Chapter 2 Ribonucleosides in Biological Fluids by A High-Resolution Quantitative Rplc-UV Method

Chapter 2 Ribonucleosides in Biological Fluids by A High-Resolution Quantitative Rplc-UV Method

C41 CHAPTER 2 RIBONUCLEOSIDES IN BIOLOGICAL FLUIDS BY A HIGHRESOLUTION QUANTITATIVE RPLC-UV METHOD KENNETH C. KUO1, DAT T. PHAG, NATHAN WILLIAMS', an...

6MB Sizes 0 Downloads 0 Views

C41

CHAPTER 2 RIBONUCLEOSIDES IN BIOLOGICAL FLUIDS BY A HIGHRESOLUTION QUANTITATIVE RPLC-UV METHOD KENNETH C. KUO1, DAT T. PHAG, NATHAN WILLIAMS', and CHARLES W. GEHRKE' ]Department of Biochemistry, University of Missouri and Cancer Research Center, Columbia, MO 65201 (USA) 2Hewlett Packard Corporation, Avondale. PA ( U S A )

TABLE OF CONTENTS 2.1 Introduction 2.2 Experimental . 2.2.1 Chemicals 2.2.2 Ultrafiltration Procedure . 2.2.3 Phenylboronate Gel Column Procedure . 2.2.4 Procedure for Isolation of Urinary Nucleosides . 2.2.5 Procedure for Isolation of Serum Nucleosides . 2.2.6 Preparation of Internal Standard Solutions . 2.2.7 Determination of Adenosine Deaminase Activity in Serum and Urine . 2.2.8 HPLC Instruments and Conditions . 2.2.9 Creatinine Analysis by HPLC-UV . 2.3 Results and Discussion 2.3.1 Chromatography . 2.3.2 Identification of Urine and Serum Ribonucleoside Peaks . . 2.3.3 Quantitation of Nucleosides . 2.3.4 Internal Standard 3-Methyluridine . 2.3.5 Prechromatography Sample Preparation Procedure 2.3.6 Recovery of the Method 2.3.7 Precision of the Method . 2.3.8 Stability of Nucleosides . 2.3.9 Ribonucleotides and Oligoribonucleotides in Normal and Cancer Serum .

.

.

.

. . . .

. . . .

C42 c43 c43 c44 c45 c45 c47 c47

.

C48

. c49

. .

C50 C51 . C51 c54 c57 C60 C65 C7 1 c73 c77

.

C78

C42 2.3.10

2.4 2.5 2.1

Analysis of Creatinine in Urine and Serum by a Modified HPLC Method . . C80 2.3.1 1 A Comparison of Nucleoside Levels in Random and Total 24 Hour Human Urine Collections . . C83 2.3.12 Serum and Urine Ribonucleoside Levels in Normal Populations . . C85 2.3.13 Clearance Values of Nucleosides . C92 2.3.14 Adenosine Deaminase Activity in Serum and Urine . C93 2.3.15 Serum Nucleosides in Canines with Osteosarcoma. . c93 2.3.1 6 Serum Nucleosides in Leukemia and Lymphoma Patients . C98 2.3.17 Polynuclear Aromatic Hydrocarbon (PAH) CarcinogenRibonucleoside Adducts in the Urine of Fish and Rat , C99 C105 Summary . References. C107

INTRODUCTION In the late 1970's we developed analytical protocols for the quantitative measurement of nine ribonucleosides in urine (refs. 1-3). The high resolution, speed, and sensitivity of reversed-phase high performance liquid chromatography combined with the selectivity of phenylboronate gel affinity chromatography (refs. 1-4) of this method have been of value to many researchers who have adopted or modified this methodology and used i t in their laboratories and clinical studies on urinary nucleosides as A comprehensive review of potential biological markers (refs. 5-14). research activities in the field of the modified nucleosides as biomarkers is presented in the Introduction on "Nucleoside Markers for Cancer" and the thirteen chapters presented in Part C of this treatise. The latest experimental approaches and technologies of measurement are described in these chapters. Promising results from these studies on the urinary nucleosides (refs. 1550) stimulated interest on investigations of a larger number of ribonucleosides and especially for ribonucleosides in serum (refs. 5 1-64). The concentrations of the modified ribonucleosides in serum are ca. 100-

c43

fold lower (ppb levels) than in urine and the high protein concentration in serum makes serum a most difficult matrix to analyze. Previous serum nucleoside methods were limited either by sensitivity or selectivity which allowed measurement of only a very few of the serum nucleosides in high concentrations ( Y , I, and U). In 1982 we developed a high resolution HPLC-UV method for quantitative measurement of low picomole amounts of nucleosides in tRNAs (ref. 62). However, this technology for the quantitation of total serum nucleosides was not achieved until after the new sample pretreatment protocols were developed (ref. 64). Using this new sample pretreatment and chromatography methodology, we demonstrated that twenty known nucleosides in urine or serum (hU, Y, ncmSU, mlA, I, X, PCNR, mlI, m*G, ac4C m2G, m2m2G, t6A, m6A, mt6A, ms2t6A, C, U, G and A) and ten unidentified nucleosides can be quantified in a single 35 minute chromatographic run. Further, the precision, speed, sensitivity and ruggedness of the methods are well suited for clinical research applications. In this chapter we have described for the investigators fully validated and reliable methodologies for the analysis of nucleosides including pre-chromatography sample preparation techniques of biological samples. Each chromatographic protocol is designed for high resolution, selectivity, and speed of analysis. Comprehensive information is also presented on investigations on the metabolism of ribonucleic acids and their relationships as biologic markers of cancer.

2.2 EXPERIMENTAL 2.2.1 Chemicals The methanol and acetonitrile solvents used were RPLC grade either of B & J Brand from American Scientific Products (McGaw Park, IL) or OmniSolv from EM Chemicals (Cherry Hill, NJ). RPLC water was obtained through a three-step purification process. The first step was reverse osmosis using an RO-Pure apparatus (Model D0640, Barnstead Company, Boston, MA). The second step of purification was accomplished with a Nanopure four cartridge system (Model D1794, Barnstead) composed of one charcoal cartridge for adsorption of organics, two mixed bed ion-exchange cartridges for removal of anions and cations, and one filtration cartridge capable of removing particulates larger than

c44

0.22 pm. In the third step, the nanopure water was distilled in an all-glass still with teflon tubing connections (Model AG-11, Corning Glass Works, Corning NY). Ammonium phosphate, zinc sulfate, and sodium acetate were purchased from J.T. Baker Chemical Co., (Phillipsburg, NJ). Ammonium hydroxide and phosphoric acid were from Mallinckrodt Co., (St. Louis, MO). The modified ribonucleoside reference standard compounds used were from several sources including Sigma Chemical Co. (St. Louis, MO), Mann Research Labs (New York, NY) and Vega Biochemicals (Tucson, AZ). Nuclease P1 was purchased from Boehringer Mannheim Biochemicals (Indianapolis, IN). Bacterial alkaline phosphatase (BAP) from E. coli Type I11 was purchased from Sigma Chemical Co. , product No. P-4252, ( St. Louis, MO). The bacterial alkaline phosphatase must be pretested for possible contamination of adenosine deaminase. The above enzymes are the only sources that we have tested which are free of adenosine deaminase activity under our hydrolysis protocol. An enzyme blank must also be run for each newly purchased enzyme lot to observe possible RNA and DNA contamination. All of the transfer ribonucleic acids (tRNAs) as listed were purchased from Boehringer Mannheim Biochemicals (Indianapolis, IN), unfractionated tRNAs from brewer's yeast (Cat. No. 109 517), unfractionated tRNAs from calf liver (Cat No. 647 576), and unfractionated tRNAs from E. coli MRE 600 RNase negative (Cat. No. 109 541). Amino acid specific tRNAs from E. coli MRE 600, N-formylmethinoine-specific (Cat. No. 109 584), glutamic acid-specific I1 (Cat. No. 109 609), phenylalaninespecific (Cat. No. 109 673), tyrosine-specific (Cat. No.109 703) and valinespecific I (Cat. No. 109 720), and tRNA phenylalanine-specific from brewer's yeast (Cat. No. 109 657). 2.2.2 Ultrafiltration Procedure To one (1.0) ml of serum, 0.50 nanomoles of internal standard 3methyluridine (m3U) in 100 p1 of water was added and mixed well. The sample was then filtered through a micropartition system (MPS-1, with a 25,000-30,000 molecular weight cut off, type YMT membrane; Amicon Co., Danvers, MA). A 30' fixed angle rotor centrifuge (IEC HN SII Centrifuge equipped with model 1 1/80 rotor International Equipment Co., Needham

c45

MA) at 1500 g was used for centrifugation. After 3 to 4 hours 600 to 700 pl of filtrate should be obtained. If cost is not of major concern, divide the sample into two MPS-I filtration units and centrifuge at 1500 g for 45 to 60 min. In this way 800-900 p1 of filtrate can be recovered. 2.2.3 Phenylboronate Ge1 Column Procedure Bio-Rad Affigel 601, (Cat. No. 153-61-01, from Bio-Rad Laboratories, Richmond, CA 94804), an immobilized boronic acid based gel, is used. The cleanup procedures for urine and serum are described in detail as follows: Cleaning and Conditioning New Affigel Gel 601: The boronate gel (1 g) is placed in water (ca. 25 ml), allowing a contact time of five minutes to permit the gel to swell. The gel is then alternately washed with methanol and water for at least ten cycles. Following this procedure, the gel is washed two times with 25 ml of 0.1 M NaC1, followed with 3 x 25 ml of 0.1 N HCOOH, 3 x 25 ml of 0.25 M CH3COONH4, 3 x 25 ml of 50% CH30H in water, 3 x 25 ml of 0.1 N HCOOH in 50% CH30H in water, 2 x 25 ml 0.1 M NaCI, and then resuspend the gel in 50 ml of 0.1 M NaCI. The gel is now prepared for placement into the cleanup columns. Column Dimensions: The column length is ten cm and 0.3 cm i.d.. The borosilicate glass column is fitted with a 5 ml reservoir and fine tip plugged with glass wool. Packing the Gel Column: a. Pack the gel column to a height of 3 cm with the washed and conditioned gel in 0.1 M NaC1. b. Just prior to sample cleanup, equilibrate the gel column by passing through the column 15 ml of 0.25 M CH3COONH4, pH 8.8. Be sure that all air bubbles have been removed. Air pockets can be removed by gentle stirring of the gel bed with a glass rod. 2.2.4 Procedure for Isolation of Urinarv Nucleosides 1. Aliquot exactly 250 p1 of urine into a 1.5 ml polypropylene microcentrifuge tube. 2. Add 100 pl of 2.5 M CH3COONH4, pH 9.0.

C46

3. Exactly add internal standard (m3U), 100 p1 of a 100.0 nmol/ml solution (10.0 nmo1/250 p1 urine). 4. Mix well for a few seconds using a vortex mixer. 5. Centrifuge at 12,000 rpm for three minutes. 6. Transfer the urine sample onto the boronate gel column with a Pasteur pipet and allow free flow, taking care not to disturb the precipitate in the sample tube. 7. Add as rinse 500 p1 of 0.25 M CH3COONH4, pH 8.8, to each sample tube. 8. Mix well for a few seconds using a vortex mixer. 9. Centrifuge at 12,000 rpm for three minutes. 10. Transfer the rinse onto the gel column, taking care not to disturb the precipitate in the sample tube. 11. Wash the gel column with 3 ml of 0.25 M CH3COONH4, pH 8.8, allowing free flow. 12. Wash the gel column with 300 p l of 50% methanol/water (v/v), allowing free flow. 13. Elute the nucleosides with 5.0 ml 0.02 N HCOOH in 50% methanol in water (v/v). Collect in a 10 ml polyethylene tube. 14. Remove the methanol from the eluate using a Speed Vac Concentrator (Savant Instruments Inc., Hicksville, N.Y.) with a water aspirator as vacuum source. When the sample volume is reduced to less than half, essentially all the methanol has been removed. 15. Freeze the sample with the tube in a slanted position. 16. Lyophilize the sample to dryness using the Speed Vac Concentrator with a mechanical vacuum pump and -50°C cold trap. 17. Redissolve the sample in 500 p l of distilled water; vortex for a few seconds. 18. Inject 100 p1 onto HPLC. 19. Prior to column re-use, wash the gel column with 10 ml of 0.02 N formic acid in 50% methanol/water (v/v). This removes strongly adsorbed substances. 20. Wash column with 10 ml 50% methanol/water (v/v). 21. Wash column with 10 ml 0.02 N HCOOH in water.

c47

22. Store the gel columns in 0.02 N HCOOH each day in water. The gel column can be reused up to fifteen times. If the gel column is not used for several days, store in 0.1 M NaC1. Recovery tests must be conducted to establish quantitation of the cleanup procedure. 2 3 . Prior to use, re-equilibrate the gel column with 15 ml of pH 8.8, 0.25 M CH3COONH4. 2.2.5 Procedure for Isolation of Serum Nucleosides 1. Aliquot exactly 1.0 ml of serum into a 1.5 ml polypropylene centrifuge tube. 2. Add internal standard (m3U), exactly 100 p l of a 5.00 nmol/ml solution (0.500 nmol/l .O ml of serum). 3. Mix well for several seconds using a vortex mixer. 4. See "Ultrafiltration Procedure Section". This step will take approximately four hours and give ca. 600-700 p1 of filtrate from 1.0 ml of serum. 5. Add 250 p1 of 2.5 M CH3COONH4, pH 9.0 to the ultrafiltrate and mix well. 6. Transfer the sample onto a washed, conditioned and preequilibrated boronate gel column. Steps 7 to 16 are the same as for the urinary nucleoside isolation procedure. 17. Redissolve the sample in 200 p1 of distilled water; vortex for a few seconds. 1 8 . Inject 180 p1 of sample onto HPLC. 2.2.6 Preparation of Internal Standard Solutions A 0.500 pmol/ml stock solution of the internal standard (m3U) is made by weighing an accurately known weight (mgs) of m3U (P-L Biochemicals, Milwaukee, WI 53205) and dissolving the m3U (molecular weight = 258.23) in a calculated amount of water (grams) to make the final concentration 0.500 pmol/ml. For example, the weight of m3U is 1.350 mg (5.228 pmol), then 10.45 g of water (by balance) is used to make a concentration of 0.500 pmol/ml. Appropriate dilutions were made from the stock solution to obtain the desired concentrations for the working solutions, 10.0 nmol/nil for

C48

urine and 0.500 nmol/ml for serum analysis. Each new dilution of working solution should be checked by HPLC for accuracy (compare with the established value of arednmol of m3U). All solutions are maintained frozen at -20°C in small aliquots and thawed out only just prior to use. 2.2.7 Determination of Adenosine Deaminase Activity in Serum and Urine In Serum: 1.0 ml of pooled normal human serum was pipetted into each of six 1.5 ml polypropylene microcentrifuge tubes and the substrate and enzyme were added as follows to study adenosine deaminase activity on A and m l A . 1. a control sample not spiked , no adenosine deaminase added. 2. sample not spiked, add 10 p1 adenosine deaminase (2.5 units). 3. sample spiked with 0.5 nmol adenosine, no adenosine deaminase added. 4. spiked with 0.5 nmol adenosine, add 2.5 units adenosine deaminase. 5. spiked with 0.5 nmol mlA, no adenosine deaminase added. 6. spiked with 0.5 nmol m1A , add 2.5 units adenosine d e am i n a s e . Vortex and centifuge. Incubate over night at 37 "C. Analyze on HPLC.

