Capillary Electrophoresis of RNA Oligonucleotides: Catalytic Activity of a Hammerhead Ribozyme

Analytical Biochemistry 266, 93–101 (1999) Article ID abio.1998.2942, available online at http://www.idealibrary.com on

Capillary Electrophoresis of RNA Oligonucleotides: Catalytic Activity of a Hammerhead Ribozyme Jan Saevels, Ann Van Schepdael, and Jos Hoogmartens Laboratory for Pharmaceutical Chemistry and Drug Analysis, K.U.Leuven, Van Evenstraat 4, B-3000 Leuven, Belgium

Received July 6, 1998

Ribozymes are sequences of catalytic RNA that are being evaluated as possible antisense therapeutics. This paper describes how capillary electrophoresis (CE) could be used to measure the catalytic rate of a synthetic hammerhead ribozyme in cleaving its substrate. This substrate was a synthetic full-RNA 17-mer, whereas the ribozyme was made up of a mixture of 37 2*-OH and 2*-OCH 3 RNA nucleotides. After experimental conditions to exclude ribonuclease contamination were successfully met, different CE modes were tried out to separate the ribozyme from its substrate. Only the combination of chemical and thermal denaturation was adequate to disrupt strong secondary structures and to inhibit comigration of the two molecules. Cleavage kinetics were measured by continuous injection from the reaction vial into a polymer-filled capillary, and by determination of the area of the shrinking substrate peak. Compared to the well-established slab gel electrophoresis, CE is at least one order of magnitude faster, may be completely automated, allows easier and more precise quantitation of results, and, due to the small scale and self-contained nature of the apparatus, reduces health risks from dangerous chemicals. Unfortunately, UV detection in a 100-mm internal diameter capillary lacked the sensitivity to perform assays in the nanomolar range, which was necessary for a full Michaelis–Menten analysis. © 1999 Academic Press

Until the past two decades, it was thought that only proteins could act as catalysts in biochemical reactions, and that nucleic acids only contained genetic information, or played a structural role. During the early 1980s, however, it was shown that some RNA sequences have the ability of catalyzing chemical transformations within the same molecule (cis-acting) or within another RNA molecule (trans-acting) (1). The discovery that not only proteins but also these catalytic RNA molecules or ribozymes could enhance biochemi0003-2697/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

cal reactions is one of the landmarks in biochemistry of the past 15 years. Trans-acting ribozymes have potential therapeutic applications, comparable to antisense oligodeoxynucleotides (ODNs), 1 because in theory they can be directed against any particular RNA sequence, based on complementary Watson–Crick basepair formation (2– 4). Substrate cleavage is achieved by catalyzing the breakage of a specific phosphodiester bond, after which the ribozyme molecule has the ability to dissociate from its substrate. This feature is in contrast to conventional antisense oligonucleotides, where only one target molecule can be blocked sterically at a time. Moreover, ribozymes do not rely on the host cell’s RNase H (ribonuclease H) machinery to destroy formed duplexes (5). These characteristics make ribozymes attractive compounds in the treatment of viral infections and illnesses due to uncontrolled expressions of genetic information, such as cancers (6, 7). At least five distinct classes of ribozymes have been identified so far, but the discovery of additional RNA molecules possessing catalytic activity is certain and the exploitation of their enzymatic properties in biological systems is inevitable (4). Hammerhead ribozymes, named according to their remarkable two-dimensional structures, were first identified in tobacco ringspot virus. For therapeutic applications, they are particularly interesting because they can be produced chemically using the classical oligonucleotide synthesis pathways that are used in numerous laboratories. Moreover, hammerhead ribozymes can be designed to cleave on the 39-side of any NUX triplet (N is any nucleotide; X is A, C, or U) (8). A notable problem that must be dealt with is the low stability of these compounds, due to the 1

Abbreviations used: BGE, background electrolyte; CE, capillary electrophoresis; CPSE, capillary polymer-sieving electrophoresis; DEPC, diethyl pyrocarbonate; HEC, hydroxyethyl cellulose; HPMC, hydroxypropyl methyl cellulose; ODN, oligodeoxyribonucleotide; Taps, Tris(hydroxymethyl)methylaminopropanesulfonic acid. 93

