Chiral Stationary Phase Based on a Biostable l-RNA Aptamer

aptamer, was created. It was shown that this mirror-image approach was a very simple and powerful strategy to develop a highly stable stationary phase...
0 downloads 0 Views 116KB Size
Anal. Chem. 2005, 77, 1993-1998

Chiral Stationary Phase Based on a Biostable L-RNA Aptamer Agne`s Brumbt, Corinne Ravelet, Catherine Grosset, Anne Ravel, Annick Villet, and Eric Peyrin*

De´ partement de Pharmacochimie Mole´ culaire UMR 5063 CNRS, Institut de Chimie Mole´ culaire de Grenoble FR 2607, Universite´ Joseph Fourier, UFR de Pharmacie de Grenoble, Avenue de Verdun, 38240 Meylan, France

An immobilized anti-L-arginine D-RNA aptamer, used as a target-specific chiral stationary phase (CSP), was found to be very quickly degraded by RNases under usual chromatographic utilization and storage. To overcome this severe limitation for a practical use, a CSP based on the L-RNA aptamer, that is, the mirror image of the D-RNA aptamer, was created. It was shown that this mirror-image approach was a very simple and powerful strategy to develop a highly stable stationary phase due to the intrinsic insensitivity of L-RNA to the RNase degradation. In addition, such an approach allowed one to reverse the enantiomer elution order relative to that obtained with the corresponding D-RNA CSP. Due to their molecular recognition abilities, competing with that of antibodies, aptamers have found recently several applications in various analytical fields.1 Notably, DNA aptamers have been used successfully as affinity stationary phases in liquid chromatography (LC) and capillary electrochromatography (CEC) for the purification/separation of a variety of molecules, from compounds as large as proteins to compounds as small as amino acids. Romig et al.2 presented in 1999 the first work concerning the use of an immobilized DNA aptamer as stationary phase. An aptamer specific for human L-selectin was applied in the chromatographic purification of recombinant L-selectin-Ig fusion protein from Chinese hamster ovary cell-conditioned medium. McGown and co-workers used, in open-tubular CEC, an antithrombin DNA aptamer and a similar oligonucleotidic sequence that form intramolecular G-quartet structures, to separate various nontarget molecules.3-5 Kennedy and co-workers described a chromatographic stationary phase, based on a DNA aptamer selected against adenosine, to resolve adenosine and related compounds.6 This aptamer affinity column was also used for the adenosine quantification in complex mixtures such as tissue extracts.7 * Corresponding author. E-mail: [email protected]. (1) Jayasena, S. D. Clin. Chem. 1999, 45, 1628. (2) Romig, T. S.; Bell, C.; Drolet, D. W. J. Chromatogr., B 1999, 731, 275. (3) Kotia, R. B.; Li, L.; McGown, L. B. Anal. Chem. 2000, 72, 827. (4) Charles, J. A. M.; McGown, L. B. Electrophoresis 2002, 23, 1599. (5) Rehder, M. A.; McGown, L. B. Electrophoresis 2001, 22, 3759. (6) Deng, Q.; German, I.; Buchanan, D.; Kennedy, R. T. Anal. Chem. 2001, 73, 5415. (7) Deng, Q.; Watson, C. J.; D.; Kennedy, R. T. J. Chromatogr., A 2003, 1005, 123. 10.1021/ac048344l CCC: $30.25 Published on Web 02/08/2005