In Urine: Pipet 250 p1 of urine into each of six 1.5 ml polypropylene microcentrifuge tubes. Add 100 p1 0.1 M KH2PO4 to adjust acidic urine to pH 7.5. Label the centrifuge tubes 1-6 and add substrate and enzyme as follows: 1 A control sample, not spiked with adenosine and no adenosine deaminase added. 2. Sample not spiked, add 1.25 units of adenosine deaminase. 3. Sample spiked with 3 nmol adenosine, no adenosine deaminase added. 4. Sample spiked with 3 nmol adenosine, add 1.25 units adenosine deaminase.

c49

5. Sample spiked with 3 nmol mlA, no adenosine deaminase added. 6. Sample spiked with 3 nmol m l A , add 1.25 units adenosine deaminase. Vortex and centrifuge. Incubate over night at 37°C. Analyze on HF'LC.

..

2.2.8 HPLC Instruments and Conditions A fully automated LC instrumentation system consisting of a n H P 1090M (Hewlett Packard, Avondale, PA). The HP-1090M system was made up of a DR5 ternary solvent delivery system, variable-volume autoinjector, autosampler, diode-array detector, and heated column compartment. The liquid chromatography workstation is based on an HP model 310 computer supported by Rev. 4.05 operation software; HP-HIL 512 x 400 color monitor with bit-mapped display; and HP-9133H 20 mb Winchester disc drive with 3.5" 710 kb micro floppy disk. A Think-Jet printer and H P 7475A plotter were used for hard copy data presentation. The cooling coil of the heated column compartment was circulated with refrigerated ethylene glycol based antifreeze by a Haake model FJ circulating bath (Saddle Brook, NJ). The cooling bath was positioned inside a small refrigerator and the antifreeze was also circulated through a 10 ft, 114 in. coiled copper tubing which was positioned inside the freezer compartment for additional cooling. Detailed chromatographic conditions are as follows: Column: Supelcosil LC-18s 15 cm x 4.6 mm, with 2.0 x 4.6 mm LC18s-Supelcosil Guard column, (Supelco, Inc., Bellefonte, PA). Flow-rate: 1.O ml/min. Column Temperature: 26 It: 0.5 "C. Elution Buffers: A: 2.50% methanol in 0.010 M NH4H2P04; pH 5.3. B: 20.0% methanol in 0.010 M N a H 2 P 0 4 ; pH 5.3. C: 50.0% acetonitrile in water. The elution gradient is presented in Table 2.1.

C50

Table 2.1 HPLC Elution Gradient for Separation of Ribonucleosides in Urine and Serum Step No.

1 2

Buffer A

Step Time (min) 7.2 4.8 3.0 4.2 7.8 8.0 0.5 4.5

Composition,% B C 0.0

100.0 90.0 75.0 40.0 0.0 0.0 0.0 0.0 ~

~

~

_

_

_

10.0 25.0

60.0 100.0 60.0 0.0 0.0 _

_

0.0 0.0 0.0 0.0 0.0 40.0

100.0 100.0

Gradient Type

Isocratic Linear Linear Linear Linear Linear Linear Isocratic

_

Equilibrate the column with 100.0% Buffer A for 15 min between runs.

2.2.9 Creatinine Analysis by HPLC-UV

Sample Preparation: 1. Aliquot exactly 100 p1 of serum and/or urine into a 1.5 ml polypropylene microcentrifuge tube. 2. Add 250 p1 of acetonitrile to the sample. 3 . Vortex for 10 seconds, hold at 0 "C for one hour. 4. Centrifuge at 12,000 x g for 2 minutes. For Serum Creatinine Analysis: 5 . Aliquot exactly 100 p1 of the supernatant and transfer to a glass WISP insert. 6. Evaporate to dryness using a Speed Vac concentrator and water aspirator. 7 . Redissolve the sample in 100 p1 of HPLC water and mix well. 8. Inject 50 pl onto HPLC column. For Urine Creatinine Analysis: 5 . Aliquot exactly 20 pl of the supernatant and transfer to a glass WISP insert. 6. Evaporate to dryness using a Speed Vac concentrator and water aspirator.

C51

7. Redissolve the sample in 200 pl of HPLC water and mix well. 8. Inject 50 p1 onto HPLC column. HPLC Conditions for Creatinine Analysis: Column: Whatman SCX-10, 25 x 0.46 cm with 2 x 4.6 cm Whatman SCX Guard column. Elution Buffer: 0.10 M NH4H2P04, pH 4.8. Flow-rate: 2.0 ml per min, isocratic. Column Temperature: 40 "C. Detection: 254 nm. Using the above conditions on a new column, the retention time for creatinine is between 4.5 and 5.0 min. As the column is used (ca. 100 sample runs), the retention time of the creatinine decreases. In general, the retention time can be restored to 4 to 4.5 min after the column is When the retention time of washed with 70% methanol in water. creatinine decreases to 3 to 3.5 min, the eluant salt concentration is decreased to 0.05 M. A standard creatinine solution is analyzed with each ten samples to re-calibrate the response factor. 2.3 RESULTS AND DISCUSSION 2.3.1 ChromatograDhv The chromatographic conditions that we used for the reversed-phase liquid chromatographic separation of ribonucleosides in urine and serum was essentially the same as the high speed chromatography method for nucleosides in RNA hydrolysates (ref. 66), except the run time was reduced to 35 minutes (steps 7 and 8 in Table 2.1 were used to wash the strongly retained compounds from the column). We did not customize a separation gradient specific for the nucleosides in physiological samples as we wanted to apply the qualitative and quantitative parameters established for RNA hydrolysates directly to these physiological sample analyses. Chromatograms are presented in Figures 2.1-a and 2.1-b from a 254 nm signal of urine and serum nucleosides in a leukemia patient. In these chromatograms twenty-one known nucleosides were identified. The array

C52

3

Human Serum

4

E

Fig. 2.1 Chromatogram of modified nucleosides in human serum and urine of a leukemia patient. See Experimental for chromatographic details.

of nucleosides in physiological fluids is considerably different from the patterns for tRNAs (ref. 6 6 ) . More than 60 known modified nucleosides were observed in tissue tRNAs but only 15 of these were found in serum and urine. The fate of those missing ribonucleosides is yet unknown. Ten unidentified peaks in serum and urine were classified as unknown ribonucleosides because not only do they possess a ribose moiety but they also have UV spectra similar to the nucleosides. They are either metabolic products of modified nucleosides, minor components of modified nucleosides in tRNAs, or catabolites of other ribose related biomolecules. Since the nucleoside pattern and composition in physiological samples are different from the tRNA hydrolysates a customized HPLC separation protocol for urine and serum nucleosides should be developed. There is a need for an HPLC protocol to increase the separation in two regions of the chromatogram. One region is at the beginning of the run to the guanosine peak. Decreasing the pH of buffer A from 5.3 to 4.5 - 4.0

c53

10

3

Q

E s c4c

0

0

20

10

40

30

50

15

3

< E

0

0

20

10

30

Time (rnin)

Fig. 2.2-a Chromatography of nucleosides in serum (0.20 ml) on a 25 cm x 2.1 mm minibore column. Fig. 2.2-b Chromatography of nucleosides in All other serum (1.0 ml) on 15 cm x 4.6 mm analytical column. chromatographic conditions were the same.

c54

and decreasing the concentration of the methanol to 1% will increase the separation of the y~ peak from the background peaks and improve the separation of m7G, I, T, X, and G peaks. The second region is between the l-methylinosine ( m l I) peak and the 2-methylguanosine (m2G) peak, especially the separation between m2G and 4-acetylcytidine (ac4C) peaks. In serum, the ratio of ac4C and m2G is high and often results in separation insufficient for quantitation. Decreasing the buffer gradient ramp slope between 19 to 24 minutes will improve the separation of the m l I and m l G peaks from their adjacent peaks as well as the separation of ac4C from m2G. Also, the step-time of the gradient step 1 should be correspondingly reduced to maintain the elution efficiency of the late eluting peaks. A major improvement would be the use of minibore (2.1 mm id) or microbore (1.0 mm id) columns. Minibore and microbore columns will greatly improve the sensitivity of the method. Preliminary comparisons of a 15 cm x 4.6 mm analytical column and a 25 cm x 2.1 mm minibore column for serum analysis are shown in Figure 2.2. One ml of serum was injected into the analytical column (Fig. 2-b) and only 0.2 ml of the same serum sample was injected into the minibore column (Fig. 2-a). A fourfold increase in mass sensitivity (peak area or height per unit weight of nucleoside) was observed for the minibore column as compared to the regular analytical column. Still greater increases in sensitivity can be achieved with a 15 cm length minibore column instead of the 25 cm minibore column shown here. With the minibore column, the serum volume needed for the analysis could be reduced to 250 p l or less. This reduction in serum sample volume allows a significant decrease in sample preparation time and simplification of automation of the total analysis process. High quality minibore columns are available from many suppliers and many HPLC instruments are available for the minibore column application. Further, more than a 20-fold increase in sensitivity could be expected from a microbore column as routine gradient applications of microbore HPLC are available. 2.3.2 Identification of Urinarv and Serum Ribonucleoside Peaks Nucleoside peaks from urine and serum were identified by comparing their chromatographic retentions and UV spectra with known reference nucleosides. In addition, some major urinary nucleoside HPLC

c55

Table 2.2 Day-to-Day HPLC Retention Time Reproducibility Chromatography Retention Nucleoside

hU Y

C ncm5~ U m 3 ~ mlA m 5 ~ I G m 3 ~ mlI mcmR~ mlG ac4~ m2G A m2m2G mcm5s2~ t6A m6A mt6A ms2t6A

Mean

3.02 3.21 4.19 4.43 6.03 6.56 7.81 8.61 12.80 14.19 17.21 19.23 19.49 19.84 20.62 20.92 21.68 24.06 25.68 26.16 28.93 30.67 32.42

Time,

of Nucleoside

min

SD 0.035 0.043 0.061 0.070 0.093 0.488 0.273 0.164 0.213 0.203 0.200 0.0467 0.3 12 0.144 (N = 1) 0.125 0.182 0.181 0.243 0.131 0.431 0.0573 0.262

RSD,% 1.16 1.36 1.45 1.33 1.54 0.732 3.50 1.91 1.67 1.43 1.16 0.243 1.60 0.723 0.604 0.838 0.754 1.10 0.500 1.49 0.187 0.807

High Speed Nucleoside Chromatography, a Supelcosil LC-18s 150 x 4.6 mm column was used with 2.0 cm LC-18s guard column. Data were collected from 29 HPLC runs over 3 days.

peaks, such as v, PCNR, mlI, mlG, and m2m2G were collected structure confirmed by mass spectrometry (ref. 67). In routine a combination of retention time, peak shape and the A254tA280 ratio, in general, are sufficient for positive identification (Figures

and their operation, absorption 2.2, 2.3).

C 56

Table 2.3 RPLC-UV Response Factors of Ribonucleosides in Serum pet. Nucleoside

hU Y

C ncms~ U m 3 ~ mlA m 5 ~ m7G I X

G

PCNR m 3 ~

Q

mlI rncrns~ mlG ac4~ m2G A m2m2G mcm5s2~ t6 A Br8G m2A m6A mt6A ms2t6A

Time(min)

AJuLwuz4.70 5.03 6.55 7.21 9.28 10.21 11.87 13.20 19.60 19.90 20.84 21.77 22.80 27.90 28.31 30.73. 30.98 31.64 32.62 33.14 34.70 38.33 41.04 41.88 42.90 45.20 46.65 48.80 52.70

3.02 3.21 4.19 4.43 6.03 6.56 7.81 8.61 12.80 12.80 13.52 14.19 14.78 17.10 17.95 19.23 19.49 19.84 20.72 20.92 21.68 24.06 25.68 26.16 26.80 28.86 28.93 30.67 32.42

RMR-Rr8G

RMR-IIL~Y

254/280

Urine and

MR 254/280/21(

0.37810.505 0.40910.275 0.38 910.561

0.59411.73 0.64210.949 0.61211.94

0.52510.280 0.30510.627 0.67110.244 0.29310.627 0.63010.620 0.68010.250 0.40910.450 0.83710.667 0.879/0.432 0.63610.283 0.52510.531 0.56010.225 0.26610.272 0.83710.695 0.31110.217 0.91810.680 0.781/0.184 0.949/0.811 0.219/1.11 0.41310.629 1.0011.00 0.73310.184 0.61610.713

29011 14 0.82510.977 0.47912.16 1691256 1.0510.842 371199.6 0.46112.16 1621256 0.99012.14 3481253 1.0710.863 3761102 2261184 0.64311.55 1.3212.30 4631272 1.5111.72 4861177 1.0011.00 35211191364 0.82511.83 2901217 0.88010.776 3 10191.9 0,41810.939 14711 1 1 1.3212.40 4621284 0.48910.749 172188.6 1.4412.35 5071278 1.2310.635 431175.1 1.4912.80 5241331 0.34413.84 1211453 0.64912.17 2281257 1.5713.43 5531408 1.1W0.635 406175.1 0.96812.46 3401291

2091206 2261112 2151229

to be re-established to be re-established

HRNC: High Resolution Nucleoside Chromatography; Supelcosil LC-18s 250 x 4.6 mm column with 2.0 cm LC-18s guard column. HSNC: High Speed Nucleoside Chromatography; Supelcosil LC-18s 150 x 4.6 mm column with 2.0 cm LC-18s guard column. R M R - B r 8 G : Relative Molar Response; 8-Bromoguanosine as internal standard. R M R - m 3 U : Relative Molar Response; 3-Methyluridine as internal standard. M R : Molar Response; in units of arealnmol, area is the counts that were obtained from HP-1090M liquid chromatography work station. 254/280/210: 2 5 4 n m l 2 8 0 n m 1 2 1 0 n m . hU: peak area from 210 nm. P C N R : l-Ribosylpyridin-4-one-3-carboxamide

c57

The retention times for 23 nucleosides, which were established in our laboratory, using the experimental protocols presented by Gehrke and Kuo (ref. 62) are given i n Table 2.2. Figure 2.3 shows a 254 nm chromatogram of nucleosides from a pooled urine sample. Thirty-six known and unknown peaks were observed and their corresponding HPLCUV spectra are presented i n Figures 2.4-a and 2.4-b. These spectra were routinely used for additional peak identification and confirmation of the purity of the peaks.

2.3.3 Ouantitntion of Nucleosides The urine and serum nucleosides were quantified by using the dual wavelength (254 and 280 n m ) internal standard method, except for hU for which only 210 nm was used. The relative molar response factors (RMR) using Br*G or m3U as internal standard, and molar response factor (MR) of Molar 27 nucleosides that we obtained are presented in Table 2.3. response factors can only be applied when both an HPLC data system and chromatographic conditions identical to ours are used. The relative molar

I"""""""""'""'""'~"""""'""""~"''

0

10

20

30

40

50

Time (rnin)

Fig. 2.3 HPLC of ribonucleosides in human urine with HPLC-UV spectral identification. Refer to Experimental for chromatographic details.

C58

Fig. 2.4-a HPLC-UV spectra (200 to 350 nm) of ribonucleosides in urine.

c59

I

I

*m m ,

.........