94

SAEVELS, VAN SCHEPDAEL, AND HOOGMARTENS

free 29-hydroxyl groups (9). Moreover, the synthesis of all-RNA ribozymes is expensive, the yield is low, and the synthesis products are extremely labile. The synthesis of DNA/RNA chimeric ribozymes is becoming more and more investigated because this should combine greater stability, easier synthesis, and in some cases improved dissociation kinetics (10 –12). Because the essence of a ribozyme is its enzymatic activity, the determination of the kinetic parameters K m and k cat is an important step in the evaluation of synthetic ribozymes (13). At present, mostly slab gel electrophoresis is used to determine the kinetic parameters of designed ribozymes (14, 15). After incubation of the ribozyme with its substrate RNA fragment under appropriate conditions, the reaction is stopped and offline analysis is performed using denaturing PAGE in the presence of 5 to 8 M urea. Sensitive detection is only possible with radioactive labeling using 32P or 35S, followed by scintillation counting. The goal of this paper is to check whether capillary electrophoresis can be used for the analysis of RNA fragments, in the same way as described for the analysis of DNA fragments (16, 17). For the analysis of RNA oligonucleotides, special precautions need to be taken, since oligoribonucleotides are known to form persistent secondary structures, and anomalous migration might be expected to occur. Whether CE will allow measurement of the catalytic rate of a selected synthetic hammerhead ribozyme will be investigated. Compared to slab gel electrophoresis, expected advantages of using CE are smaller sample size, ease of automation, increased analysis speed, and on-capillary UV detection. MATERIALS AND METHODS

Oligonucleotide Preparations We have used a generic, catalytically active, nuclease stable 37-mer hammerhead ribozyme motif that contains only five ribose residues; the remaining residues all consist of 29-OCH 3 nucleotides for improved stability. A short synthetic 17-mer full-RNA substrate was used to assess the ribozyme activity. The two oligonucleotides were custom synthesized and purified by Eurogentec (Seraing, Belgium). RNA Stability All RNA preparations arrived as lyophilized powders in sterile tubes. Eppendorfs that were used for storage of the RNA oligonucleotides were washed with an RNase-removing detergent (RNase ERASE, ICN Biomedicals, Aurora, OH), rinsed several times with fresh Milli-Q water (Millipore, Milford, MA), and left to dry under sterile laminar air flow. All other material that made contact with RNA when the stock solutions were

prepared was treated in the same way. RNA was dissolved at a concentration of 1 mg/ml in TE buffer (10 mM Tris (Acros Organics, Geel, Belgium) 1 1 mM EDTA (Carlo Erba, Milan, Italy); to pH 7.5 with HCl; 0.2 mm filtered) and stored in 10-ml fractions in pretreated Eppendorfs at 280°C. Disposable latex gloves were worn for all manipulations of the RNA preparations. For migration studies, reconstitution from the stock solutions to the final concentration was carried out with freshly filtered TE buffer or as indicated, and samples were stored at 220°C until they were used, which was definitely within 1 week. Once taken from the 220°C freezer and taken to room temperature, samples were evaluated within 12 h. Capillary Electrophoresis For all experiments, a SpectraPHORESIS 1000 (Thermo Separation Products, Fremont, CA) capillary electrophoresis unit was used, operated under PC 1000 software (Thermo Separation Products). J&W DB-17 silicone-coated capillaries with a 100-mm i.d. were purchased from Alltech (Laarne, Belgium) and cut to a total length of 44.0 cm (36.0 cm to the detector). Each morning, the capillary was washed with water, methanol, water, and background electrolyte (BGE), each time for no less than 5 min. UV detection was performed at 260 nm. Hydrodynamic injection at 0.75 psi was used throughout (10 –30 s). CE Using HPMC Hydroxypropyl methyl cellulose (HPMC, Acros Organics, Geel, Belgium) was tried as sieving agent in the analysis of substrate and ribozyme. This particular type of HPMC displayed a viscosity of 6.0 6 1.5 mPa z s at 2%. The BGE was prepared by adding an appropriate amount of HPMC to Milli-Q water, after which the solution was stirred for 30 min at room temperature. Finally, 20 mM Taps (Tris[hydroxymethyl]methylaminopropanesulfonic acid, Sigma, St. Louis, MO) and 7 M urea (ICN Biomedicals) were added, and the pH was adjusted to 7.5 with solid Tris. Samples were prepared by adding 40 ml TE buffer to the 10 ml RNA 1 mg/ml stock solution. Electrophoresis was run at 230 kV and the capillary was thermostated at 25°C. Each sample was analyzed three times and between runs, the capillary was flushed with the BGE during a period equivalent to three times the fill time. CE Using HEC Hydroxyethyl cellulose (HEC EP09, Union Carbide, Antwerp, Belgium) with a viscosity of 9 mPa z s at 2% was also tried as a polymer additive for the size-based separation of the RNA oligonucleotides. Twenty-five milliliters of the BGE was prepared as follows: 1.0 g of

CAPILLARY ELECTROPHORESIS OF HAMMERHEAD RIBOZYME TABLE 1

Background Electrolytes Using HEC as the Sieving Agent a. b. c. d. e. f.

