Nucleotide sequence determination of bacteriophage T4 leucine transfer ribonucleic acid.

Abstract The nucleotide sequence of T4 tRNAleu, one of several transfer RNAs specifically coded for by bacteriophage T4, has been determined using 32P-labeled material from T4-infected cultures of Escherichia coli. The purified RNA species which has been sequenced has been shown to have leucine acceptor activity, and to hybridize well to T4 DNA. The sequence is: pGCGAGAAUGGUCAAADDGmGDAAAGGCACAGCACUNAAA * A ψ GCUGCGGAAUGAUUUCCUUGUGGGTψCGAGUCCCACUUCUCGCACCA—OH. The 87 nucleotide length is the same as that of the two E. coli leucine tRNAs, the sequences of which have been reported. The molecule can be arranged in the classic cloverleaf pattern. The sequence further shows that the anticodon of the T4 tRNAleu is -N-A-A-, in which N is a modified form of U. Thus, the molecule might be expected to recognize the leucine codons UUA or UUG, or both together, but trinucleotide binding studies by Scherberg and Weiss ((1972) Proc. Nat. Acad. Sci. U. S. A. 69, 114) indicate that it recognizes only UUA.


Infection of Escherichia coli wit,h bacteriophage
T4 results in two kinds of change in the tRNh complement of the infected cell. .4t least one of the normal host species, a tRNALeU which recognizes the codon CUG, is degraded by endonucleolytic cleavage after T4 infection (3)(4)(5).
In addition to this inactiration of a host tRNA, T4 itself produces some 7 or 8 new tRN&\ species, including a tRNA,L"" which are transcribed directly from its own genome (s-12).
Polyacrylamide gel electrophoresis of 32P-labeled tRNX estracted from T4-infected cells results in the cheracterist,ic band pattern seen in Fig. 4 (2, 13). Bands 3, 4, and 6 of this pattern have each been found to be a single pure species of tR.NA and therefore can be easily isolated and purified.
Band 3 is T4 * This work was supported by Grant C-4 10984 from the Xational Cancer Institute and in part by a grant from the Cancer Research Coordinating Committee, University of California. Preliminary reports of this work have been presented elsewhere (1,2).

Ezytnes
Spleen phosphodiest,erase, pancreatic ribonuclcase, snake venom phosphodiesterase, and bacterial alkaline phosphatase were obtained from Worthington Biochemical Corporation. Ul ribonuclease (14,15), an enzyme with the same properties as Tl ribonuclease, was donated by C. Dekker. Tl and T2 ribonucleases were obtained from Calbiochem.
Wild type bacteriopha.ge T4D was obtained from S. Brenner.

Low Phosph.ate Growth Media
The low phosphate medium (Medium A) used in the preparation of 32P-labeled bacterial tRN.4 was as described in Landy et al. (24).
Phosphate-free pcptone was prepared by dissolving 20 g of Difco Bacto-peptone in approximately 100 ml of distilled water, and bringing to pH 10 with concentrated NHiOH.
After addition of 1.0 ml of 1 br XgCIZ, the precipitate contaiuiug magnesium-ammonium-phosphate was removed by centrifugation and the supcrnatant was brought to neutral pH with concentrated HCl and autoclaved.
Final volume varied from 120 to 170 ml.

Radioisotopes
Carrier-free l-I,V'04 was obtained from New England Nuclear Corporation.
[W]Leucine uniformly labeled at 312 mCi per mmole was obtained from Amersham/Searle Corporation.
Jluterials for Column Chromatography DEAE-cellulose Whatman DE 32 was prepared as directed by the manufacturer.
BD-cellulose4 was prepared as in Gillam et al. (25). Material for RPCX (26) was a gift of G. D. Novelli and B. Void.

Bu$ers
Standard DEAE-buffer for use with DEAE-cellulose columns was 0.01 M Tris pH 7.4, 0.002 M mercaptoethanol.
NaCl was added as specified.
Standard BD-buffer for use with BD-cellulose columns was 0.01 M sodium acetate pH 5.0, 0.01 M MgC&, 0.002 M mercaptoethanol.
NaCl and ethanol were added as specified. Standard RPC buffer for use with RPC2 columns was 0.01 nr sodium acetate, pH 5.0, 0.01 M h4gC12, 0.001 M EDTA, 0.002 JI mercaptoethanol.
NaCI was added as specified.

Gel Materials
Acrylamide was obtained from Matheson. It was purified by twice repeated recrystallization from ethyl acetate, washed with hexane, and vacuum dried. N, N'-Methylene bisacrylamide and N, N, N', N'-tetramethylethylenediamine were obtained from Eastman Kodak.
In addition, cellulose acetate strips (2.5 x 55 cm) were obtained from Schleicher and Schuell.
Thin layer cellulose plates (20 x 20 cm) were obtained from Eastman Kodak Company.
Test tubes and small sample tubes were acid washecl, rinsed, and baked overnight, and then filled with or immersed in a lyO solution of dichlorodimethylsilane in toluene for 5 s, and baked again for several hours or overnight.
Eastman Kodak RP14 x-ray film (14 X 17 inches) was used for making autoradiographs of 32P-labeled materials fractionated on paper, thiu layer DEAE-cellulose plates, and polyacrylamide gels.

