[13] Preparation of isotopically enriched RNAs for heteronuclear NMR

300

DNA AND RNA STRUCTURE

[13] P r e p a r a t i o n

of Isotopically Enriched Heteronuclear NMR

[131

RNAs for

B y ROBERT T. BATEY, JOHN L. BATTISTE, a n d JAMES R . WILLIAMSON

Introduction The ability to overexpress proteins isotopically labeled with 13C and 15N is a cornerstone of the current NMR methodology used to solve solution structures of large proteins. 1,2,3Recently, similar heteronuclear NMR techniques have been applied to R N A structures in solution. 4,5,6 A number of laboratories have developed techniques for the synthesis of isotopically labeled ribonucleotide triphosphates as precursors for the preparation of any R N A of defined sequence 7's'9,1° and for the efficient assignment of l a b e l e d R N A s . 11,12J3,~4 This article presents a detailed procedure for the production of 13C- and/or 15N-labeled R N A from inexpensive and readily available sources of these isotopes. The procedure, outlined in Fig. 1, is general and can be adapted to the preparation of 13C- and/or 15N-labeled RNA. Bacterial cells are grown in a minimal salts medium containing isotopically labeled carbon and/or nitrogen substrates. E. coli are grown to produce 15N-labeled nucleotides, whereas ~3C-labeled or 13C/15N nucleotides are best produced by growing Methylophilus methylotrophus on 13C-methanol, which is an economical source of 13C. The cells are harvested and lysed by detergent, the proteins are removed by phenol-chloroform extraction, and the total nucleic acids are precipitated with isopropanol. The total cellular nucleic acids are hy1 G. M. Clore and A. M. Gronenborn, Prog. Nucl. Magn. Reson. Spectroc. 23, 43 (1991). 2 M. Ikura, L. E. Kay, and A. Bax, Biochemistry 29, 4659 (1990). 3 G. M. Clore, L. E. Kay, A. Bax, and A. Gronenborn, Biochemistry 30, 12 (1991). 4 E. P. Nikonowicz and A. Pardi, Nature 355, 184 (1992). s E. P. Nikonowicz and A. Pardi, J. Am. Chem. Soc. 114, 1082 (1992). 6 j. Santoro and G. C. King, J. Mag. Reson. 97, 202 (1992). 7 E. P. Nikonowicz, et al., Nucleic Acids Res. 20, 4507 (1992). 8 R. T. Batey, M. Inada, E. Kujawinski, J. D. Puglisi, and J. R. Williamson, Nucleic Acids Res. 20, 4515 (1992). 9 M. J. Michnicka, J. W. Harper, and G. C. King, Biochemistry 3332, 395 (1993). t0 j. V. Hines, S. M. Landry, G. Varani, and I. Tinoco, J. Am. Chem. Soc. 116, 5823 (1994). 11 E. P. Nikonowicz and A. Pardi, J. Mol. Biol. 232, 1141 (1993i. 12j. p. Marino, et al., J. Am. Chem. Soc. 116, 6472 (1994). i3 V. Sklenar, R. D. Peterson, M. R. Rejante, E. Wang, and J. Feigon, J. Am. Chem. Soc. 115, 12181 (1993). 14V. Sklenar, R. D. Peterson, M. R. Rejante, and J. Feigon, J. Biomol. N M R 4, 117 (1994).

METHODS IN ENZYMOLOGY, VOL. 261

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

[13]

301

PREPARATION OF ISOTOPICALLYENRICHED R N A s

Cell growth

.. .

I

]

C,gluco~

is N-ammonium

I

M. methylotrophus:

I 13 C.methanol

M. 13methylotrophus: C-methanol 15 N.ammonium

I

Cell l y s l a ~ ' ~ , ~ , ~ total nucleic acids

Nucleic add hydrolysis NMPs and dNMPs

Ntmleotlde oeparatlon NMPs

Enzymatic phoaphorylatlon

~r crude NTPs

Nucleotlde purification

~r pure NTPs

In vitro

transcription RNA

Fl6. 1. Flowchart outlining the procedure for generating isotopically labeled nucleoside triphosphates.

drolyzed to nucleoside monophosphates using nuclease P1, and the deoxyribonucleotides are separated from the ribonucleotides using a boronate affinity column. The ribonucleoside monophosphates are then enzymatically converted to nucleoside triphosphates, which can be used in transcription reactions to prepare any desired R N A of defined sequence. Note: This procedure has been optimized to a great extent; departures from the described protocol should be avoided without due consideration. It is recommended that these techniques be reproduced with unlabeled materials prior to embarking on a labeled preparation.

302

DNA AND RNA STRUCTURE

[ 131

Methods

Growth of Methylophilus Methylotrophus Reagents Methylophilus methylotrophus (ATCC 53528) (American Type Culture Collection) 99% 13C-methanol (Cambridge Isotope Laboratories) 15N-ammonium sulfate (Cambridge Isotope Laboratories)

Equipment 10 L Microferm Fermentor (New Brunswick Scientific) JA-10 rotor for Beckman J2-21 low-speed centrifuge

Procedure 1. ATCC Medium 154515 (Table I) was prepared as a one liter solution containing only the phosphate salts with the pH adjusted to 6.8 and then autoclaved. Individual 100x stock solutions of ammonium chloride, calcium chloride, and magnesium chloride were made and autoclaved separately. A 1000x trace metals solution was made from concentrated stock solutions of the individual metals and filter sterilized. After the phosphate solution had cooled, these solutions, as well as the methanol, were added. For plate medium, 15 grams of agar were added to the phosphate solution prior to autoclaving. 2. The ATCC lyophilized stock of Methylophilus methylotrophus was revived in 5 ml of ATCC Medium 1545 with shaking at 30° for 2-3 days until the cultures become very cloudy. The revived culture was plated out on ATCC Medium 1545 plates and incubated for 2 days at 37 °. 3. Methanol-minimal salts medium (Table I) was prepared by autoclaving one liter of a solution containing only the phosphate salts, with the pH adjusted to 6.8. Ammonium sulfate and magnesium sulfate were made as individual 100x solutions and autoclaved separately. A 1000x trace metals solution was made from concentrated stock solutions of the individual metals and filter sterilized. After the solution had cooled, these solutions, along with methanol, were added. 4. A 5 ml overnight culture in methanol-minimal salts medium was inoculated with an individual colony and grown for 12-16 hours with shaking at 37 °. Growth in isotopically labeled medium was performed in three stages: an unlabeled 50 ml culture, a labeled 500 ml culture, and a labeled 15 j. F. Hamel, personal communication.

