Recognition of Novel and Divergent Higher Plant Chloroplast

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 27, Issue of July 8, pp. 18223-18231, 1994 Printed in U.S.A.

Recognition of Novel and Divergent Higher Plant Chloroplast Ribosomal Proteins by Escherichia coli Ribosome during in Vivo Assembly*

(Received for publication, November 3, 1993, and in revised form, January 24, 1994)

Mikhail G. Bubunenko and Map R. SubramanianS From the Mar-Planck-Znstitut fur Molekulare Genetik, Zhnestrasse 73, 14195 Berlin-Dahlem, Germany

Architecture of higher plant chloroplast ribosomes in- volves additional protein domains over that found in the Escherichia coli ribosome, although the rRNAs in these two kinds of ribosomes are very similar in length and sequence (Subramanian, A. R. (1993) -ends Biochem. Sci. 18, 177-180). Here, we show that two chloroplast- specific protein domains (a novel chloroplast ribosomal protein of the 30 S subunit, called Psrp-1 or 522, and a divergent protein of the 50 S subunit with long terminal extensions and low homology to its E. coli counterpart, L21) are both incorporated in E. coli ribosomes and polysomes when their gene constructs are expressed in E. coli. Also, the 67-residue NH,-terminal extension in chloroplast L21 by itself is incorporated. Thus, our results indicate preexisting binding sites for novel chloroplast- specific ribosomal proteins/domains on eubacterial ribosomes. Additionally, we observed cleavage of the chloroplast-targeting transit peptide (present in the expressed Psrp-1 precursor), indicating protease(s) of the required specificity in E. coli cells. The expression of chloroplast L21 with its NH,-terminal extension was inhibitory to E. coli growth, suggesting a drastic effect of the latter on some property of L21. Expression of Psrp-1 was neutral, consistent with a function only in chloroplast translation. Based on analysis of the assembly of Psrp-1 and various L21 fragments in E. coli ribosomes, a general model for studying ribosomal protein-ribosome interactions is suggested.

Chloroplast ribosomes have many similarities to eubacterial ribosomes (1,2), but the most striking difference between them is in their protein moiety (3-6), which, in higher plants, is supplemented by 5-10 additional proteins (6). Five of these have been characterized to date, variously at protein (7-9),’ cDNA (7, 8, lo), and gene (11) levels. In four cases, no signifi- cant homology to any known ribosomal protein (RP)’ or other sequenced proteins is found (7, 8, lo), suggesting their evolution specifically for the chloroplast translational system. In the fifth case, homology to a polypeptide present in several ther- mophilic bacteria (12) was observed (9). All of these novel chloroplast RPs are encoded in the nuclear genome (2).

Over 35 further chloroplast RP sequences, each homologous to a corresponding Escherichia coli RP, have also been reported

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1332; Fax 49 30 8413 1380. $To whom correspondence should be addressed. Tel.: 49 30 8413

A. R. Subramanian, unpublished results. The abbreviations used are: RP, ribosomal protein; NTE, NH,-ter-

minal extension; CTE, COOH-terminal extension; IPTG, isopropyl-l- thio-P-D-galactopyranoside; PCR, polymerase chain reaction.

(reviewed in Refs. 13-16). Many of these RPs contain NH,- and COOH-terminal extensions (NTE, CTE) of considerable length beyond the homologous sequence regions (2, 6). Nothing is known at present about the origin or role in translation of chloroplast-specific RPs and extensions, but it is possible to envision that they help generate the structural and functional uniqueness in chloroplast ribosomes. In contrast, the rRNA of higher plant chloroplast ribosomes is very similar in chain length and primary/secondary structure to eubacterial rRNA (17, 18).

To understand the unique features of chloroplast ribosomes at the biochemical level, we have examined the behavior of two such proteins (and their derivatives) at the assembly stage ofE. coli ribosomes. Psrp-1 is an acidic, relatively high M, protein present as a stoichiometric component in two-dimensional gels of sucrose gradient-purified spinach chloroplast 30 S ribosomal subunits (7). I t has been isolated and sequenced, and its cDNA clones (7,8) and a genomic clone have been characterized (11). The Psrp-1 (designated S22 in Ref. 11) sequence showed no detectable homology to any of the known eubacterial, archae- bacterial, mitochondrial, or cytosolic RPs (7, 8). The second protein we chose was chloroplast L21, which is double the size of its E. coli counterpart and contains a 67-residue NTE and 30-residue CTE (19, 20). Its cDNA clones and a genomic clone have been characterized (19-21). The central 103-residue region of chloroplast L21 is homologous to the entire E. coli L21. The extensions (present only in higher plant chloroplast L21; see Ref. 19) have no homology to any reported protein sequences in data bases.