4% HEC in 20 mM 4% HEC in 20 mM with Tris 4% HEC in 20 mM with Tris 4% HEC in 20 mM 4% HEC in 20 mM 7.5 with Tris 4% HEC in 20 mM 7.5 with Tris

Taps, 7 M urea, to pH 7.5 with Tris Taps, 7 M urea, 10 mM EDTA, to pH 7.5 Taps, 7 M urea, 20 mM EDTA, to pH 7.5 Taps, 8 M urea, to pH 7.5 with Tris Taps, 7 M urea, 30% formamide, to pH Taps, 3.5 M urea, 30% formamide, to pH

the HEC was added to 17 ml of Milli-Q water at 50°C. After 10 min of continuous stirring, a clear solution was obtained and Taps, urea, and EDTA or formamide (Merck, Darmstadt, Germany) was added. Finally, the pH was adjusted to 7.5 with Tris. Table 1 shows the set of buffers that was prepared. Samples were prepared in TE or Tris–Mg buffer (50 mM Tris, 10 mM MgCl 2 (Merck), to pH 7.5 with HCl), as in Table 2. Electrophoresis was run at 215 kV, and a temperature of 25 or 60°C. Inbetween runs, the capillary was flushed with the BGE for 5 min. Cleavage Kinetics The catalytic rate of the hammerhead ribozyme was measured under multiple turnover conditions (18, 19). Reactions were initiated as follows: appropriate dilutions of ribozyme and substrate were made in TE buffer (10 times more concentrated than the final reaction concentration), 5 ml of each was added to 30 ml of a 80 mM Tris buffer at pH 7.5, this was heated to 95°C for 5 min to disrupt potential aggregates, equilibrated to room temperature, and the reaction was started with the addition of 10 ml of a 51 mM MgCl 2 solution. The final incubation conditions were thus: 50 mM Tris, pH 7.5, in the presence of 10 mM free Mg 21. Reactions were run in the nonthermostated autosampler of the CE apparatus, at room temperature, which was around 23°C. Hydrodynamic injections (30 s at 0.75 psi) from the incubation vial directly into the separation capillary were performed at regular time intervals. Kinetic analyses were performed with ribozyme concentration fixed at 200 nM, whereas substrate concentration was varied from 1 to 20 mM. RESULTS AND DISCUSSION

(DEPC), which inactivates the RNase, followed by autoclaving to decompose the DEPC. This laborious method is used with apprehension, since DEPC is a toxic substance. On the other hand, detergents to destroy RNase also exist, and we have chosen to use this in combination with ultrapure water. Because RNases utilize the ribose 2-hydroxyl group to cleave the phosphodiester bond, the full-RNA 17mer should definitely be stored at 280°C. The 37-mer is a hybrid of RNA and RNA 29-OCH 3 (more nuclease resistant) and has an extra protection on the 39-end. Nevertheless, it is treated with the same precaution as for the full-RNA preparations. As a result, no spontaneous breakdown was seen and the ribozyme kept its activity over a period of more than two months. Ribozyme–Substrate Complex The structure of our ribozyme–substrate complex using the uniform numbering system proposed by Hertel et al. (20) is given in Fig. 1. The sequences of ribozyme and substrate are based on literature reports of catalytically active hammerhead structures (21–23). The central core motif contains 17 nucleotides with only 5 purine ribose residues, namely G5, A6, G8, G12, and A15.1. The other residues are 29-OCH 3. 29-Methoxy substitution confers stability to the hammerhead ribozyme; it is more nuclease resistant than 29-OH, even more than phosphorothioate. Cleavage takes place by nucleophilic attack of the oxygen atom of the 29-hydroxyl group of A17; release of the 59-hydroxyl group then proceeds via an intermediate with a trigonal– bipyramidal arrangement of the oxygen atoms around the central phosphorus atom, yielding a 29,39-cyclic phosphate at A17 and a free 59-hydroxyl group to A1.1 (24). Thus, the expected reaction products are 59-AUG GAG AU-39 (product I) and 59-CAG GGA UUA cP-39 (product II, cP stands for a 29,39-cyclic phosphate on A17).