[1 3]

PREPARATION OF ISOTOPICALLY ENRICHED R N A s TABLE I PREPARATION OF MEDIA

ATCC Medium #1545 (per liter) CH3OH Na2HPO4 KH2PO4 NI-I4C1 CaCI2 MgSO4 • 7H20 Trace metals Fe-EDTA FeSO4.7H20 NaMoO4 • 2HEO ZnSO4 • 7H20 MnCI2 • 4HzO H3BO3 COC12 • 6H20 NiClz "6H20 EDTA

5 ml 0.33 g 0.26 g 0.5 g 0.2 g 1.0 g 5.0 mg 500/zg 2.0 mg 400 ~g 20 tzg 15 ~zg 50/zg 10 tzg 25O tzg

Methanol-minimal salts medium (per liter) CH3OH K2HPO4 NaH2PO4 (NH4)2SO4 MgSO4 • 7H20 Trace metals FeSO4 • 7H20 CuSO 4 " 5H20 MnSO4 - 5H20 ZnSO4 - 7H20 CaCI2 • 2H20 COC12 H3BO3 NaMoO4

1 ml 0.95 g 0.78 g 0.36 g 0.2 g 50 mg 100 tzg 50 ~g 50/~g 1.3 mg 10/xg 7/zg 10/xg

Glycerol-minimal salts medium (per liter) glycerol KH2PO4 NaH2PO4 Na. citrate (NH4)2S04 MgSO4 • 7H20 Trace metals CaClz • 2HzO NazEDTA FeC13 • 6H20 CuSO4 • 5HzO MnSO4 - 5H20 ZnSO4 • 7HRO CoCI 2

21.5 g 1.6 g 5.3 g 0.5 g 0.7 g 0.3 g 750/zg 30 mg 25 mg 240 tzg 180/zg 27 ~g 270 p~g

303

304

DNA AND RNA STRUCTURE

[13]

10 L culture. At all stages during this procedure, it is imperative to keep the cultures actively growing and not allow them to reach stationary phase. Growth was monitored by reading the optical density at 660 nm. 5. A 50 ml culture containing t2C-methanol-minimal salts medium was inoculated with 1 ml of the overnight liquid culture stock and grown with vigorous shaking at 37 ° for approximately 12 hours. 6. A 500 ml culture of 13C-methanol-minimal salts medium was inoculated with one ml of the 50 ml 12C-culture to minimize isotopic dilution and incubated at 37 ° with agitation for approximately 18 hours. 7. The entire 500 ml 13C-culture was used to inoculate 10 L of ~3Cmethanol-minimal salts medium in a microfermentor. The microfermentor was set to 37 °, 5 L/min air flow rate, and 600 rpm rotor speed for agitation. Growth was monitored by taking 660 nm absorbance readings every hour (Fig. 2). When the cell density of the culture reached an absorbance of 0.4-0.45, the growth of the cells tapered off, indicating the transition to stationary phase. The temperature of the fermentor was turned down to 4° and allowed to incubate with agitation for another hour. 8. The culture was harvested by centrifugation of the culture in six 500 ml centrifuge bottles at 6400 g (6000 rpm in a JA-10 rotor) for 20 minutes. This yields 15-16 grams of pink, tightly packed wet cell pellets distributed between six 500 ml centrifuge bottles. During centrifugation, the remaining culture continued to be agitated in the fermentor at 4 °. After each centrifugation, the supernatant was poured off and more culture was

0.5

•

0.4"

0.3'

d 6

0.2' 0.1 0.0

10

Time, hours FIG. 2. Growth curve of Methylophilus methylotrophus in a 10 L fermentor. Culture was grown in m e t h a n o l - m i n i m a l salts m e d i u m with an air flow rate of 5 L/min, agitation speed of 600 rpm, and at 37 °. A b s o r b a n e e readings were t a k e n at 660 nm. Using the absorbance readings from 2 to 8 hours, the doubling time of the culture is 1.9 hours.

[13]

PREPARATION OF ISOTOPICALLYENRICHEDRNAs

305

added to each bottle. Four centrifuge runs were required to harvest 10 L of culture.

Comments. Our choice of which methylotrophic bacterial strain to use for isotopic labeling was based in part on the biochemistry of methylotrophy. Methylophilus methylotrophus represents a class of methylotrophic bacteria that incorporate methanol via the ribulose monophosphate cycle in which three equivalents of methanol are converted into 3-phosphoglycerate. 16 Other common methylotrophic bacteria such as Pseudomonas AM1 and Methylobacterium extorquens utilize the serine pathway to incorporate two moles of methanol and one mole of carbon dioxide into one mole of 3-phosphoglycerate.16 There is the potential for dilution of isotopic label via incorporation of atmospheric carbon dioxide with serine pathway organisms. However, growth of Methylobacterium extorquens on 30%J3C-metha nol did not give rise to any apparent isotopic dilution, l° Another primary consideration in isotopic labeling was rapid growth rate on methanol, and Methylophilus methylotrophus reproducibly has grown with a doubling time of 1.9 hours. The doubling rates of other methylotrophic bacteria can be considerably slower. ~° The percentage of RNA as dry mass of a cell growing in log phase is a function of the growth rate. As the doubling time of the cell decreases, the percentage of cell dry weight that is RNA increases. 17 The relatively rapid doubling time of Methylophilus methylotrophus on methanol compared to other methylotrophic bacteria maximizes the amount of RNA yielded per input gram isotopic substrate. The primary problem related to the growth of Methylophilus methylotrophus has been the production of a heteropolysaccharide sheath during logarithmic growth. 8 This growth characteristic generates bacteria that yield loosely packed white pellets when centrifuged. These bacteria produced a great deal of non-nucleic acid material during isopropanol precipitations and resulted in very low yields of cellular nucleic acids. 8 We have empirically found that after cells have stopped growing, they rapidly lose this sheath of material if they are continued to be agitated. By lowering the temperature of the cells from 37° to 4° just before they reach the stationary phase, while continuing to agitate the cells for an additional hour, we obtained wellpacked cells that do not yield non-nucleic acid material while still maintaining a high yield of total cellular RNA. Agitation of the cells in a 16C. Anthony,"The Biochemistryof Methylotrophs."AcademicPress, London,1982. a7D. Herbert,"The ChemicalCompositionof Micro-Organismsas a FunctionoftheirEnvironment," 11, 391-416. UniversityPress, London,1961.