Preparation of l\'ucleic &ids
Bacteriophage T4 DNA-T4D DNA was phenol-extracted as needed, from 5 ml of phage at 2 x lOI per ml according to procedures described by Thomas and Abelson (30).
3?P-Labeled T4 tRYA--il 400.ml culture of CA274 or TY3110 4 The abbreviations used are: BlXcellulosc, benzoglated DEAEcellulose; IWC:, revr:rsed phase chromatography. was grown at 37" in Medium B to an A650 of 0.3. Fifty micrograms per ml of n-tryptophan were added, and the culture was infected with T4 uvz tA3 at a nominal multiplicity of 10 phage per bacterium.
At 2 min after infection, 100 mCi of H332P04 were added.
At 8 min after infection, 50 kg per ml of chloramphenicol were added.
Growth was continued, now at 30", for 90 to 120 min.
The cells were harrestcd by centrifugation and the bacterial pellets mere resuspended in a total volume of 5.0 ml of 0.01 M Tris, pH 7.4, 0.1 M NaCI, 0.002 M mercaptoethanol.
Normally, 80-907, of the radioactivity added to the infected culture was recovered.
The resuspended cells were extracted twice with an equal volume of phenol, and the aqueous layer transferred to a siliconized screw-capped tube. It was brought to 0.2 M NaCl and the nucleic acids precipitated with twice the volume of absolute ethanol.
The ethanol precipitate was chilled in a freezer for at least 30 min.
The ethanol precipitate was pelleted in a clinical centrifuge, and resuspended in 1.0 ml of standard DEAE-buffer without salt. The material was loaded onto a small DEAE-cellulose column in a Pasteur pipette and washed with 20 ml of 0.3 M NaCl DEAE-buffer to remove very low molecular weight material.
The tRNA was then eluted with 1.0 M NaCl DEAE buffer in small (20 to 25 drop) fractions.
The bulk of the 3zPlabeled tRNA eluted in one or two small fractions, and the two or three most radioactive fractions were pooled.
The tRNA, now in 1.0 M NaCl DEAE-buffer, was again ethanol precipitated, pelleted, and deacylated by resuspending in 1.0 or 2.0 ml of 1.0 M Tris, pH 9, 0.002 M mercaptoethanol, and incubating at 37" for 30 min. After another ethanol precipitation, the deacylated tRNA was resuspended in 1.0 ml of 0.3 M NaCl BD-buffer.
It was loaded onto a small BD-cellulose column (about 1 x 5 cm) and washed with another 10 to 12 ml of the same buffer. The 32P-labeled material was then eluted in 2.0.ml fractions.
About 85 to 90% of the radioactivity eluted from the BD-cellulose with 1.0 M NaCl BD-buffer. The remaining 10 to 15% would elute only with 1.0 M NaCl, 10% ethanol BD-buffer.
The most radioactive fractions in the 1.0 M NaCl eluate and in the 1.0 M NaCl, 10% ethanol eluate were pooled separately and referred to as the "BD-cellulose 1.0 M fraction" and the "BD-cellulose hydrophobic fraction." Carrier CA244 tRNA was added to the two BD-cellulose fractions in amounts depending on the next procedure, and they were ethanol precipitated.
For final processing of [3zP]tRNA for sequencing, see below.
32P-LnbeEed E. co& tRNd-32P-labeled bacterial tRNA was prepared by growing E. coli CA274 in a YIedium A in the presence of 0.1 mCi per ml of 32P and allowing growth to continue for several generat,ions.
The cells were harvested by centrifugation and resuspended in 5.0 ml of 0.01 M Tris buffer, pH 7.4. The tRSh was phenol-extracted, purified on DEAE-cellulose, stripped, and fractionated on BD-cellulose as described above for T4 [32P]tRNA.
It was further purified in 200.mg lots by phenol extraction and batchwise DEAEcellulose chromatography followed by standard ethanol precipitation.

Reversed Phase Column Chromatography
To the "BD hydrophobic fraction" of 32P-labeled T4 tRNA, 2.4 mg of carrier tRS*1 were added. The sample was ethanol by guest on March 17, 2020 http://www.jbc.org/ Downloaded from precipitated, pelleted, and resuspended in 3.0 ml of 0.4 M NaCl RPC buffer.
It was then fractionated by elution from an RPC2 column (26) using a lOOO-ml salt gradient, from 0.4 to 1.0 M NaCl RPC buffer. The column (1 x 100 cm) was maintained at 30". Fractions of approximately 8 ml were collected, at a flow rate of 10 min per fraction.
Pooled fractions of radioactive material, diluted to less than 0.3 M NaCl, were reconcentrated by siphoning onto a small DEAE-cellulose column and eluting in small volumes of 1.0 M NaCl, 0.01 M sodium acetate pH 5, 0.002 M mercaptoethanol. sis, BD-cellulose chromatography of acylated and derivatized tRNA--was prepared for sequencing by several cycles of ethanol precipitation from 0.2 M sodium acetate, pH 5, to reduce residual NaCl.
Where necessary, carrier tR?jA was added to at least 20 pg per ml to ensure precipitation.
The final pellet was gently air-dried and finally resuspended in 50 t,o 100 ~1 of distilled HzO.

Sequencing
Polyacrylamide Gel Eleetrophoresis Ten per cent polyacrylamide slab gels (20 x 40 cm) were prepared, and samples of rZP]tRNA prepared and loaded according to the procedures of De Wachter and Fiers (31). Electrophoresis was at 400 volts for 12 to 16 hours. Elect.rophoresis buffer was 0.10 M Tris-acetate pH 8.5. An autoradiograph of the resulting gel was used to make a template to cut out radioactive bands. Material from gel bands was eluted by homogenizing with 0.3 M NaCl, using a total volume of 3.0 ml. Gel material was removed by centrifugation, and the supernatant filtered through a 0.45-p Millipore filter.
In some experiments, first dimensional electrophoresis of total and partial enzymatic digests of T4 tRsALeU were run on 55-cm strips of cellulose acetate obtained from Schleicher and Schuell, instead of on 85.cm strips cut from rolls obtained from 0x0, Ltd.
Enzymic Digestions-U1 ribonuclease which has the same specificity as Tl ribonuclease was used instead of the latter for fingerprints and partial digestions.

Aminoacylation
of tRNA Digestions of oligonucleotides with silkworm nuclease were carried out in 10 ~1 of the enzyme at 125 pg per ml in 0.5 M Na&03-NaHC03, pH 10.5, 0.1 M NaCl, 0.005 M magnesium acetate.
Incubation was at 37" for 60 min. Combined silkworm nuclease-bacterial alkaline phosphatase digestions were performed by including 5 to 10 pg of alkaline phosphatase in the digestion mix. Digestion products in either case were separated by electrophoresis at pH 3.5 on DE-\E-paper. s2P-Labeled bacterial and T4 tRNA were acylated using the E. coli synthetases and an excess of nonradioactive leucine in the same reaction mix as above.
The modified pancreatic enzyme cleaves most frequently at -Cph or -Up-l sequences (20).
Sham acylat.ions of 32P-labeled tRNAs were carried out exactly as above but in the absence of amino acid.
Partial pancreatic ribonuclease digestions were carried out in 0.01 M Tris, pH 7.4, 0.02 M hlgCle, 0.1 mg per ml of bovine serum albumin at 4", usually for 30 min. Xn enzyme to substrate ratio of 1: 500 was used.

Phenoxyacetyl Derivatization of Leucyl-tRNA
The phenoxyacetyl ester of N-hydroxysuccinimide was prepared as described by Gillam et al. (32).
Acylated tRNA was recovered from the above reaction mixes by phenol extraction and ethanol precipitation.
The final pellet wa,s resuspended in 0.5 ml of distilled Hz0 and reacted with the phenoxyacetyl ester as described in Gillam et al. (32). After two ethanol precipitations in the presence of 0.2 M sodium acetate, pH 5.0, the derivatized material was resuspended in 0.3 M NaCl BD-buffer and rechromatographed on BD-cellulose. Elution was with BD-buffers containing 0.3 M NaCl; 1.0 M NaCl; 1.0 M NaCl, 10% ethanol; 1.5 M NaCl, 15y0 ethanol.
Enzyme to subst,rate ratios of l:lOO, 1:50, and 1:20 were used. Incubations were for 30 min at room temperature.