36.1

42.6

I..

.

.

I

Fig. 2.4-b HPLC-UV spectra (200 to 350 nm) of ribonucleosides in urine.

C60

response factors can be used when the same HPLC chromatographic conditions are employed. All of the factors were confirmed using pure known sequenced tRNAs. 2.3.4 Jnternal Standard 3-Methvluridine Due to the multiple sample preparation steps, the internal standard method is essential for the accuracy of quantitation. The internal standard must be added to the sample prior to any sample manipulation. The most desirable internal standard; i) should if possible, be a ribonucleoside or a compound that has the same chemical and physical properties as the ribonucleosides, ii) must not be present in the sample, iii) stable throughout the entire analytical-chromatographic procedure, iv) elute at a position without interference peaks so that it can be integrated accurately. 3-Methyluridine (m3U) clearly met all of the above criteria. It was reported not to occur in tRNAs and only at a very low level in rRNAs. m3U is present at only trace levels in urine and serum from both normal and cancer patients (Tables 2.4 and 2.5). The endogenous m3U in serum and urine is less than 1.5% and 0.91%, respectively, of the m3U added as internal standard to the samples. Thus the presence of endogenous m3U can result in no more than a 1.5% and 0.91% positive bias in the analysis of serum and urine. An unknown nucleoside eluted just before m3U and its concentration varies in different samples. Thus, the HPLC separation must be sufficient to separate this unknown from m3U. Also, 2'-methyluridine (Urn)is a major modified nucleoside in tRNAs, and it is usually present in urine and serum at high concentration. The present HPLC protocol does not separate m3U from Urn. Thus, caution must be taken in using the correct phenylboronate gel isolation steps to ensure a complete removal of Urn. As Urn is not a cis diol it passes through the boronate gel column. The performance of m3U as an internal standard was demonstrated by the quantitative recovery of m3U from the phenylboronate gel columns at physiological concentration (Table 2.6) and quantitative recoveries of spiked yr, mlI, m2m2G, and t6A in serum were obtained based on m3U as the internal standard (Table 2.7). N2lN2-Dimethylguanine (refs. 1-2), 8-bromoguanosine (ref. 69), 6methylisocytosine (2-amino-4-hydroxy-6-methylpyrimidine), Tubercidin (7-deazaadenosine) (refs. 11-13), 5-hydroxymethyluridine (ref. 68),

C61

deoxyadenosine (ref. 2 6 ) , deoxyguanosine (ref. 52) and other nucleoside analogs have been used as internal standards by various investigators. We selected m3U as internal standard to measure ribonucleosides in serum and urine because m3U i s a typical modified pyrimidine and is stable chemically and biochemically. Thus, it is one of the best possible internal standards to ensure the accuracy for the internal standard method.

Table 2.4 Endogenous 3-Methyluridine

(m3U) in Human Serum

Serum

Endogenous nmol/ml

NS-I NS-2 NS-3 cs-I cs-2 cs-3

0.0075 0.006 1 0.0054 0.0074 0.0067 0.0033

Endo./IS,

%

1.5 1.2 1.1 1.4 1.3 0.7

NS = Normal Serum CS = Cancer Serum Endo./IS,% = Percent of endogenous m3U to added m3U; 0.5 nmol/ml of m3U was added.

Table 2.5 Endogenous 3-Methyluridine Urine NU-1 NU-2 NU-3

cu-1

cu-2 cu-3

(m3U) in Human Urine

Endogenous n m o I / ni 1 0.132 0.104 0.364 0.028 0.360 0.028

Endo./IS,

%

0.33 0.26 0.91 0.07 0.90 0.07

NU = Normal Urine CU = Cancer Urine Endo./IS, % = Percent of endogenous m3U to added m3U; 40 nmol/ml of m3U were added.

C62

Secondly, m3U has not been found in tRNAs. It is only found in very low amounts in rRNA. We have examined m3U levels in pooled and individual serum and urine from both normal and cancer patients. Very low amounts of m3U were detected in all the samples. In addition, the amount of m3U does not increase in cancer urine and serum as does the other modified nucleosides (See Tables 2.4 and 2.5). And thirdly, m3U elutes at an open area of the chromatogram and can be integrated accurately. It elutes in the middle of the chromatography run and also in the middle of a gradient ramp, thus the variation between runs can be more readily compensated by relative retention times and relative response factors to m3U. With this internal standard the reliability of peak identification and accuracy of quantitation are enhanced. Over the past two years, the m3U internal standard method has been rigorously validated and applied in our laboratory. More than 500 serum samples and 250 urine samples have been analyzed. The results obtained have been very satisfactory. The performance of this internal standard method was demonstrated by the quantitative recovery of m3U from the phenylboronate gel columns at physiological concentration (Table 2.6). This shows the stability of m3U, and m3U gave essentially an identical recovery as for the other modified nucleosides. Further, the precision of recovery from the boronate affinity column was excellent. Using m3U as the internal standard, quantitative recoveries of four nucleosides mlI, m2m2G, and t6A) in serum were obtained (Table 2.7). This proved that m3U can quantitatively compensate for the loss of endogenous ribonucleosides during ultrafiltration. A slightly lower recovery (85%) of m6A is due to incomplete elution from the gel column. To improve the recovery of m6A, the elution volume of 0.02 N HCOOH in 50% methanol should be increased from 4.5 ml to 5.0 ml. Additional evidence in support of the use of m3U as the internal standard was obtained by comparing physiological nucleoside values that we obtained with corresponding values obtained by other high quality methods. The recent report by John T. Bernert, Jr. et al. (ref. 68) on an in serum and HPLC-UV internal standard method for quantitation of urine provides this comparison. Bernert et al. used 5-hydroxmethyluridine (omsU) as internal standard. The average values that they obtained from 19 normal urine and' serum samples were 2.77 nmol/ml for

(w,

w

w

C63

Table 2.6 Recovery of Standard Reference Ribonucleosides Phenylboronate Gel* Column

Added to

Recovery,% N

Conc.(pmol/ml)

Mean

SD

RSD,%

w

19 20

1150 1150

91.7 91.5

4.46 4.66

4.86 5.09

mlI

19 20

270 270

99.8 104

2.38 2.20

2.38 2.19

m 3 ~

19 20

474 474

95.3 93.9

2.43 2.69

2.54 2.86

m 2m2G

19 20

480 480

2.56 4.30

2.51 4.17

t6A

19 20

967 967

2.46 4.35

2.62 4.42

Nucleoside

102 103 94.0 98.7

1.0 ml of nucleoside standard solution was added to each of the columns. Phenylboronate Gel: Affigel 601 from Bio-Rad Laboratories. Two groups of gel columns were tested approximately 6 months apart. Each group consisted of 19 and 20 columns, respectively.

Table 2.7 Recovery of Added Nucleosides from Human Serum Recovery

%

Run No.

Y

m11

m 2m 2G

t6A

m6A

1

90.5 97.2 97.7 91.5 90.0 94.6 89.8 96.2 93.4 94.0 90.0 96.1

120 99.5 108 96.3 102 99.3 103 101 109 103 108 98.2

91.0 99.7 97.9 91.7 98.0 93.2 99.7 91.3 93.7 102 95.7 92.0

101 104 101 93.9 100 98.5 92.9 101 100 97.5 96.8 95.4

87.2 82.3 85.2 87.7 90.2 89.0 86.9 79.4 88.8 84.2 83.7 79.3

2 3 4 5 6 7 8 9 10 11 12

C64

Table 2.7 (continued) Mean

SD RSD, %

93.5 2.90 3.10

95.5 3.85 4.03

104 6.46 6.22

98.5 3.31 3.36

85.2 3.65 4.28

Recovery % = [(Cspk - c ) X loo]/ (Cadd) C = determined mean nuclcoside concentration value, in pmol/ml, in pooled normal serum. Cadd = pmol/ml of nuclcoside added to the pooled normal serum. Cspk = determined mcan nucleoside concentration value, in pmol/ml, of nucleoside in spiked pooled normal serum. Pooled normal serum obtained from blood bank. The nucleoside concentration in pmol/ml was dctermined by our nucleoside rncthod:

picomoleslml Y

Mean (N = 5) 1980 SD 67 3.4 RSD, %

m11

m 2 ni 2 G

t6 A

m 6A

73.9 4.0 5.4

24.8 1.1 4.4

39.4 2.1 5.3

<1

---

Table 2.8 Comparison of Human Urine and Serum Pseudouridine Values Bernert's y~ V a l u e s Serum Urine

2.77 nrnol/ml (N=19) 375* mmo1/24 hr (N=19)

Our y~ Values 2.84 nmol/ml (N=94) 331 mmo1/24 hrs. (N=18)

*Calculated from Tablc 1 of rcf. 68. (203.2 x 1.851 = 375 mmo1/24 hrs.)

serum and 375 mmo1/24 hours for urine. These values were in good agreement with the values that we obtained (Table 2.8). These data show that the internal standards n:3U nnd hm5U performed equally well.

C65

2.3.5 Prechromatogaphy Sample Preparation Procedure Two selective isolation protocols, ultrafiltration to remove high molecular weight biopolymers, and phenylboronate gel affinity column chromatography, were used for selective isolation of ribonucleosides from other components of serum and urine. Ultrafiltration is not required for urinary nucleoside analysis. However, the phenylboronate gel isolation step is needed for all physiological fluid samples. Whole blood, plasma, serum, cell culture medium, and other body fluids which contain a high concentration of protein or other high molecular weight biopolymers, require the additional ultrafiltration steps. Enzymatically hydrolyzed tissue RNAs can be directly injected on to the HPLC column without pretreatment. The flow charts of the protocols for cleanup of serum and urine are presented in Figures 2.5 and 2.6 and are detailed stepwise in the Experimental Section. Ultrafiltration of Serum: We have investigated many different physical and chemical deproteinization procedures and found that ultrafiltration gave the lowest and most reproducible background. The direct application of serum onto the phenylboronate gel resulted i n a number of high background peaks. In this case only y~ and Urd can be measured as the serum concentrations of these two nucleosides are more than 100-fold greater than the other nucleosides. Ultrafiltration, with a micropartition system (MPS-1) equipped with YMT membranes produced the lowest and most reproducible background peaks and required less labor than the use of other deproteinization procedures. The disadvantages are the high cost of the filter and a lower absolute recovery of the nucleosides which is due to incomplete filtration. One ml of serum filtered by one MPS-1 filter gave only 0.6 to 0.7 ml of filtrate after centrifugation for 4 hours at 1000 to 2000 x g. Using 0.5 ml of serum per filter usually resulted in the recovery of more than 0.4 ml of filtrate in 45 minutes. This greater recovery of sample in a shorter period of time, however, increased the analytical cost per sample as two filters were needed to obtain the filtrate from one ml of serum. Improved recovery of filtration by washing the protein precipitate resulted i n a considerably lengthened sample preparation time. Since the internal

C66

Fig. 2.5 ISOLATION of RIBONUCLEOSIDES IN HUMAN URINE 250 pl urine 100 pl 2 M NaOAc 1.0 nmole 3-Me-Urdl100 p1 water

Load on washed and conditioned Boronate gel column (Affigel 601) Gel bed: 3 cm x 0.3 cm I

I Wash with: 1 ) 3.0 ml 0.25 M NH40Ac pH 8.8 2 ) 300 ~1 50 % methanol in water

1 I

I

Elute nucleosides with 4.5 ml 0.02 N HCOOH in 50 % methanol in water

I

I Evaporate methanol Lyophilize off water redissolve in 500 pl water

I Inject 100 pl onto HPLC

C67

Fig. 2.6 ISOLATION of RIBONUCLEOSIDES IN HUMAN SERUM Add to 1.0 ml Serum 0.50 nmoles 3-Me-Urd in 100 p1 water

Filter through Amicon Centricon-10 use micro concentrator with type YMT, 25K-30K cutoff filter.

Collect filtrate, add 250 p1 2.0 M NH40Ac pH 9 solution

Load on washed and conditioned Boronate gel column (Affigel 601) Gel bed: 3 cm x 0.3 cm)

Wash with:

1 ) 3.0 ml 0.25 M NH40Ac 2 ) 300 p.1 50 % methanol in water

C68

standard (m3U) was added before filtration and as there was no selective loss of m3U during the sample preparation, accuracy of the measurement was still maintained. We chose the simplest and most cost-effective procedure; namely, using one ml of serum per filter without additional washing. The recoveries of the five selected nucleosides (w,mlI, m2m2G, t6A, and m6A) from 12 independent determinations, are presented in Table 2.7. These nucleosides were added to pooled normal human serum at 2 to 3-fold their physiological concentrations. The concentrations of the endogenous nucleosides in the pooled serum were determined and are presented as a footnote to the table. Quantitative recovery of all the added nucleosides was obtained, The average recovery of w, mlI, mZmZG, and t6A from 12 independent determinations ranged from 94% to 104%, with a RSD of 3 to 6%. An average recovery of 85 k 4% was obtained for m6A. Phenylboronate Gel Affinity Chromatography Isolation of Ribonucleosides in Urine and Serum: The phenylboronate nucleoside isolation procedure that Uziel and we developed earlier for urine (refs. 14) cannot be directly applied to serum because the resulting background prevents quantifying the low levels (picomol/ml) of serum nucleosides. We have made several improvements in the new protocol to reduce this background. We employed alternate water and methanol washing steps which successfully removed the impurities from the gel. Repeated swelling of the gel in water followed by shrinking of the gel with methanol forced out most of the UV-absorbing materials which were trapped inside the sponge-like gel. The reagent background was reduced by decreasing the size of the gel bed to 210 p1; this decreased the volumes of reagents used for equilibration, washing, and elution. The gel was washed with 50% methanol in water to remove molecules which were retained by the gel matrix through hydrophobic interaction, and 0.02 N HCOOH in 50% methanol was used for the elution of nucleosides. This minimized the cleavage of the amide linkage between the phenylboronate group and the gel matrix which reduced the chromatographic background. The high selectivity of our prechromatography sample preparation procedure is illustrated in Figures 2.7 and 2.8, in which very few nonribonucleoside molecules were observed. A comparison was made of the 254 nm chromatogram and the 210 nm chromatogram for both urine (Figure 2.7) and serum (Figure 2.8). A peak with a very high A2io/A254

C69

I1

SCC Patient 1.0 ml Sample

2 5 4 nrn. 2 0 mAb

~

~

0

"

,

,

"

"

10

"

"

~

"

"

~

"

"

~

20

~

~

30

~

'

~

"

"

~

40

~

~

"

I

"

~

"

"

50

Tirndmin)

Fig. 2.7 Modified nucleosides and UV-absorbing components in human urine.

1"""""""""'""""""""'"'"'"""''''"'

0

10

20

30

40

50

Timdmin)

Fig. 2.8 Modified nucleosides and UV-absorbing components in human serum.

