,\NALYTICAL
68. 485-493
BIOCHEMISTRY
Separation
of Low
( 1975)
Molecular
Polyacrylamide FREDERICK
Received The weight absence
effect
Gel
VARRICCHIO
January
13.
of electrophoresis
Weight
AND
1975:
co/i tRN
an improved 2-dimensional weight RNAs. A first
followed
an identical
by
aration rated
of E. CW/; and into approximately
terns
than
previously and a pH
E.
c,f,/i
As.
on
Combinations
second rat tRNAs
the
amide gel electruphuresi~ changes in the ionization gel system\ for specific
of these
in
this
probably of \omc purposes
new
1975
separation
of
electrophoresis
gel n~itlr~~t
which used dimension.
20.
low
at pH 5.X or pH 8.3. the relative migration
urea
depend\ on base\. The uw is al\o discu\\ed.
gave
were separating at pH X.3
the gr-eatest
sep-
rat liver tRNA was gave more complex
sepapat-
Z-dimensional
a pH 5.X gel The separation
molecular presence or of specific
conditions
gel system for gel with 7 M urea
cytoplasmic tRNAs. The spots. Euharyutic tRNAs both
reported system. X.3 gel in the second
May
polyacrylamide dimension. 16%
dimension
liver JS
J. ERNST
HENRY
accepted
conditions
used to find low molecular
by
Electrophoresis
RNAs has been studied. Electrophoresis of 7 M urea. and heating all influence
Esclwric~hirr
RNA
system in the first oftRNAs
conformational of this and
other
and
in
the
dimension gel hy pulyacr-ylditferences ?dimensional
and
Polyacrylamide gel electrophoresis has proved to be a very sensitive method for separating ribonucleic acids. Ribosomal t I) and transfer (3) RNAs and their precursors and messenger RNA (3) have been separated by this technique. The introduction of Z-dimensional systems has greatly increased the resolution of complex mixtures such as ribosomal proteins (4). nuclear proteins (5), low molecular weight RNAs (6,7), and the products of enzymic digestion of RNA (8,9). One recently introduced Z-dimensional system used for separating low molecular weight RNA involves a 10% gel in the first dimension and a 30%~gel in the second dimension ( IO). Two other systems involve running the first- and second-dimension gels at different pHs (6,9). The system described by Stein and Varricchio (6) was shown to be capable of resolving the approximately 56 Escherichicl coli tRNAs into 70 spots. We have investigated the effect of several variables on the separation of specific RNAs. This has led to the development of another ?-dimensional system for separating low molecular weight RNAs. These two systems resolve tRNAs from eukaryotic sources into a more elaborate profile with over 30 spots. 4x5 CopyrIght All rights
,I. lY75 hy Academic Presr. Inc. of reproduction in any form rererved.