TABLE 2

Sample Preparation and Treatment for Analyses Using HEC as the Sieving Agent a. b. c. d.

RNA Stability The main issue for maintaining intact RNA is its susceptibility to degradation by RNases. Appropriate measures should be taken to destroy these enzymes. Many researchers prepare their RNase-free solutions by pretreating the water with diethyl pyrocarbonate

95

e. f. g.

10 ml RNA (1 mg/ml) 1 40 ml Tris–Mg buffer 10 ml RNA (1 mg/ml) 1 40 ml TE buffer Composition as in b, and heated at 95°C for 5 min before analysis Composition and heating as in c, and quenched on ice before analysis 10 ml RNA (1 mg/ml) 1 40 ml TE buffer 1 50 ml formamide Composition as in e, and heated at 95°C for 5 min before analysis Composition and heating as in f, and quenched on ice before analysis

96

SAEVELS, VAN SCHEPDAEL, AND HOOGMARTENS

FIG. 1. Two-dimensional representation of the hammerhead–substrate complex.

Secondary Structures It is generally known that RNA is more difficult to denature than DNA because secondary structures are more easily formed. Also, RNA/RNA duplexes are more stable than RNA/DNA and DNA/DNA duplexes. Whereas 7 M urea is normally sufficient to denature DNA, stronger measures may be required for RNA (25). Moreover, 20 years ago, it had been described that the persistence of secondary structures during electrophoresis does influence the electrophoretic mobility of RNA fragments (26). The mobility of low-molecularweight RNA molecules is different from that of DNA, and it has been suggested that an interaction of the RNA molecules with the entangled polymer network exists. As a result, the migration models that have been described for the movement of DNA through gels and entangled polymer networks (i.e., Ogston and reptation model) may not be directly applicable for the separation of low-molecular-weight RNA (27, 28). Conditions for denaturing nucleic acids in slab gel electrophoresis are not always sufficient in capillary electrophoresis. It might be that heating of the slab gels during electrophoresis contributes to denaturation. It might be necessary to do runs at elevated temperature (heat labile secondary structures), to incorporate extra denaturing agents, or to preheat the oligonucleotides at 95°C for 5 min (29, 30). The formation of intra- or intermolecular secondary structures such as hairpin loops or duplexes of sufficient stability may result in shorter electrophoretic migration times in capillary gel electrophoresis (31), and in general anomalous migration is observed if sufficient denaturation is not accomplished, which means that denaturation is essential to obtain uniform and identifiable peaks.

Looking at the sequence of both the substrate and the ribozyme, we attempt to propose possible secondary structures: for the 17-mer, the sequence A16.3 U16.2 U16.1 A17 A1.1 U1.2 from 59 to 39 can form a duplex with another strand with the sequence U1.2 A1.1 A17 U16.1 U16.2 A16.3 from 39 to 59. As far as the 37-mer is concerned, intramolecular hydrogen bonds between c11.1– g10.1, c11.2– g10.2, g11.3– c10.3, and g11.4 – c10.4 (this is equivalent to helix II in the substrate–ribozyme complex) make the molecule adopt a more compact orientation, comparable to the conformation in the complex. A triplet of intramolecular basepairing can be formed between a15.2– u7, u15.3–A6, and c15.4 –G5. In addition, other double adjacent hydrogen bonds are possible, as well as some terminal loops. These suggestions of secondary structure formation should further demonstrate the need for strong denaturing conditions. Migration Studies Hydroxypropyl methyl cellulose was tried as sieving matrix in capillary polymer-sieving electrophoresis (CPSE) because it is known to produce good results for RNA segments ranging from 100 to 2000 bases in length (32). A low-viscosity grade is preferred here because this is generally more suited for separation of shorter fragments. Eight different concentrations of HPMC were assessed for fill time on the capillary and migration times for the 17- and the 37-mer. Figure 2 summarizes the obtained results. Above a HPMC concentration of 6%, the fill time increases more steeply than between 1 and 5%. If we take fill time in the capillary to be proportional to the viscosity of the solution, the viscosity increases drastically above 6%, most

97

CAPILLARY ELECTROPHORESIS OF HAMMERHEAD RIBOZYME

FIG. 2. Concentration of the HPMC solution versus fill time on the capillary (right ordinate) and versus migration time of the 17- and 37-mer (left ordinate).