~

C70

ratio most likely is not a ribonucleoside peak. In the urine sample, we observed one large 210 nm absorption peak at 8.8 min and two small absorption peaks at 27 to 28 minutes (between the t6A and m6A peaks). In the serum sample, we observed some large 210 nm absorption peaks before w, peaks at 9.0, 10.1, 10.9, 23.2, and between the ms2t6A and N39.9 peaks. None of these non-nucleoside peaks interfered with the measurement of the nucleoside peaks. In our laboratory, using this cleanup procedure, low chromatographic backgrounds were consistently obtained from hundreds of urine and serum samples. Occasionally we did observe that some of the new gel columns gave a higher background on the first serum sample analyzed. These background peaks appeared between the m l G and m6A peaks with the peak size around 10 to 30 pmol. These background peaks did not interfere with the modified nucleosides of interest such as a d C , m2rnzG and t6A, but care should be taken in assigning the baseline for correct peak integration. For all new gel columns, a pooled serum sample was routinely run on the new columns to verify the column background before use with actual samples. The miniaturization of the gel column to 3 x 0.3 cm greatly increased the speed of analysis. In our laboratory, an analyst can operate a group of 20 gel columns at one time and process 40 samples per day. The exact usable life of the gel column has not been determined. During routine analysis, a pooled control serum and a control urine with known nucleoside values were analyzed with each group of 10 to 20 samples to monitor the performance of the analysis. The columns still performed well after 15 or more urine or serum samples were processed on each column. As a rule, the gel columns were replaced after 15 sample applications or when the columns were stored for a long period of time (several months) for precautionary purposes. The nucleosides exhibit a hydrophobic interaction with the polyacrylic gel matrix, and elute from the Affigel 601 column in the same order as from C-18 reversed-phase chromatography (unpublished data). 8-Bromoguanosine (BrgG) is strongly retained by the Affigel 601 and after 40 column volumes of 0.5 N HCOOH wash, only 60% was recovered. We have not observed the presence of strong hydrophobic nucleosides such as m6m6A, Y, and i6A in most human urine and serum samples. The question of whether they are not in the sample or simply are not eluted from the

C71

Unsplked Serum

1""""'"""""""""""""""""""""""

0

10

20

30

40

60

Fig. 2.9 Recovery of nucleosides added to serum. gel should be investigated. Lowering the pH of the elution reagent (by increasing the concentration of HCOOH from 0.02 N to 0.05 N or 0.1 N) or adding a few % of 1,2-propylenediol to the elution solvent should be tested. Also, experiments should be made to investigate the use of 100 % methanol as the elution solvent. This would eliminate the time consuming lyophilization step which is essential for automation of this method. For an in-depth description of phenylboronate affinity chromatography, and an excellent review of boronate ligands in biochemical separations, see reference (ref. 65).

2.3.6 Recovery of the Method The performance of the phenylboronate gel procedure is presented i n Table 2.6. One nil of an aqueous solution containing five selected nucleosides at concentrations of 1150 to 270 pmol/ml were applied to two groups of 19 and 20 newly prepared gel columns over a time period of six months. Absolute recoveries ranging from 92 to 104 % were obtained. A

C72

Table 2.9 Recovery of Urinary Nucleosides by HPLC Recovery, Nucleoside

w

U

%

Run 1

Run 2

Run 3

Run 4

Ave.

97.1 101.3 105.0 105.8 105.5 97.7 107.8 100.5 99.8 104.5

91.4 101.5 101.5 110.4 105.5 101.9 109.5 93.5 105.8 102.8

94.1 108.9 99.8 112.9 106.2 102.7 108.9 104.0 102.3 107.9

93.7 103.0 115.0 107.2 113.3 107.4 106.9 99.3 106.5 100.2

94.1 103.6 105.3 109.1 107.1 102.5 108.3 99.3 103.6 103.8

~

_

_

_ ~~~~~~~

~

Spike level added as nmoles per 0.25 ml of urine: w 23.5; U 20.5; mlA 1.87; I 1.76; mlI 1.47; m l G 0.905; m2G 1.39; A 0.615; m h 2 G 1.13; and m6A 1.29.

Table 2.10 Recovery of Nucleosides from Human Serum Nucleoside

Std.-

83.4 103.2 96.4 98.2 106.4 104.7 100.9 99.8 103.8 108.1

Std.

-

2(a)

87.3 101.5 94.1 99.4 106.4 106.4 101.9 98.6 104.0 109.3

Serum-1

Serum-2

87.4 106.6 71.7 116.7 102.7 104.6 (b) 24.7 94.5 134.6

81.5 99.5 77.1 92.1 103.6 101.3 95.8 15.4 86.0 123.9

a) 1.0 ml of water spiked with nucleosides processed as a serum sample. Spike level added as nmoles per 1.00 ml of serum: w 4.81; U 4.17; mlA 0.361; I 0.372; mlI 0.306; m'G 0.184; the m2G 0.277; A 0.123; m2m2G 0.226; and m6A 0.253. b) High ac4C peak, which interfered in the integration of the ni2G peak.

c73

slightly lower recovery of Y (92%) was expected as Y has the lowest affinity to the gel and this loss occurred during the wash step of 300 pl of 50% methanol. However, this wash step is essential to reduce the background peaks. Recovery of m3U at 95% may indicate a slight loss of m 3 U on the gel colurnn. Since the loss is only about 5% and the contribution of endogenous m3U is about 1 to 2% (Tables 2.4 - 2.5), the error in measurement should only be about +3%. We chose not to correct the results from the recovery data of the internal standard because a 3% bias is acceptable for quantitation of biological samples. Figure 2.9 presents the chromatograms of a spiked and an unspiked serum sample from a breast cancer patient. Tables 2.9 and 2.10 present the recovery of 10 nucleosides from urine and serum, respectively. The recovery for all the nucleosides from both urine and serum were essentially quantitative. The recovery data i n Table 2.10 were obtained from a nucleoside performance standard and human serum and shows the effects of a serum matrix on the recovery of the nucleosides. No loss of nucleosides was observed in the aqueous performance standards, but ca. 80% of A and 25% of m l A were lost from the two serum samples. The corresponding high recovery of I indicates the conversion of A to I by deaminase in the serum, and the high recovery of m6A indicates the transmethylation of m l A to m6A at alkaline pH. There is no increase in the concentration of mlI. This implies that m*A does not deaminate to form m l I by adenosine deaminase in serum. Also, this is in agreement with our results from the adenosine deaminase study (see section on adenosine deaminase activity in serum and urine). The origin of the high concentration of m l I in body fluids is still an unanswered question.

2.3.7 Precision of the Method The precision of the method is illustrated in Figures 2.10 and 2.11. In Figure 2.10 a normal male urine sample was analyzed independently in triplicate. The chromatogram obtained at 254 nm from each analysis was plotted. There are more than 30 identified and unidentified peaks which all match with respect to peak height and retention time. Figure 2.11 shows the duplicate analysis of serum from a cancer patient with untreated small cell carcinoma of the lung. The excellent reproducibility of

c74

RUN 2 RUN 3

t

r-vr-v----Tl

0

10

1

20

8

7

7

7

r

I

1

9

-7-r7-l-v-

.

.

I

60

40

30

...

TIME (min)

Fig. 2.10 Reproducibility of HPLC analysis of ribonucleosides in human urine from three independent runs.

RUN 1 -

254 nrn. 2 0 mAU

1 ' ' ' ' ' ' 1 ' ' 10 ' ' ' ' ' ' ' ' ' ' ~20' 1

0

~ ' ' ' ' ' ' ' ' ' ' ' ' ' r ' ' ~ ' ' l " ' ' ' " ' ' ' ' '

30

40

50

Tlrnebnin)

Fig. 2.1 1 Reproducibility of HPLC analysis of ribonucleosides in human serum from two independent runs.

c75

the two independent runs for serum analysis was again demonstrated by the exact match of all the peak heights and retention times between the two runs. The non-reproducibility of the peaks at about 10 minutes is due to the fact that the peaks at 9.6 and 10.5 minutes retention are not nucleosides. They are retained by the gel column resulting from adsorption and not covalent bonding. They are present in the final sample due to variations of washing of the gel column in the cleanup step. A pooled serum sample spiked with nucleosides was used as a control sample to monitor the day to day performance over time of the HPLC-UV nucleoside chromatographic method. This control sample was analyzed with each group of ten to twenty samples. Also the pooled serum was spiked with m l I and m 2 m 2 G at levels so that the change of the respective peak height ratios to the m3U (internal standard) could be visually noted during the chromatographic run. In this way by simply observing the chromatogram the analyst will know the performance of the cleanup step. The concentration of 13 nucleosides from each of the 18 control serum samples analyzed over a period of five weeks is presented in Table 2.11. The results are presented in 3 groups. Runs 1 to 6 were obtained in the first week, runs 7 to 12 in the second week, and runs 13 to 18 five weeks after the sample was pooled. The samples were stored at 4 "C between analyses. The relative standard deviations (RSD, %) of the data for the seven nucleosides (w, U, X, m l I , a&, m2m2G, and t6A) ranged from 3.6% to 9.0% and is quite good. It was also observed that the concentration of seven nucleosides (mlA, I, G, PCNR, ac4C mlG and m6A) changed with storage time, indicating the biological and chemical instability of these nucleosides at 4 "C. High RSD, % values were observed for m l A and m6A, which was caused by the conversion of m l A to m6A during sample preparation. Concentrations of I, G, PCNR, and m l G were affected by the storage time at 4°C. The concentration of I decreased with time, and the concentration of G showed a slight increase in the second week and a decrease after five weeks. PCNR showed a slight increase after one week. The increase in concentration of m l G was caused by an increase of the interference peak N20.3, which eluted as a shoulder at the front of the m l G peak.

Table 2.11 Day-to-Day

0

Precision

of

HPLC

G

PCNR

mlI

mlG

ac4C

2 m2G

t6A

m6A

29.4 26.6 74.5 73.1 85.4 69.1

1050 999 932 894 912 792

59.9 59.6 55.1 53.8 59.9 53.3

45.3 40.5 46.8 41.6 42.5 37.2

45.7 54.4 57.0 52.4 45.6 44.9

340 327 329 341 342 299

19.3 25.3 28.4 25.0 21.9 19.2

125 103 155 123 147 117

425 431 450 450 444 398

52.6 62.9 54.5 49.9 50.7 50.0

37.3 35.2 9.7 12.0 11.9 10.6

7920 7870 7610 7880 7680 7830

53.3 30.1 26.4 38.8 25.9 54.0

1010 1020 969 999 999 1020

96.5 96.9 90.0 87.6 98.2 94.1

66.5 66.3 57.2 54.1 67.8 63.5

63.8 67.2 54.4 60.4 58.1 63.2

329 337 342 340 337 326

34.7 35.7 27.2 31.2 30.6 28.7

111 172 128 163 131 138

417 438 419 433 427 435

50.8 49.5 49.7 47.8 55.3 53.2

46.5 35.9 37.5 39.2

3530 3940 3260 3840 3430 3590

7570 7390 7490 7350 7170 7220

68.3 i i i i 81.4

887 527 614 511 529 539

93.3 85.2 87.2 82.4 43.9 77.8

56.8 41.9 46.0 46.5 25.1 44.2

61.5 55.7 55.2 61.7 71.1 57.0

330 356 370 356 349 354

32.2 39.7 39.0 41.6 43.9 38.0

154 103 106 122 119 106

405 421 436 433 409 408

57.4 54.1 51.9 52.7 50.6 54.0

9.2 22.0 28.7 23.9 22.3 12.3

3815 343 9.0

7792 488 6.3

51.3 22.3 43.5

969

89.2 6.69 7.50

51.9 9.64 18.6

56.9 7.40 13.0

338 15.6 4.6

34.9 5.26 15.1

129 21.4 16.6

426 15.5 3.6

52.8 3.58 6.77

27.1 12.6 46.5

u

1 2 3 4 5 6

4140 4540 3960 3660 3700 3580

8650 8430 8420 8500 8470 7580

7 8 9 10 11 12

3770 3620 3920 3720 3740 3900

13 14 15 16 17 18

RSD,%

U Q1

Analysis

X

Y

SD

Nucleoside

I

Run No.

Mean

Serum

m

l

~

55.5

7.5

All values are in units of pmollml of serum. Pooled serum from both normal and cancer patients was divided were obtained within the first week after the serum was pooled 13 through 18 were obtained after the pooled serum was stored Inosine (1) is not stable during storage at 4 "C. Only the values

25.3

39.6

into 1.0 ml aliquots and stored at 4°C until analysis. Runs 1 through 6 , runs 7 through 12 were obtained during the following week, and runs for five weeks. for runs 1 through 12 were used.

c77

Table 2.12 Storage Stability of Human Serum Nucleosides at -20°C After 10 days and 6 Months

-20°C for 10 Davs Nucleoside Y

U mlA I X G

PCNR rnlI

mlG ac4C

m2m * G t6A

-20°C for 6 Mo nths

O-Day JO-Davs 0-Dav 10-DavS 5020 7680 3.8 220 i 33 24 130 36 350 50 94

4820 7290 3.0 170 i 23 30 160 40 i 49 97

2640 8840 11 360 150 65 36 100 50 240 46 76

Serum

O-Dav6-Mos O-Dav

2560 9040 14 100 234 6.9 29 120 57 260 47 74

1010 6600 15 360 86 47 92 41 81 860 170 28

914 6810 12 50 150 11 98 48 100 680 180 26

1630 6040 23 9700 160 130 140 63 130 1100 22 36

D

L-Mos 1610 6170 14 2440 270 40 120 76 130 860 24 37

All values are in units of prnol/ml of serum. i = interference peak. PCNR = l-Ribosylpyridin-4-one-3-carboxamide.

2.3.8 Stabilitv of Nucleosides The storage stability of urinary ribonucleosides was reported earlier. (ref. 21). Eight nucleosides, mlA, PCNR, mlI, m l G , a d C , m2G and m 2 m 2 G are all stable at -20 "C or -70 "C for at least one month. Since one month is more than sufficient for accomplishing the analysis, a longer storage time was not studied. Urine is generally sterile, thus, the concentration of nucleosides are not altered biologically. Serum, however, contains many enzymes, and biological alteration of the nucleosides upon storage is a major concern. From long term precision data (Table 2.11), we observed that the concentration of certain serum nucleosides change over a period of 5 weeks at 4 "C. In another study four serum samples from cancer patients were selected for storage. After establishing their original (0 day) nucleoside concentrations, the samples were stored at -20 "C until the second analysis. Two samples were analyzed 10 days later and the other two were analyzed after 6 months. The results are presented in Table 2.12. w, m l G , m2m2G, and t6A were

w,

c7a

found to be stable at -20 "C for six months. In general, the other nucleosides either increased or decreased to some extent in concentration over time, with some of the changes being inconsistent within the individual samples. PCNR, a&, and m l I concentrations changed less than 20% over the six months. At this time, we do not have the data for serum stored at -70 OC. However, it is strongly recommended that serum samples should be held at -70 "C to minimize alteration of nucleoside concentration as a result of enzymatic action in the serum.