© 2005 American Chemical Society

More specifically, the enantioselective properties of DNA aptamers selected against a target enantiomer were accounted by our group to create a new class of target-specific chiral stationary phases (CSPs).8,9 These DNA aptamer CSPs were found to be very useful for the chromatographic separation of the enantiomers of an oligopeptide (arginine-vasopressin),8 a nucleoside (adenosine),9 and an amino acid derivative (tyrosinamide).9 An apparent enantioseparation factor of ∼3.5 (at 20 °C) was observed for the anti-D-adenosine aptamer CSP while a very high enantioselectivity was obtained with the anti-D-vasopressin and anti-L-tyrosinamide aptamer CSPs (the target enantiomers were significantly retained by the columns while the nontarget enantiomers were eluted roughly in the void volume). Most of the aptamers reported in the literature are related to RNA sequences (70% of aptamers are RNAs).10 Dissociation constants as low as 6 nM for a small ligand (tobramycin)11 and 200 pM for proteins (VegF,12 bFGF13) have been obtained with RNA aptamers, demonstrating that they can bind their targets with very high affinity. The ability of RNA aptamers to bind stereoselectively their target has been also widely observed.14-18 At the present time, chromatographic or electrophoretic applications using RNA aptamers as target-specific CSPs have not been explored. Only two papers by Clark and Remcho19,20 described the use an immobilized RNA aptamer in open-tubular CEC to separate flavin mononucleotide and thiourea or anthracene.20 The well-known inherent instability of RNA, which is significantly higher than that of DNA due to the ability of the 2′-hydroxyl groups to act as intramolecular nucleophiles in both base- and enzyme-catalyzed hydrolysis,21-23 is expected to be the major (8) Michaud, M.; Jourdan, E.; Villet, A.; Ravel, A.; Grosset, C.; Peyrin, E. J. Am. Chem. Soc. 2003, 125, 8672. (9) Michaud, M.; Jourdan, E.; Ravelet, C.; Villet, A.; Ravel, A.; Grosset, C.; Peyrin, E. Anal. Chem. 2004, 76, 1015. (10) http://aptamer.icmb.utexas.edu. (11) Wang, Y.; Rando, R. R. Chem. Biol. 1995, 2, 281. (12) Jellinek, D.; Green, L. S.; Bell, C.; Janjic, N. Biochemistry 1994, 33, 10450. (13) Jellinek, D.; Lynott, C. K.; Rifkin, D. B.; Janjic, N. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 11227. (14) Connell, G. J.; Illangesekare, M.; Yarus, M. Biochemistry 1993, 32, 5497. (15) Famulok, M. J. Am. Chem. Soc. 1994, 116, 1698. (16) Majerfeld, I.; Yarus, M. Nat. Struct. Biol. 1994, 1, 282. (17) Famulok, M.; Szostak, J. W. J. Am. Chem. Soc. 1992, 114, 3990. (18) Geiger, A.; Burgstaller, P.; von der Eltz, H.; Roeder, A.; Famulok, M.; Nucleic Acids Res. 1996, 24, 1029. (19) Clark, S. L.; Remcho, V. T. Anal. Chem. 2003, 75, 5692. (20) Clark, S. L.; Remcho, V. T. J. Sep. Sci. 2003, 26, 1451. (21) Trawick, B. N.; Daniher, A. T.; Bashkin, J. K. Chem. Rev. 1998, 98, 939. (22) Kaukinen, U.; Lyytikainen, S.; Mikkola, S.; Lonnberg, H. Nucleic Acids Res. 2002, 30, 468.

Analytical Chemistry, Vol. 77, No. 7, April 1, 2005 1993

Figure 1. (a) Sequence and secondary structure of the anti-Larginine 44-mer RNA aptamer. The 6-mer and 10-mer consensus sequences are shown in boldface type (from ref 25) (b) Scheme of the RNA aptamer immobilization strategy.