likely reflecting the entanglement threshold. Under this value, the migration time of the 17-mer increases with increasing HPMC concentration, while the results for the 37-mer are not very repeatable. The migration does not seem to be much influenced by the polymer concentration and subsequent analyses show much variation in migration time. At entanglement concentration, the 17- and the 37-mer are comigrating. Concentrations higher than 8% were not tried out because this would give rise to impracticable fill times exceeding 5 min. At low HPMC concentrations, peak shapes are very bad (not shown). This improves substantially above entanglement concentration, although peaks for the 37-mer stay relatively broad. Due to nonrepeatability of the results, bad peak shapes, and comigration of the 17- and the 37-mer above the entanglement concentration, HPMC was not the polymer of choice for the size-based separation of the ribozyme and its substrate. Analysis of the 17- and the 37-mer by CPSE with hydroxyethyl cellulose as the sieving agent was also tried out. The 4% HEC solution in 20 mM Taps and 7 M urea is known to give good results in discriminating DNA oligo’s of different length (16, 17, 33). Four percent is well above the entanglement threshold, and results are repeatable. Fill times on the capillary are between 2.5 and 3.0 min for all HEC buffers described under Materials and Methods. A concentration higher than 4% becomes too viscous to load in an acceptable time period, while a concentration of 3% does not improve resolution between the fragments (data not shown). The pH was 7.5 for all BGEs because this is the pH that was successfully used for separation of oligodeoxyribonucleotides. As a rule of thumb, 7 M urea is included in the BGE as a denaturing agent (BGE a, see Table 1). Under these circumstances, the two fragments are comigrating under all conditions of sample preparation (Fig. 3,

left). Sample preparation a comprises the medium in which catalytic activity is tested, but it contains a high concentration of Mg 21 ions which promotes secondary structure formation. The preparations were then made in pure TE buffer (sample preparation b) and heated to 95°C to break possible RNA aggregates (sample preparation c). With (sample preparation d) or without rapid cooling, no difference in migration was seen. In analogy with slab gel electrophoresis where samples from stability studies are collected over time, quenched in formamide and analyzed simultaneously after the last sample is taken, we prepared samples according to sample preparations e, f, and g, but apart from smaller peaks, no difference in migration was seen (Table 3). Comigration of the 17- and 37-mer is probably caused by strong secondary structure formation (cf. supra). Because it is known that cations can promote secondary structure persistence, and because the HEC we are using is only technical grade and can contain some metal-ion impurities, EDTA was incorporated in BGEs b and c. Currents were higher compared to BGE a, but no difference in migration was seen. Stronger denaturing conditions were then sought. Raising the urea concentration to 8 M (BGE d) was tried, but did not improve the separation. Increasing the temperature of analysis is another way to enhance denaturing conditions, and this was tried out in combination with BGE d. At 60°C, a current of 28 mA was produced, migration was faster TABLE 3

Migration Times for 17- and 37-mer and Observed Current for Different HEC Background Electrolytes, Temperatures, and Sample Treatments Migration time

BGE

Temperature (°C)

Current (mA)

Sample preparation

17-mer (min)

37-mer (min)

a

25

20

b c d e

25 25 25 60 25

38 56 19 28 38

f

60 25

57 32

60

45

a b c d e f g b b b b b c b b c a b c

13.3 13.3 13.1 13.2 13.4 13.3 13.2 14.1 16.2 15.4 10.6 16.6 16.8 10.2 18.1 17.1 11.6 11.4 11.7

13.2 13.2 13.2 13.3 13.3 13.2 13.0 14.0 16.3 15.2 10.8 16.5 16.8 11.8 18.0 17.1 13.6 13.3 13.8

98

SAEVELS, VAN SCHEPDAEL, AND HOOGMARTENS

FIG. 3. Typical electropherograms of CPSE using HEC as the sieving agent. Capillary: J&W DB-17 100 mm i.d., 44.0 cm total length, 36.0 cm effective length; BGE: (left) 4% HEC EP09 in 20 mM Taps, 7 M urea, to pH 7.5 with Tris; (right) 4% HEC EP09 in 20 mM Taps, 3.5 M urea, 30% formamide, to pH 7.5 with Tris; sample: 0.2 mg/ml TE buffer; voltage, 215 kV (20 – 45 mA); left, 25°C; right, 60°C; UV detection at 260 nm.