Hybridization
of RNA with DNA RNA-DNA hybridization was carried out on nitrocellulose filters as described by Gillespie and Spiegelman (33).
Acylated and derivatized [32P]tRNA recovered from BD-cellulose chromatography in 1.5 M NaCI. Ethanol BD-buffer (15%) was ethanol precipitated, stripped of amino acids, and finally resuspended in distilled Hz0 for use in hybridization experiments.
Identi$cation of ModiJied Nucleotidesj-Modified nucleotides frequently found in tRNAs were identified essentially as in Sanger et al. (27). T, q', D, and GmG were identifiable by their mobilities at pH 3.5 on both Whatman 52 and DEAE-papers, and T, 'k, and GmG were further identified by paper chromatography in isopropanol-NH3 (27). Nucleotide A*, 2-methylthio-N6-(r, y-dimethglallyl) adenosine, was identified by its electro-5 In referring to standard nucleotides and the usual modified nucleotides found in tRNAs. A. C. G. U. and D. T.\k. Gm are used  (r,r-dimethylallyl) adenosine is also used to refer at times to the nucleotide.
N is used to indicate an unknown but specific nucleotide in the anticodon.
X is used to indicate any nucleoside and is used in notations pXp and Xp. The word "nucleotide" is used in places in the text, tables and figures to mean di-, tri-, and oligonucleotide as well as mononucleotide. azP-labeled T4 tRNA was chromatographed on BD-cellulose, and the 1.0 M NaCl, 10% ethanol "hydrophobic fraction" was acylated with leucine and phenoxyacetyl derivatized as described under "Materials and Methods." A derivatized, sham-acylated sample was also prepared. The two samples were then rechromatographed on BD-cellulose with elution buffers as shown. The shaded peak in the lower leucinecharged sample, eluting in 1.5 M NaCl, 15% ethanol BD-buffer is T4 tRNAneU. The peak still eluting at 1.0 M NaCI, 10% ethanol is T4 tRNAser plus some unacylated T4 tRNALeU. The peak eluting at 1.0 M NaCl is assumed to be nonhydrophobic material incompletely eluted in the original BD-cellulose fractionation, or hydrophobic material altered by experimental manipulation. phoretic mobility at pH 3.5 on Whatman No. 3MM paper, and at pH 3.5 on DEAE-paper (34, 35). All nucleotides were further characterized by running on cellulose thin layer plates developed with isobutyric acid-NH3 or 2-propanol-HCI (46). Some nucleotides were also run in a 1-butanol-formic acid system (36). In addition, GmG was digested with spleen phosphodiesterase and the products obtained were identified as G and Gm on a cellulose thin layer developed with 2-propanoLHC1.
we at first employed the procedures developed by Gillam et al. (32). ["P]tRNA was isolated from T4-infected cells as described in "Materials and Methods." The [32P]tRNA was fractionated on a BD-cellulose column. This gave two fractions: the RNA eluting from the column with 1 M NaCl, and the "hydrophobic fraction" which requires 1 M NaCl plus 10% ethanol for elution. To determine which fraction contains the tRNALeU we acylated each fraction with leucine. A control sample from each fraction was subjected to a sham acylation. These samples were then reacted with the phenoxyacetyl ester of N-hydroxysuccinimide as described by Gillam el al. (32) and rechromatographed on BD-cellulose. Fig. 1 shows the result for the hydrophobic fraction. It can be seen that a substantial portion of this fraction has been made more hydrophobic as a result of acylation with leucine and phenoxyacetyl derivatization. The 1.5 M NaCI, 15% ethanol fraction was designated as T4 tRNALeU. No difference between the leucine-charged and sham control was observed for the fraction which had eluted from the BD-cellulose column with 1 M NaCl.
To characterize the [32P]tRNALeU eluted from the BD-cellulose column with 1.5 M NaCl, 15 ye ethanol, we performed RNA-DNA hybridization experiments and fingerprint analysis. Fig. 2 shows that 70% of the T4 [a2P]tRNALeU can hybridize to T4 DNA whereas E. coli [32P]tRNALeU purified by the same procedure did not hybridize. In experiments not shown here, we found that bulk E. coli tRNA does not compete in the hybridization of T4 [azP]tRNALeU with T4 DNA.
A Ul fingerprint of the T4 tRNALeU, prepared as described above, was similar to the one presented in Fig. SA, but these preparations often contained contaminating oligonucleotides which were later identified as originating from T4 tRNASer. Yields in this acylation and derivatization procedure were low.
To obtain pure preparations of T4 tRNALeu, we resorted to chromatography on the reversed phase system RPC2 (26). T4 [32P]tRNA was chromatographed on BD-cellulose and the "hydrophobic" fraction was further fractionated by RPC2 chromatography as shown in Fig. 3. It can be seen that three peaks, Peaks 1, II, and III, are obtained, all of which elute from the column after most of the E. coli tRNA. Acylation with leucine and derivatization of Peak I resulted in a much larger percentage (about 60%) of the material, now-eluting from BD-cellulose in 1.5 ~\2 NaCl, lSa/, ethanol. Peak I is thus a less contaminated form of T4 tRNALeU than the original "hydrophobic" fraction used in Fig. 1. Some T4 tRNALeU is also found in Peak II. The remainder of Peak II is T4 tRNASer, as is Peak III (see below). Carrier tRNA (2.4 mg) was added to the 1.0 M NaCl, 10% ethanol "hydrophobic fraction" obtained from BD-cellulose chromatography of T4 lazP]tRNA. The sample was chromatographed by gradient elution from an RPC2 column as described under "Materials and Methods." Fractions of about 8 ml were collected at a flow rate of 10 min per fraction. Percentage of transmission at 260 nm was continuously monitored. Radioactivity of evennumbered fractions from 12 through 88 was assayed by counting lo-p1 aliquots in a Nuclear Chicago scintillation counter. --, percentage of transmission; Cl---0, counts per min X 10-a of *zP in lo-p1 aliquots of each fraction (e.g. Fraction 54 in Peak I contained about 2 X lo6 cpm per ml); ---, NaCl gradient.
This result is reminiscent of those obtained by Waters and Novelli (9) who demonstrated two new peaks of tRNALeU in RPC2 chromatography following T4 infection. These peaks eluted at higher salt concentrations than any of the E. coli tRNALeU species.