306

DNA aND RNA STRtJCrURZ

[13]

fermentor seems to be significantly more effective than shake flasks with respect to this problem.

Growth of E. coli Reagents E. coli strain ATCC 15244 (American Type Culture Collection) 15N-ammonium sulfate (Cambridge Isotope Laboratories)

Equipment JA-10 rotor for Beckman J2-21 low-speed centrifuge

Procedure 1. Glycerol-minimal salts medium 15(Table I) was prepared by autoclaving one liter of a solution containing glycerol and the phosphate salts with the pH adjusted to 7.0. Sodium citrate, ammonium sulfate, and magnesium sulfate were made as individual 100× solutions and autoclaved to sterilize. A 1000× trace metals solution was made as a mixture of all of the individual metals and filter sterilized. After the glycerol/phosphate solution had cooled, the other nutrients were added. 2. A 5 ml culture of E. coli was grown on glycerol-minimal salts medium containing 1aN-ammonium sulfate overnight at 37° with agitation. 3. One ml of this culture was used to inoculate a one liter culture of glycerol-minimal salts/phosphate medium containing ~SN-ammonium sulfate in a 4 L Erlenmeyer flask and incubated at 37 ° with vigorous agitation for 15 hours. 4. The cells were harvested by centrifugation at 6400 g (6000 rpm in a JA-10 rotor) for 20 minutes. Each liter of cell culture yielded approximately 4 g of wet-packed cells.

Comments. The yields for various labeling schemes are shown in Table II. For the production of 15N-labeled nucleotides, E. coli is the recommended organism because it is simpler to grow than Methylophilus methylotrophus, although the yields of NMPs per input gram of isotope are most likely comparable between the two organisms. Cell Lysis Reagents STE buffer (0.1 M NaC1, 10 mM Tris-C1 [pH 8.0], 1 mM EDTA [pr~ S.0]) 10% Sodium dodecyl sulfate, pH 7.2

[13]

PREPARATION OF ISOTOPICALLYENRICHEDRNAs

307

TABLE II YIELDS OF LABELEDRIBONUCLEOT1DES

Organism E. coli E. coli M. methylotrophus

M. methylotrophus M. methylotrophus

g cells g isotope

mg NMPs g cells

m8 NMPs g isotope

13C6-glucose + (15NH4)2SO4 (15NH4)2504 13C-CH3OHa

2.9

20

58

6.4 2.5

7 10

45 25

13C-CH3OHb 13C-CH3OH+ (15NH4)2SO4b

1.9 1.5

27 32

54 48

Limiting nutrient

methylotrophus grown in 1 L shake flasks. M. methylotrophus grown in a 10 L fermentor.

M.

0.1 M Tris-C1, p H 8.0 Phenol containing 0.1% 8-hydroxyquinoline Chloroform solution containing 24 : 1 chloroform: isoamyl alcohol 3 M sodium acetate, p H 5.2 Isopropyl alcohol Equipment

Waring blender JA-10 rotor for Beckman J2-21 low-speed centrifuge Procedure

1. Phenol containing 0.1% 8-hydroxyquinoline was prepared from solid reagent grade phenol by equilibrating against STE buffer, is The solid phenol was liquefied by incubating in a 65 ° water bath. The liquid phenol was extracted twice against 0.1 M Tris, p H 8.0, and twice against STE buffer in a one liter separatory funnel. 0.1% w/v 8-hydroxyquinoline was added to the buffered phenol and the phenol solution was stored in an amber bottle at 4 °. Caution: Phenol is an extremely caustic compound and appropriate protection (gloves, labcoat, eye and face protection) should be used in all steps utilizing phenol. 2. Approximately 4 g of wet-packed cells were resuspended thoroughly in 5 ml of STE buffer, ensuring that there were no clumps of cells in the suspension. 18j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning: A Laboratory Manual." Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1989.

308

DNA AND RNA STRUCTVRE

[13]

3. This solution was added slowly to a mixture of 90 ml STE buffer and 5 ml 10% SDS at 37 ° with vigorous stirring. The cell solution was allowed to stir for 15 minutes. 4. The cell solution was added to 200 ml of a 1 : 1 phenol : chloroform solution that was preheated to 65 °. This mixture was incubated for 30 minutes at 65° with occasional vigorous mixing to keep the organic and aqueous phases emulsified. 5. The aqueous and organic phases in the emulsion were separated by centrifuging for 10 minutes at 6400 g (6000 rpm in a JA-10 rotor) in 500 ml polypropylene (or other chemically compatible) centrifuge bottles. The upper aqueous layer was removed with an aspirator, taking care not to disrupt the white protein inclusion layer. 6. The organic and inclusion layers were extracted twice against 100 ml of STE buffer in a Waring blender set on high for 5 minutes, and the aqueous layer was removed after centrifugation. 7. The aqueous layers from the extractions were pooled, yielding a total of 300 ml volume, and extracted with an equal volume of chloroform solution. The phases were allowed to separate completely without centrifugation before the aqueous layer was removed. 8. The nucleic acids in the aqueous layer (approximately 300 ml) were precipitated by adding 30 ml of 3 M sodium acetate, pH 5.2, and 330 ml isopropyl alcohol to the aqueous phase and incubating overnight at - 2 0 °" 9. The solution was centrifuged at 6400 g (6000 rpm) for 30 minutes in 500 ml centrifuge bottles in a JA-10 rotor, the liquid decanted, and then the pellet was air dried and slowly resuspended in 25 ml of STE buffer. Four grams of wet cells should yield approximately 3500 A260units of crude nucleic acids for both Methylophilus methylotrophus and E. coli.