2.3.9 Ribonucleotides a nd Olieoribonucleotides i n Normal and Cancer

Serum

The possibility of observing nucleotides or oligoribonucleotides in serum was a concern, especially in serum from cancer patients who have an abnormally high turnover rate of tRNAs and a highly permeable cell wall. Serum samples from both chronic lymphoblastic leukemia and nonHodgkin's lymphoma have very high concentrations of serum nucleosides. A normal serum sample was used for the control and unfractionated calf liver tRNAs were added to another aliquot of the normal serum to ensure that there were no inhibitors of the snake venom phosphodiesterase (SVP) and bacterial alkaline phosphatase (BAP) in the ultrafiltrate of the serum. The corresponding recoveries of major and modified nucleosides from the SVP and BAP hydrolysate of the normal serum filtrate spiked with calf liver tRNAs indicate that serum does not have inhibitors of the enzymes. Results from the two cancer serum samples are shown in Figures 2.12 and 2.13. In each figure, the lower chromatogram shows the free nucleosides in the serum filtrate, and the upper chromatogram is from the same serum filtrate, but incubated with s n a k e venom phosphodiesterase and bacterial alkaline phosphatase at 37 "C for 16 hours. In both hydrolyzed samples only four major ribonucleosides, Cyd, Urd, Guo, and Ado were increased, but the increased levels were low and less than one nanomole per ml. No changes were noted in the concentration of modified nucleosides in the hydrolyzed samples. A decreased concentration of m l A and an increased concentration of m6A after hydrolysis was due to the alkaline pH of the serum filtrate ( c a . pH 9) which converted m l A to m6A. Some unknown peaks which decreased in size after hydrolysis probably are the oligonucleotides, and appeared as

c79 IN'S

LYMPHOMA

Hydrolyzed ( P D A - L B A P, 1 6 h)

Tine Cninl

Fig. 2.12 RNA oligomers and nucleotides in non-Hodgkin's lymphoma patient serum.

CHRONIC

LYMPOBLASTIC

LEUKEMIA

BKG

Hydrolyzed (PDA L BAP. 16 h)

U

Fig. 2.13 RNA oligomers and nucleotides in chronic lymphoblastic leukemia patient serum.

C80

the small amount of major nucleosides in the hydrolyzed samples. We did not observe any significant amount of oligonucleotides or nucleotides in the normal or cancer serum samples. 2.3.10 Analysis of Creatinine in Urine and Serum bv a Modified HPLC

Method From our earlier collaboration with Dr. Waalkes and the late Dr. Borek it became apparent that the urinary nucleoside levels in normal subjects was a function of the muscle mass of the individual. Therefore, we decided to determine the urinary nucleoside levels in random samples and not in 24 h urine collections. The concentration of creatinine in urine is a function of muscle mass. Thus, we related the nucleoside levels to the creatinine level in urine (refs. 69, 70). Most of the studies on urinary nucleosides in cancer marker investigations have used the nucleoside to creatinine ratio.

Table 2.13 Day-to-Day Matrix Dependent Creatinine in Serum

Precision

mgldl

of

HPLC

Analysis

mgldl

Sample

Run 1

Run 2

Sample

Run 1

Run 2

1 2 3 4

8.57 7.26 5.76 7.61 7.81 9.60 8.23 6.08 8.22 9.60

8.09 7.44 5.37 7.68 8.01 9.66 8.54 6.05 7.98 9.64

11 12 13 14 15 16 17 18 19 20

5.20 8.13 4.66 8.4 1 5.56 9.12 6.54 7.84 7.80 7.4 1

5.60 8.33 4.76 8.47 5.72 8.89 6.39 8.04 7.72 7.38

Mean

7.48 mg/dl 0.24 3.2

5

6 7 8 9 10

SD RSD,%

Data obtained from routine serum creatinine analyses over a period of 2 weeks. A duplicate run was made with each set of 20 samples. SD = {[C(X~-X2)12/2P}o~s

of

C81

-

*

URINE

N

Run 2 1.16 mglml I?

*?

w

.4 STD

1.05 mdml rn N

*

E

c 0)

SERUM STD 6.65 vg/ml 1 -4 0

w

F Fig. 2.14 HPLC-UV chromatographic analysis of creatinine in urine and serum.

C82

Q

T

1""""'""""""""""""""""""'"""''

0

10

20

30

40

so

Time ( m i d

Fig. 2.15 Consistency of urinary nucleoside profiles from three normal males. When we measured the nucleoside levels in serum of non-cancer patients it was noted for a number of cases that very high levels of nucleosides were obtained for those individuals that have kidney malfunction. This observation agreed with those of our collaborators in Naples. Salvatore's group has proposed that a nucleoside/creatinine ratio (defined as nucleoside coefficient) should be used to correct this discrepancy. They reported that a higher correlation to clinical status was obtained when this ratio was used (refs. 71, 72). In our studies, we used the reported creatinine data provided by the clinical laboratories. In general, their data were obtained by an automated Jaffe colorimetric creatinine method. The sensitivity and precision of the Jaffe method for serum creatinine was poor and only one significant figure was given for the measurement. We obtained the same results of imprecision and positive bias in our laboratory when we made the measurements using the colorimetric method. The accuracy and precision of the nucleoside measurements were often compromised by the creatinine method and

C83

data used. Thus, we modified an HPLC ion-exchange-UV method of Chiou et al. (ref. 73). This modified HPLC method gave good separation (Fig. 2.14) of creatinine from background material. In our laboratory a day-today matrix dependent precision was achieved with a RSD,% of 3.2 (Table 2.13). These data were collected over a two week period using results from 20 independent duplicate analyses. The recovery of creatinine added to serum and urine were in the range of 94.9 to 103%. 2.3.11 A Comparison of Nucleoside Levels in Random and Total 24 Hour Human Urine Collections The level of urinary nucleosides from normal male and female populations were separately examined. Random and 24 hour urine collections were obtained from fifteen males and 10 females, ranging in The random age from 19 to 58 and 20 to 50 years, respectively. specimens were caught at 8:00 am, 1O:OO am and 3:OO pm and the 24 hr samples were collected on the next day. The nucleoside concentrations in urine were expressed as nanomoles of nucleoside per micromole of cre at inine .

Table 2.14 A Comparison of Nucleosides from Random and Total 24 Hour Human Urine Collections Ratio of Nucleoside in Random to 24 Hour Samples Nucleoside Y

mlA PCNR mlI

8 AM 1.03 0.98 0.82 0.99

(20.7%) (28.3%) (38.2) (48.4%)

10 AM 1.07 0.86 1.08 0.96

(17.2%) (49.0%) (25.6%) (55.6%)

3 PM 1.00 0.95 0.95 0.95

(19.5%) (16.4%) (35.5%) (34.7%)

24 Hour 1.00 1.00 1.00 1.00

(19.4%) (27.5%) (31.1%) (35.5%)

m2G

1.13 (19.8%)

1.15 (39.0%)

0.97 (34.0%)

1.00 (38.9%)

m2m2G

0.90 (26.3%)

0.90 (31.6%)

0.94 (31.9%)

1.00 (27.5%)

Nucleoside determined as nmol of nucleoside per pmol of creatinine. Value in ( ) = RSD,%.

C84

Table 2.15 Nucleoside levels in Normal Human Serum ~

pmol of Nucleoside per ml of Serum ~~

Nucleoside

Mean

SD

RSD,%

Max.

Min.

Y

2850 5570 93 1560 85 221 64 65 26 144 25 44 15

755 1340 38 2700 44 387 25 21 8.8 54 5.4 9.9 5.0

26.5 24.1 40.7 173 52 1880 39 32 33 38 22 23 33

4994 8712 270 14000 350 1840 232 110

1820 2990 31 29 16 13 19 11 11 18 17 24 6.3

U

mlA I

X

G

PCNR

mlI mlG ac4~

m2m2G t6A m6A

51

272 41 61 29

N = 94,age range from 19 to 84,51 females and 43 males. PCNR = l-Ribosylpyridin-4-one-3-carboxamide.

Table 2.16 Nucleoside Creatinine Ratios in Normal Human Serum nmol Nucleoside per p m o l Nucleoside Y

U

mlA I

X

G PCNR

mlI mlG ac4~

m2m2G t6A

Mean

43.7 84.3 1.22 25.2 1.19 3.59 0.92 0.98 0.38 2.14 0.361 0.638

Creatinine

SD

RSD, %

Max

Min

14.4 30.5 0.48 45.4 0.81 6.79 0.35 0.42 0.14 0.87 0.11 0.15

33.0 36.1 39.7 180 67.9 189 38.1 41.5 37.0 40.5 29.5 23.9

111 185 2.81 234 6.2 38.3 2.06 2.19 0.92 4.15 0.93 1.15

22.1 30.8 0.368 0.323 0.253 0.254 0.159 0.171 0.150 0.339 0.186 0.309

C85

Table 2.16 ( c o n t i n u e d ) m6A

0.222

0.069

30.9

0.39

0.091

N = 94, age range from 19 to 84, 51 females and 43 males. PCNR = l-Ribosylpyridin-4-one-3-carboxamide.

Figure 2.15 demonstrates the remarkably constant excretion of urinary nucleosides from three healthy normal subjects. The chromatograms were plotted after normalization of the m2m2G peaks. The only difference noted is for the peak at a retention time of 9.6 minutes, and as pointed out earlier in the serum precision study (Figure 2.11) this peak most likely is not a nucleoside. 2.3.12 Serum and Urine Ribonucleoside Levels in Normal PoDulationS A number of research groups have reported the nucleoside levels in urine and serum from normal populations (refs. 17, 18, 26, 34, 37, 39, 40, In these reports, only a few modified 44, 47, 51, 53, 5 8 , and 63). nucleosides were reported and at times from a relatively small number of individuals. Our new HPLC technology allowed the quantitation of 21 known ribonucleosides, hU, w, C, ncmSU, U, mlA, I, X, G, PCNR, m3U, m l I , m l G , a d C , m2G, m2m2G, mcm5s2U, @A, m6A, mt6A, and ms2t6A in urine and serum. We made a concerted effort using this methodology to establish a reference level of these nucleosides in a normal population of males and females and of different age groups, and from hospitalized patients who have diseases other than cancer. In collaborative programs with Professor F. Salvatore and his research group at the University of Naples Medical school, Dr. Edith Mitchell in the Department of Oncology, University of Missouri Medical School, and Dr. John McEntire of the Cancer Research Center, Columbia, MO, we collected 94 samples of serum from normal healthy donors and 47 serum samples from non-cancer male patients. The normal healthy population consisted of 5 1 males and 43 females ranging in age from 19 to 84. Thirteen serum modified nucleosides and creatinine were quantified and the data are presented in Tables 2.15 and 2.16. Table 2.15 gives the results of the nucleoside levels in pmol/ml of serum and Table 2.16 presents the serum nucleoside/creatinine ratio in units of nmoles

C86

nucleoside/pmole of creatinine. Data are given for 13 serum nucleosides with mean values, RSD, %, and range values. The RSD, % for most of the nucleosides ranged from 22 to 39% for the population. Inosine, xanthosine, and guanosine showed a high RSD, which was expected, and was discussed earlier in this paper as due to the instability of these molecules and The narrow possibility of re-absorption during the excretion process. distribution (RSD, %) of each nucleoside in the 94 samples was essentially the same whether the data were expressed as pmol/ml or as the nucleoside/creatinine ratio. This indicates a stringently controlled metabolic rate of nucleic acids for healthy subjects. A comparison of normal serum nucleoside values by age and sex are presented in bar graph form in Figures 2.16-a, 2.16-b and 2.17-a, 2.17-b. There is no age and sex dependency for any of the nucleosides studied.

Table 2.17 Nucleoside levels in Human Serum from Diseases* Other than Cancer nanomoles/ml Nucleoside Y

U mlA I

X

G PCNR mlI mlG ac4~ m2m2G t6A

m6A

Mean 3.52 7.62 0.100 0.243 0.087 0.033 0.044 0.106 0.034 0.237 0.038 0.048 0.017

SD

RSD, %

1.01 2.16 0.069 0.490 0.040 0.023 0.023 0.039 0.019 0.095 0.014 0.014 0.013

28.6 28.3 69.6 20 1 45.5 73.0 52.1 36.7 54.5 39.9 36.0 29.4 74.9

DOTC* = diseases other than cancer. Population (N) of 47 males, age range 27-83. PCNR = l-Ribosylpyridin-4-one-3-carboxamide.

Max 6.14 13.4 0.498 2.35 0.167 0.114 0.161 0.215 0.111 0.598 0.069 0.098 0.064

Min 2.06 4.40 0.005 0.013 0.016 0.0010 0.0091 0.068 0.01 1 0.088 0.014 0.027 0.002

200

IAGE

AGE 19-29(37) AGE30-39(29) AGE40-49(9) AGE 50-59(9) II] AGE60-84(8)

19-29(37) AGE 30-39(29) AGE40-49(9) AGE50-59(9) AGE60-84(9)

5

\ VI

0

k

.-

a 100

pseU

Urd NUCLEOSIDE

1

C

rnlA

x

PCNR m 1 I rnlG ac4C m Z m X t 6 A NUCLEOSIDE

Fig. 2.16-a and 2.16-b. Comparison of serum nucleoside values by age groups. parentheses gives the number of individuals for that age group.

m6A

The number in

0

200

W W

n

Male(51) Female(43)

Male (51) Female(43)

100

0 pseU

Urd Ino NUCLEOSIDE

mlA '

X

'PCNR ' m l I ' mlG 'ac4C m2m2G t 6 A ' m 6 A NUCLEOSI DE

'

Fig. 2.17-a and Fig 2.17-b Comparison of serum nucleoside values of normal males vs females. number in parentheses gives the number of males and females.

The

C89

Table 2.18 Nucleoside Creatinine Ratios for Human Serum from Diseases Other than Cancer nmol of Nucleoside per pmol Nucleoside Y

U mlA I X G PCNR mlI mlG ac4~ m2m2G t6A m6A

Mean 39.0 85.9 1.16 2.99 0.969 0.376 0.484 1.16 0.384 2.68 0.429 0.536 0.193

SD 10.6 30.0 1.01 6.47 0.477 0.327 0.220 0.412 0.231 1.25 0.167 0.158 0.167

RSD, % 27.3 35.0 87.2 216 49.3 86.8 45.6 35.4 60.2 46.9 39.0 29.6 86.5

Creatinine

Max

Min

63.2 193 7.16 32.8 2.22 1.66 1.23 2.58 1S O 7.95 0.790 0.919 0.851

20.6 40.9 0.041 0.178 0.230 0.015 0.069 0.512 0.090 0.939 0.106 0.281 0.023

Population of 47 males, age range from 27-83.

Table 2.19 Nucleoside Levels in Normal Human Urine pmoles per 24 hours Nucleoside hU Y

U mlA

X

G PCNR *lI

mlG

ac4~ m2G A

Mean 58.6 331 2.08 24.1 0.63 0.88 11.7 21.1 10.6 15.9 6.05 3.45

SD 24.5 112 1.25 8.90 0.43 0.55 4.71 10.8 4.87 6.62 3.00 2.45

RSD, % 41.9 33.8 60.3 37.0 69.0 62.5 40.2 51.0 45.8 41.8 49.5 71.0

Max

Min

123 648 6.08 48.4 1.18 2.15 25.1 52.4 23.9 29.6 11.7 8.33

27.9 186 0.61 7.89 0.06 0.22 6.17 5.62 2.30 1.68 0.47 0.23

c90

Table 2.19 (Continued) A* m2m2G t6A m6A

3.50 16.1 10.6 0.58

1.76 6.44 3.52 0.53

50.2 40.1 33.2 90.4

7.21 35.9 21.9 1.95

1.44 4.83 6.41 0.15

N = 18; 7 males and 11 females, age range from 25 to 53. PCNR = l-Ribosylpyridin-4-one-3-carboxamide. A* = Unknown, probably is a modified A, calculated as A. Creatinine: Mean = 9.77; SD = 3.66; RSD,% = 37.5; Max. = 16.9; Min. 24 hr).

