Homology between Chloroplast and Prokaryotic Initiator tRNA NUCLEOTIDE SEQUENCE OF SPINACH CHLOROPLAST METHIONINE INITIATOR tRNA*

The nucleotide sequence of a chloroplast methionine initiator tRNA from spinach has been determined. Al-though from a eukaryotic organism, this tRNA strongly resembles prokaryotic initiator tRNAs. Spinach chloroplast tRNA?' has a much higher sequence homology with prokaryotic initiator tRNAs (81 to 84%) than with eukaryotic initiator tRNAs (64 to 69%). In addition, it possesses the two unique features of prokaryotic initiator tRNAs, lacking a base pair between the 5'-terminal residue and the fifth nucleotide from the 3'-end and containing a T-\k-C-A sequence in loop IV. Also, like prokaryotic initiator tRNAs, the chloroplast tRNAfMe' is 77 nucleotides long and has few modified nucleosides (2'-O-methylguanosine, dihydrouridine, 7-methylgu- anosine, ribothymidine, and pseudouridine). This chloroplast initiator tRNA is strikingly different in sequence homology (55 to 62%), number of residues, and structure from mitochondrial initiator tRNAs. Restriction enzyme mapping techniques have shown that the chloroplast tRNAfM't hybridizes to spinach chloroplast DNA. A set of characteristic chloroplast tRNA features seems to be

The nucleotide sequence of a chloroplast methionine initiator tRNA from spinach has been determined. Although from a eukaryotic organism, this tRNA strongly resembles prokaryotic initiator tRNAs. Spinach chloroplast tRNA?' has a much higher sequence homology with prokaryotic initiator tRNAs (81 t o 84%) than with eukaryotic initiator tRNAs (64 t o 69%). In addition, it possesses the two unique features of prokaryotic initiator tRNAs, lacking a base pair between the 5'-terminal residue and the fifth nucleotide from the 3'-end and containing a T-\k-C-A sequence in loop IV. Also, like prokaryotic initiator tRNAs, the chloroplast tRNAfMe' is 77 nucleotides long and has few modified nucleosides (2'-O-methylguanosine, dihydrouridine, 7-methylguanosine, ribothymidine, and pseudouridine). This chloroplast initiator tRNA is strikingly different in sequence homology (55 t o 62%), number of residues, and structure from mitochondrial initiator tRNAs. Restriction enzyme mapping techniques have shown that the chloroplast tRNAfM't hybridizes to spinach chloroplast DNA.
A set of characteristic chloroplast tRNA features seems to be emerging from a comparison of this tRNAP' and several other chloroplast tRNAs which have been completely or partially sequenced. All have a 2'-O-methylated G-G sequence in the dihydrouridine loop, and the sequence T-9-C-A, as opposed to T-\k-C-G, is predominantly found in loop N. This is the reverse of the situation encountered in the overall non-chloroplast tRNA population. Substantial and well defined differences occur in the initiation processes of protein synthesis in prokaryotes and eukaryotic cytoplasm. In particular, there are specific and characteristic differences between initiator methionine tRNAs from prokaryotic and eukaryotic sources (1-7). Recently, two mitochondrial initiator tRNA sequences were determined and marked differences between them and all other initiator tRNAs were observed (8,9). Since the structures of initiator tRNAs are so indicative of their prokaryotic or eukaryotic origins, a comparison of initiator tRNA sequences from both mitochondria and chloroplasts with those from prokaryotic and eukaryotic cells may aid in the evaluation of the possible functional and/or evolutionary relationships of these organelles to prokaryotes, eukaryotes, or to each other. We have * The costs of publication of this article were defrayed in part by the P a p e n t of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. therefore determined the sequence of a chloroplast initiator methionine tRNA.