Comments. This cellular lysis procedure yields total nucleic acids, which necessitates the separation of deoxynucleotides from ribonucleotides at a later stage in the protocol. Alternative procedures have been worked out in other laboratories where D N A and R N A are separated from one another at the cellular lysis stage. Using a phenol/SDS solution of low pH, D N A can be partitioned with the protein into the organic and white inclusion layers, allowing R N A to be selectively removed with the aqueous layer. 1° Residual D N A was removed with RNase-free DNase I prior to ethanol precipitation. Cellular lysis has also been achieved using methods that do not involve organic extraction of cells. Gentle disruption of cells with lysozyme and SDS followed by selective salting out of proteins, DNA, and SDS with NaC1 yields an aqueous phase highly enriched in cellular RNA, which was

[13]

PREPARATION OF ISOTOPICALLY ENRICHED R N A s

309

then collected by ethanol precipitation. 9 Another alternative is to disrupt cells with a French press and perform a ribosome preparation of the supernatant. 7 Since the majority of cellular RNA of logarithmically grooving eells is ribosomal RNA, this yields approximately 80% of the total possible cellular RNA. Also, this method does not yield tRNA, a significant source of modified ribonucleosides, although this has never been a significant problem using other protocols.

Nucleic Acid Hydrolysis Reagents Nuclease P1 (Boehringer Mannheim) Calf intestinal alkaline phosphatase (Boehringer Mannheim) Phosphatase buffer (50 mM Tris-HCl, pH 8.5, 0.1 mM EDTA, pH 8.5) 83.3 mM triethylammonium phosphate, pH 6.0 in 4% methanol 3 M sodium acetate, pH 5.2 40 mM ZnSO4

Equipment 250 × 4.6 mm Vydac C18 HPLC column (Vydac)

Procedure 1. The resuspended nucleic acids from a single lysis procedure (32503500 A260 units in 25 ml STE) were denatured in a boiling water bath for 5-6 minutes. 2. The solution was cooled in an ice bath before adjusting to 15 mM sodium acetate, 0.1 mM ZnSO4, pH 5.2 by adding 125 txl 3 M sodium acetate, pH 5.2, and 62.5 t~l 40 mM ZnSO4. 3. Nucleic acids were hydrolyzed by adding 6 units of nuclease P1 and incubating at 650.19 After one hour, an additional 6 units of P1 nuclease were added to ensure complete hydrolysis of the nucleic acids. 4. Hydrolysis of the nucleic acids to nucleotides was monitored by reverse phase HPLC using a C18 column. For efficient analytical separation of ribonucleotides and deoxyribonucleotides, the nucleotides were dephosphorylated with calf intestinal alkaline phosphatase prior to injection. A sample of 20-30 tzg of nucleoside monophosphates was dephosphorylated by incubating with 2 units of calf intestinal alkaline phosphatase for 30 19 S. L. Haynie and G. M. Whitesides, Appl. Biochem. Biotechnol. 23, 205 (1990).

310

DNA ANDRNA STRUCTURE

C/U

[13]

G -0.05 au

A

I

0

. . . . . . . . . . . . . . . . . . .

4

8

12

16 20 24

I

28 32 36 40

Minutes FIo. 3. Reverse phase HPLC chromatogramof nucleoside mixtures from a culture Methylophilus methylotrophusgrown in a 10 L fermentor. A total of 250/~g of nucleoside monophosphates were dephosphorylated with calf intestinal alkaline phosphatase prior to injection onto an Vydac C18 reverse phase column. Elution was isocratic, with methanol-83.3 mM triethylammonium phosphate, pH 6.0 (4 : 96 v/v), a flow rate of 1.0 ml/min, and detection at 268 nm at 0.5 absorbance units full scale (aufs).

minutes at 37 ° in phosphatase buffer, followed by another 2 units of phosphatase and continued incubation for 30 minutes. 5. H P L C analysis of the reaction mixture was performed on a C18 column using an isocratic elution at a flow rate of 1 ml/min. 2° Sample volumes of 1 0 - 2 0 / z l were injected and nucleosides were detected at 268 nm with a U V detector, as shown in Fig. 3.

Comments. The use of H P L C at this stage also allows the R N A / D N A ratio to be measured, which was useful for optimizing growth conditions of bacteria. Using a previously cited protocol for the growth of Methylophilus methylotrophus, a R N A : D N A ratio of 2.4 was observed. 8 Improvements in the growth of this strain by using a fermentor growth showed a significant increase in the R N A : D N A ratio to 4.8 (Fig. 3). The composition of the ribonucleotide mixture was estimated from the peak areas in Fig. 3, correcting for absorbance differences at 268 and 260 nm, as 22% AMP, 24% CMP, 31% GMP, and 23% UMP. 20C. K. Lim and J. J. Peters, J. Chromatogr.461, 259 (1989).