=

3.88 (in mg per

Table 2.20 Nucleoside levels in Normal Human Urine mg per 24 hours

Nucleoside hU Y

U mlA X G

mlI mlG ac4~ m2G A A* m2m2G mcm5s2~ t6A m6A

Mean 14.4 80.8 0.508 6.78 0.178 0.250 5.96 3.15 4.54 1.80 0.922 0.995 5.01 0.808 4.37 0.164

SD

RSD, %

Max

Min

6.02 27.3 0.305 2.50 0.123 0.156 3.05 1.45 1.89 0.893 0.655 0.471 2.00 0.462 1.45 0.148

41.9 33.8 60.3 37.0 69.0 62.5 51.0 45.8 41.8 49.5 71.0 47.2 40.1 52.7 33.2 90.4

30.2 158 1.48 13.6 0.335 0.610 14.8 7.10 8.45 3.48 2.23 1.93 11.2 1.85 9.03 0.548

6.86 45.4 0.148 2.22 0.016 0.062 1.59 0.683 0.480 0.139 0.062 0.385 1S O 0.037 2.64 0.04 1

N = 18; 7 males and 1 1 females, age range from 25 to 53. PCNR = l-Ribosylpyridin-4-one-3-carboxamide. A * = Unknown, probably is a modified A, calculated as A. Creatinine: Mean = 9.77; SD = 3.66; RSD,% = 37.5; Max. = 16.9; Min. (in mg per 24 hr).

=

3.88

c91

T a b l e 2.21 Nucleoside Creatinine Ratios in Normal Human Urine nmol Nucleoside/p mol Creatinine Nucleoside hU Y

U

mlA

X

G PCNR mlI mlG ac4~ m2G A A* m2m2G mcm5s2~ t6A m6A

Mean

6.13 35.1 0.217 2.51 0.663 0.093 1.25 1.99 1.01 1.56 0.579 0.348 0.356 1.66 0.240 1.13 0.071

SD

1.47 7.17 0.105 0.557 0.062 0.050 0.348 0.534 0.331 0.573 0.208 0.217 0.105 0.350 0.091 0.23 1 0.081

RSD, %

Max

Min

24.0 20.4 48.5 22.2 93.5 53.1

8.18 53.2 0.495 4.14 0.171 0.204 1.97 3.10 1.54 2.86 0.952 0.847 0.603 2.50 0.376 1.74 0.31 1

2.72 24.0 0.087 1.66 0.008 0.031 0.654 0.807 0.331 0.218 0.120 0.034 0.21 1 1.10 0.019 0.675 0.016

28.0

26.8 32.8 36.6 35.9 62.3 29.5 21.1 37.7 20.4 114

N = 18; 7 males and 1 1 females, age range from 25 to 53. PCNR = l-Ribosylpyridin-4-one-3-carboxamide A* = Unknown, probably is a modified A, calculated as A. Creatinine: Mean = 9.77; SD = 3.66; RSD,% = 37.5;Max. = 16.9; Min. (in mg per 24 hours).

=

3.88

Thirteen serum modified nucleosides in patients with a number of diseases other than cancer (DOTC) were also investigated. This study The data are included 47 males with ages ranging from 27 to 83. expressed in units of nmol/ml of serum (Table 2.17) and nmoles of nucleoside per pmol of creatinine (Table 2.18). The nucleoside values for the DOTC patients were essentially the same as for the normals. Sixteen urinary nucleosides and creatinine were measured in 24 hour collections of urine from 18 normal healthy donors (7 males, 11 females, ages 25 to 50). The standard deviation and RSD, % of a population of individuals are prcsented as pmo1/24 hour, mg/24 hour, and nmol of nucleosides per pmole of creatinine in Tables 2.19, 2.20 and 2.21,

c92

respectively. A narrow distribution of each nucleoside was again observed in the urine of normal healthy subjects as for serum in healthy subjects. We have recently reported an estimation and "reference values " of a series of serum and urinary modified nucleosides by HPLC. A comprehensive statistical evaluation was made of the serum nucleosides to estimate the "normality" of the data distribution to determine the effect of sex, age, and ethnic origin (Italian vs. American) groups on this distribution. The values for all of the nucleosides tested showed a normal distribution. A slightly higher serum nucleoside concentration was obtained for the American group as compared to the Italian group. There was also a small sex (male vs. female) difference observed for m l I and m2m2G. 2.3.13 Clearance Values of Nucleosides It is very interesting to note that the observed serum nucleoside profiles are clearly different from the profile of urinary nucleosides, either expressed in concentration units or as a ratio to creatinine. This can only be explained in that the renal excretion for each nucleoside is different. Using the equation, Clearance (ml of serum/ min) = (U/1440)/S U = pmol of nucleoside/24 hour of urine collection S = pmol of nucleoside/ ml of serum we calculated the clearance values for 10 nucleosides and creatinine from 18 healthy subjects (Table 2.22) and 19 untreated leukemia and lymphoma patients (Table 2.23). Similar clearance values for the healthy and cancer population were observed for the respective nucleosides. However, a larger RSD, % of the clearance values was found in the cancer population. This may be due to the larger variation in renal function of the patients. The much lower clearance values of U, G, and X as compared to creatinine indicates the re-absorption of these nucleosides. The lower clearance value for ac4C is caused by the instabliltiy of this compound in urine, thus, a lower than true value of urinary ac4C was obtained for the calculation. It is interesting that PCNR, mlI, mlG, rnZmZG, and t6A gave much higher clearance values than for creatinine. The clearance of mlA is so high, (> 1000 mYmin) that its serum concentration often is too low to be

c93

measured accurately for the calculation. The high clearance for these nucleosides can be explained only in that secretion of these molecules in addition to filtration is taking place. 2.3.14 Adenosine Deaminase Activity in Serum and Urine To assess if adenosine deaminase will cause a change of A to I and m l A to m l I in urine and serum, we spiked serum with m l A and A. Both A and m l A in serum and urine are converted to I and m lI , respectively in the presence of adenosine deaminase. The rate of conversion for m l A to m l I is approximately 20% of the rate of conversion of A to I. 2.3.15 Serum Nucleosides in Canines with Osteosarcoma Figures 2.18 and 2.19 give correlations of the change in serum nucleoside levels in dogs with osteosarcoma to treatment with 1 5 3 s m e t h y 1en ed i ami n e t e tr a m e t h y 1p h o sp h ate ( 53 Sm-EDTMP) . This work was done (1988) with Dr. D. McCaw of the School of Veterinary Medicine, UMC

Table 2.22 Clearance of Nucleosides and Creatinine in Normal Human Subjects Nucleoside Y

U X

G PCNR mlI mlG ac4~ m 2rn 2G t6A m6A Creatinine

Mean 82.8 0.268 8.15 17.5 140 207 330 124 496 177 28.5 98.7

SD

RSD.%

Max

Min

24.9 0.24 0.088 17.9 60.6 95.3 132 83.8 154 51.3 24.7 32.5

30.0 87.7 1.08 102 43.4 46.0 40.1 67.7 31.0 29.0 86.9 33.0

138 1.2 8.2 70.9 303 398 613 356 904 321 97.2 144

46.5 0.069 8.1 1.4 58.5 68.0 88.0 4.82 187 106 7.2 476

Population of 7 males and 11 females, age range 25-53. Values are expressed as ml of serum/min.

c94

Table 2.23 Clearance of Nucleosides and Creatinine in Cancer Patients Nucleoside Y

U X G

PCNR rnlI mlG ac4~

m2rn2G t6A m6A Creatinine

Mean

SD

RSD.9

Max

85.1 0.82 62.0 43.0 131 174 351 76.9 389 163 25.6 109

54.0 1.5 80.3 64.6 94.0 167 3 19 81.4 322 103 52.6 66.0

63.5 180 130 150 71.7 95.8 90.8 106 82.9 63.6 206 60.5

206 6.4 222 287 428 640 1100 319 1179 406 246 276

Min 15.8 0.01 2.7 0.79 22.9 157 6.7 1.1 2.1 43.0 0.07 17.6

There are no data on the sex and age of the above population. Values are expressed as rnl of serumhin.

(ref. 74). Figure 2.18 shows 4 panels on the change in nucleoside levels from a dog that responded to the treatment. Panels A and B show that the levels of eight modified nucleosides, w, m5C, mlI, a&C, @A, mzrnZG, PCNR and N20.3 increased immediately after the treatments (day 1 and day 7), then decreased from day 20 to day 40. Finally the concentration of all 8 nucleosides remained constant and at low level from day 40. Figure 2.19 shows the change in serum nucleoside levels for 3 dogs, one that responded, the second partially responded, and a third dog with progressive disease. The change in serum modified nucleoside levels for eight nucleosides (four nucleosides are shown in this figure) are closely correlated to the clinical prognosis of the dogs under treatment. For the dog that responded the levels of the nucleosides decreased after 20 days and remained at a lower level. For the dog that partially responded the levels of the nucleosides initially decreased (day 2 to day 10 or 20) then all continuously incI;eased until death. For the dog with progressive disease, three of the four nucleosides were continuously increasing until death of the dog. The unknown nucleoside N14.9 level decreased during treatment over time for all 3 dogs. This decrease in the level of N14.9 in all

(')Nucleoside level in percent relative to prelreatmentvalue (day 0).

Fig. 2.18 Four panels on the change in the relative concentration of serum nucleosides for a dog. The nucleoside level is given in percent relative to the pretreatment value (day 0).

c)

ID

m

200

A

Responded

M

-

7

a,

> W

100

a,

U

.r u)

0 W

"1

7

U

3

z

0

0

40

20

60

80

100

-

- I Days

Days

120

-

=

7

.nn

IW

W

Ih\

B

Partially Responded

/

h

120

D

N14.9 in All Three Dogs

I

Q

Partially Rospondod

> W

7

W

U

80

.r u)

0 W

7

V

60 -a- AC4C

3

z

0

P

4)

80

7 n .. . -0

Days

. . . -

P

4

0

I

8

0

-

I

80

-

a

1

la,

Days

(')Nucleoside level in percent relalive lo prelrealrnenl value (day 0).

Fig. 2.19 Correlation of serum nucleosides in three dogs with osteosarcoma to 153Sm-EDTMP treatment for a dog that responded (panel A), a dog that partially responded (panel B), and a dog with progressive disease (panel C). Panel D gives the response of unidentified nucleoside N14.9 in all three dogs.

c97

three dogs over the course of the treatment is most interesting. The possibility of using N14.9 as a marker for assessing immunological activity should be investigated. Figure 2.20 presents a serum nucleoside chromatogram of a dog with osterosarcoma. A new major modified nucleoside (C*) was identified as 5hydroxylmethylcytidine (om5C) (ref. 75). omSC has only been observed in the serum of dog and cat. The average concentration of om5C in normal canine is 22 nmol/ml which is more that 10 fold higher than the levels of the other serum nucleosides. Also, m5C which was not observed in human serum and urine is present in high concentration in dog serum. In addition, two unknown nucleosides N14.9 and N2c.3, were found in dog serum which are yet to be identified. N14.g has not been observed in human body fluids and the unknown N20.3 is a minor modified nucleoside observed in human serum. This difference in the dog serum nucleoside profile as compared to human indicates a difference in metabolism of RNAs in dog and man.

10 9

OSTEOSARCOMA

0

7

6 3

a E

5 4

3 2

1

0 0

10

20

Time

(min.)

30

40

50

Fig. 2.20 HPLC-UV chromatography of nucleosides in serum of dog with osteosarcoma. For chromatographic details refer to Experimental.

C 98 2.3.1 6 Serum Nucleosides in Leukemia and Lymphoma Patients In collaboration with Professor F. Salvatore's group at the University of Naples, Italy, serum from pretreatment leukemia and lymphoma patients were collected and analyzed. Brief preliminary results are presented as bar graphs. Figure 2.21 shows a comparison of the normal serum nucleoside levels to the levels found in acute lymphocytic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphocytic leukemia, and chronic myeloid leukemia (CML). Figure 2.22 gives a comparison of the normal serum nucleoside levels to the levels found in Hodgkin's lymphoma (HL) and non-Hodgkin's lymphoma (NHL). We found that the level of modified nucleosides from the patients with all types of leukemia and lymphoma are significantly higher than the normal values. Acute lymphocytic leukemia patients have much higher levels than other leukemias and lymphomas. This indicates the excellent diagnostic value of modified nucleosides for leukemia and lymphoma. The preliminary data also show that the modified nucleoside profiles of some leukemias are different from others.

Fig. 2.21 Serum nucleoside levels in leukemia patients. parentheses gives number of subjects in the study.

Number i n

c99

Fig. 2.22 Serum nucleoside levels in lymphoma patients. parentheses gives number of subjects in the study.

Number in

2.3.17 Polynuclear Aromatic Hvdrocarbon Carcinogen-Ribonucleoside Adducts in the Urine of Fish and Rat Monitoring the catabolites of polynuclear aromatic hydrocarbons (PAH) carcinogen-RNA adducts, specifically ribonucleoside adducts (RNadducts) in body fluids as "an indicator" for occupational exposure and environmental contamination has many obvious advantages over the measurement of PAH-DNA adducts. Measuring RN-adducts in body fluids is non-invasive: The samples of urine and blood are easily obtainable while DNA-adducts generally require isolation of DNA from the tissue of the subjects. Secondly, the analytical method for RN-adducts is simpler and faster, thus a much lower cost of analysis than for the DNA-adducts (32P post-labelling 2D-TLC method): Ribonucleoside adducts are the product of RNA metabolism and are present in blood and urine, while

ClOO

DNA-adducts require isolation of the DNA from tissue followed by a hydrolysis of DNA to the nucleotides. Using the methods that we have described in this chapter, RN-adducts can be isolated in high purity and yield from body fluids and then measured by well established techniques such as HPLC-UV, HPLC-fluorescence, GC-MS erc.. Much larger amounts of RN-adducts should be found in urine than in DNA (one adduct per 109 bases) because not only does urine accumulate RN-adducts from RNA turnover over time, but also a higher amount of RNA-adducts should be present initially than for DNA-adducts in the tissue. The enzymatically activated PAHs (epoxides) react directly with the RNAs in the cytoplasm and thus do not need to be transported across the membrane of nuclei as in the reaction with DNA, therefore there should be considerably more RNA-adducts formed with RNA than with DNA. Gerhard Schoch et al. (ref. 7 6 ) calculated that there are about equal amounts of RNA and DNA in most of the eukaryotic cells and that the whole-body turnover of RNA in human adults is about 100 mg/kg body-wt /day. An average adult body weighs about 60 kg, thus the turnover of RNA is about 6 grams per day. With a conservative assumption that the level of PAH-RNA adducts in RNA is the same as for PAH-DNA adducts in DNA, i.e. one adduct per lognucleobases, this would calculate that there should be more than 20 picomol of each RN-adduct excreted into the urine each day. This amount is within the sensitivity of the measurement of modern chromatography-spectrometry techniques. In collaboration with Dr. Mark Smith of the Cancer Research Center, Columbia, Mo., a preliminary experiment was conducted in our laboratory to investigate RN-adducts in the urine of rat and fish that have been exposed to benzoIa1pyrene (BaP). Three female rats, body wt. ca. 200 g each, were injected i.p. with 300 pl dimethylsulfoxide (DMSO) containing 0.30 mg (1.2 pmol) cold BaP and 80 x 106dpm (ca. 0.7 nmol) of 3H-BaP (53 Wmmole). The 3H activity found in the 24 hour urine collected from each of the three rats for three days is presented in Table 2.24. Also, the nonadducted benzo[a]pyrene metabolites were determined in the ethyl acetate extract of the urine, and the 3H activity found is assumed to be the B[a]P metabolites (Table 2.25). Only 6 to 8% of the total 3H activity was found in the first 24 hours of urine collected. The amount of activity in the urine decreased significantly from day 1 to day 3.