EXPERIMENTAL PROCEDURES AND RESULTS
The r3H]methionine (5.3 Ci/mmol) was obtained from New England Nuclear. Calcium folinate (leucovorin) was a gift of Lederle Laboratories, Pearl River, NY. The sources of all other materials have been previously described (10, 11).
Preparation of Spinach Chloroplast tRNA?-Chloroplasts were purified from freshly harvested, depetiolated leaves of commercially field-grown, mature Spinacia oleracea L. var. 424, and total RNA was extracted from these chloroplasts as described (12). Crude spinach chloroplast RNA (256 mg) was purified by DEAE-cellulose column chromatography as described (11). The initiator tRNA was isolated from total chloroplast tRNA by chromatography on two successive RPC-5 columns as described in Fig. 1.
Methionine Acceptance and Formylation Assays-To assay for methionine acceptance, the aminoacylation reactions were done essentially as described (13), except that the reaction volume was 20 pl and r3H]methionine was used. To discriminate between the methionine initiator tRNA, and methionine elongator tRNAs, a modification of the procedure of Schofield (14) was employed. A 2O-pl assay mixture containing 0.1 to 0.75 pg of tRNA, 0.1 M Hepes'/KOH, pH 7.6, 10 mM MgC12, 0.5 mM EDTA, 5 mM ATP, 0.5 mM calcium folinate, 0.57 nmol of r3H]rnethionine, and 0.125 p1 of crude Escherichia coli extract containing aminoacyl-tRNA synthetases and transformylase (13) was incubated at 37°C for 30 min. Then, 300 p1 of 25 mM CuS04 was added and the mixture was incubated for an additional 5 min. to hydrolyze any unformylated methionyl-tRNA while leaving the N-formylmethionyl-tRNAf" intact (15). E. coli tRNAfM", isolated as described (16), and tRNA corresponding to pooled fraction A ( Fig. lA ), were used as controls in this experiment. After precipitation with cold trichloroacetic acid, the acid-insoluble material was collected on Whatman GF/A filters, dried, and counted in Omnifluor/toluene. Sequence Analysis of Spinach Chloroplast tRNA?-The modified nucleosides in spinach chloroplast tRNAfMet were determined using the method of Randerath and Randerath (17) and modified nucleotides were identified by the procedure of Silberklang et al. (18). The methods used to determine the nucleotide sequence of spinach chloroplast tRNAfM" consisted of RNA sequence gels, a modification of the methods of Fractions of 1 ml were collected a t a flow rate of 0.19 ml/min. Fraction 84 was used for the sequence determination and formylation assays. These assays showed that the methionine tRNA in Fraction 84 could be formylated while the methionine tRNA in pooled fraction A of Fig. 1A could not be formylated. Stanley and Vassilenko (19) and "mobility shift" analysis (10, 20, 21). RNA sequence gels, using partial enzymatic reaction conditions previously described (10, l l ) , were run on both 3' and 5'-end labeled spinach chloroplast t R N A p . A short electrophoretic run on a 20% polyacrylamide gel of partial digestion products of [5"""P]tRNAP gave the sequence of

-U-X-G-(D)-A-G-X-U-C-G-Y-A-A-G-G-
long electrophoretic run of partial digestion products of [3" "2P]tRNA?et gave the sequence of residues 9 to 60 as A-G-A-

G-C-A-G-U-U-U-X-G-Y-A-G-X-U-C-G-C-A-A-G-G-X-U-C-A-U-A-A-Y-C-U-U-G-A-G-X-U-Y-A-Y-G-G-G-X-(\k)-C-A-
A-A. A shorter electrophoretic run of the partial digestion products of [3"32P]tRNA~"t gave the sequence of residues 50

to76asA-X-G-G-G-X-X-C-A-A-A-U-X-X-U-G-U-X-U-X-X-
In addition to RNA sequence gels, the procedures of Stanley and Vassilenko (19) and Gupta and Randerath (22), involving analysis of Y-"P-labeled partial formamide-derived fragments, established the identity of all residues in this tRNA except residues 1, 74 to 77, and the Gm at residue 19. Fig. 2 shows the analysis of residues 43 to 71 using these procedures. That position 19 is the locus of the Gm was shown by the fact that the product of a complete ribonuclease T.L digestion of a formamide fragment corresponding to residue 19 had a mobility identical to that of [5'-:''PP]pGmpGp in these solvent systems. This and other [5'-"'P]formamide-derived oligonucle-G-C-A-A-C-C.  above (A,B). Numbers refer to positions in the nucleotide sequence. otides which had modified nucleotides a t their 5'-termini were also digested with nuclease PI, and the resulting [5":"P]nucleoside 5'-monophosphates were fractionated by two-dimensional thin layer chromatography (TLC) on cellulose plates (18). Fig. 2Cshows the autoradiograms of three such analyses.
Hybridization of Spinach Chloroplast tRNArM"-The [5'-"'PItRNA?" was hybridized to restriction fragments of spinach chloroplast DNA which were generated by the enzymes Sal I and Pst I, using methods previously described (11,(23)(24)(25). Spinach chloroplast [5"""P]tRNA?" hybridized to the same Sal I and Pst I fragments that have been shown to contain the genes for two spinach chloroplast methionine tRNAs, Metl and Metn (26). Therefore, the spinach chloroplast tRNA?" must be encoded on the spinach chloroplast genome.