We have taken the previously cloned (7,19) coding regions of spinach chloroplast Psrp-1 and L21 and have expressed various constructs made from them in E. coli. The presence of the expressed proteins in ribosomal particles was assayed by sucrose gradient fractionation followed by Western blotting/ immunostaining with antibodies specific to chloroplast RPs. Thus, incorporation of these “foreign” proteins by the E. coli ribosome during assembly and utilization in protein biosynthe- sis, i.e. participation in polysome formation, could be deter- mined. The results are presented and discussed in this paper.

MATERIALS AND METHODS Spinach Chloroplast RP cDNAs-cDNAs of spinach (Spinacia olera-

cea, cv. Matador) chloroplast Psrp-1 and L21 (7, 19), cloned in h g t l l as previously described (231, were used for this study.

DNA Manipulations-All DNA manipulations were done according to standard procedures (24, 25). DNA fragments were isolated from agarose gels after electrophoresis using Geneclean I1 (BIO 101, Inc., Vista, CAI. E. coli strains, XL1 and JMllO (Stratagene), and WK6 (26) were used for transformation and expression.

Psrp-1 Constructs for Its Expression-The Psrp-1 cDNA insert in h g t l l recombinant phage was subcloned into pT7PT3-19U (Pharmacia Biotech Inc.) in the EcoRI site and used as the starting material. The AsnI-EcoRI fragment of the insert, containing almost the whole cDNA sequence (without 26 base pairs of 5’-untranslated region and the first

18223

18224 Novel and Divergent Chloroplast Ribosomal Proteins 26 base pairs of the transit peptide region), was cloned into NdeI and EcoRI sites of the thermoinducible E. coli expression vector pJLA502 (Medac, Hamburg, Germany) after end-filling (using Klenow fragment) the AsnI and NdeI ends. It yielded an open reading frame with the AUG codon and ribosome binding site of the vector to produce plasmid pMBSl (see Fig. 1). pMBS5 contains the 3'-terminal portion of the cDNA (NdeI-EcoRI fragment) subcloned into the NdeI and EcoRI sites of pJLA502 (after filling NdeI ends), pMBS13 is an altered form of pMBSl with an introduced termination codon (TAA). It was made by cutting at the internalNco1 site, end-filling, and religation. In pMBS14, the acidic region of Psrp-1 is deleted by cutting out the BclI-BamHI portion of pMBSl (Fig. 1).

L21 Constructs for Expression-The spinach rpl21 constructs were made using appropriate restriction sites or PCR amplification of defined parts of L21 cDNA in the h g t l l clone (19). For PCR amplification, the following deoxyribonucleotide primers were used. The forward primers include: oligo 1, 5'-dTCAACTAACCATGGCGTCTGCAACT at position 39-63 (Ref. 191, which introduces a NcoI site (underlined); oligo 2, 5'-dCACTTCAATTGAATGTTCCACCTCT at position 122-146 (Ref. 19), which introduces a MunI site (underlined); oligo 3, 5"dCTCCCCATG- GCCGCCAAACGAC at position 202-223 (Ref. 19), which introduces a ScoI site. The reverse primers include: oligo 4, 5'-dTGTCCA"TAACAT- GAAGCTG at position 1043-1061 (Ref. 19); oligo 5, 5"dAGGT- GCTAGCWGTTCCCACTAG at position 535-559 (Ref. 19), which introduces a TAA stop codon.