drawback of such RNA aptamer CSPs. However, no data on the RNA stationary-phase stability have been reported, and the precise lifetime of such affinity columns is still unexplored. In this paper, we report the use of immobilized RNA aptamer as target-specific CSP. A 44-mer D-RNA aptamer specifically selected against L-arginine15 was immobilized on a chromatographic support via a biotin-streptavidin bridge8,9 (Figure 1) and the RNA-modified particles were packed in-house into a microbore column. The enantioselective properties of such a RNA aptamer stationary phase were studied over the 4-17 °C temperature range. When not in use, the column was conditioned in the aqueous buffer (used as mobile phase) as previously described for the DNA CSPs.8,9 To evaluate the D-RNA CSP lifetime under these conditions, the stationary-phase stability was assessed by comparing the L-arginine apparent retention factor during almost three weeks in the same operating conditions. Additional experiments were carried out in order to evaluate the origins of D-RNA CSP instability. Finally, an approach based on the mirror-image strategy was applied to design a RNA aptamer CSP resistant to the enzymatic degradation. EXPERIMENTAL SECTION Reagents and Materials. D-Biotin and D- and L-arginine were obtained from Sigma Aldrich (Saint-Quentin, France). Na2HPO4, NaH2PO4, NaCl, and MgCl2 were supplied by Prolabo (Paris, France). The RNase inhibitor, ProtectRNA, was supplied by Sigma Aldrich. This new RNase inhibitor can be used successfully for in situ hybridization studies24 and then is expected to maintain the physicochemical properties of the oligonucleotide. However, the chemical nature of this RNase inhibitor was not communicated by the supplier. Water was obtained from an Elgastat option water purification system (Odil, Talant, France) fitted with a reverse osmosis cartridge. The D-RNA and L-RNA oligonucleotides were synthesized and 5′-biotinylated by Eurogentec (Herstal, Belgium) or CureVac (Tubingen, Germany). Biotin phosphoramidite containing a 16-atom spacer arm based on triethylene glycol was used for the aptamer biotinylation. The oligoribonucleotides were purified by gel electrophoresis (Eurogentec) or HPLC (CureVac), and their identity was confirmed by MALDI-TOF mass spectrom(23) Li, Y.; Breaker, R. R. J. Am. Chem. Soc. 1999, 121, 5364. (24) https://www.sigmaaldrich.com.

1994 Analytical Chemistry, Vol. 77, No. 7, April 1, 2005

etry. At the 1-µmol scale, the yields ranged from 37 to 56 OD 260 units for the RNA samples from CureVac to 16-19 OD 260 units for the RNA samples from Eurogentec. The streptavidin POROS bulk media (20-µm polystyrene particles) was purchased from Applied Biosystems (Courtaboeuf, France). POROS perfusion chromatography particles have large throughpores that transect the particles and short diffusive pores that branch off from the throughpores. PEEK tubing (0.76- or 0.51-mm i.d.), PEEK end fittings (unions), and microbore frits were obtained from CIL Cluzeau Info Labo (Sainte-Foy-La-Grande, France) or Applied Biosystems. Stationary-Phase Preparation. Prior to immobilization, the biotinylated aptamers were renaturated by heating oligonucleotides at 85 °C for 5 min in an aqueous buffer (25 mM phosphate buffer, 25 mM NaCl, 5 mM MgCl2 adjusted at pH 7.3) and left to stand at room temperature for 30 min. Immobilization of aptamers was attained by mixing ∼60 nmol of oligonucleotide in the aqueous buffer per 1000 µL of the streptavidin media slurry (1:5 wet settled media volume) during 3 h at ambient temperature. Unbound RNA was removed by washing with the same buffer. The amount of oligonucleotide coupled to the chromatographic support was estimated by subtracting the UV absorbance, at 260 nm, of the unbound RNA solution from the initial solution. From ∼14 to ∼20 nmol of biotinylated oligonucleotide was bound per 100 µL of support media in relation to the RNA samples used. Column Packing. The RNA-modified particles were packed in-house into five PEEK microbore columns (370 × 0.76 mm for the D-RNA CSPs 1 and 4 and the L-RNA CSP 5, and 340 × 0.51 mm for the D-RNA CSPs 2 and 3) using the previously described slurry-packing procedure.9 Nonmodified streptavidin particles and streptavidin-biotin complex particles were also packed into two PEEK microbore columns (370 × 0.76 mm) as control stationary phases, following the procedure reported previously.9 Storage Conditions. When not in use, the columns containing the RNA CSPs 1, 2, 4, and 5 were stored in the aqueous buffer. In contrast, the RNA CSP 3 column was stored in the aqueous buffer containing the RNase inhibitor (2 mL for 1000 mL of mobile phase as indicated by the supplier).24 This was achieved by pumping at least 20 column volumes of the aqueous buffer containing the RNase inhibitor through the column. The storage conditions were at 4 °C for all the columns. Before the chromatographic experiments, the RNA CSP 3 was washed with the aqueous mobile phase until the baseline was stable in order to eliminate the RNase inhibitor from the column. For a significant comparison of the stability, the D-RNA CSPs 2 and 3 were made using the same D-RNA sample from Eurogentec and the RNA CSPs 4 and 5 were made using the same supplier (CureVac) for the synthesis of D- and L-oligonucleotides. The same reagents and equipment were used for the immobilization/packing procedures. These columns were used in identical operating conditions: same mobile phase, chromatographic equipment, and period of experiments (column temperature equal to 4 °C). Apparatus. The HPLC system consisted of a LC Shimadzu pump 10AT (Sarreguemines, France), a Shimadzu SIL-10AD autoinjector, a Shimadzu SPD-10A UV-visible detector (detection at 208 nm, cell volume 140 nL), a Shimadzu SCL-10A system controller with Class-VP software (Shimadzu), and an oven Igloocil (Interchim).