The sequence analysis to be presented has not given evidence for two distinct sequences, therefore we assume that the two peaks of tRNALeU are caused by heterogeneity in the modification of the tRNA. Such heterogeneity has been observed in E. coli tRNATyr produced in @OdSuIII-infected cells (34).
Some of the sequence analysis was performed with material obtained from RPC2 Peak I, but during the course of the work \\\\ \ it was found that a much simpler purification could be obtained Fm. 4. Polyacrylamide gel electrophoresis of T4 [s2P]tRNA.
by electrophoresis of T4 [32P]tRNA on 10 % polyacrylamide gels. Unfractionated T4 la*P]tRNA purified only through the DEAE- Fig. 4 shows the result obtained when unfractionated T4 cellulose step was electrophoresed on a 10Yc polyacrylamide gel as described under "Materials and Methods." Distance migrated [asP]tRNA is subjected to electrophoresis on a 10% polyacryl-is shown in centimeters on the lejt. Bands of radioactivity are amide gel. The autoradiograph of the gel reveals seven sharp numbered 0 to 6'. BPB indicates the position of migration of the bands. Six of these bands (0, 1, 2, 3, 4, 6) are pure species of bromphenol blue dye marker RNA. Band 5, which has the same mobility as bulk E. coli tRNA, is a mixture of four species of tRNA. All of these RNA species hybridize specifically to T4 DNA (37). Fig. 5 shows an acrylamide gel electrophoresis of T4 [azP]tRNA and of various fractions obtained from BD-cellulose chromatography. The 1 M NaCl fraction is shown in Sample b. The hydrophobic fraction, Sample c, consists only of Bands 3 and 4. The T4 tRNALeU obtained in 1.5 M NaCl, 15% ethanol in Fig. 1, is shown in Sample d. It is mostly band 4 with some contaminating band 3.
RPC2 Peak I (Fig. 3) migrates as Band 4 in polyacrylamide gel electrophoresis. Peak II gives a mixture of Bands 3 and 4, and Peak III migrates as Band 3. Band 3 RNA has been sequenced by M&lain and Barrelll and is T4 tRNASer.
Mutants of T4 can be obtained in which the T4 tRNASer is not made, such as T4 pSulwa and T4 pSubx (38). Such mutants are convenient to use in the preparation of T4 tRNALeu because the "hydrophobic fraction" of their tRNA contains only T4 tRNALeU, and gives only band 4 on polyacrylamide gels as can be seen for T4 pSul-in Fig. 6. Sequence Analysis of TI. tRNALeU the products were separated using the standard two-dimensional system developed by Sanger et al. (27). Fig. 7A shows the resulting fingerprint, while Fig. 7B identifies the oligonucleotides corresponding to each spot.
Analysis of the pancreatic digestion products and their sequence is shown in Table I. All but five nucleotides, ~10, ~12, ~16, p17 and ~18, were sequenced by the combined data from alkaline and Ul ribonuclease digestions. Definitive sequences for ~10, ~16, ~17 and ~18 were established by analysis of 3'dephosphorylated versions of the four nucleotides. Material eluted from an ordinary pancreatic fingerprint was incubated with bacterial alkaline phosphatase to remove the 3'-phosphates. The dephosphorylated oligonucleotides were purified by electrophoresis on DEAE-paper at pH 3.5. Separate aliquots of each were subjected to alkaline digestion which produces 3'nucleotides, and to digestion with snake venom phosphodiesterase which produces 5'-nucleotides. The results in Table I establish the sequences of the four nucleotides as shown.
Nucleotide p12 contains the modified base A*, 2-methylthio-N6-(y,y-dimethylallyl) adenosine (39-41). As indicated in the legend to Fig. 7, nucleotide ~12, AAA*,4\k, occurs in several forms with distinctlv different mobilities. This effect has been Pancreatic Ribonuclease Digestion Products-T4 [32P]tRNALeU observed before by Gefter and Russell (34)) and is due to heterowas digested to completion with pancreatic ribonuclease, and geneity in the modification of A*. It is thought to account for from BD-cellulose chromatography; d, 1.5 M NaCl, 159& ethanol fraction from BD-cellulose rechromatography of leucinecharged and derivatized material (see Fig. 1).
the low yield of A*p in the alkaline digestion products of ~12 shown in Table I, and for the low yield of p12 itself seen in Table  II. The sequence shown in Table I for ~12, AAA*A\k, depends on later results (Table V-D) obtained from the silkworm nuclease digestion of the Ul oligonucleotide u15C which contains ~12. Nucleotide (~5) is subsumed under p4 in Table I and together they are shown as having the variable sequence AAA:.
Both spots usually appear in fingerprints of oligonucleotides produced by pancreatic RNAase digestion as seen in Fig. 7, A and B, but their combined molar yield is only 1.1 (see Table II). There is only one AAA"U" sequence in the entire tRN-4, as shown by pancreatic digestion of Ul nucleotides, and the Ul nucleotide u14 (see Table III) in which it occurs cannot possibly contain 6 Ap residues. We conclude that our preparations of T4 tRNALeU contain some unmodified U in this position.
Experimentally determined molar yields of the pancreatic nucleotides of Fig. 7 and Table I are shown in Table II. As discussed above, the most discrepant result is that shown for p12 for which the experimental yield is very low.
Ul Ribonuclease Digestion Products-A standard two-dimensional fingerprint of the Ul ribonuclease digestion products of T4 tRNALeU is shown in Fig. 8, A and B. Nucleotide u5, the 3'-end of the molecule, CACCA--OH, does not show on the fingerprint, but can be seen in Fig. 8E. Fig. 8, C and D, shows a Ul fingerprint in which the second dimension has been developed by homochromatography on a thin layer DEAE-cellulose plate. Fig. 8E is a Ul fingerprint of a T6 tRNA which is identical with the T4 tRNALeU both in its Ul fingerprint pattern and in the pancreatic digest products of its Ul nucleotides. The relative position of u5 is the same as in T4 fingerprints. Fig. 8F is a standard fingerprint of a combined Ul-bacterial alkaline phosphatase digest. As shown in the schematic diagrams and in Fig. 8F, the only large oligonucleotides resulting from complete Ul ribonuclease digestion of T4 tRNALeU are ~13, ~14, u15A, and ul5C.S Of the four, u15C is the least reproducible. It contains the pancreatic digestion product ~12, AAA*A\k, and also the modified base N, and like p12 it shows much variable mobility and hence low yield. A similar problem arose during the sequencing of the tyrosine suppressor tRNA (41) in which the analogous Tl nucleotide contained AA*Aq. Table III shows the composition and sequence of the Ul ribonuclease digestion products, as far as they could be determined by initial analysis of their alkaline and pancreatic ribonuclease digestion products. It also shows the alkaline and total snake venom phosphodiesterase digestion products of the 3'. dephosphorylated nucleotides.