than at 25°C (decreased viscosity), but both molecules were still comigrating. Because 60°C is the maximal temperature allowed by the CE equipment, two other denaturing buffers were tried, BGEs e and f, which contained a combination of urea and formamide. Formamide is considered a stronger denaturant than urea, but is less used, probably because it reduces the lifetime of gel-filled capillaries in capillary gel electrophoresis. At 25°C, neither 7 M urea 1 30% formamide (BGE e) nor 3.5 M urea 1 30% formamide (BGE f) could discriminate between the two compounds, and this was tried without preheating (sample preparation b) and with heating at 95°C for 5 min (sample preparation c). At 60°C, however, migration of the 37-mer was definitely slower than the 17-mer, and this phenomenon was seen for both BGEs e and f. Observed currents were lower for BGE f (less Joule heating), which was also more easy to prepare, and that is why preference was given to this buffer (Fig. 3, right). The effect of sample preparation was further checked for BGE f. When the sam-

ples were dissolved as described in sample preparation a, no substantial difference in migration was seen when compared to sample preparation b. At 60°C, efficiency was somewhat lower than at 25°C, probably due to diffusional brand broadening. About 50,000 plates were attained at 60°C, whereas 100,000 plates can be reached for the substrate when working at 25°C. Quantitative Performance Using BGE f and sample preparation a, linearity, detection and quantification limits were determined for 30 s hydrodynamic injections from a 50-ml insert containing the 17-mer. The detector response at 260 nm was tested at 7 concentrations between 570 nM and 36.2 mM, and mean corrected areas plus or minus the estimated standard deviation (analysis in triplicate) are given in the left panel of Fig. 4. The concentration range between 2.3 and 18.1 mM of injected 17-mer seemed linear (r 2 5 0.995) (Fig. 4, right). Deviation

FIG. 4. Detector response for injected 17-mer between 0.57 and 36.2 mM (left); linear portion between 2.3 and 18.1 mM (right).

CAPILLARY ELECTROPHORESIS OF HAMMERHEAD RIBOZYME

99

FIG. 5. Typical electropherograms showing different stages in the ribozyme-mediated catalytic breakdown of the 17-mer RNA oligonucleotide. Capillary: J&W DB-17 100 mm i.d., 44.0 cm total length, 36.0 cm effective length; BGE: 4% HEC EP09 in 20 mM Taps, 3.5 M urea, 30% formamide, to pH 7.5 with Tris; voltage; 215 kV (45 mA); 60°C; UV detection at 260 nm. Between two runs, the capillary was flushed with BGE for 5 min.

from linear behavior was seen above 18.1 mM and under 2.3 mM. At a signal to noise ratio of 12.4, quantification limits were set around 2 mM, whereas detection limits were around 500 nM at a signal to noise ratio of 3.9. Hydrodynamic injection was used throughout to make sure sample composition would not change during injection. No sample workup was carried out (desalting), although the ionic strength of the sample seemed higher than that of the used BGE f. However, ionic impurities in the technical grade HEC account for an increased ionic strength of the sieving buffer, and this probably keeps the injected sample zones sharp enough. These impurities also explain the generated currents, which are relatively high for organic Good-type buffers. The influence of sample ionic strength on the efficiency of the peaks was assessed by performing injections from a sample solution of reduced ionic strength. When samples were made in TE buffer instead of Tris–Mg buffer, efficiencies and signal to noise ratios improved only slightly.

CE of Cleavage Products Figure 5 shows four typical electropherograms of the continuous injections from the incubation vial containing 0.2 mM ribozyme and 9.1 mM substrate, with the first injection 5 min after mixing. Disappearance of the substrate together with the appearance of reaction products I and II could be observed in a time-dependent manner. The fragment containing the 29,39-cyclic phosphate (product II) migrates a little slower than a 39-phosphate would, but faster than a 29,39-OH would, as was already noticed by Kuimelis et al. (34). Kinetic Analysis Kinetic analysis was performed in the range where the substrate showed linear detector response (i.e., 2, 5, 10, 15, and 20 mM), plus one assay at 1 mM. Ribozyme rates were quantified by measuring the corrected area of the remaining uncleaved substrate peak at regular time intervals, and the natural logarithm of this value was plotted versus incubation time for the

100

SAEVELS, VAN SCHEPDAEL, AND HOOGMARTENS

FIG. 6. Cleavage rates for highest and lowest substrate concentration. Observed rate constants are calculated from the first exponential decay. See text for experimental conditions.