Chromatographic Operating Conditions. Solute samples were prepared in the mobile phase and injected (100 nL) in triplicate. The injected solute concentration varied from 50 to 1000 µg/mL. The mobile-phase flow rate was 50 µL/min for the RNA CSPs 1, 4, and 5 and the control streptavidin columns (nonmodified streptavidin and streptavidin-biotin complex stationary phases) and 25 µL/min for the RNA CSPs 2 and 3. The apparent retention factor k was determined for the lowest injected solute concentration (50 µg/mL) using the following relation:

k ) (tR - t0)/t0

(1)

where tR is the retention time of the enantiomer and t0 the retention time of the unretained species. The column void time was determined using the sodium nitrate peak.9 The retention times and column void time were corrected for the extracolumn void time. They were assessed by injections of solute onto the chromatographic system when no column was present. It can be noted that the extracolumn void time was lower for the experiments using the RNA CSPs 2-5 than for the RNA CSP 1 experiments due to the replacement of some tubing and union pieces between the columns and the injector and detector. The chromatographic resolution Rs was also calculated using the following relation:

Rs ) [2(tR2 - tR1)]/(w2 + w1)

(2)

where w is the peak width at the baseline of each enantiomer. The suffixes 1 and 2 refer to the first (nontarget) and last (target) eluting enantiomers. The efficiency of the column was characterized by estimating the reduced plate height h for the lowest injected solute concentration as follows:

h ) L/Ndp

(3)

N ) 5.54(tR/δ)2

(4)

with

where N is the number of theoretical plates (δ is the peak width at half-height), L is the column length, and dp the average particle diameter. The asymmetry factor As was determined for the lowest injected solute concentration by calculating the ratio of the second (or right part) of the peak over the first (or early part) of the peak at 10% of the peak height. Kinetics of the RNA Degradation. The apparent retention factor is classically described as follows:8

k ) mLK/VM

(5)

where mL is the active binding site number in the column, VM is the void volume of the column, and K is the association constant between the solute and the CSP. Therefore, for a given chromatographic system, a variation of the retention factor in identical operating conditions reflects a change in the number of the active binding sites in the column. So, the kinetics of RNA degradation can be evaluated indirectly by monitoring the target retention

factor versus time. The data were fit to the following relationship:

ln k ) -St + C

(6)

