ABSTRACT
A method has been developed for the microscale determination of 5,6-dihydrouridine, the most common post-transcriptional modification in bacterial and eukaryotic tRNA.
The method is based on stable isotope dilution liquid chromatography-mass spectrometry (LC/MS) using [1,3
-
15N
2
]dihydrouridine and [1,3
-
15N
2
]uridine as internal standards.
RNA samples were enzymatically digested to nucleosides before addition of the internal
standards and subsequently analyzed by LC/MS with selected ion monitoring of
protonated molecular ions of the labeled and unlabeled nucleosides.
Sample quantities of
~
1 pmol tRNA and 5 pmol 23S rRNA were analyzed for mole% dihydrouridine.
Dihydrouridine content of
Escherichia coli
{t R N A} sub {V G A} sup {S e r}
and
{t R N A} sub {G G U} sup {T h r}
as controls were measured as 2.03 and 2.84 residues/tRNA molecule, representing
accuracies of 98 and 95%
.
Overall precision values for the analyses of
E.coli
{t R N A} sub {V G A} sup {S e r}
and
E.coli
{t R N A} sub {G G U} sup {T h r}
, unfractionated tRNA from
E.coli
and 23S rRNA from
E.coli
were within the range 0.43-2.4%.
The mole% dihydrouridine in unfractionated tRNA and 23S rRNA from
E.coli
were determined as 1.79 and 0.0396%, corresponding to 1.4 and 1.1 residues/RNA
molecule respectively.
Modified nucleosides occur in DNA (
1
), but are particularly characteristic of tRNA, rRNA and eukaryotic mRNA (
2
). More than 79 different nucleosides are presently known in tRNA, the most
highly modified of the RNAs from all sources (
2
). The modified nucleosides show considerable structural variety, from simple
methylation of either the base or the O-2' hydroxyl of ribose to much more complex types of modification in
the base.
5,6-Dihydrouridine (D) is a post-transcriptionally modified nucleoside first reported as a naturally
occurring constituent of RNA by Madison and Holley in 1965, in tRNA
Ala
from yeast (
3
). It functions to promote conformational flexibility (leading references in
4
) and is the single most common form of post-transcriptional modification in tRNA from bacteria and eukaryotes (
5
,
6
), where it is found at conserved positions of the D loop in numbers up to 5
residues/tRNA. It occurs less commonly at position 47 of the variable loop of
tRNA and recently has been identified in the peptidyl transferase loop of 23S
rRNA from
Escherichia coli
(
7
). Dihydrouridine is characteristically absent from the RNA of most archaea
(archaebacteria) and is present only in trace amounts in the few archaea in
which it is found (
8
).
The method of choice for accurate quantitation of nucleosides in RNA
hydrolysates has traditionally been the measurement of HPLC chromatographic
peak heights or areas using UV detection and comparison with data from weighed
amounts of authentic nucleoside standards (
9
,
10
). Because D, unlike all other natural nucleosides, possesses no significant
chromophore, HPLC analysis with UV detection is not practical due to poor
sensitivity. Its previous quantitation in RNA has been achieved by several
methods. Gehrke and Kuo (
11
) quantitated dihydrouridine in six isoaccepting tRNAs from yeast and
E.coli
by monitoring absorbance at 210 nm, requiring 5 [mu]g tRNA. Cerutti
et al
. (
12
) treated tRNA with sodium borotritiide, followed by characterization and
quantitation of the labeled reduced trialcohol products. Magrath and Shaw (
13
) converted D of RNA to [beta]-alanine by alkaline treatment, followed by quantitation of [beta]-alanine with an amino acid analyzer. Molinaro
et al
. (
14
) estimated D in RNA by measuring the time-dependent loss of A
235
in 0.1 N KOH. Jacobsen and Hedgcoth (
15
) utilized a colorimetric assay for dihydropyrimidine after conversion of D in
RNA to its open ring form (
N
-ribosyl-3-ureidopropionic acid) and also TLC analysis of radioactively
labeled RNA digests for the quantitation of dihydrouridine. Randerath
et al
. (
16
) developed a tritium derivative method, which was limited by nucleoside
recovery losses, to semi-quantify modified nucleosides in RNA. Johnson and Horowitz (
17
) utilized the latter method for estimating D content in tRNA and 23S rRNA from
E.coli
.
Although the assays described above are generally reliable for the detection of
dihydrouridine, they all have notable limitations with regard to accurate
quantitative measurements. These include harsh reaction conditions with
potential for base loss, lack of sensitivity, selectivity and accuracy of
identification, sample loss and speed of analysis. In the case of reversed
phase HPLC-based methods, D is the earliest eluting nucleoside, resulting in
potential loss of selectivity due to minor UV absorbing impurities that elute just after the void volume. To overcome these limitations we have developed a
rapid, sensitive and accurate assay based on stable isotope dilution liquid
chromatography-mass spectrometry (LC/MS) with selected ion monitoring for the direct
chemical measurement of dihydrouridine in enzymatic hydrolysates of RNA.