The quantitative data from alkaline and pancreatic digestion alone were sufficient to establish the results shown in the fourth B, diagram identifying pancreatic nucleotide products. Broken-line circle below p10 is a derivative of ~10. Second broken-line circle below ~10 is a variant form of ~12, AAA*A?zl. Broken line circle near p.20 is also C, probably C cyclic. Products pl, " LJ," include D and q, as well as U.  Fig. 7 were analyzed by standard alkaline and enzymatic digestions. The heading "Nucleotide" of the first column refers to any pancreatic digestion product, either mononucleotide or oligonucleotide.
Dashes between bases indicate phosphodiester bonds. Relative yields were determined by measuring the radioactivity of the paper containing the nucleotide products in a Nuclear Chicago scintillation counter; or, where shown as integers, by visual inspection of intensities of spots on autoradiographs.
b Products were identified by their electrophoretic mobilities at pH 3.5 on Whatman no. 52 paper. Identifications of modified bases Tp, Pp, Dp, and Gmp Gp were verified by isopropanol-ammonia paper chromatography.
c All products other than mononucleotides were analyzed by alkaline digestion.
e Nucleotide pl, more than one spot in most fingerprints, includes qp and Dp.
' Nucleotides p4 and (~5) are shown here as one variably modified product.
See also Table II and text. 0 Base A*p, 2-methylthio-Ne-(r,r-dimethylallyl) adenosine, was identified by its electrophoretic mobility at pH 3.5 on Whatman No. 3MM paper ( RAMP = 1.35) and by its electrophoretic mobility on DEAE-paper at pH 3.5 ( RAMP = 0.78). These values agree with those reported for an authentic sample of A*p by Gefter (35). For the sequence determination of ~12 see text and   In the experiments shown in Table III, however, the two were not resolved and the free U's of ~14 arc therefore shown as "II." As ment.ioned earlier, the D modification is present in variable yield, The final deduced sequence of nucleotide u15C is CACIX-AAA*.4qG, and here the quantitative results from alkaline and pancreatic digestion are quite discrepant.
The alkaline digest results show only 4 Ap residues. The pancreatic digests of u15C, however, clearly and consistently show an A# spot and an AC, indicating the presence of a total of 5 Ap residues.
The pancreatic product AAA*Aq from u15C has been repeatedly shown by alkaline digestion to coiltain 4 Ap residues, and it.s mobili@ at both pH 3.5 and in 7% formic acid on DEAE-paper indicates the same thing.
The consistent appearance of .4&Y and AC together in u15C forces the conclusion t,hn.t, 111X contains 5 A 6335 residues rather than 4. In addition to this, the free U products in u15C are shown as "U" in these data for reasons similar to those for ~14. Alkaline digestion of u15C should generate both free Up and free Ip, but in these experiments those t.wo bases were not resolved.
On the other hand, base N did appear in alkaline digests as a "slow C," that is, a product migrat~ing with about one-third the mobility of Cp. In pancreatic digests of ~1% the Q appeared as part of product A@ but free Up and free Np could not be distinguished.
-4nalyses of the 3'-dephasphorylated Ul nucleotides yielded complete sequences7 for nucleotides ul through ~12. Dephosphorylated versions of u4 and u8 were prepared from the corresponding spots of ordinary Ul fingerprints. Alkaline and snake venom digestions of these two led to the sequences shown. Dephosphorylated nucleotides u10 through u15C were obtained from combined Ul-phosphatase fingerprints as shown in Fig. $6'. Blkaline and snake venom digestions of u10 and ull confirmed the already established sequences. Similar digestions plus paper chromatography of (Tp f \Irp) and p9? establish ~12 as T(*,C)G.
Partial digestion of u12 with U4 ribonucleaae, a 5'-exonuclease with properties similar to spleen phosphodiesterase, yielded the product CpGp which establishes the sequence as mCG.
Alkaline and snake venom digests of ~113 unequivocally established the single Ap residue as the 5'-base of this nucleotide and hence AU as the 5'-dinucleotide.
Quantitative results for the alkaline and snake venom digestion products of nucleotides ~14, u15A, and u15C showed various anomalies which could only be resolved by detailed sequence analysis.
Tables IV and V show the results of further experiments needed to establish the sequences of the four major Ul nucleatides.
Sequence Analysis 0~" ul.!?-Partial digestions with snake venom phosphodiesterase of 3'-dephosphorylated ~13 obtained from fingerprints such as shown in Fig. SF were carried out as described in "Materials and Methods," using an incubation time of 15 min at 37". Separation and analysis of the partial digestion products were carried out as described by Barrel1 (29). The products are listed in Table IV. Combined with the over-all composition seen in Table III, this series of products establishes the sequence of ~13, since only one U and the terminal G remain t,o he added to the sequence of partial product number (5). The complete sequence of u13 must be AUUUCCUUG.
The sequence of ~13 was confirmed by analysis of the products obtained from digestion with silkworm nuclease (16,17). Since its sequence was already known, it served as a model for analysis of the silkworm nuclease digestion products of the other major Ul nucleotides.
Two basic classes of products result from silkworm nuclease digestion of a Ul-phosphatase nucleotide. The enzyme cleaves 3'-phosphodiester bonds leaving 3'-OH and 5'.phosphate products. Possible kinds of products are: type A, which include the 5'.end of the nucleotide, and type B which do not. Type A products from ~13, shown in Table V-A, are ApU, ApUpU,