ClOl Table 2.24 3 8 Activity Found in Urine from Benzo[a]pyrene Treated Rats 3H Rat No.

Activity,

Day 1

Day 2

1

46.0

2

6 1 . 8 ~lo5 (7.8)

3

47.0

x

(dpm)

lo5 (5.8)

105 (5.9)

Day 3

32.8

105 (4.1)

46.9

x lo5 (5.9)

13.0 x lo5 (1.6)

25.8 x 105 (3.3)

9.18 x 105 (1.1)

9.35

105 (1.2)

~

(

) gives the percentage of 3H activity of the total injected excreted per

day.

Table 2.25 Non-adducted

Benzo[a]pyrene Percentage

Metabolites in Rat Urine of

Non-adducted

B[a]P

Metaboliteda)

Rat No.

Day 1

Day 2

Day 3

1

15.4

17.8

15.5

2

21.8

25.8

16.1

3

21.1

22.4

22.4

(a) Percentage of the 3H activity extracted into ethyl acetate phase.

One-half of the 24 hour rat urine collection was evaporated to 2 ml volume and adjusted to p H 9 with 1 N NaOH. Two nmol of internal standard (m3U) were added and the total sample was loaded onto a 3 x 0.6 cm Affigel column. The column was then washed with 10 ml of 0.25 M CH3COONH4, followed with a 2 ml wash of 50% methanol in water. All of the ribonucleosides were eluted from the column in 25 ml of 0.2 M HCOOH in 50% methanol in water. (There was no 3H activity found in the 2nd 25 ml eluate of 0.2 M HCOOH in 50% methanol/H20. This indicates that all of

c102

the RN-adducts were completely eluted in the first 25 ml). The first 25 ml eluate of 0.2 M HCOOH was evaporated to dryness and re-dissolved in 2.0 ml of water. 100 pl of the sample were injected onto the HPLC. The HPLC conditions used were the same as for modified ribonuclesides in biological samples (see Experimental Section) except that elution solvent C was replaced with 60/40 C H 3 C N / H 2 0 and the run time was extended to 60 minutes to elute the BaP-nucleoside adducts. The HPLC eluate was collected in 0.5 ml fractions and 100 p1 aliquots from each fraction were counted for 3H activity. Only 10% (4800 dpm) of the total 3H activity was found in the 0.2 M HCOOH eluate fractions. This calculates as 7 nmol of ribonucleoside-adducts excreted in rat urine in the 1st 24 hours. Thus 35 nmol of RN-adducts per kg body wt. per day were excreted by a rat, or there is one RN-adduct in 105 of the nucleobases. We have synthesized a number of reference ribonucleoside-BaP adducts to assist in the identification of the RN-adducts. The major activated BaP metabolite, Benzo[alpyrene-trans-7,8-dihydrodiol-9,1O-epoxide (BPDE), (Lot No. CSLO85-008-10-20, was purchased from Midwest Research Institute, Kansas City, MO.) and reacted with each of the four major ribonucleosides. 4 0-

11A

131) (IS)

1%

30-

3

a E

PCNR

206

-

24 48 mli (ace R8. 2.23.b

I El-

Fig. 2.23-a. HPLC-UV chromatogram of ribonucleosides in urine from a rat injected i.p with cold and 3H-labelled benzo[a]pyrene. See text for details of chromatography and experimental.

C103 :m-: 82:

-

*

3

r!OU BPDE

~

a: a

Adenosine Adducts

-

LC!

2c+

32

3 i

36 ?!me

38

Irnin.1

413

42

44

Fig. 2.23-b. Panels A, B, and C show the HPLC-UV chromatograms for synthesized reference BPDE-adducts of Ado, Cyt, Urd, and Guo, respectively. Panels show only the sections of the HPLC chromatograms from 30 to 45 minutes. Panel D is the 3H radiogram between 30 to 45 minutes from Fig. 2.23-a.

The reaction conditions of Jennette et al. (ref. 77) were used with minor modification to prepare BPDE ribonucleoside adducts (BPDE-cytidine, 100 y1 of nucleoside solution (1.0 uridine, -guanosine and -adenosine). mg/ml in water) was reacted with 50 pl of 1.0 mg/ml of BPDE in acetone. After the reaction mixture was incubated at 37 "C for four hours, an additional aliquot of 50 yl of the BPDE solution was added. The reaction was allowed to continue at 37 "C overnight with no pH adjustment. After

C104

extraction of the non-adducted metabolites with ethyl acetate the aqueous layer was analyzed with HPLC-UV using the same chromatography conditions as for the rat urine samples. Figure 2.23-a shows the HPLC-UV chromatogram of the ribonucleosides in rat urine collected on day one from a rat injected with cold and 3H labelled benzo[a]pyrene. The HPLC eluate was collected after the UV detector in 0.5 ml fractions for 3H activity counting. Only the fractions between 30 to 45 min contained any 3H activity and all of our synthesized reference ribonucleoside-BPDE adducts also eluted in this section (30 to 45 min) of the chromatogram. Figure 2.23-b shows a comparison of the HPLC-UV chromatograms from four of our synthesized RN-BPDE adducts to the HPLC-3H activity radiogram of urine from a rat with injected BaP. At least four RN-adducts at about the same level of concentration were observed in the rat urine. The peak at 39.5 min most likely is the N2, BPDE-guanosine adduct. The peak of 37.5 min could be

m3U

/

U

\

I

\

i I6A

?

B

J ri

10

I

254 nm -r

20 irnc

( r n i r i .

>

30

40

5u

Fig. 2.24. HPLC-UV chromatogram of ribonucleosides in urine from a channel catfish injected with 3H labelled benzoralpyrene. See text for details of chromatography and experimental.

C105

either the N6, BPDE-adenosine adduct or a BPDE adduct of uridine. Peaks at 32.5 and 34.0 are not nucleoside adducts of BPDE. They are the adducts of other active benzo[a]pyrene metabolites, perhaps from the 4,5-epoxide. In collaboration with Drs. Chris Schmitt and Brian Steadman at the National Fisheries Contaminant Research Center (NFCRC), urine from a benzo[a]pyrene treated (oral injection) channel catfish was collected. The results of HPLC-UV analysis of the urine is presented in Figure 2.24. It is very interesting to note that high levels of the four major nucleosides were found in the urine, and I and t6A were the only two major modified nucleosides excreted. Three adducts were found in the fish urine, the major one eluting at 39 min, probably N2-guanosine-BPDE. The two unknown minor adducts that eluted at 40.0 and 42.5 min have only onetenth of the 3H activity of the 39 min adduct peak, also their retention times do not match with any of the reference RN-BPDE adducts available (see Fig. 2.23-b).

Summary During the last two years we have improved and extensively validated our method for quantitation of ribonucleosides in biological samples. This technology represents a significant advancement over the The precision, speed, methods that we reported earlier (refs. 1, 2). sensitivity and ruggedness of our methods are well suited for use in clinical research applications. With the described chromatography protocols, twenty known nucleosides in urine or serum (hU, Y, ncm5U, mlA, I, X, PCNR, mlI, mlG, ac4C m2G, rn2mzG, t6A, m6A, mt6A, ms2t6A, C, U, G and A), and more than ten unidentified nucleosides can be measured in a single 35 minute chromatographic run. The precision and ruggedness of the method was ensured with the introduction of a new internal standard, 3-methyluridine (m3U), which is added to the urine or serum sample before prechromatography treatment. Also, the accuracy of the method was improved by employing a UV diode-array detector and multiwavelength quantitation protocols. The within-day relative standard deviation (RSD, %) obtained on five nucleosides (y,U, mlI, m2m2G, and t6A) from a pooled human serum is under 5%. Long term (day to day) analytical precision for the five nucleosides (w, U, m11, m2m2G, and t6A) in a pooled serum sample over a period of five weeks (N = 15) gave a RSD in 2.4

C106

the range of 3.6 % to 9.0 %, and the long term recoveries for five representative nucleosides, (w, mlI, m2m2G, t6A and m6A) spiked at physiological levels in human serum, analyzed over a period of three weeks (N = 12) were 94 f 3 %, 104 f 6 %, 96 f 4 %, 99 f 3 % and 85 f 4 %, respectively. In our laboratory this method has been applied to approximately 500 human serum samples and 200 urine samples with consistently satisfactory results. U, mlA, I, X, G, PCNR, Thirteen human serum nucleoside levels mlI, mlG, a&, rnZmZG, t6A, and m6A) and 17 human urinary nucleoside levels (hU, w, U, mlA, I, X, G, PCNR, m*I, mlG, ac4C. m2G, A, m2m2G, ncm5s2U, t6A, and m6A) were established on analysis of a large number of samples from a normal population. In addition, preliminary studies on serum nucleosides as potential biological markers for small cell lung carcinoma, leukemias and lymphomas were achieved. Some significant correlations were noted between the levels and profiles of serum nucleosides and different neoplasias. The broad applicability of this method was demonstrated by the analysis of nucleosides in human plasma, whole blood, and other biological samples. Nucleosides in serum and urine from dog, cat, rat, mouse, monkey, fish and cell culture media have also been successfully measured. This high efficiency chromatography protocol also can be used for the enrichment of PAH and alkylated carcinogen-ribonucleoside adducts in urine and serum, then their measurement by either microbore HPLC with laser-induced fluorescence detection or capillary GC-MS. Recently, we have used this method for the identification and characterization of benzo[alpyrene-ribonucleoside adducts (BaP-RN-adducts) in the urine of fish and rat. High levels of BaP-ribonucleoside-adducts were found in the urine of BaP treated rat and fish. About twenty percent of the BaP inetabolites in rat urine are non-adducted (free), 10% of the metabolites were found in the urine as BaP-RN-adducts, and 70% have not been identified and are probably BaP-protein related adducts. The BaP-RN adducts were found at a level of one .(1) adduct per 105 nucleobases in RNAs. This was calculated from the number of BaP-RN adducts excreted in and the average whole-body turnover for the urine per day (7 nmol) tRNA of 61.3 mg/kg body-wt/day, and for rRNA of 477 mg/kg body-

(w,

C107

wtlday (ref. 76). Further, we did not take into account the BnP-RN adducts in the feces which is considered to be at much higher levels than in urine. Even with this conservative calculation, the BaP-RNA adducts in the cells are at least 10 to 100 fold higher than the BaP adducts in DNA. These results confirm our initial concepts that there are much higher levels of PAH-RNA adducts than those for DNA. Our findings strongly support that measurement of ribonucleoside adducts can serve as important endpoints in monitoring occupational exposure, environmental contamination, and the roles of RNA-adducts in chemical carcinogenesis.

Ack n owled gm en t We wish to gratefully acknowledge Marion Laboratories of Kansas City, MO, Supelco., Inc. of Bellefonte, Pa., and the University of MissouriColumbia and the State of Missouri for their financial support of a number of research projects reported in this three-volume series. 2.5 1.

2.

3. 4 5.

6.

References C. W. Gehrke, K. C. Kuo, G . E. Davis, R. D. Suits, T. P. Waalkes and E. Borek, Quantitative high-performance liquid chromatography of nucleosides in biological materials, J. Chromatogr., 150, (1978) 455476. G. E. Davis, R.S. Suits, K.C. Kuo, C.W. Gehrke, T.P.Waalkes, and E. Borek, High performance liquid chromatographic separation and quantitation of nucleosides in urine and some other biologic fluids, Clin. Chem. 23, (1977) 1427-1435. K.C. Kuo, C.W. Gehrke, R.A., McCune, T. P. Waalkes, and E. Borek, Rapid, q u a n t i t at i ve high pe rforma nc e liquid chromatogr aphy of pseudouridine, J. Chromatogr. Biomed. Applic. 150, (1978) 455-476. M. Uziel, L.H. Smith and S.A. Taylor, Modified nucleosides in urine: selective removal and analysis, Clin. Chem., 22, (1976) 1451-1455. M. Zakaria, P. R. Brown, M. P. Farnes and B. E. Barker, HPLC analysis of aromatic amino acids, nucleosides, and bases in plasma of acute lymphocytic leukemics on chemotherapy, Clin. Chim. Acta, 126, (1982) 69-80. A.M. Krstulovic, R.A Hartwick, and P.R. Brown. High performance liquid chromatographic determination of serum UV profiles of normal subjects and patients with breast cancer and benign fibrocytic changes, Clin. Chim. Acta 97, (1979) 159-170.

C108

7.

8.

9.

10. 11.

12.

13.

14

15

16 17.

R.A. Hartwick, A.M. Krstulovic, and P.R. Brown. Identification and quantitation of nucleosides, bases and other UV-absorbing compounds in serum using reversed-phase high-performance liquid chromatography. 11. Evaluation of human sera, J. Chromatogr. 186, (1979) 659-664. R.C. Simpson and P.R. Brown, High-performance liquid chromatographic profiling of nucleic acid components in physiological samples, J. Chromatogr., 379, (1986) 269-31 1. E. Hagemeier, K. Kemper, K.-S. Boos and E. Schlimme, Development of a chromatographic method for the quantitative determination of minor ribonucleosides in physiological fluids. Characterization and quantitative determination of minor ribonucleosides in physiological fluids, J. Clin. Chem. Clin. Biochem., 22, (1984) 175-184. E. Schlimme, K.S. Boos, E. Hagemeier, K. Kemper and U. Meyer, Direct clean-up and analysis of ribonucleosides in physiological fluids. J. Chromatogr. 378, (1986) 349-360. G. Schoch, J. Thomale, M. Lorenz, M. Suberg and U. Karsten, A new method for the simultaneous analysis of unmodified and modified urinary nucleosides and nucleobases by high-performance liquid chromatography, Clin. Chim. Acta, 108, (1980) 247-257. H. Topp, G. Sander, G. Heller-Schoch and G. Schoch, A high performance liquid chromatography method for the determination of pseudouridine and uric acid in native human urine and ultrafiltered serum, Anal. Biochem., 150, (1985) 353-358. G. Sander, J. Wieland, H. Topp, C. Heller-Schoch, N. Erb and G. Schoch, An improved method for the simultaneous analysis of normal and modified urinary nucleosides and nucleobases by high-performance liquid chromatography, Clin. Chim. Acta, 152, (1985) 355-361. R.A. De Abreu, J.M. van Baal, CHMM de Bruyn, JAJM, Bakkeren, EDAM Schretlen. High-performance liquid chromatographic determination of purine and pyrimidine bases, ribonucleosides, deoxyribonucleosides and cyclic ribonucleosides in biological fluids, J. Chromatogr., 229, (1982) 67-75. T. P. Waalkes, M. D. Abeloff, D. S . Ettinger, K. B. Woo, C. W. Gehrke, K. C. Kuo and E. Borek, Modified ribonucleosides as biological markers for patients with small cell carcinoma of the lung, Eur. J. Cancer Clin, Oncol., 18, (1982) 1267-1274. T. P. Waalkes, M. D. Abeloff, D. S. Ettinger, K. B. Woo, C. W. Gehrke, K. C. Kuo and E. Borek, Biological markers and small cell carcinoma of the lung, Cancer, 50, (1982) 2457-2464. C. W. Gehrke, K. C. Kuo, T. P. Waalkes and E. Borek, Pattern of urinary excretion of modified nucleosides, Cancer Res., 39, (1979) 1150-1153.