where S is the observed decay rate of the target retention factor, t the time, and C a fit parameter. RESULTS AND DISCUSSION Separation of the Arginine Enantiomers Using an Immobilized Anti-L-arginine D-RNA Aptamer as a Target-Specific CSP (D-RNA CSP 1). The RNA Aptamer Sequence Used in this Study. The 44-mer RNA aptamer sequence used in this study has been selected using L-arginine as target.15 This oligonucleotide is able to discriminate between L- and D-arginine.15 The dissociation constants, determined by equilibrium gel filtration, are in the micromolar range: Kd ) ∼56 µM for the L-arginine-aptamer association and ∼412 µM for the D-arginine-aptamer association, corresponding to a “true” enantioselectivity of ∼7.4 and a ∆∆G of ∼4.6 kJ/mol. The secondary structure of this 44-mer RNA aptamer is presented in Figure 1a. The anti-L-arginine aptamer contains secondary structure folds defined by two asymmetric internal loops flanked at each end by stem segments. Famulok and co-workers have demonstrated that the largest internal loop is critical for the L-arginine binding (Figure 1a).25,26 No Interactions between the Streptavidin Chromatographic Support and the Arginine Enantiomers. It was shown previously that streptavidin was able to discriminate among the enantiomers of some analytes such as warfarin,27 trimipramine,27 or adenosine.9,28 To evaluate possible enantioselective properties of the immobilized streptavidin toward D-,L-arginine, a racemic mixture of arginine was injected onto microbore columns packed with nonmodified streptavidin or streptavidin-biotin complex POROS particles (eluent: phosphate buffer 25 mM, NaCl 25 mM, MgCl2 5 mM, pH 7.3). The enantiomers were not separated and eluted roughly in the void volume whatever the column temperature (data not shown). This result indicates that the streptavidin chromatographic support was inert toward the arginine enantiomers and did not affect the solute retention. Separation of the Arginine Enantiomers using the D-RNA Aptamer CSP 1. In the first stage of this work, the retention and chiral discrimination abilities of a D-RNA aptamer CSP (D-RNA CSP 1) were characterized. The enantioseparation was analyzed at various column temperatures (4-17 °C) using an aqueous buffer as eluent (phosphate buffer: 25 mM, NaCl 25 mM, MgCl2 5 mM, pH 7.3). The target enantiomer was significantly more retained by the column than the nontarget enantiomer, as expected from the chiral discrimination properties previously discovered.15 Figure 2a shows the chromatographic resolution of arginine for a column temperature equal to 4 °C (10 ng of D-,L-arginine injected). Apparent retention factor for the L-enantiomer and resolution are presented in Table 1 for all the column temperatures, in initial conditions. (25) Burgstaller, A. T.; Kochoyan, M.; Famulok, M. Nucleic Acids Res. 1995, 23, 4769. (26) Yang, Y.; Kochoyan, M.; Burgstaller, A. T.; Westhof, E.; Famulok, M. Science 1996, 272, 1343. (27) Tanaka, Y.; Terabe, S. Chromatographia 1999, 49, 489. (28) Ravelet, C.; Michaud, M.; Ravel, A.; Grosset, C.; Villet, A.; Peyrin, E. J. Chromatogr., A 2004, 1036, 155.

Analytical Chemistry, Vol. 77, No. 7, April 1, 2005

1995

Figure 2. Chromatograms for the resolution of arginine using the D-RNA CSP 1 in initial conditions. Amount of D-,L-arginine injected: (a) 10 and (b) 100 ng. Column: 370 × 0.76 (i.d) mm; mobile phase phosphate buffer 25 mM, NaCl 25 mM, MgCl2 5 mM, pH 7.3; column temperature 4 °C; injection volume 100 nL; flow rate 50 µL/min; detection at 208 nm. Table 1. Chromatographic Resolution of the Arginine Enantiomers on the Target-Specific Anti-L-arginine D-RNA Aptamer CSP 1a T (°C)

kLb

Rs

4 8 11 14 17

2.20 1.73 1.43 0.99 0.83

1.70 1.30 1.26 1.18 1.02

a D-,L-arginine amount injected, 5ng. b k is the retention factor of L the last (target) eluting enantiomer in initial conditions. Relative standard deviation of the solute retention factors was less than 6%.