ApUpUpUpC.
Type B products can be either internal products like pUpUpC, or 3'-end products like pCpUpUpG.
Type A (5'-end) products are both 5'-OH and 3'-OH, and can be analyzed by alkaline and snake venom digestion in the same way as the partial snake venom products of Table IV. Both kinds of type B products (internal and 3'-end) are 5'- 7 We have assumed the sequences -CG and -CCA-OH at this point. ?lX?CG has been sequenced as described immediately below. The sequence of u5 is verified by partial digestion products (see Table VII).  3.4A + l.@C f 3 1 "U"e t 0.8GmG 2.9 A t 1.0 C t i.8 "U"" t 1.1 GmG b All mononucleotides shown as alkaline digestion products are 3'-phosphate. In ~12 the notation (T + \k) indicates one spot containing both Tp and 'kp. In ~14, GmG is the dinucleotide GmpGp.
In the snake venom digestion product of ~12 shown in the fifth column following "SW"; the notation shown means 0.7 FJ*. c All bases shown arc 3'.phosphate.
All products other than mononucleotides were analyzed by alkaline digestion. Snake venom digestion products shown following "SV" are 5'"phosphate.
In other experiments, Gm and G Y-phosphates were resolved and shown to be in equimolar yields.
e The notation "U," as explained in the text, is unresolved Up and Dp, or pU and pD. As also noted in the text, however, pancreatic digests with U and D were resolved on DEAE-cellulose, pH 3.5. f The notation "U," as explained in the text, is unresolved Up and qp, or Up and Np, or pU and p'E. g In alkaline digestions base Np had 0.3 times the Inobility of Cp. In pancreatic 1tNase digestions Np was indistinguishable from Up at pIi 3.5 on l)EAK-paper.
In snake venom digestions, bases with unusual mobilities of 0.4 and 2.9 times the mobility of pC were seen. The latter may have been a dinucleotide involving N. = Prodrlcts identified and yields determined as in Table I. Alkaline digestion yields 3'.mononucleotides for all except the 3'terminal nucleoside of the partial product.
b Products identified and yields determined as in Table I. Snake venom phosphodiesterase digestion yields 5'.mononucleotides for all except the 5'.terminal nucleoside of the partial products.
c Partial products are 3'-OH. d JZ = z/c/l/ where y = distance from origin to the digestion product and z = the distance from the digestion product to the next smaller degradation product as discussed in Brownlee and Sanger (42).
>I does not fall in the ranges for terminal nucleotides as determined by Brownlee and Sanger (C 0.6 to 1.2, A 2.1 to 2.9, U phosphate and 3'-OH.
The 5'-base of a type B product appears as a nucleoside diphosphate pXp after alkaline digestion.
The pXp nucleotides have characteristic mobilities, and identification of a given pXp is verified by the appearance of the 5'-mononucleotide pX after snake venom digestion.
An example is product pUpCpC in Table V-A. Alkaline digestion yields pup plus Cp. Snake venom yields 2 pC residues and 1 pU, which verifies the identification of pup. In ad&ion to ordinary silkworm nuclease digestions, one can also perform combined silkworm-bacterial alkaline phosphatase digestions.
In this case, type A, 5'.end products are identical with type -1 products from silkworm digestion alone. Type B products, however, are now lacking their 5'-phosphates. This facilitatex analysis in various ways. Type il products can be identified a~ those which, from either digestion, never yield a pXp on alkaline digestion.
Also, identical 5'-end sequences from both digestions obviously have identical mobilities.
On the other hand, type I), products from silkworm digestion alone yield pXp nucleoticles on alkaline digestion, whereas the corresponding silkworm-phosphatase products with the same base sequence do not. Iu addit.ion, the two type B products have completely different mobilities.
This makes identification of type B products more cert'ain. Also the alkaline and snake venom digestions of silkworm-phosphatase products reinforce or clarify the inter-1.7 to 1.9, and G 2.6, to 4.4 (42)) for all of the products listed. However, it is obvious from the data presented that the sequence determinations are correct.
Others have observed that the M values for the terminal nucleotides do not always fall within the expected ranges (41).
p Partial products are listed in order of decreasing mobility for each Ul nucleotide.
Numbers followed by periods are arbitrary. f Data shown in brackets do not support the deduced sequence shown.
g Alkaline and snake venom digestion products were not quantitated, but were determined by visual estimation from the autoradiograph film.
pretations of plain silkworm digestion products which involve pXp nucleotides.
Silkworm digestion products from u13 in column "SW" of Table V-A are shown in order, from the 5'-end to the %-end, rather than by mobility.
Silkworm-phosphatase products are in column "SWP." Type A products in both columns had identical mobilities.
Pairs of type U products with the same sequence had different mobilities but are shown on the same line in bot.h columns for clarity.
The deduced sequence AUUUCCUUG verifies that based on partial snake venom products.
Sequence Analysis of ul&-Partial snake venom digest.ions of ~14 were carried out using an incubation time of 60 min at 37". These results (Table IV) establish the sequence of the first six bases in ~14.
Table V-B shows theresultsof silkworm andsilkworm-phosphatase digestions of u14. ,4s with ~13, a series of pairs of Y-end products with identical mobilities could be identified, and sequenced by alkaline and snake venom digestion.
In the 3'-end products it was difficult to ascertain just how many U or D residues precede GmG. The sequence is established firmly, however, by the isolation of (Up, Dp)GmpGp as a product of U2 digestion of ~14. Since U2 cleaves following purines, the final sequence of u14 must be UCAAADDGmG.
Furthermore, pancreatic digestion of the U2-derived oligonucleotide, DDGmG, by guest on March 17, 2020 http://www.jbc.org/ Downloaded from Ul nucleotides from fingerprints as in Fig. 8F were digested with silkworm nuclease and combined silkworm nuclease-bacterial alkaline phosphatase; products were then separated as described under "Materials and Methods." Products were analyzed and sequenced by alkaline and snake venom phosphodiesterase digestion.
A a SW = products of digestion with silkworm nuclease alone. Products are listed in order from 5'-end to 3'-end of the Ul nucleotide, rat'her than by mobility.
b SWP = products of combined silkworm-phosphatase digestion.
The digest and a plain silkworm nuclease digest were run side by side at pH 3.5 on DEAE-paper.
The 5' end products in both columns had identical mobilities.
Other products with the same base sequence in both columns had different mobilities (see text). c Alkaline and/or snake venom digestion products could not be accurately quantitated. d A U2 ribonuclease digestion product with the same base sequence as these products was isolated from 3'-phosphate Ul nucleotide u14 (see text).
e As noted in text, the 3'-end sequence depends on these products. f See text.
followed by separation of the products on DEAE-paper, pH 3.5, showed the ratio of pyrimidines to be 80 $& D to 20 y0 U, indicating that both U nucleotides can be modified to D in any given molecule.
In addition, some pancreatic digests of ~14 followed by separation of the products on DEAE-paper, pH 3.5, showed free U to be in approximately 300/ greater yield than free D, indicating 850/, modification of the pyrimidine adjacent to GmG. When the product AAAD was further analyzed by alkaline digestion and product identification on Whatman No. 52 paper, the ratio of pyrimidines was 75% D and 257, U. Thus it seems that either or both of the U nuclaotides on the 5'-side adjacent to GmG can exist in the modified or unmodified state with the former predominant.
Identification of GmG was accomplished as stated in "Materials and Methods," and the mobilities for GmG and degradation products obtained from it are given for several electrophoretic and chromatographic systems as described in Table VIII. In addition, c" and G 5'-phosphates were obtained when dephosphorylated p14 was digested by snake venom phosphodiesterase and the products separated by electrophoresis on Whatman No. 52 paper at pH 3.5.
Sequence Analysis of ulBA-Partial snake venom digestions of by guest on March 17, 2020 http://www.jbc.org/ Downloaded from u15A were carried out using an incubation time of 30 min at 37". The results are shown in Table IV. Sufficient products were obtained to sequence the first 9 bases of the 5'.end, although the quantitative results from the alkaline digestion of partial product number 5, shown in brackets, do not fit the deduced sequence shown, and the data for partial products 6 and 7 are visual estimations from autoradiograph films. Silkworm and silkworm-phosphatase digestions of u15A, shown in Table V-C, generated a series of 5'-end products which could be recognized as such and sequenced.
Internal products could be fitted to a consistent over-all sequence.
The 3'-end sequence of u15A depends on silkworm-phosphatase products alone. The last silkworm phosphatase product, UpCpG, could be sequenced by alkaline and snake venom digestions.
This permitted the sequencing of the next longest product, CpUpCpG.
Although the 5'-phosphate analog of the latter, pCpUpCpG was found, it could only have been sequenced as pCp(Up, Cp)G without the information from the smaller silkworm phosphatase product. The final sequence of u15A is as shown, UCCCACUUCUCG.
Sequence analysis of ulSC-In addition to problems with variable mobility and yield, nucleotide u15C also proved unamenable to various enzymatic digestions.
As seen in Table IV, no partial snake venom digest products were obtained from u15C in sufficient quantity to analyze, although the nucleotide could be digested under conditions of total snake venom digestion.
Digestion of dephosphorylated u15C with U2 ribonuclease generated the products CA, *pG, (C,U2)A in 0.3 molar yield, and [C,U,N(X)]A.* \kpG gave \kp as the T2 digestion product and a snake venom digestion product of pG. The other U2 oligonucleotides were also analyzed by T2 digestion. Separation of the T2 digestion products of [C, U, N (X)]A by electrophoresis on Whatman No. 52 paper, pH 3.5, gave molar yields of C, U, and A and half-molar yields each of slow C and fast C as well as some unidentified products in low yield with various mobilities. Alkaline hydrolysis of the same U2 oligonucleotide gave the same products but with even smaller yields of the unidentified products. It seems likely that the slow and fast C products are different stages of modification of nucleotide N and that the fast C is the same as nucleotide X obtained in the silkworm phosphatase product &own in Table V-D. Furthermore, the presence of the U2 product, (C,U,)A suggests that N is indeed a modified U. It should be noted that [C,U,N(X)]A ran on DEAE-paper in 7 y0 formic acid with a mobility like that of an oligonucleotide containing only C, U, and A (i.e. only 1 U). It seems plausible that N is basic enough to cause oligonucleotides containing it to run faster on DEAE-paper.
Results from the silkworm and silkworm-phosphatase digestions of u15C were somewhat meager, as seen inTableV-D.
Two 5'-end products were obtained, however, which establish the sequence of the first four bases, CACU.
A silkworm-phosphatase product was obtained for which alkaline digestion yielded A*p, Ap, and \kp, and snake venom digestion gave pA, p@, and pG. This implies the sequence A*pAp@pG and establishes the position of A* within AAA*Aq.
Another silkworm-phosphatase product is shown in Table V-D as XpA.
Here base Xp in the alkaline digestion had the "fast C" mobility observed before in the U2 product [C,U,N(X)]A.
The sequence data on u15C establish the 5'-end as CACU and the 3'-end as AAA*A*G.
The sequence of the U2 product containing N must then be CUNA and the final sequence for ul5C must be CACUNAAA*A*G. 8 X refers to a possible alternate form of modified nucleotide N and has a mobility of a "fast C" as discussed in the text. Relative yields of products were obtained in fingerprints such as those in Fig. 8   Experimentally determined molar yields of the Ul ribonuclease digestion products are shown in Table VI.