[ 13]

PREPARATION OF ISOTOPICALLYENRICHEDRNAs

311

Separation of Deoxyribonucleotides and Ribonucleotides Reagents Affi-gel 601 boronate-derivatized polyacrylamide gel (Bio-Rad) TE buffer (10 mM Tris-HC1, pH 8.0, 1 mM EDTA, pH 8.0) 1 M triethylammonium bicarbonate (TEABC), pH 9.5

Equipment 20 × 2.5 cm glass column (Bio-Rad) Fraction collector (Gilson)

Procedure 1. 1 M triethylammonium bicarbonate (TEABC), pH 9.5, was prepared by bubbling CO2 through 141 ml triethylamine in 700 ml H20 at 5° until the pH dropped to 9.5, then H20 was added to a final volume of one liter. To adjust the pH of 1 M triethylamine to 9.5 required 2-4 hours when CO2 was bubbled through a glass Pasteur pipette. 2. To prepare the affinity chromatography column, 5 g Affi-gel 601 was hydrated in TE buffer and packed in a 20 × 2.5 cm glass column. The column was equilibrated with 1 M TEABC at 5 °. 3. The nucleotide solution (-<6000 A260 units) from the P1 digestion was lyophilized and resuspended in 10 ml of 1 M TEABC. 4. The sample was applied to the column and then washed with 1 M TEABC while collecting 5 ml fractions until the A260 of the eluant dropped below 0.1 (Fig. 4a). 21 The deoxyribonucleotides, salts, and other impurities washed through the boronate column while the ribonucleotides remained covalently bound to the boronate ligand. 5. To elute the bound material, the column was washed with H20 that had been acidified to pH 4-5 by bubbling with CO2. Elution was continued until the A260 of the eluant dropped below 0.1, which usually occurred 100 ml after the start of elution (Fig. 4a). 6. Fractions of low pH eluant >0.1 A260 were pooled, lyophilized, and resuspended in 5 ml of H20. Complete separation of dNMPs from NMPs using this method was verified by C18 reverse-phase HPLC. 8

Comments. The boronate chromatography procedure allows one to quantitatively and reproducibly separate deoxyribonucleotides from ribonucleotides. This chromatography, however, must be performed carefully if one is to achieve these results. First, all boronate chromatography must be carried out at 5° since ribonucleotides have a markedly reduced affinity 2xH. Schott, E. Rudloff, P. Schmidt, R. Roychoudhury, and H. Kossel, Biochemistry 12, 932 (1973).

312

DNA AND RNA STRUCTURE

[13]

40,

(a)

dNMPs 0 tO 04

e0 ¢: ¢=

== 0 W

30-

rNMPs 20-

10

<

0

IL...........L

......

100

200

Fraction

(b)

=

20'

0 N

0 rm

10-

0 r~

<

0 0

10

20

30

40

50

60

Fraction Fie. 4. Purification of ribonucleotides on an Affi-gel 601 boronate affinity column. The column contained 5 g of hydrated Affi-gel 601 in a 20 × 2.5 cm glass column, and loading, washing, and elution stepwise was done using gravity elution. The nucleotides were bound and washed with 1 M TEABC, pH 9.5, and eluted with H20 that had been acidified to pH 4-5 with CO2. The absorbance of individual fractions was monitored at 260 nm. (a) Separation of deoxyribonucleotides and ribonucleotides. The total nucleotide pool after P1 nuclease hydrolysis was applied to the boronate affinity column while collecting 2 ml fractions. Elution with acidified water began with fraction 58. (b) Desalting of ribonucleoside triphosphates. NTPs were applied to the boronate affinity column, collecting approximately 7 ml fractions. The NTPs were eluted with acidified water and applied beginning with fraction 13.

for the boronate ligand at room temperature. Second, a wide column was important for good flow rates because of resin volume changes with pH and ionic strength. Although this property is seemingly inconvenient, it provides a visual gauge of the progress of elution. Immediately after elution is initiated by washing the column with acidified water, the column matrix begins to visibly swell. During this time, the rNMPs do not appreciably

[13]

PREPARATION OF ISOTOPICALLY ENRICHED R N A s

313

elute off the column. After further washing, the column matrix begins to shrink and continues to shrink until it reaches about half of its original volume. During this time the matrix bed turns visibly darker and ribonucleotides begin to elute. The completion of this process correlates to the completion of nucleotide elution from the column. Furthermore, it has been observed that the beginning of the elution peak is enriched in pyrimidines, whereas the tail of the elution peak is enriched in purines. Third, it is very important to load significantly less than the advertised binding capacity to minimize ribonucleotides eluting in the wash. Typically, we load approximately 200 mg of rNMPs (6000 A260 units) at a time onto the column. The use of alternative published techniques for cell lysis circumvents the need to separate deoxyribonucleotides and ribonucleotides. 4'9'~°

Separation of Individual Ribonucleotides Reagents HPLC Solvent A (0.05 M NH4COOH, pH 3.0) HPLC Solvent B (0.5 M NH4COOH, pH 2.5)

Equipment 100 × 7.8 mm HP-PEI HPLC column (Interaction Chemicals) Fraction collector (Gilson)

Procedure 1. The pooled ribonucleotides in 5 ml H20 were injected onto the HPPEI HPLC column in 0.5 ml aliquots containing no more than 70 mg of ribonucleoside monophosphates (-2100 Az60 units). Separation was carried out using a two-step profile: isocratic elution at 100% solvent A for 20 minutes and then a linear gradient of 0 to 100% solvent B over 20 minutes at a flow rate of I ml/min. Nucleotides were detected at 300 nm with a UV detector and 2 minute fractions were collected (Fig. 5b). 2. Fractions corresponding to each nucleotide were pooled and lyophilized to dryness. To remove completely trace amounts of ammonium formate, each pool of nucleotides was resuspended in 20 ml of H20 and lyophilized to dryness.

Comments. Ion exchange HPLC rapidly separates the four individual ribonucleoside monophosphates with linear scale-up capacity and baseline resolution. This column can resolve all four ribonucleotides with baseline resolution between >100/zg and 70 mg of total ribonucleotides without a significant loss in resolution (Fig. 5a,b). Reinjection of the pool of GMP onto the column reveals no cross-contamination with the other three nucleotides

314

D NA AND R N A STRUCTURE (a)

(b)

C

A

U G

[13]

(c)

0.2 au I A

I

0

,

,

,

. . . . .

I

10 20 30 40 Minutes

.

.

.

.

.

.

.

.

10 20 30 40 Minutes

I

~,

0

. . . . . . .

I

10 20 30 40 Minutes

FIG. 5. Ion exchange HPLC separation of the individual ribonucleoside monophosphates. Ribonucleoside monophosphates were injected onto a HP-PEI ion exchange HPLC column and eluted using the scheme described in the text with a flow rate of 1 ml/min. (a) Injection of approximately 100/zg of total ribonucleoside monophosphates in a volume of 0.5 ml with detection at 268 nm, 2 aufs. (b) Injection of approximately 65 mg of total ribonucleoside monophosphates in a volume of 0.5 ml with detection at 300 nm, 2 aufs. (c) Fractions from injection (b) corresponding to the guanosine monophosphate peak were pooled and 100/zg reinjected in a volume of 0.1 ml, with detection at 268 nm, 2 aufs.