For the PCR reaction to amplify the coding and 3"nontranslated regions, 50 ng of h g t l l DNA, 1 p~ primers (oligos 1 and 4), 200 p~ each of four dNTPs, and 3-4 units of Vent polymerase (New England Biolabs Inc.) in a final volume of 100 pl was subjected to 30 cycles of 90 s at 100 "C, 80 s at 63 "C, and 90 s at 72 "C. The final extension step was for 10 min at 72 "C. The amplified DNA fragment was purified by electrophoresis/Geneclean, restricted with NcoI and AsnI, and cloned into NcoI and BamHI sites of pJLA502 after end-filling the AsnI and BamHI ends (plasmid pMBS19). The L21 cDNA was also cloned in the IPTG-inducible expression vector pTrc99A (Pharmacia); the NcoI-Sal1 fragment from pMBS19 was cut out and cloned into the NcoI-Sal1 site of pTrc99A (pMBS23).

All of the other L21 gene derivatives were also constructed in pTrc99A. The EcoRI-AsnI fragment of the L21 cDNA was cloned into EcoRI-SmaI sites to obtain pMBS18. pMBS26 is a derivative of pMBS18 produced by restriction with EcoRI and NheI and religation after end-filling, which restores the reading frame in the remaining NheI-AsnI portion in the clone. pMBS27 is a derivative of pMBS23 with a TAG stop codon introduced by restriction with XbaI and religation after end-filling. pMBS32 has the same translated region as pMBS27, but the 3' portion downstream of the PstI site (see Fig. 2) is absent; it was made by restricting pMBS32 with PstI and religation. pMBS30 was produced by deleting the NcoI-EcolO9I portion of pMBS23 and religat- ing the ends after filling, followed by additional restriction, end-filling, and religation of the restored NcoI site. The reading frame was restored by the last step. pMBS33 is a derivative of pMBS30, obtained by restriction with XbaI and NheI and religation of their compatible ends. pMBS52 is also a derivative of pMBS30, obtained by restriction with NheI and religation after end-filling to generate the TAG stop codon.

pMBS41 and pMBS42 contain gene fragments amplified from L21 cDNA using either oligo 2 or oligo 3 as the forward primer and the same reverse primer (oligo 5) under conditions described above. The PCR product with oligo 2 was digested with MunI and NheI and cloned into EcoRI-NheI-digested pMBS18 to obtain pMBS41. The PCR product with oligo 3 was restricted with NcoI and NheI and cloned into NcoI- and NheI-digested pMBS18 to obtain pMBS42. L21 cDNA fragments in pMBS43 and pMBS44 were also obtained by PCR amplification with forward primers oligos 2 and 3, respectively, and reverse primer oligo 4. These fragments were cloned into EcoRI-SmaI sites (pMBS43) or NcoI- SmaI sites (pMBS44) of pTrc99A. pMBS48 and pMBS49 are derivatives of pMBS41 and pMBS42, respectively, produced by restriction with XbaI and religation after end-filling to introduce the TAG stop codon.

Expression of Psrp-I and L21 and Their Derivatives in E. coli- Expression of Psrp-1 and its derivatives was performed in E. coli WK6 and that of L21 (and derivatives) in E. coli XL1. Cells were grown in LB broth (24) supplemented with 150 pg/ml ampicillin. Transformation was by the standard CaCl, procedure (24). A single colony of freshly transformed E. coli was inoculated into 5-10 ml of LB broth and grown overnight at 28 "C (WK6) or 37 "C (XL1). Portions of the overnight culture (0.25-0.5 ml) were used to inoculate 25 ml of LB broth, and bacteria were grown at the indicated temperatures. Cell growth was monitored by measuring Ah6, absorption (27). pJLA502-based constructs (pMBS1, -5, -13, -14, and -19) with a thermoinducible promoter

and grown at 28 "C were induced by shifting the culture temperature from 28 to 42 "C at 0.2-0.25 AEWnm. Protein expression from the pTrc99A-based constructs was induced by adding IPTG (0.04%), when bacterial culture reached 0.5-0.6 A,, nm. At indicated times, cultures were taken and immediately centrifuged (Eppendorf) at 4 "C. The cell pellet was suspended in SDS cracking buffer and heated for 10 min at 100 "C. The proteins were analyzed by SDS-gel electrophoresis in 15% acrylamide (281, followed by staining with Coomassie Blue or by immunostaining after Western blotting.