6361
The figures are based on a compilation of data from both Ul and Ul-phosphatase experiments.
The yield shown for u15C (0.9) is deceptively high as it was based largely on results from a series of Ul-phosphatase experiments such as in Fig. 8F, in which yields of ul5C were relatively good. Yields of u5 and u7 were based on results from one or two fingerprints only. The experimental molar yields agreed fairly well with those expected on the basis of the finally determined structure.

Partial Enzymatic Digestion Products and Over-all
Sequence of T4 tRNA-" Partial digestions of T4 [32P]tRNALeU, using Ul, pancreatic, and modified e-carboxymethyllysine-41-pancreatic RNases, were carried out and the products fractionated as described under "Materials and Methods." Table VII shows a series of such partial digestion products.
Many other partial products were characterized than those shown in Table VII which shows more than sufficient examples to establish the over-all sequence of the molecule.
In addition, the total digestion products were analyzed by further digestion with the opposite enzyme, or with alkali, or both.
The combined data usually, but not always, led to a definitive sequence for the partial product.
Where a partial product cannot be independently sequenced, however, the results from other partials can be applied to order it. Partial product 606-U12, for instance, is CACUNAAA4*A\kGCUG.
It consists of u15C and ~8. Total Ul digestion yields these two products, which arc completely identifiable.
Inspection will show however that the pancreatic digestion products of order u15C-u8 are identical with those of u8-u15C.
The two Ul nucleotides cannot be ordered.
However, several partial products in Table VII show definitively that u15C is immediately preceded by u4, CACAG. Furthermore, product 605.MPll, AAA*A*GCUGC, shows the 3'-end of u15C linked to CUG.
Therefore the order of BOB-U12 must be as show-n.
Total digestion of 605-MP20 with pancreatic RNase yielded AAAGGC, AC, and AGC, standard products ~10, p19, and p3. They were identifiable by mobility in 7 y0 formic acid. However, the identity of AAAGGC was verified by both alkaline and Ul RNase digestion, AC was verified by alkaline digestion, and AGC was verified by Ul digestion.
Theoretically the three pancreatic nucleotides can be arranged in six possible orders.
Total digestion of 605-Ml'20 with Ul RNase yielded products AAAG, G, CAC4G, and C. Once these were identified by mobility and alkaline or pancreatic digestion, it was clear that BAAG and C are not normal U1 nucleotides from the intact molecule; they must therefore be the terminals of the modified pancreatic partial.
It follows at once that C must be the 3'.terminal since the enzyme cleaves following pyrimidines, and, therefore, AAAG is the 5'-terminal.
This leads to AAAG(G,CACAG)C, which on the basis of the pancreatic digest leads to the sequence as shown.
571.LIP9 is a less trivial example.
At this point information about the 5'.end of u15C from silkworm digestions (Table V-D) must be invoked to arrive at AAXGGChCXGCACnU.
By the time the experiment has reached this stage, not enough radioactivity remains to determine how many 3'.U nucleotides are present.
Since Up and Np are indistinguishable in this digest, the partial product may or may not include N. Hence N is shown in Table VII in  parentheses. Other long Ul and pancreatic partial products such as 381-P12, 263-U30, and 263-U31 could not be completely ordered on the basis of the original data. Information obtained from later partial digests made it possible to establish the sequences as shown.
Partial product 571-MPl, AAADDGmGD, yielded separate pancreatic digestion products in 7% formic acid which a.lkaline digestion showed to be AAAU, AAAD, U, D, and GmGD. Ul digestion yielded one long product plus D. The long Ul product again gave, on pancreatic digestion, the products AAAU, AAAD, U, D, and GmG.
These data again indicate that both the "U" residues preceding GmG in Ul nucleotide u14 can be . . either U or D. This region of the molecule is shown as AAADD-GmG on the ground that this is the fully modified form. Partial products 381-Pll and 605.MP12 confirm the sequence of U5 as CACCA-OH by virtue of the terminal product, CAC, which both yielded upon Ul digestion.
The final sequence is shown both in Table VII and in the familiar cloverleaf configuration in Fig. 9.