six substrate concentrations, as shown in Fig. 6 for the highest and lowest concentration. The fraction of substrate remaining uncleaved as a function of time could be fitted to a double exponential curve. Because the first exponential decay accounted for 70% of total degradation, this portion was taken for the calculation of the rate constant k obs. As other authors have suggested, the second, slower segment of substrate cleavage presumably reflects ribozyme–substrate complexes in alternative (inactive or less active) conformations (23). One analysis was conducted where no Mg 21 was added because some hammerheads have been reported to support minimal cleavage in the absence of a metal cofactor (34). Also, the stability of the substrate under the same conditions without adding the ribozyme nor magnesium was determined (accounts for spontaneous cleavage). In both cases, no substantial degradation was seen after 300 min (substrate .95% intact when no ribozyme or Mg 21 was added, and .90% intact when only ribozyme was added). Finally, whether metal-catalyzed hydrolysis of the RNA substrate in the absence of ribozyme had occurred, as was reported previously by Pyle (3), was checked. No degradation was seen however in the presence of Mg 21 and the absence of hammerhead ribozyme.

A Michaelis–Menten plot was constructed, plotting substrate concentration versus observed rate constant. As can be seen in the left panel of Fig. 7, the rate constants in the tested substrate concentration range do not change significantly, i.e., they are all situated on the Michaelis–Menten plateau. Apparently, the Michaelis constant K m of the system must be even lower in concentration than the tested 1 mM limit. Unfortunately, the system lacks concentration sensitivity for a full kinetic analysis. If only the acquired data are used to construct a double reciprocal plot (Fig. 7, right) a tentative K m value of 370 nM and a k cat of 0.0150 min 21 can be determined from the intercept on the abscissa and ordinate, respectively. Catalytic efficiency as expressed by a specificity constant k cat/K m of 40,540 M 21 min 21 or 675 liters mol 21 s 21 is about a factor 5 lower than what has been described for equivalent hammerhead ribozymes (21–23). CONCLUSIONS

A generic modified hammerhead ribozyme and its short-chain full-RNA substrate were custom synthesized. As far as the stability of the preparations was concerned, the precautions that were taken to prevent

FIG. 7. Michaelis–Menten (left) and Lineweaver–Burk plots (right) for the ribozyme-mediated cleavage of the 17-mer.

CAPILLARY ELECTROPHORESIS OF HAMMERHEAD RIBOZYME

RNA degradation were adequate, no spontaneous breakdown was seen and the ribozyme kept its activity over a period of more than 2 months. The substrate and the ribozyme are structurally very similar: they are both negatively charged, and they even possess the same charge density: separation needs to be size-based, if we want them to exhibit different electrophoretic mobilities. For CPSE of RNA oligonucleotides, stronger denaturing conditions than for DNA oligos are needed; the combination of thermal and chemical denaturation is mandatory to eliminate anomalous migration. The developed CPSE method can be used to follow the catalytic activity of hammerhead ribozyme in real time. Injecting directly from the reaction vial allows continuous monitoring of the enzymatic assay without the inconvenience of aliquoting samples at various time intervals for a later analysis by slab gel electrophoresis. Also, the on-line detection system does not require an indirect postelectrophoretic chemical reaction such as staining. Determination of the kinetic parameters is limited by the detection and quantification limits of CE, a concentration insensitive separation technique. Due to the small detection path length, concentrations in the nanomolar range cannot be determined with conventional CPSE with UV detection. Possible solutions are the use of laser-induced fluorescence detection (native fluorescence or with labeled fluorophore), or a preconcentration step by isotachophoresis, followed by CE. Both techniques lower detection limits about 100-fold, and they should allow to measure substrate concentrations in the nanomolar range, where most of the hammerhead ribozymes display their Michaelis constant (9). REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Cech, T. R. (1987) Science 236, 1532–1539. von Ahsen, U., and Schroeder, R. (1993) BioEssays 15, 299 –307. Pyle, A. M. (1993) Science 261, 709 –714. Kiehntopf, M., Esquivel, E. L., Brach, M. A., and Herrmann, F. (1995) J. Mol. Med. 73, 65–71. Grassi, G., and Marini, J. C. (1996) Ann. Med. 28, 499 –510. Zaia, J. A., Chatterjee, S., Wong, K. K., Elkins, D., Taylor, N. R., and Rossi, J. J. (1992) Ann. NY Acad. Sci. 660, 95–106. Bashkin, J. K., Frolova, E. I., and Sampath, U. S. (1994) J. Am. Chem. Soc. 116, 5981–5982. Marschall, P., Thomson, J. B., and Eckstein, F. (1994) Cell. Mol. Neurobiol. 14, 523–538. Christoffersen, R. E., and Marr, J. J. (1995) J. Med. Chem. 38, 2023–2037. McCall, M. J., Hendry, P., and Jennings, P. A. (1992) Proc. Natl. Acad. Sci. USA 89, 5710 –5714.