Western Blotting and Immunostaining-After electrophoresis, proteins were blotted onto nitrocellulose membrane (BA85, Schleicher & Schuell) essentially as previously described (29). The membrane was incubated at room temperature for 1-2 h with rabbit antiserum against pool 43 of spinach chloroplast RPs. This pool contains proteins Psrp-1 and L21 as previously documented (7, 19). The antiserum was diluted 1:10,000 with 3% bovine serum albumin, 0.3% Tween 20 in phosphate- buffered saline. Incubation with alkaline phosphatase-conjugated secondary antibodies (goat anti-rabbit IgG, Sigma), diluted 1:7,500 in 0.3% Tween phosphate-buffered saline, was for 1-1.5 h. Color development was done in the dark with 5-bromo-4-chloro-3-indolyl phosphate (0.175 mg/ml) and 4-nitrotetrazolium chloride (0.45 mg/ml) in 100 mM Tris- HC1, pH 9.5, 100 mM NaC1, 50 mM MgC1, for 10-60 min (7, 19). The image of the blots were computer-recorded (Cybertech CS1) and the data processed (WinCam program) to obtain the relative band intensi- ties (peak areas) for estimating the percentage of incorporation in ribosomes.

were prepared as previously described (30). Maize (FR9CMS x RF37, Ribosomes for Immunoscreening-Spinach chloroplast ribosomes

grown in a growth chamber) chloroplast ribosomes were similarly isolated, and maize 80 S ribosomes were from the postmitochondrial supernatant. Synechocystis PCC 6803 was cultured as previously described (29). Chlamydomonas reinhardtii ribosomes were a gift from Drs. N. W. Gillham and J . E. Boynton (Durham, NC).

Polysome Isolation and Analysis-E. coli cells transformed with the appropriate plasmids were grown in 100 ml of LB broth and induced by IPTG or temperature shift at 0.2-0.25 A,,,,,. After 6 h of expression, cells were harvested by centrifugation and lysed by the freeze-thaw lysozyme method (31), and the lysate was used for polysome isolation (10-40% sucrose gradient in 50 m Tris-HC1, pH 7.8,300 mM NH,Cl, 10 mM MgCl,, 5 mM 2-mercaptoethanol, and 200 p~ phenylmethylsulfonyl fluoride; Beckman SW 40 rotor, 21,000 rpm, 10 h, 4 "C). 50 pl of the gradient fractions were loaded on SDS-acrylamide gels and electrophoresed. The separated proteins were analyzed by Western blotting as described.

Effect of L21 and Psrp-1 Expression on E. coli Cell Growth-Single colonies ofE. coli XL1 cells, freshly transformed with pMBS30, -33, -42, -44, or -49 plasmids, were streaked onto LB agar plates supplemented with 100 pg/ml ampicillin (and when required, also 40 pg/ml IF'TG) and grown at 37 "C for 18 h to observe the relative colonial growth. Cells from the same colonies were used to determine growth curves (in LB broth containing ampicillin with or without IPTG added at the begin- ning). After 8 h of growth, aliquots of the cultures were taken and analyzed for the presence of the expressed proteins by Western blotting. Cells transformed with Psrp-1 plasmids were grown at 28 or 42 "C.

RESULTS AND DISCUSSION

Expression of Psrp-1 Protein and Its Deletion Derivatives in E. coli-Three derivatives of Psrp-1 and the complete cDNA encoding the mature protein and transit peptide were expressed from the thermoinducible promoter of pJLA502 (Fig. 1A). Since some of the constructs carried the amber (TAG) stop codon, all were expressed in the E. coli strain WK6, which lacks amber suppressor tRNA (26). In no case was a distinct Coo- massie Blue staining band observed; therefore, all further analyses were performed by immunostaining. Evidently, all four constructs were expressed in E. coli (Fig. 1B) albeit at a level not detected by dye staining.