Identification and Characterization of Jlodified Nucleotides
Modified nucleotides were identified and further characterized as described under "Materials and Methods." The mobilities of the modified and unmodified nucleotides in the various systems employed are summarized in Table VIII. All modified nucleotides in the molecule with the exception of N are identified and their mobilities in the various systems agree with published values (29). The identity of nucleotide N remains unknown.
It is a basic modification of uridylic acid. DISCUSSION We believe that the data presented above unambiguously prove the nucleotide sequence presented in Table VII and Fig. 9. The uncertainties in the structure all involve the identification of the modified nucleosides. This is always a problem in sequence determinations in which only radiochemically pure tRN-4 is used. It is impossible to obtain spectra of the nucleosides a.nd one must rely on chromatographic and electrophoretic mobilities for identification. This is of no use when the nucleoside has not already been well described.
The problem is compounded when, as in our case, the tRNA is isolated from phage-infected cells in which modification may often be incomplete (35). pancreatic, or Tl digestions of the parent Tl or pancreatic oligonucleotides, except that Gpm was derived from a total spleen phosphodiesterase digestion of dephosphorylated ~14, GpmG was derived from alkaline digestion of dephosphorylated ~14, and pGm and pG were derived from total snake venom digestions of dephosphorylated ~14.
c Mobilities were obtained for nucleotides which were first isolated by electrophoresis on Whatman No. 540 paper after alkaline digestion of the parent Tl or pancreatic oligonucleotide.
An exception to this rule is that the mobility of Gum was determined directly from the spleen phosphodiesterase digestion of GmGp for which the reaction products were spotted onto the cellulose thin layer and run in Solvent B.
d The Ru values for Gp"G, pGm, and pG were determined on Whatman No. 52 paper instead of Whatman No. 540 paper.
0 System A is ascending thin layer cellulose chromatography with isobutyric acid and 0.5 M NHnOH (5:3, v/v) as the solvent. RF for Up = 0.53.
h System D is descending paper chromatography on Whatman No. 1 paper with 2-propanol, water, and concentrated NHaOH (70:X): I, v/v/v) as the solvent,.
The modified nucleotides D, T, \k and GmG were identified by their mobilities in electrophoresis at pH 3.5 and by chromatography in several systems. Their familiar positions in the cloverleaf model support these identifications.
We suggest that the nucleoside A* in the anticodon loop is 2methylthio-N6- (7, y-dimethylallyl) adenosine. A* has the same electrophoretic mobility on both Whatman No. 3MM and DEAE-paper at pH 3.5 as A* in E. coli tyrosine tRNA (34,35,41). It is also found in the same position next to the anticodon. This nucleotide has been detected in the position next to the anticodon of several tRNAs recognizing codons beginning with U. It is this nucleotide alone that is responsible for the hydrophobic nature of tRN ATYz in E. coli (35) and thus presumably of / all tRNAs recognizing codona beginning with U (with the possible exception of tRNAPhe).
T4 tRNALeU recognizes UUA (43) and elutes in the hydrophobic fraction on BD-cellulose. Heterogeneity in the modification of A* in tRNATY' (41) led to problems in the mobility of oligonucleotides containg A*. Similar difficulties have been encountered in this study.
Np is another modified nucleotide in the anticodon loop. The fact that the NpA bond is cleaved by pancreatic RNase indicates that N is a pyrimidine.
In alkaline digestions of ul5CN is present as a nucleotide with one-third the mobility of C at pH 3.5 on Whatman No. 540 paper.
In the separation of pancreatic RNase digestion products on DEAE-paper at pH 3.5, Np must have the same mobility as Up. This and the fact that both CUUA and CUNA are obtained as U2 products from u15C suggests that N is a modified form of U. N is in the "wobble position" of the anticodon. Scherberg and Weiss (43) have shown that T4 tRNALeU recognizes only the codon UUA. This is a violation of the wobble hypothesis (44) that has been observed before in cases where there is a 2-thiouridine derivative in the wobble position of the anticodon (45,46). In E. coli the 2thiouridine derivative that is found is 2-thio-5-methylaminomethyl uridine (36). Dr. J. Carbon has provided us with an authentic sample of 2-thio-5-methylaminomethyl uridylic acid. Np turned out to have different mobilities in several different chromatographic systems than did the standard. Thus we must leave Np as an unknown with the suspicion that it could be a 2-thiouridine derivative of some sort. In their original discovery of T4 tRN*4, Hsu et al. (6) reported that the T4 tRNA could be labeled with 35SO*'. If the above speculation is correct both A* and N contain a thio modification.
In most E. coli tRNAs where there is a U in position 8, it is a 4thiouridine.
T4 tRNALeU has a U in position 8 but none of the experiments we have done would be capable of telling whether or not this U is a 4-thiouridine or not. In Fig. 9 we have compared the T4 tRNALeU sequence with those of E. coli tRNALeU I and II (47,48).
The residues that all three molecules share uniquely have been indicated.
It can be seen that the three molecules are quite dissimilar.
It is unlikely that E. coli contains a tRNALeU which is any closer in sequence to T4 tRNALeU. E. coli tRNA does not compete in DNA-RNA hybridization between T4 tRNALeU and T4 DNA. We have purified the E. coli [32P]tRNA Le" eluting in the hydrophobic fraction on ED-cellulose. This tRN,4 does not hybridize to T4 DNA (Fig. 2), and it has a completely different Ul fingerprint from that of T4 tRNAL"*.
Although quantitative rate measurements have not been performed, it is clear from the work of several groups that T4 tRNA4L"" is acylated by the E. coli leucyl-tRNA synthetase (8,9). Thus all three tRNA sequences which are compared in Fig.  9 are recognized by the same enzyme.
Although these comparisons do not allow us to define a synthetase recognition site uniquely, we note that the dihydrouridine loop is nearly identical in all three molecules.
Four nucleotides in the 3'-5' stem region are also the same. This region has been implicated as important in the recognition of E. co& tyrosine tRNA by its synthetase (49, 50).
It is also of interest that all three molecules are exactly the by guest on March 17, 2020 http://www.jbc.org/ same length (87 nucleotides) . The length of the tRNA molecule seems to be a factor in the recognition of tRNh by the yeast phenylalanyl-tRNA synthetase (51). T4 is closely related to two other bacteriophages, T2 and T6. In Fig. 8fl it can be seen that the Ul fingerprint of T6 band 4 RNA is identical with the T4 tRNALeU Ul fingerprint.
Further analysis of the Ul products of T6 band 4 RNA indicate that this molecule is identical in sequence with T4 tRNALeU. T2 also produces a band 4 RNA species with a sequence closely related to T4 tRNALe", but this conclusion is only supported by the similarity of the two Ul fingerprints.
These data suggest that T2 and T6 also produce a tRNALCU and that the sequences of all three molecules are very closely related, if not identical.
AeknowEedgments-We wish to thank our many colleagues listed in the text for their generous gifts of enzymes and bacterial strains.
We thank W. McClain and B. Barrel1 for communicating their unpublished results to us. The following is an incomplete list of the many people who have helped us in our work and given us valuable suggestions and encouragement: P. Lipke, P. Johnson, A. Blank, E. Smith, D. Wicker, B. Told, D. Winkelman, L. Paddock, and D. Nierlich.
We are especially indebted to Frank Mullet, Ruby Johnson, and the staff of the radiology department at Scripps Memorial Hospital for developing literally hundreds of x-ray films for us. We give special thanks to June bliller for patiently helping us to prepare the manuscript.