(Fig. 5c). H P L C separation of NMPs also has been successfully performed using a Nucleogen D E A E column to separate the individual nucleotides. 9 A n alternative to H P L C is anion-exchange liquid chromatography using a A G 1 - X 2 or A G 1 - X 8 resin. 7'~° The nucleotides are eluted from the column using either a linear gradient of increasing salt 9 or step gradient of decreasing pH. 7

Enzymatic Phosphorylation Reagents Phosphoglycerate mutase from rabbit muscle (Boehringer Mannheim) Nucleoside monophosphate kinase from beef liver (Boehringer Mannheim) Enolase from baker's yeast (Sigma) myokinase from chicken muscle (Sigma) pyruvate kinase from rabbit muscle (Sigma) guanylate kinase from porcine brain (Sigma)

[ 13]

PREPARATION OF ISOTOPICALLYENRICHEDRNAs

315

3-phosphoglycerate (barium salt) (Sigma) AG 50W-X8 ion exchange resin (H + form, 200-400 mesh) (Bio-Rad) Argon HPLC Solvent A (0.045 M NH4COOH, pH 4.6) HPLC Solvent B (0.5 M NaH2PO4, pH 2.7) Equipment

Jenco p H / O R P Controller, Model 3671 (Markson) Peristaltic pump (Gilson) Vydac Nucleoside Analysis HPLC column (Rainin) Procedure

1. The sodium form of AG 50W-X8 was prepared by treating 25 g of resin with three 50 ml exchanges of 1 M NaC1 for 15 minutes for each exchange, followed by three exchanges with H20. 2. The sodium form of 3-phosphoglycerate was prepared by vigorously stirring 2 g of barium 3-phosphoglycerate with 15 ml of a 50% slurry of Na÷-AG 50W-X8 resin for 30 minutes. The resin was removed by filtration and washed three times with 5 ml H20. The pH of the combined filtrates was adjusted to 7.5 using 1 M NaOH. The final concentration of sodiumexchanged phosphoglycerate was approximately 0.2 M. 3. Enzymatic phosphorylation of approximately 83 mg of nucleoside monophosphates to nucleoside triphosphates was performed in a 50 ml three-necked round bottom flask flushed with argon at ambient temperature, using a modification of a procedure developed by Whitesides. 19 4. The NMPs were dissolved in a solution containing 15 mM KC1, 75 mM MgCl2,15 mM dithioerythritol, and 10 mM sodium 3-phosphoglycerate 22 to give a final concentration of 10 mM total NMPs. The pH of the solution was adjusted to 7.5 with 1 M HCI and was maintained during the course of the reaction with a pH controller delivering 0.1 M HC1 via a peristaltic pump. 5. The phosphorylation reaction is initiated by first generating ATP. Synthesis of ATP was initiated by adding 1/xM ATP, 10 units phosphoglycerate mutase, 200 units myokinase, 100 units enolase, and 200 units pyruvate kinase. 6. After 3 hours, when ATP represents >90% of the adenosine nucleotide pool, 0.5 units of guanylate kinase and 1.0 unit nucleoside monophosphate kinase were added, and the concentration of 3-phosphoglycerate was increased to 20 mM. 7. Conversion of NMPs to NTPs was monitored by HPLC, and several chromatograms during the course of a typical phosphorylation are shown 22E. S. Simon, S. Grabowski, and G. M. Whitesides,J. Org. Chem. 55, 1834 (1990).

316

DNA

RNA

AND

[13]

STRUCTUPJ~

in Fig. 6. Populations of NMPs, NDPs, and NTPs were analyzed with the Vydac Nucleotide Analysis column using a linear gradient from 0 to 100% solvent B in 10 minutes and a flow rate of 2.0 ml/min. Sample volumes of 20/zl were injected, and nucleotides were detected at 268 nm with a UV detector. 8. During the synthesis of CTP, GTP, and UTP, the concentration of 3-phosphoglycerate was increased by 5 mM every 3-4 hours. Complete phosphorylation of all nucleotides occurs within 12-14 hours, and the reaction mixture was stored at - 2 0 °. Comments. This procedure has been used successfully to phosphorylate up to 500 mg of nucleotides by linearly scaling up the volume of the reaction while the concentration of the individual components was kept constant. The phosphorylation of GMP to GTP was performed using the same protocol as described previously, with a few modifications. The concentration of catalytic ATP was increased to 0.3 mM; a significant concentration of ATP is required to drive the reaction forward at an appreciable rate. Also, nucleoside monophosphate kinase was omitted from this reaction while

n

~i

t = 0 hrs

t =3 hrs

jj• /

I

,

,

,

0246810 Minutes

,

I

246810 Minutes

t =6 hrs

t-ll

hrs

t= 14hrs

UUL-

4~

0246810 Minutes

I . . . . 0246810 Minutes

I

t .... 0246810 Minutes

Fxc. 6. HPLC chromatograms of the reaction time course of an enzymatic nucleotide phosphorylation reaction. At the zero time point, a catalytic quantity of ATP was added to the reaction, along with myokinase, to initiate the reaction. After 3 hours, when the vast majority of AMP had converted to ATP, guanylate kinase and nucleoside monophosphate kinase were added to initiate the phosphorylation of the other nucleoside monophosphates. The chromatograms are of 15 nmol of nucleotides injected onto a Vydac nucleotide analysis column using the elution scheme described in the text at a 2.0 ml/min flow rate and detecting at 268 nm at 0.2 aufs.