c109

18. J. Speer, C. W. Gehrke, K. C. Kuo, T. P. Waalkes and E. Borek, tRNA breakdown products as markers for cancer, Cancer, 44, (1979) 21202123. 19. E. Borek, 0. K. Sharma and J. I. Brewer, Urinary nucleic acid breakdown products as markers for trophoblastic diseases, Am. J. Obstet. Gynecol., 146, (1983) 906-910. 20. E. Borek, T. P. Waalkes and C. W. Gehrke, Tumor markers derived from nucleic acid components, Cancer Detect. Prev., 6, (1983) 67-71. 21. C. W. Gehrke and K. C. Kuo, Major and modified nucleosides, RNA, and DNA, in : L. J. Marton and P. M. Kabra (Eds.), Liquid Chromatography in Clinical Analysis, Humana Press Inc., Clifton, New Jersey, 1981, pp. 409-443. 22. H.A. Scoble, J.L. Fasching, and P.R. Brown, Chemometrics and liquid chromatography in the study of acute lymphocytic leukemia, Anal. Chim. Acta. 150, (1983) 171-181. 23. J. Thomale, G. Nass, Elevated urinary excretion of RNA catabolites as an early signal of tumor development in mice, Cancer Letter 2, (1982) 149-159. 24. G . Nass, J. Thomale, A. Luz and U. Friedrich, Excretion of modified nucleosides by malignant cells in vivo and in vitro and its clinical relevance. manuscript in preparation. 25. E. Schlimme, K.-S, Boos, B. Wilmers and H.J. Gent, Analysis of ribonucleosides in body fluids and their possible role as pathobiochemical markers, in: F. Cimino, G.D. Birkmayer, E. Pimentel, J.V. Klavins, F. Salvatore (Eds.), Human Tumor Markers, Walter de Gruyter and Co., Berlin (1987) pp. 503-518. 26. R. W. Trewyn, R. Glaser, D. R. Kelly, D. G. Jackson, W. P. Graham I11 and C. E. Speicher, Elevated nucleoside excretion by patients with nasopharyngeal carcinoma. Preliminary diagnostic/prognostic evaluations, Cancer, 49, (1982) 2513-2517. 27. D. A. Heldman, M. R. Grever, J. S. Miser and R. W. Trewyn, Relationship of urinary excretion of modified nucleosides to disease status in childhood acute lymphoblastic leukemia, J. Natl. Cancer Inst., 71, (1983) 269-273. 28. D. A. Heldman, M. R. Grever and R. W. Trewyn, Differential excretion of modified nucleosides in adult acute leukemia, Blood, 61, (1983) 291296. 29. R. W. Trewyn and M. R. Grever, Urinary nucleosides in leukemia: Laboratory and clinical applications, CRC Lab. Sci., 24, (1986) 7 1-93. 30. J.W. Mackenzie, R.J. Lewis, G.E. Sisler, W. Line, J. Rogers and I. Clark, Urinary catabolites of ribonucleic acid as cancer markers: A preliminary report of their use in patients with lung cancer, The Ann. of Thorac. Surg., 38, (1984) 133-139.

CllO 31. I. Clark, J.W. Mackenzie, J.R. McCoy and W. Lin, Comparison of urinary

32. 33. 34.

35. 36. 37.

38.

39. 40 41.

42.

modified nucleosides and bases between rats with hepatomas and nephroblastomas, in: G. Nass (Ed,), Recent Results in Cancer Research, Springer-Verlag, Berlin and New York, 84, (1983) 388-400. W. Lin, J. W. Mackenzie and I. Clark, Excretion of RNA catabolites by rats with hepatoma transplants, Cancer Letters, 22, (1984) 187-192. W. Lin, J.W. Mackenzie and I. Clark, Effect of transplanted osteogenic sarcoma on urinary RNA catabolites, Cancer Letters, 35,( 1987) 47-57. T. Rasmuson, G.R. Bjork, L. Damber et al. Tumor markers in bronchogenic carcinoma. An evaluation of carcinoembryonic antigen, tissue polypeptide antigen, placental alkaline phosphatase and pseudouridine. Acta Radiol. Oncol. 22, (1983) 209-214. T. Rasmuson and G.R. Bjork. Pseudouridine: A prognostic marker in non-Hodgkin's lymphoma. Cancer Chemother. Report, 59, (1975) 721727. T. Rasmuson and G. R. Bjork, Excretion of pseudouridine in urine as a tumor marker in malignant diseases, Bull. Mol. Biol. Med., 10, (1985) 143-154. K. Nakano, K. Shindo, T. Yasaka and H. Yamamoto, Reversed-phase liquid chromatographic investigation of nucleosides and bases in mucosa and modified nucleosides in urines from patients with gastrointestinal cancer, J. Chromatogr., 332, (1985) 127-137. K. Nakano, K. Shindo, T. Yasaka and H. Yamamoto, Reversed-phase high-performance chromatographic investigation of mucosal nucleosides and bases and urinary modified nucleosides of gastrointestinal cancer patients, J. Chromatogr., 343, (1985) 21-33. G. Mills, F. C. Schmalstieg and R. M. Goldblum, Urinary excretion of modified purines and nucleosides in immunodeficient children, Biochem. Med., 34, (1985) 37-51. A. Fischbein, 0. K. Sharma, I. J. Selikoff and E. Borek, Urinary excretion of modified nucleosides in patients with malignant mesothelioma, Cancer Res., 45, (1983) 2971-2974. G. Sander, H. Topp, J. Wieland, G. Heller-Schoch and G. Schoch, Possible use of urinary modified RNA metabolites in the measurement of RNA turnover in the human body, Hum. Nutr. Clin. Nutr., 40C, (1986) 103118. H. Topp, G. Sander, G. Heller-Schoch and G. Schoch, A high-performance liquid chromatographic method for the detemination of pseudouridine and uric acid in native human urine and ultrafiltered serum, Anal. Biochem., 150, (1985) 359-368.

Clll

43. H. Topp, G. Sander, G. Heller-Schoch and G.T. Schoch, Determination of 7-methylguanosine, N2,N2-dimethylguanosine and pseudouridine in ultrafiltered serum of healthy adults by high-performance liquid chromatography, Anal. Biochem., 161, (1987) 49-56. 44. S. Tamura. J. Fujii, T. Nakano, T. Hada and K. Higashino, Urinary pseudouridine as a tumor marker in patients with small cell lung cancer, Clin. Chim. Acta, 154, (1986) 125-132. 45. S. Tamura, Y. Amuro, T. Nakano, F. Fujii, Y. Moriwaki, T. Yamamoto, T. Hada, and K. Higashino, Urinary excretion of pseudouridine in patients with hepatocellular carcinoma, Cancer (Phila.) 57, (1986) 1571-1575. 46. P. Vreken and P. Tavenier, Urinary excretion of six modified nucleosides by patients with breast carcinoma, Clin. Biochem., 24, (1987) 598-603. 47. F. Oerlemans and F. Lange, Major and modified nucleosides as markers in ovarian cancer: A Pilot study, Gynecol. Obstet. Invest. 22, (1986), 212-217. 48. J.G. Maessen, G.J. van der Vusse, M. Vork, and G. Kootstra. Nucleotides, nucleosides, and oxypurines in human kidney measured by use of reversed-phase high-performance liquid chromatography, Clin. Chem. 34/6, (1988) 1087-1090. 49. GPJM Gerrits, AAM Haagen, RA De Abreu, LAH Monnens, JM, Gabreels, JM, Grijbels, ALM, Theeuwes and JM, van Baal. Reference values for nucleosides and nucleobases in cerebrospinal fluid of children, Clin. Chem. 34/7, (1988) 1439-1442. 50. L. Buhl, C. Dragsholt , P. Svendsen. et al Urinary hypoxanthine and pseudouridine as indicators of tumor development in mesotheliomatransplanted nude mice, Cancer Res. 45, (1985) 1159-1162. 51. G. Apell, F.L. Buschman and O.K. Sharma, Improved and rapid method for quantitation of modified nucleosides in urine and sera with radialpack cartridge, J. Chromatogr. 374, (1986) 149-154. 52. A. Colonna, T. Russo, F. Cimino and F. Salvatore, Modified nucleosides in biological fluids of cancer patients by high performance liquid chromatography, J. Clin. Chem. Clin. Biochem., 19, (1981) 640. 53. A. Colonna, T. Russo, F. Esposito, F. Salvatore and F. Cimino, Determination of serum pseudouridine and other nucleosides in human blood serum by high-performance liquid chromatography, Anal. biochem., 130, (1983) 19-26. 54. T. Russo, A. Colonna, F. Esposito, F. Salvatore and F. Cimino, Detection and estimation of several modified nucleosides in serum of cancer patients, It. J. Biochem., 31, (1982) 75-78. 55. F. Salvatore, A. Colonna, F. Costanzo, T. Russo, F. Esposito and F. Cimino, Modified nucleosides in body fluids of tumor-bearing patients, Recent Results Cancer Res., 84, (1983) 360-377.

c112

56. Y. Amouro, H. Nakaoka, S . Shimomura, S . Tanmura. T. Hada and K. Higashino, Rapid high-performacne liquid chromatography for pseudouridine assay in serum and urine, Clin. Chim. Acta, 172, (1988) 117-122. 57. S.Tamura, J. Fujii, TR. Nakano, T. Hada. Serum pseudouridine as a biochemical marker in small cell lung cancer. Cancer Res. 47, (1987) 6138-6141. 58. F. Cimino, F. Costanzo, T. Russo, A. Colonna, F. Esposito and F. Salvatore, Modified nucleosides from transfer ribonucleic acid as tumor markers, in: E. Usdin, R. T. Borchardt, C. R. Creveling (Eds.), The Biochemistry of S-Adenosylmethionine and Related Compounds, MacMillan, London, (1982) 409-412. 59. F. Salvatore, A. Colonna, F. Costanzo, T. Russo, F. Esposito and F. Cimino, Modified nucleosides in body fluids of tumor-bearing patients, Recent Results Cancer Res., 84, (1983) 360-377. 60. F. Salvatore, T. Russo, A. Colonna, L. Cimino, G. Mazzacca and F. Cimino, Pseudouridine determination in blood serum as tumor marker, Cancer Detec. Prev., 6, (1983) 531-536. 61. L. Sacchetti and F. Salvatore, Novel aspects of blood serum nucleosides and isoenzymes as tumor markers, J. Tumor Marker Oncol., 2, (1987) 245-253. 62. K. C. Kuo, F. Esposito , J. E. McEntire and C. W. Gehrke, Nucleoside profiles by HPLC-UV in serum and urine of controls and cancer patients, in: F. Cimino, G. D. Birkmayer, J. B. Klavins, E. Pimentel, F. Salvatore (Eds.) Human Tumor Markers, Walter de Gruyter & Co., Berlin, New York, (1987) 520-542. 63. L. Sacchetti, F. Pane and F. Salvatore, Modified nucleosides in biological fluids: An overview of their role as biochemical signals of human neoplasia, J. Tumor Marker Oncol.,( 1989) 133-147. 64. C.W. Gehrke, K.C. Kuo, R.A. McCune, and P.F. Agris, Quantitative enzymatic hydrolysis of tRNAs: Reversed-phase high-performance liquid chromatography of tRNA nucleosides, J. Chromatogr. Biomed. Applic., 230, (1982) 297-308. 65. S. Fulton and E. Carlson, Boronate Ligands in Biochemical SeparationsApplications, Method. Theory of Matrix Gel PBA Amicon Corporation, Scientific Systems Division 17 Cherry Hill Drive, Danvers, MA 01923. 66. C.W. Gehrke and K. C. Kuo, Ribonucleoside analysis by high performance reversed-phase liquid chromatography, in: C.W. Gehrke and K.C. Kuo (Eds) Chromatography and Modification of Nucleosides, Part A, Elsevier Chromatography Library Series, Amsterdam, 1989, in press. also J. of Chromatogr., 471, (1989), 3-36.

C113

67. G.E.

Davis, Isolation and analysis of major and modified ribonucleosides by high performance liquid chromatogrpahy, Ph.D Dissertation, University of Missouri, Columbia, 1978. 68. J.T. Bernert Jr., C.J. Bell, J. Guntupalli, and W. Harryn Hannon, Pseudouridine is unsuitable as an endogenous renal clearance marker, Clin. Chem. (1988) Vol. 34. No. 6, 1011-1017. 69. E. Borek, Transfer RNA and its by products as tumor markers, in: S. Sell, Cancer Markers, The Humana Press, Clifton, New Jersey, (1980)

445-462. 70. C.W. Gehrke, K.C,.Kuo, T.P. Waalkes, E. Borek, Patterns of urine excretion of modified nucleosides. Cancer Research, 39, (1979) 11501153. 71. M. Savoia, T. Russo, E. Rippa, L. Bucci, F. Mazzeo, F. Cimino and F. Salvatore, Serum pseudouridine, Its evaluation as a biochemical signal of neoplasia, J. Tumor Marker Oncol., 1, (1986) 61-68. 72. F. Salvatore, M. Savoia, T. Russo, L. Sachetti, and F. Cimino, Pseudouridine in biological fluids of tumor-bearing patients, in Human Tumor Markers, F. Cimino et al. (eds) Walter de Gruyter and Co., Berlin, (1 987), 45 1-462. 73. W.L. Chiou, M.A.F. Gadalla, and C.W. Peng, Simple, rapid and micro high-pressure liquid chromatographic determination of endogenous "true" creatinine in plasma, serum, and urine, J. Pharmaceut. Sci, 67,

(1978), 182. 74. D.L. McCaw, K.C. Kuo, C.W. Gehrke, Serum levels of transfer RNA

modified nucleosides in normal and tumor bearing dogs. in preparation, Dec. 1989. 75. C.W. Gehrke, J.A. Desgres, G. Keith, P.F. Agris, H. Sierzputowska-Gracz, M. Tempesta and K. C. Kuo, Structural elucidation of unknown nucleosides, in: C.W. Gehrke and K.C. Kuo (Eds) Chromatography and Modification of Nucleosides, Part A, Elsevier Chromatography Library Series, Amsterdam, 1989, in press. 76. G. Schoch, G. Sander, H. Topp and G.H. Schoch, Modified nucleosides and nucleobases in urine and serum as selective markers for the wholebody turnover of tRNA, rRNA and mRNA-cap - Future prospects and impact, in: C.W. Gehrke and K.C. Kuo (Eds) Chromatography and Modification of Nucleosides, Part C, Elsevier Chromatography Library Series, Amsterdam, 1989, in press. 77. K.W. Jennette, A.M. Jeffrey, S.H. Blobstein, F.A. Beland, R.D. Harvey and I.B. Weinstein, Nucleoside adducts from the in vifro reaction of benzo[a]pyrene-7,8-dihydrodiol-9,lO-oxide or benzo[a]pyrene 4,5oxide with nucleic acid. Biochemistry, 5, (1977) 16.