) ligation
`of 27-mer, 41-mer, and aminoacylated 8-mer RNA blocks: sample,
`250 ng of product dissolved in 25 mmol/L EDTA solution. Other
`conditions as in Figure 1.
`
`yields even at high coupling efficiencies per individual cycle as
`well as contaminations with failure sequences that are difficult to
`remove from the target, the total synthesis of tRNA is very
`challenging. Hence,
`it is strongly advisable to assemble the
`biopolymers by the conjunction of smaller pieces of oligomers
`that are easier to synthesize and purify. The preparation of
`aminoacyl transfer RNA described in this example includes the
`synthesis and purification of two RNA blocks and an aminoacyl-
`RNA fragment and their subsequent ligation as outlined schemati-
`cally in Figure 4.
`Figure 5 illustrates the IP-RP-HPLC-ESI-MS analysis of 250
`ng ((cid:24)10 pmol) of the transfer RNA product after purification by
`polyacrylamide gel electrophoresis. The sample was separated at
`70 (cid:176) C with a gradient of 4-56% acetonitrile in 25 mmol/L BDMAB
`in 10 min. While both slab gel electrophoresis and chromatogra-
`phy using UV detection revealed only a single band or peak,
`respectively, mass spectrometry readily identified additional
`components in the sample mixture. The deconvoluted spectrum
`of the product peak that eluted at 6.85 min showed the target
`product with a molecu