I

[13]

PREPARATION OF ISOTOPICALLYENRICHEDRNAs

317

increasing the guanylate kinase added to 3 units. Similar modifications of the CMP and UMP reactions have been reported by other laboratories. 9"1° Nucleotide Purification Equipment Centrifugal microconcentrators, 10,000 molecular weight cut-off (Amicon) JA-10 rotor for Beckman J2-21 low-speed centrifuge Procedure 1. The phosphorylation reaction mixture was lyophilized, resuspended in 10 ml ice cold 1 M TEABC, pH 9.5, and purified using the boronate affinity column procedure described in the separation of ribonucleotides and deoxyribonucleotides. 2. The lyophilized NTPs were resuspended in 2 ml H20. Any remaining high molecular weight contaminants were removed by passing the nucleotide solution through a Centricon 10,000 molecular weight cut-off microconcentrator at 2200 g (4000 rpm) in a JA-17 rotor and collecting the filtrate. 3. The concentration of the NTP solution was estimated by UV absorbance at 260 nm. Comments. This step desalts the nucleotides and removes any high molecular weight impurities, which is critical for their function in transcription reactions in vitro. However, other protocols have simply used ethanol precipitation to recover the nucleotide triphosphates from the enzymatic phosphorylation reaction and used them directly in transcription reactions without further purification. 9,1° In vitro Transcription of R R E RNA Procedure. RNA was synthesized by in vitro transcription with T7 RNA polymerasez3 from an oligonucleotide template with a single-stranded template region and a double-stranded promoter region. Oligodeoxynucleotides were synthesized by standard phosphoramidite chemistry using an automatic DNA synthesizer. RRE RNA uniformly labeled with I3C was synthesized in three 50 ml transcription reactions. Each was incubated for 4 hours at 37° in 80 mM K +. HEPES, pH 8.1, 1 mM spermidine, 5 mM DTT, 16 mM MgCIE, 0.01% Triton X-100, 80 mg/ml polyethylene glycol (8000 molecular weight), 10.0 mM NTPs (~2.5 mM each), 400 nM each 23 j. F. Milligan, D. R. Groebe, G. W. Witherell, and O. C. Uhlenbeck, Nucleic Acids Res. 15, 8783 (1987).

318

D N A AND R N A STRUCTURE

[131

55 GG

(a)

aQucua ccaaAC

S0

CaCAGC% acGuca A

GI A U 70

75

IA 641

58

C1'

(b)

C4'

I Fully 13C labeled RRE RNA

I

'~,'D"° ""

,O

o

o

o~

G4S Hl'le]", Q

bD O

t •

o

~O

o.. -r"

,,

o g~Ooo

,i

•

o o eL'

•

~__.',

)

peaks

peaks

o.

positive, negative

•

m.+ 92.0 m.o 88.o

(c)

O'

G71 HI' [~" ' - . .

o

~

O ~ o

88.o m.+

82.o

8o.o

Guanosine mC labeled RRE RNA '

0*

G71 HI'~]~ ' - .

°

-

C~ H I ' [ ] "

D

-r

i

v-

i i

(--', ) positive• negalive peaks peaks

9~,.0

9'~.0

90.0

8ko

86.0 8Lo

82.0

q ¢.D

86.0

13C PPM FIG. 7. Fully 13C-labeled and guanosine laC-labeled RRE RNA were prepared as described in the methods section. A 1 : 1 complex of each RRE RNA with a basic peptide from the Rev protein of HIV was formed as described.2s The final concentrations were 1.5 mM for fully 13C-labeled and 1.2 mM for guanosine 13C-labeled RNA-peptide complex in 500/zl D20 (10 mM NaPO4, pH 6.5, 50 mM NaC1, 0.1 mM EDTA). (a) Sequence of RRE RNA used for NMR studies. The discontinuity in numbering at adenosine 58 is to keep the numbering in the internal loop region the same as wild-type RRE. (b) C1'/C4' region of an HSQC-CT spectrum of fully 13C-labeled RRE RNA in complex with unlabeled Rev peptide. The constant time interval was set to 1/Jcc = 25 ms with sweep widths of 5500 and 5000 Hz in the proton

[131

PREPARATION OF ISOTOPICALLYENRICHEDRNAs

319

D N A strand, and 0,063 m g / m l T7 R N A polymerase. 24 R R E R N A labeled with 13C-GTP was transcribed in a similar reaction, but with 2 m M ATP, CTP, and UTP, and 4 m M laC-GTP and 0.075 m g / m l T7 R N A polymerase. Purification and preparation of the R N A for N M R was p e r f o r m e d as described elsewhere. 2s Comments. The concentration of Mg 2+ in the transcription reaction needs to be reoptimized for each new template and each new preparation of NTPs, since some Mg 2+ may copurify with the NTPs during the preparation. The optimal added MgC12 concentration for transcription of the R R E template differed significantly for commercial N T P s (36 m M added MgCI2) and isotopically labeled cellular NTPs (16 m M added MgC12). This seems to be a general characteristic of nucleotides generated using this and other protocols. For a n u m b e r of different R N A s , the amount of Mg 2÷ added to the transcription is markedly less. 8'9a°

Applications A complete overview of the m e t h o d o l o g y for using isotopically labeled R N A in N M R structure determination is presented elsewhere in this volume. 26 H e r e , we present some examples of heteronuclear N M R applied to the sequential assignment of H I V R e v Responsive E l e m e n t ( R R E ) R N A .

Heteronuclear Single Quantum Coherence Since the pattern of N O E s between base and H I ' protons is used for sequential assignment, determination of the chemical shifts of all H I ' protons is an important first step for assignment of R N A molecules. A useful experiment for assignment of ribose protons is a variation of the H e t e r o nuclear Single Q u a n t u m Coherence ( H S Q C ) experiment that uses the constant-time m e t h o d ( H S Q C - C T ) to r e m o v e carbon-carbon couplings. 6 W h e n the constant time interval is set to an odd integer multiple of 1/Jcc, where 24j. Grodberg and J. J. Dunn, J. Bacteriol. 170, 1245 (1988). 25j. R. Wyatt, M. Chastain, and J. D. Puglisi, Biotechniques 11, 764 (1991). 26A. Pardi, this volume, Chapter [15].

and carbon dimensions, respectively.6 1024 and 112 complex points were acquired in t2 and tl, respectively, with 32 scans per slice. Positive and negative contours were plotted the same, but are indicated by being either to the left or right of the drawn dotted line. Note the two cross-peaks that resonate in the C4' carbon chemical shift range, yet have positive crosspeaks. (c) This spectrum is the same as in (b) except that the RNA is labeled only with 13C guanosine.