[1,3-
15
N
2
]Uridine was purchased from Cambridge Isotope Laboratories (Woburn, MA). [1,3-
15
N
2
]Dihydrouridine was synthesized in 95% yield by the hydrogenation of [
15
N
2
]uridine under atmospheric pressure using 5% rhodium on alumina catalyst in
aqueous media (
18
) as follows. [
15
N
2
]Uridine (10 mg) was dissolved in 1 ml water, 7 mg 5% rhodium alumina was added
and the mixture was hydrogenated. After 4 h, the catalyst was removed and the
filtrate was purified by reversed phase HPLC using 25 mM NH
4
HCO
3
containing 1% acetonitrile, pH 6.5 (flow rate 1 ml/min,
t
R
4.3 min). The identity of [
15
N
2
]dihydrouridine was verified by LC-MS:
t
R
3.2 min (
19
), [M + H]
+
249. Structures of the labeled nucleosides are shown in Figure
1
.
Isoaccepting {t R N A} sub {V G A} sup {S e r} and {t R N A} sub {G G U} sup {T h r} from
E.coli
were purchased from Subriden Inc. (Rolling Bay, WA). Unfractionated tRNA from
E.coli
was purchased from Boehringer Mannheim Inc. (Indianapolis, IN). 23S rRNA from
E.coli
was isolated and purified as described (
7
).
RNA was completely hydrolyzed to nucleosides using nuclease P
1
, snake venom phosphodiesterase and bacterial alkaline phosphatase as previously
reported (
21
).
Analysis of nucleosides in RNA digests was carried out with a mass spectrometer
consisting of a non-commercial quadrupole mass analyzer, with a thermospray HPLC interface
(Vestec Corp., Houston, TX), controlled by a Vector/One data system (Teknivent
Corp., St Louis, MO). HPLC separations were made using a Supelcosil LC-18S column (4.6 * 250 mm) and a 3 cm Brownlee Spheri-5 C
18
precolumn, thermostatted at 31oC. The HPLC gradient elution system of Buck
et al.
(
9
), with 0.25 M ammonium acetate, pH 6.0, and acetonitrile, was used with minor
modifications in the gradient profile (
19
). Mass spectra for the region
m
/
z
244-250 were acquired every 0.36 s during the 10 min chromatographic elution of D and U. The instrument, procedures and
interpretation of data for qualitative LC/MS analysis of nucleosides in RNA
hydrolysates have been described in detail (
19
).
Using weighed amounts of authentic nucleosides, standard curves (nmol nucleoside
versus A
254
) for pseudouridine ([psi]), cytidine (C), uridine (U), 5-methyluridine (m
5
U), guanosine (G), 7-methylguanosine (m
7
G) and adenosine (A) were constructed based on HPLC chromatographic peak
heights, using UV detection (data not shown). These calibration curves were
then used in conjunction with HPLC chromatograms of
E.coli
tRNA digests to calculate the molar proportion of uridine in tRNA from
E.coli
.
To verify the spectroscopically determined concentrations of the isotopically
labeled internal standard solutions, they were calibrated by reverse isotope
dilution using primary standards of unlabeled U and D. In addition, this
procedure provides a test of the accuracy of mass spectrometrically measured
isotope ratios in mixtures of the labeled and unlabeled nucleosides. For this
standardization, four samples were prepared for each nucleoside by mixing known
amounts of primary standard and labeled internal standard solutions to achieve
four different optimum isotope ratios (
m
/
z
245/247 for U,
m
/
z
247/249 for D) in the mixtures. These solutions were then analyzed by LC/MS.
The peak area ratios were calculated automatically by the data system of the
mass spectrometer. Contributions from natural abundance heavy isotopes were
taken into account through the calibration curve. Each measurement was
performed in triplicate for statistical purposes. The data were then subjected
to a linear least squares analysis.
The quantitative assay consists of the addition of isotopically labeled U and D
to the RNA digest (1 [mu]g tRNA, 3 [mu]g 23S rRNA) prior to analysis by LC/MS. Quantitation of U and D in RNA
is accomplished by selected ion monitoring of their MH
+
ions and those of the corresponding isotopically labeled internal standards.
The mass values of these ions are: U, 245; D, 247; [
15
N
2
]uridine, 247; [
15
N
2
]dihydrouridine, 249. Although dihydrouridine and [
15
N
2
]uridine possess the same mass, their signals are distinguished by a difference
in retention time. The peak area ratios (245/247 for U, 247/249 for D) are used
to derive the amounts of U and D in the RNA. Since the mole% of U in the RNA is
known [as calculated from sequence data (
5
,
22
) or as measured from HPLC chromatographic peak heights], the mole% of D can be
determined from the U:D ratio. This approach is similar in principle to the
GC/MS method earlier developed for quantitative determination of 5-methylcytosine in DNA (
23
).
The calibration curves for U and D are shown in Figure
2
. Verification that the calibration solutions of labeled and unlabeled
nucleosides did not contain interfering species of equal mass was achieved by
MS analysis of samples containing labeled or unlabeled nucleoside in the
absence of the other. The enzyme solutions used for hydrolysis were similarly
tested for absence of interfering ions at the appropriate retention times.
The authors thank J.L.Stauffer for his instrumentation expertise. This work was
supported by NIH grant GM 29812. JJD was supported by NIH Trainee Fellowship 5
T32 GM08537.
REFERENCES
Return