With pMBSl plasmid, which contains the whole cytoplasmic precursor of Psrp-1, four specifically immunoreacting major bands were observed (Fig. lB). The highest M, band corresponded to the size of Psrp-1 precursor ( M , = 33,752) (71, whereas the lowest of the four bands migrated with about the same mobility of the mature protein ( M , = 26,791). Therefore, the latter must have arisen in E. coli by an approximately

Novel and Divergent Chloroplast Ribosomal Proteins 18225

FIG. 1. Gene constructs of the novel Psrp- 1 chloroplast ribosomal protein Psrp-1 and expression in E. coli. A, schematic diagram of the four constructs described in this work. Top, coding region of the precursor cDNA (Psrp-1) and essential pJLA502 features of the expression vector used (pJLA502), showing its promoter (PR, PJ, transcription termination signal (fd-tt), and ribosome binding (GGAGA) and translation intitiation (ATG) sites. MCS, multiple cloning site. The depicted constructs can express the following: pMBSI, the nearly whole precursor (without the NH,-terminal 8 residues); pMBS13, a polypeptide internally termi- pMBS5 nated (TAA) and thus lacking the acidic and COOH-terminal regions; pMBS5, a polypeptide containing mainly the acidic pMBS 14 and COOH-terminal regions; pMBS14, an internally deleted polypeptide lacking the acidic region. Open bur (pMBS131, unexpressed part of Psrp-1; dotted segment (pMBS14), deleted coding region. See text for further details. B, immunostained Western blot of expressed Psrp-1 derivatives. Expression was induced by temperature shift (32-42 "C). Cells were harvested 5 h after induction, lysed in SDS, and electrophoresed in 15% acrylamide gel. The numbers above electrophoretic lanes indicate pMBS clone numbers. Con- trol lunes, pool 43 (which was used to raise antiserum) and lysate of E. coli containing the vector. Arrow shows the position of mature Psrp-1 in pool 43. An un- expected apparent processing of chloroplast transit peptide by E. coli cytoplasm is revealed in pMBSl lane. M,, prestained protein markers.

pMBS 1

pMBS 13

Asnl fl "r-7 acidic rcslon

transit peptide m a t u r e p r o t e i n

T q A I

-I-

-:'::::::::::::::In"

k D a

- 43 - 29

- 18

- 14

- 6

specific removal of the transit peptide. Processing of the transit peptides of certain chloroplast-targeted proteins by E. coli membrane signal peptidase has been previously reported (32). However, unlike these cases where the processed protein is secreted, Psrp-1 occurs in the cytoplasm (see below). Therefore, its transit peptide is cleaved off by an E. coli cytoplasmic protease. The other two major bands (of intermediate size) found in this experiment probably represent processing intermediates.

Protein expression from pMBS14, which lacks the 56-residue acidic region in Psrp-l(7), was significantly lower; the processing intermediates (if present) were hardly visible (Fig. 1B). With the pMBS5 and pMBS13 encoding the COOH-terminal and NH,-terminal regions, respectively, there is evidence that the higher molecular mass bands (Fig. 1B) are derived from translational read-through at the UGA stop codon (data not shown). With pMBS5 expression product that includes the acidic region, an apparent anomalous electrophoretic migration was noted; the mobility corresponded to a significantly higher M, than the 14,270 of the expressed polypeptide (Fig. 1B). Such anomalous mobility is reminiscent of the acidic RNA-binding domain of E. coli S1 (33).

Incorporation of Psrp-1 in E. coli Ribosomes-The Psrp-1 protein and its derivatives were examined for binding to E. coli ribosomes by isolating ribosomes and polysomes from sucrose gradients (centrifuged in high-salt buffer conditions to avoid non-specific binding). With both the Psrp-1 precursor and its processed forms including the mature protein, a relatively ef-

ficient binding was observed, i.e. 84% of the expressed protein was present in the ribosomal particles (Fig. 2A) . The precursor was present in polysomes (fractions 6-9) as well, while the mature protein was present in monosomes (MBS1, fractions 10-11) and subunits, a typical separation profile being shown in Fig. 4A. The presence of Psrp-1 in fractions containing 50 S subunits may arise from the overlapping of the 70 and 30 S peaks and the dissociation (from ribosomes) in the gradient. The NH,-terminal (pMBS13) and COOH-terminal (pMBS5) parts of Psrp-1 bound less strongly (Fig. 2, C and D), i.e. 55 and 47% of the expressed protein, respectively, as would be expected for a protein whose binding sites are organized by both parts of the molecule interacting with the ribosome.

The interaction of the polypeptide with the deleted central region is of special interest (pMBS14, Fig. 2B). The expressed protein (>go%) was found mainly in 70 S ribosomes and 30 S subunits (fractions 10 and 14). Its preferential affinity for the small ribosomal subunit is in line with the mode of binding of intact Psrp-1. The strong binding of this deletion construct argues that the deleted central region (with its unusual sequence characteristics in the context of the entire Psrp-1 polypeptide chain discussed in Ref. 7) is involved in some other function than interaction with ribosomes.