101

11. Taylor, N. R., Kaplan, B. E., Swiderski, P., Li, H., and Rossi, J. J. (1992) Nucleic Acids Res. 20, 4559 – 4565. 12. Sawata, S., Shimayama, T., Komiyama, M., Kumar, P. K., Nishikawa, S., and Taira, K. (1993) Nucleic Acids Res. 21, 5656 – 5660. 13. Fedor, M. J., and Uhlenbeck, O. C. (1990) Proc. Natl. Acad. Sci. USA 87, 1668 –1672. 14. Fu, D., Benseler, F., and McLaughlin, L. W. (1994) J. Am. Chem. Soc. 116, 4591– 4598. 15. Hertel, K. J., Herschlag, D., and Uhlenbeck, O. C. (1994) Biochemistry 33, 3374 –3385. 16. Saevels, J., Huygens, K., Van Schepdael, A., and Hoogmartens, J. (1997) Anal. Chem. 69, 3299 –3303. 17. Saevels, J., Van Schepdael, A., and Hoogmartens, J. (1997) J. Cap. Elec. 4, 167–172. 18. Hendrix, C., Mahieu, M., Anne´, J., Van Calenbergh, S., Van Aerschot, A., Content, J., and Herdewijn, P. (1995) Biochem. Biophys. Res. Commun. 210, 67–73. 19. Jankowsky, E., and Schwenzer, B. (1996) Nucleic Acids Res. 24, 423– 429. 20. Hertel, K. J., Pardi, A., Uhlenbeck, O. C., Koizumi, M., Ohtsuka, E., Uesugi, S., Cedergren, R., Eckstein, F., Gerlach, W. L., Hodgson, R., and Symons, R. H. (1992) Nucleic Acids Res. 20, 3252. 21. Beigelman, L., McSwiggen, J. A., Draper, K. G., Gonzalez, C., Jensen, K., Karpeisky, A. M., Modak, A. S., Matulic-Adamic, J., DiRenzo, A. B., Haeberli, P., Sweedler, D., Tracz, D., Grimm, S., Wincott, F. E., Thackray, V. G., and Usman, N. (1995) J. Biol. Chem. 270, 25702–25708. 22. Beigelman, L., Karpeisky, A., Matulic-Adamic, J., Haeberli, P., Sweedler, D., and Usman, N. (1995) Nucleic Acids Res. 23, 4434 – 4442. 23. Burgin, A. B., Gonzalez, C., Matulic-Adamic, J., Karpeisky, A. M., Usman, N., McSwiggen, J. A., and Beigelman, L. (1996) Biochemistry 35, 14090 –14097. 24. Sczakiel, G. (1995) Angew. Chem. Int. Ed. Engl. 34, 643– 645. 25. Satow, T., Akiyama, T., Machida, A., Utagawa, Y., and Kobayashi, H. (1993) J. Chromatogr. A 652, 23–30. 26. Kramer, F. R., and Mills, D. R. (1978) Proc. Natl. Acad. Sci. USA 75, 5334 –5338. 27. Katsivela, E., and Ho¨fle, M. G. (1995) J. Chromatogr. A 717, 91–103. 28. Katsivela, E., and Ho¨fle, M. G. (1995) J. Chromatogr. A 700, 125–136. 29. Konrad, K. D., and Pentoney, S. L. (1993) Electrophoresis 14, 502–508. 30. Rosenblum, B. B., Oaks, F., Menchen, S., and Johnson, B. (1997) Nucleic Acids Res. 25, 3925–3929. 31. Cordier, Y., Roch, O., Cordier, P., and Bischoff, R. (1994) J. Chromatogr. A 680, 479 – 489. 32. Skeidsvoll, J., and Ueland, P. M. (1996) Electrophoresis 17, 1512–1517. 33. Khan, K., Van Schepdael, A., and Hoogmartens, J. (1996) J. Pharm. Biomed. Anal. 14, 1101–1105. 34. Kuimelis, R. G., and McLaughlin, L. W. (1995) J. Am. Chem. Soc. 117, 11019 –11020.