As D-arginine was very weakly retained on the aptamer CSP, the kD and apparent enantioselectivity values may have major errors. So, the retention factor for the D-enantiomer and the apparent separation factor were not reported. Here, it can be noted that valuable and significant values for the D-arginine retention factor (and then for the apparent enantioselectivity) could be attained via the increase in the number of active sites per surface unit. A “direct” aptamer covalent immobilization, as previously described in CEC,3-5,19,20 could result in a more efficient surface coverage of the oligonucleotide. However, the RNA aptamer binding to the chromatographic support via the biotin-streptavidin bridge was chosen in this work because of the following advantages: (i) the oligonucleotide immobilization procedure was easy and rapid and (ii) the biotin-streptavidin bridge is known to maintain efficiently the receptor (aptamer) binding ability.6-9,29 At a flow rate of 50 µL/min, the reduced plate height h for L-arginine ranged from 35 to 60 in relation to the column temperature. In addition, the asymmetry factor As was found to be between ∼1.2 and ∼2. As D-arginine was not significantly retained by the RNA CSP (see below), nonspecific interactions with the immobilized RNA are negligible in this chromatographic system; i.e., only one type of site is expected to govern the target enantiomer retention. Therefore, it is likely that slow adsorptiondesorption kinetics of the L-arginine binding to the RNA-specific site contribute to the poor efficiency and peak tailing (homogeneous kinetics tailing).8 As it appears in Table 1, the L-arginine retention factor decreased significantly when the RNA CSP 1 column temperature increased. For example, in initial conditions, kL varied from 2.20 ( 0.03 at 4 °C to 0.83 ( 0.02 at 17 °C, corresponding to a solute (29) Liu, X.; Farmerie, W.; Schuster, S. Anal. Biochem. 2000, 283, 53.

1996 Analytical Chemistry, Vol. 77, No. 7, April 1, 2005

Figure 3. The ln(k/k0) versus time plots for the target enantiomer using all the RNA CSPs evaluated in this study (see Table 2).

retention reduction of ∼60%. This indicates that the target retention mechanism was enthalpically driven. This is in accordance with the earlier experiments on the anti-D-adenosine and anti-L-tyrosinamide DNA aptamer CSPs.9 Such observation is also consistent with the fact that hydrogen bonds and stacking interactions, both characterized by negative enthalpic contributions,8 play a preponderant role in the L-arginine binding to the D-RNA-specific pocket.26 The target retention factor enhancement dependent on the column temperature decrease was associated with an improvement in the enantiomeric resolution (Table 1). Additional experiments were carried out by injecting a high amount of D-,L-arginine onto the D-RNA CSP 1 column (100 ng). A very important peak tailing was observed for L-arginine due to the limited binding capacity of the microbore column. However, as shown in Figure 2b, baseline resolution was still achieved for a column temperature equal to 4 °C. Study of the Temporal Stability of the D-RNA CSP (D-RNA CSP 1). When not in use, the column was stored in the aqueous buffer used as mobile phase at 4 °C.8,9 The stability of the D-RNA CSP 1 column was evaluated by comparing the L-arginine retention factor during almost three weeks in the same conditions. The ratio of the retention factor at day D (k) from the retention factor in initial conditions (k0) was determined at a column temperature of 4 °C and ln(k/k0) was plotted against time (Figure 3). The performances of the D-RNA CSP 1 decreased very strongly with time, as exemplified by Figure 3 and the decay rate S (eq 6) reported in Table 2. A loss of ∼65% of the target retention was observed after 8 days. Such results indicate that the number of

Table 2. RNA Aptamer CSP Stability As Evaluated by Monitoring the Target Retention Factor versus Time (see Experimental Section) RNA CSPs

chiral configuration

storage procedure

S (103h-1)c

1, 2,a 4b 3a 5b

D

aqueous buffer RNase inhibitor aqueous buffer

4.3, 3.0, 7.0 0.3