Effect of Psrp-1 Expression on E. coli Growth-Expression of Psrp-1 or its derivatives did not reveal any effect on E. coli cell growth (data not shown), even though in all cases except that of pMBS14 the chloroplast protein is incorporated in E. coli polysomes, i.e. ribosomes actively participating in protein synthe-

18226 Novel and Divergent Chloroplast Ribosomal Proteins 6 7 8 9 J,Q 11 12 13 14 15 16 17CL

kDa

- 43

M B S l 9

7 8 9 11 12 13 14 15 16 17 CL

- 29

- 18

- 14

MBS14

MBS13

MBS5

43

- - 29

- 18

- 14

- 6

- 29

- 18

- 14

- 6

- 29

- 18

- 14

- 6

40 sucrose gradient

10%

FIG. 2. Incorporation of a novel chloroplast ribosomal protein and its derivatives in E. coli ribosomes and polysomes. The ribosomes and polysomes from E. coli cells expressing Psrp-1 and derivatives were separated in a sucrose gradient, and the collected fractions were analyzed by SDS-gel electrophoresis, blotting, and immunostaining. A, binding of Psrp-1 precursor (pMBS1) and the forms produced in E. coli by transit peptide processing; B , binding of the deleted central (acidic) region derivative (pMBS14); C, binding ofthe derivative lacking central and COOH-terminal regions (pMBS13); D, binding of the derivative containing only the acidic and COOH-terminal parts (pMBS5). The numbers above the lanes are those of the analyzed sucrose gradient fractions. They correspond to the gradient profile shown in Fig. 4 (fraction number of the 70 S peak is underlined). CL, cell lysate.

sis. This would suggest that the function of Psrp-1 is restricted to chloroplasts and that in the E. coli ribosome it binds to a region neutral for the translational process.

Expression of Chloroplast L21 Protein and Its Derivatives in E. coli-A total of 15 clones were constructed (Fig. 3). There was not much expression of the whole L21 precursor (i.e. including the transit peptide, mature protein, and the NH,-/ COOH-terminal extensions; pMBS19 and pMBS23) from either of the two vectors used. Also, several of the L21 deletion constructs appeared to be not expressed at all (pMBS26, -27, -32 and -52; Fig. 3B). This result stimulated us to make the series of additional L21 constructs shown in Fig. 3A (in pTrc99A vector, with IPTG-inducible promoter) to examine possible determinants of such poor expression.

Analysis of expression of all the L21 constructs led to find- ings that can be summarized as follows. All constructs containing the whole transit peptide (pMBS23, -27, and -32) were not expressed; no protein products could be detected (Fig. 3B). Removal of the 5’4erminal half of the transit peptide was sufficient to allow expression (e.g. pMBS43). The expression could be modulated to some extent by its complete removal (pMBS44) as well as by further manipulations within the L21 coding region, e.g. deletion of the 3’ coding and noncoding regions andor insertion of an internal stop codon (Fig. 3, A and B ). The 5’ half of the transit peptide coding region has an inhibitory effect on L21 expression. The 3’ noncoding sequence of L21 mRNA could form (by computer simulation) a stable stem-loop secondary structure with the coding region of the transit peptide, but removal of the 3’ region (e.g. pMBS32) did not relieve inhibition. While the real spatial structure of L21 mRNA is not known, we can only speculate that the 5’ region of the transit peptide might be involved in an alternative stable secondary structure that blocks translation of this mRNA by E. coli ribosomes.

A striking point observed during this study was the apparent high accessibility of L21 protein derivatives to proteolytic degradation (as inferred by the appearance of smearing and lower M, bands), especially those containing the transit peptide or NTE (pMBS18, -19, -43, -44, and -48). Polypeptides encoded by pMBS26 and pMBS52 (two of the three smallest fragments of L21 we made) and whose expression was not detected were probably rapidly degraded. In contrast, the third smallest fragment we studied (pMBS49), the NTE of L21 by itself, was relatively stable. I t represents a rather interesting case and will be discussed later.