DlSCUSSlON
Initiator methionine tRNAs from prokaryotic and eukaryotic cytoplasm possess specific structural features which clearly distinguish them from each other and from most noninitiator tRNAs( 1-7). All prokaryotic initiator tRNAs sequenced to date lack a base pair between the 5'-terminal nucleotide and the fifth nucleotide from the 3'-end, and contain a T(or U)-\k-C-A(or G) sequence in loop IV, while eukaryotic initiator tRNAs contain a standard base pair AI:U72' and ? The numbering system used here for tRNA nucleotide positions is the standard system of Ref. 1 (Sprinzl et al., compilation of tRNA  sequences). an A-U(or \k)-C-G sequence in loop IV. Except for mycoplasma (76 nucleotides), prokaryotic initiator tRNAs are 77 nucleotides in length, with a nine-member dihydrouridine loop, while eukaryotic initiator tRNAs are 75 nucleotides long, with a seven member dihydrouridine loop. The sequences of two mitochondrial initiator tRNAs, from Neurospora crussa (8) and yeast (9), have been determined and were found to have significant differences from both prokaryotic and eukaryotic initiator tRNAs. The sequence of a methionine initiator tRNA from spinach chloroplasts has now been determined and is shown in Fig. 3. This chloroplast tRNApt shows a striking similarity to prokaryotic initiator tRNAs. It possesses both of the structural features which distinguish prokaryotic initiator tRNAs as a class from eukaryotic initiator tRNAs. Thus, it lacks the base pair between the 5'-terminal nucleotide and the fifth nucleotide from the 3'-end, and contains the T-9-C-A sequence in loop IV. Moreover, it has a high sequence homology with prokaryotic initiator tRNAs (81 to 84%), which is a much greater homology than it has with eukaryotic initiator tRNAs (64 to 69%), or mitochondrial initiator tRNAs (55 to 62%).
In addition to the sequence homology and structural similarities discussed above, as well as similar correlations such as at the AI,:U25, Dzi, and US, residues (numbered as in Fig. 3), which are all found in prokaryotic, but not in eukaryotic, initiator tRNAs, the spinach chloroplast tRNA? is similar to a typical prokaryotic tRNA,"'" in size (77 nucleotides) and in the number and nature of its modified bases.
At least 24 methionine initiator tRNAs have now been sequenced (1-6, 8, 9, 27) from a wide variety of prokaryotes, eukaryotes, and organelles. A comparison of these sequences shows that there is a set of "initiator-specific" nucleotides, ie. nucleotides which are located at the same site in all of these initiator tRNAs, but that are not invariant or semi-invariant nucleotides normally found in tRNAs. These residues, without considering post-transcriptional modifications or the anticodon, are, using the numbering system of Fig. 3, as follows: C3, Gu, Cz4, G3], G32, G O , and G71. In addition to these general features, there is one site at which organelle initiator tRNAs differ from non-organelle initiator tRNAs. Thus, non-organelle initiator tRNAs have a G30:C42 base pair, while organelle initiator tRNAs have other nucleotides in these positions. Since so many initiator tRNAs from widely varying sources have been sequenced, reasonable confidence in the generality of these nucleotide specificities appears justified. With the determination of the sequence of this chloroplast tRNA?* and its comparison with other chloroplast tRNAs (11, 27-30, and spinach chloroplast elongator methionine tRNA3 and two other partial sequences of spinach chloroplast tRNAs3), several features characteristic of chloroplast tRNAs appear to be emerging. These include a methylated G-G sequence in the dihydrouridine loop and a predominance of T-\k-C-A as compared to T-\k-C-G in loop IV, which is the reverse of the relative abundances of these sequences observed in non-chloroplast tRNAs. As more tRNA sequences from chloroplasts are determined, it will be interesting to determine how general these "chloroplast-specic" tRNA characteristics are.
After this work was completed, the sequence of a bean (Phuseolus uulguris) chloroplast tRNAf" was determined (27). This tRNA is 92% homologous to the spinach chloroplast tRNA? reported here, and is in excellent agreement with the spinach chloroplast tRNA? in terms of its homology with prokaryotic (as opposed to eukaryotic) initiator tRNAs. Bean chloroplast tRNA,"'" also conforms to all of the characteristic "chloroplast-specic" as well as "initiator-specific" tRNA features described above.