320

D N A AND R N A STRUCTURE

[13]

(a) e ........... i-

'Ip..~

G4

~

~ ~oGe4HI'~

A7

I1.

o~. o~:--:~'-

. B..o

...... i-o

::

.~-

@-41~34

O~ 3

O 0 6'.0

(b)

d~ .................... p sls 1H PPM (HI'/H5)

o 5'.o

C~GTSH8- A75 HI'

oq o G-nH8*CTsH5

O G~ He-Ar~HI'

~:) GTOH8- C~ HI' ~H8"C49H1'

O

.p

O G~He- ce4 H I ' G~He- U~HP ~C~71 1.18- A731-11'

=E O. 13. -t-

,v-

~GS4 HS- Cs6 H5

%~,.s- u~.1'

610

i 5.5

c01 ~H PPM (~2C)

510

o

[13]

PREPARATION OF ISOTOPICALLY ENRICHED R N A s

321

Jcc is the one bond carbon-carbon coupling constant, the sign of the crosspeak depends on the number of carbon-carbon bonds. The ribose CI' and C5' cross-peaks will have a different sign than the C2', C3', and C4' crosspeaks because they have one and two carbon-carbon bonds, respectively. This difference in sign can be used to help assign cross-peaks where there is an ambiguity in the carbon chemical shift. The C4'/C1' region of an HSQC-CT experiment of RRE RNA complexed with a peptide from the Rev protein is shown in Fig. 7b. All of the Hl's can be identified by their characteristic chemical shifts and the sign of the cross-peaks. Interestingly, there are two Cl's that have chemical shifts more typical of a C4', and they were originally misidentified as H4' protons until this experiment was performed. Although fully labeling an RNA with 13C can alleviate many spectral overlap problems through the utilization of three-dimensional experiments, many ambiguities may still exist with RNAs larger than 30 nucleotides. One approach to solve this problem is to selectively label an RNA with each nucleotide (G, C, A, or U) separately, which allows for unambiguous identification of ribose spin systems by nucleotide type using the standard methods for isotopic RNA. n A simple example of this is shown in Fig. 7c, which is an HSQC-CT spectrum of the R R E - R e v complex labeled only with 13C guanosine. The spectrum is greatly simplified compared to the fully labeled spectrum, and all of the HI' chemical shifts belonging to guanosine are easily identified.

X-Double Half-Filter Selective labeling of RNA molecules with individual nucleotides permits isotopic filtering experiments that select for protons attached to either 12C

FIG. 8. (a) A NOESY spectrum of unlabeled RRE RNA in complex with the Rev peptide. The sequential assignment pathway between the H8/H6 and H I ' protons in the RNA is traced out. The sequence G41-A58is a dotted line, whereas the sequence G64"C79is a solid line. The H I ' chemical shift of the nucleotide just 3' of the tetraloop (G64) is offscale at -3.75 ppm, as was observed in previous NMR studies of the G N R A tetraloops. 29The NOESY experiment was 2048 × 512 complex points in t2 and tl, respectively, with a sweep width of 5000 Hz in both dimensions and 32 scans per slice. The mixing time was 400 ms. (b) One subspectrum of a X-double-half-filtered NOESY experiment of guanosine 13C-labeled RNA-peptide complex. This subspectrum has the protons attached to 12C along the to1 axis (H5/HI' region) and protons attached to 13C along the oJ2 axis (H8/H6/H2 region). The assignments of all H8 to H5/HI' NOEs are displayed. The pulse sequence used is the same as described. 27The spectrum was 1024 × 256 complex points in t2 and h, respectively, with a sweep width of 5500 Hz in both dimensions and 16 scans per slice. Carbon decoupling was achieved by applying a 180° carbon pulse in the middle of tl and W A L T Z decoupling during t2. The 1/2J delay for the filtering was set to 3.1 ms, and the mixing time was 300 ms.

322

DNA AND RNA STRUCTURE

[13]

or 13C. A useful filtering experiment is the X-double-half-filtered N O E S Y . 27 In this experiment, the standard NOESY spectrum is edited into four different subspectra. Two subspectra contain cross-peaks where both protons involved in the NOE are attached to either lzC- or 13C-carbons, respectively. The two other subspectra contain cross-peaks where one proton involved in the NOE is attached to ~3C and the other to ~zC. The last two subspectra allow unambiguous identification of many of the internucleotide base-H1' NOEs depending on the sequence of the RNA. Figure 8a shows the NOESY spectrum of the base to H I ' region for the unlabeled RRE/ Rev complex with the sequential assignment pathway traced out. Figure 8b shows a region of a X-double-half-filtered NOESY subspectra of 13C guanosine-labeled R R E where all cross-peaks are NOEs between a G-H8 and a H I ' / H 5 on any nucleotide other than G (which must be an internucleotide NOE). Having this information reduces the possible ambiguities involved in determining the identities of the sequential H8-HI' cross-peaks. Conclusion This procedure outlines an economical and efficient method for the production of isotopically labeled RNAs for study by NMR. The practical utility of this technique has been demonstrated using R R E RNA, an R N A whose spectrum is difficult to interpret without heteronuclear methods. It is hoped that there will be an impact of heteronuclear methods comparable to that realized for protein structure on the size and diversity of RNAs that can be studied using NMR. Acknowledgments This work was supported by grants from the Searle Scholars Program of the Chicago Community Trust and from the National Institutes of Health (GM-46314 and GM-39589).

27 G. Otting and K. Wuthrich, Q. Rev. Biophys. 23, 39 (1990). 28 j. L. Battiste, R. Tan, A. D. Frankel, and J. R. Williamson, Biochemistry 33, 2741 (1994). 29 H. A. Heus and A. Pardi, Science 253, 191 (1991).