486
VARRICCHIO
MATERIAL
AND
AND
ERNST
METHODS
The composition and preparation of the pH 5.8 and pH 8.3 gel systems as well as the procedure and apparatus for running ?-dimensional polyacrylamide gels have been described (6). To compare RNAs from different sources or to compare different first dimension gel conditions, the tRNA containing sections of interior slices from each first dimension gel were positioned end to end on a second dimension gel. Escherichicr coli tRNA was purchased from General Biochemical Company. Pure E. coli tRNAs were a gift of Dr. S. Nishimura. lodinated l”zl-labeled E. co/i fMet-tRNA was a gift of Dr. W. Prensky. Yeast tRNA was purchased from Sigma. Total chicken, rabbit and rat liver RNA was prepared by phenol-cresol extraction of the minced organ in the presence of 0.5% tri-isopropylnaphthalene-sulfonic acid ( 1 1). RNA, :“P-labeled, from adenovirus type Z-infected KB cells was provided by Dr. K. Raska. RESULTS
We have examined the effect of pH, presence or absence of urea, and temperature on the migration of several pure species of tRNA. Only the pH 5.8 and pH 8.3 systems were used because these have been found to be the most useful in Z-dimensional polyacrylamide gel separations (6). There is little separation of tRNA at pH 3.6. We have tried a pH 5.0 first dimension gel because of the report of a structural transition of E. co/i tyrosine tRNA below pH 5.4 ( 12). but this did not give as good a 2dimensional separation as the pH 5.8 first dimension. Leucyl, phenylalanyl, isoleucyl, and tyrosyl” tRNAs were run individually with equal amounts of 12”I-labeled N-formyl-methionyl-tRNA. The result was that at 12°C the presence of 7 M urea at pH 8.3 retards some tRNAs, whereas it increases the mobility of others. Analogous effects were noted at pH 5.8. Separations at higher temperatures, 5O”C, generally decreased the resolution slightly. The resolving power of polyacrylamide gel electrophoresis for RNAs is demonstrated by the fact that fMet-tRNA was separated into two bands in the pH 8.3 system without urea and in the pH 5.8 system with urea. Eschericlzia co/i fMet-tRNA has been shown to be two species differing by only one base (I 3). Combinations of 16% polyacrylamide gels run at pH 5.8 and pH 8.3 with and without urea and high temperature runs (in the first dimension only). were tried to determine which would yield improved resolution of tRNAs in a second dimension. The best separation was when pH 8.3 gels containing urea were used in the first dimension and identicLl/ gels, but without urea, were used in the second dimension. Figure 1 shows the separation of rat liver RNAs on this new system with the “urea
LOW
MW
RNA
ELECTROPHORESIS
FIG. dimension
I. Separation of tRNA by two 400 pg of’ rat liver RN A was
pH
urea
8.3
8.3). The as described hpot(\)
are
gel (right), second hy at the
for
dimension Stein and upper-
right
different separated
16 hr at I.5 mA.
2-dimensional on the
90 V (pH
gel. il Ih%~ acrylamide Varricchio (6) and run of each
AND
5.81
gel
pH and
487
SEPARATION
5.8 urea
systems. gel (left)
I6 hr at I.5 mA.
In and
the
first
on the
IO0 V (pH
gel pH X.3 without urea. was prepared 13 hr at I5 mA. 225 V. The 5.5 RNA
panel.
shift” as compared with our previously described system using a “urea and pH shift.” The newer system separates mammalian tRNA into more spots: approximately 45 as compared with about 30 with the pH shift. The urea shift system is then capable of almost as complete resolution of mammalian tRNA as reversed phase chromatography. There are no large differences in the total number of isoaccepting tRNAs between E. cdi and mouse and man ( 14). Gallo and Pestka ( 14) found at least 56 species in normal and leukemic lymphoblasts, using reversed phase chromatography. This suggests that many of the spots on the seconddimension slab gel in the “urea shift” system may be single species of tRNA. This system has the disadvantage that the identity of the spots cannot be determined easily. In the pH shift system the tRNAs of each row of the second dimension can be determined by locating aminoacyltRNAs, carrying radioactively labeled amino acid, in the first dimension pH 5.8 gel ( 151. Purified E. m/i tRNAs have been used to determine the
488
VARRICCHIO
AND
ERNST
migration of many tRNAs at pH 8.3 (T. Seno, personal communication), but one does not know in which cases mammalian and prokaryotic tRNAs will behave similarly. The pH 8.3 urea gel in the first dimension does have the advantage of giving good resolution in the second dimension with larger amounts of tRNA. The maximum load of tRNA consistent with good resolution in the second dimension is in the range of 150 pg for the pH 5.8 gels. Loads of tRNA up to 500 pg have provided consistently good resolution with the pH 8.3 urea gels. Higher loads show the existence of several minor spots in the tRNA area, to a total of about 50: and also would be useful for preparitive separations. For preparative purposes it is also possible to run double thickness. 3 mm, second-dimension gels. It is interesting that rat liver tRNAs are separated into more spots by both ?-dimensional systems than those of E. coli which comprise about 56 species, i.e., about the same number as eukaryotes ( 16). We have not
FIG. 2. Separation of embryonic chicken liver and adult rat liver RNAs on the 7-dimensional pH 8.3 system with “urea shift.” Adult rat liver (left) and I S-day embryonic chicken liver RNAs (right), 400 pg, were separated on 15% acrylamide pH 8.3 gels containing 7 M urea for 17 hr at 1..S mA/gel. 100 V. The second dimension was a 16% acrylamide pH 8.3 gel without urea. Electrophoresis was for 23 hr at 15 mA. 22 V. The SS RNA spot(s) are at the upper right of each panel.
LOW
MW
RNA
ELECTROPHORESIS
AND
SEPARATION
489
separated tRNAs from a large number of sources to determine where exactly or over what range the transition from prokaryotic to eukaryotic pattern occurs, but since the yeast pattern strongly resembles that of rat liver, the dividing line between the two tRNA patterns may be a sharp one between prokaryotes and eukaryotes. We have used both two-dimensional polyacrylamide gel electrophoresis systems to look for organ differences in tRNA within a species, and to compare tRNAs from the same organ of rat, chicken, and rabbit. Using the pH 5.8 to pH 8.1 system. the differences, if any. were small. Some clearer differences could be seen using the pH X.3 urea shift system. Figure 2 shows the separation of embryonic chicken and adult rat liver tRNAs. There are differences both in spot locations and the relative intensity of common spots. Another difference between the two Z-dimensional systems is shown in Fig. I. In the pH shift system the rat SS RNA is resolved into two species in the second dimension whereas in the other Z-dimensional system the 5s RNA is not resolved. One cannot assume that this separation of 5s RNA is the separation of two different molecules. The E. co/i 5s RNA which was separated into two bands by polyacrylamide gel electrophoresis was shown to be the same molecule t 17).
FIG. 3. Identification of RNA by 7.dimensional polyacrylamide gel electrophoresis. Rabbit reticulocyte ribosomal RNA was mixed with :‘“P-labeled RNA from adenoviru\ type Z-infected KB cells and separated on the pH 5.8 to pH 8.3 system as described in the legend to Fig. I. The stained gel (left) shows some tRNA, two 5s spots, and another slower moving spot. The autoradiograph (right) shows the two 5S spots and a darker spot which is the VA RNA.
490
VARRICCHIO
AND
ERNST-
For any particular mixture of RNAs one or another pair of electrophoresis conditions may give a greater separation of the RNAs. For example, the best separation (six spots) of mouse cell mitochondrial tRNAs was achieved by running a pH 8.3 gel without urea at high temperature in the first dimension. The second dimension was a gel of the same composition as in the first dimension, but run at room temperature, 20°C (Dubin and Varricchio, unpublished). Two-dimensional polyacrylamide gel electrophoresis could provide a useful physical parameter for identifying a specific RNA. Figure 3 shows a separation of ribosomal RNA on the pH 5.8 to pH 8.3 system. The gel shows some tRNA which was associated with the rabbit reticulocyte ribosomes, 5s RNA and another RNA which is presumably the ribosomal 5.8 S RNA ( 18). A small amount of ‘sSP-labeled RNA from adenovirus type Z-infected KB cells was coelectrophoresed with the ribosomal RNA. An autoradiogram of the slab gel shows that the VA RNA ( 19) migrated slightly slower in both dimensions than the ribosomal 5.8 S RNA did. We have not attempted to extend these experiments to high molecular weight RNAs such as ribosomal RNAs and their precursors. DISCUSSION
The results show the sensitivity of RNA to changes in the conditions of polyacrylamide gel electrophoresis. The two pairs of electrophoresis conditions, which we have found to produce the largest differences in tRNA mobilities, are a pH 5.8 urea gel or a pH 8.3 urea gel in the first dimension and a pH 8.3 gel without urea in the second dimension. The Z-dimensional separation in the pH 5.8 to pH 8.3 system also gives a good separation of tRNAs if urea is not included in the first dimension gels. The separation is better and the spots are possibly sharper when urea is included as we routinely have done. The ?-dimensional effect achieved in this system presumably involves changes in the ionization of some of the bases (6). Methylation per se is apparently not involved since tRNA from a methionine-starved E. co/i relaxed mutant gives the same pattern in the pH 5.8 to pH 8.3 system as does tRNA from wild-type E. coli. On the other hand, the effect has a great deal of specificity since in a partially degraded sample of whole rat liver RNA. the new low molecular weight RNA fragments fall on a diagonal in the second dimension. Mammalian tRNAs give a more complex pattern than E. coli tRNAs in both Z-dimensional systems. The available evidence is that there is about the same number of tRNA molecules in mammalian cells as in bacteria (I 4,16). It is known that minor base content increases with the
LOW
MW
RNA
ELECTROPHORESIS
AND
391
SEPARATION
evolutionary development of the organism (20). Furthermore. the pattern of methyl substitutions differs between prokaryote and eukaryote RNAs (21). It is striking to us that the presence and absence of urea influences the relative migration rates so greatly. In a one-dimensional separation of t RNAs in the presence or absence of urea. the separation is not remarkably different except that the migration rate of tRNAs in urea-containing pH 8.3 gels is only about 505% of that in gels without urea. In the electrophoresis systems described here. the tRNAs should be in an expanded state because of the low cation concentration and the absence of divalent cations. Apparently, urea can bring about further modifications of tRNA conformation so that electrophoretic mobility is changed. Urea has previously been reported to loosen the structure of RNA by weakening hydrogen bonds ( 33.73). Of course, once the conformation has been changed the conformation-dependent changes in ionization constant could also occur (24). One wonders if the spectrum of the potential conformational changes of tRNAs as reported here, bears a relation to biologic structure-function requirements for the i/z \,ii~) functions of tRNA. On the other hand, these conformational changes may secondarily reflect features of primary structure required by other considerations. Apparently, tRNAs can be partially melted during electrophoresis. This may be useful in some situations as it was with the mitochondrial tRNAs. It was previously noted that tRNAs with a higher T,, migrated relatively faster at pH 5.X (25). Since the migration rate of low molecular weight RNA does vary greatly in polyacrylamide gel electrophoresis systems, the results presented in this paper also point out the desirability of clearly stating the system one is using. ACKNOWLEDGMENTS We mmcial Public
wish to thank Dr. Mary support was provided Health Service.
Petermann hy Grant
for carefully CA-0874X from
reading the the National
munubcript. Institutes
Partial fiof Health.
REFERENCES I.
Adesnik.
M.
and
Levinthal.
C. ( 196‘9)
1. Varricchio. F. t I97 I) irr Methods Vol. I. p. 297. Marcel Dekker. 3. Benz. 4. 5.
E. J.. Jr. and
Forget.
M. and Varricchio. R., Sitz. T. 0..
(‘~w~ttrrrr~.
56.
101 7.
Biol.
46.
Biology York.
B. C;. ( 197 I) ./. (‘lirl.
Kaltschmidt. E. and Wittman. Y’eoman, L. C.. Taylor. C. C~rmr,7rrrr. 51, Y6.
6. Stein. 7. Reddy.
./. Mol.
in Molecular Inc.. New
f/rvc~.sr.
281. (Last.
La&n.
A.. eds.).
50. 7755.
H. G. t 1970) A~trl. Bioc~lrc,vz. W.. and Busch. H. (1973)
F. ( 1973) A&. Rio(~/rc~?z. Rtt-Choi, T. S.. and Bush.
J. and
61, 1 I?. H. ( 1974)
36, 301. Bic~c~l~cvn.
Uiop/l~.\.
Rc,s.
tlio