Evidently, more experimental work is needed to clarify the problem of the poor expression of Psrp-1, L21, as well as other (34, 35) nuclear encoded chloroplast RPs in E. coli. In the case of L21, it seems likely that the low level of expression is post- transcriptionally mediated and that its main determinants are low translational efficiency of L21 mRNA and high proteolytic accessibility of L21 protein and its derivatives. The problem of low expression is intriguing when one considers the overex- pression of chloroplast RPs that are encoded in the organelle DNA, e.g. maize chloroplast S18 and L23 (34h3 The assembly of both nuclear encoded and organelle-encoded RPs into chloroplast ribosomes proceeds in the organelle where, in principle, they would be equally accessible to organelle proteases. The difference in expression in E. coli of these two groups of genes most likely reflects a basic difference in the translational efficiency of their mRNAs by prokaryotic ribosomes. I t may arise from some yet unknown difference in the structural organiza- tion between eukaryotic vis-&vis organelle (prokaryotic-like) mRNAs. The usage of rare codons in the nuclear and chloroplast RP mRNAs is similar (34). The chloroplast L21 protein and its mRNA could be a particularly useful tool for studying this problem because its gene in a lower land plant (liverwort, Marchantia polymorpha (22)) is chloroplastic, in contrast to the case in spinach and probably other flowering plants where it is nuclear (19, 20). Thus, a direct comparison of two forms of a gene from the two genome compartments, with respect to their expression in E. coli, is possible.

Incorporation of Chloroplast L21 Protein in E. coli Ribosomes as a Model to Study RP-Ribosome InteractionSpinach chloroplast L21 protein has only 30% sequence identity to its E. coli counterpart, but the conserved amino acid residues are distrib- uted in clusters over the entire homologous region (19). Thus, it can be regarded as an extensively mutagenized E. coli L21 with

3M. G. Bubunenko and A. R. Subramanian, unpublished observa- tions.


Flc. 3. Gene constructs of chloroplast ribosomal protein L21 and expression in E. coli. A, schematic diagram of 14 constructs. Top, L21 cDNA and the coding regions for its distinct peptide tracts. Clones pMBS23 and pMBS19 con- tained this DNA. The transcription/ translation signals in pTrc99A vector (cf. pJLA502, Fig. LA) is shown below it. Among the depicted constructs are the following: pMBS44, for expressing the mature L21; pMBS30, for the homologous region including CTE; pMBS27 and pMBS32, for the NTE including the transit peptide; and pMBS49, to express essentially the NTE. The remaining constructs were for expressing various other regions of the L21. Open bur, unexpressed part of L21 coding region; dotted segment, (pMBS33), deleted coding region. See text for further details. MCS, multiple cloning site. B, Western immunoblots showing expressed chloroplast protein L21 and its various derivatives. Expression was induced with IPTG. Cells were harvested 5 h after induction, lysed in SDS, and electrophoresed in 15% acrylamide gel. The numbers above electrophoretic lunes indicate pMBS clone numbers. Controls of pool 43 and cell lysate with pTrc99A are shown a t left. Arrow shows the position of mature chloroplast L21. The expressed protein level was too low to be detected in several cases (pMBS23, -26, -21, -32, and -52). pMBS19 expression was also low. In the experiment shown, a three times greater amount of sample (relative to others) has been loaded.

pMBS23 pMBS I9

pTrc99A

pMBS43

pMBS44

pMBS 18

pMBS30

pMBS26

pMBS33

pMBS27

pMBS32

pMBS41

pMBS48

pMBS42

pMBS49

pMBS52

L21 +I

Ncol (PCR) EcoRI Ecol09I Xbal NhcI Pstl TAA Asnl I

~- - .. y , . , : . , homologous region ( -r E

transit pcptidc m a t u r e p r o t c i n

Ptrc rMCS - I SS rmRTIT? -+ --WAACAGACC IATGJG... --

Munl (PCR) . :. si-^

Ncol (PCR) 'YC$y..y... -. ,-$: ". ./ .

: , I . . ; , I. , ~ "_

L I .. . . w-%?&=+" .

u. - -

m '.'

. .. . w . a "

TAG

1

TAG

TAA 5 +& . "7 .... ~ ., ,. s :

TAG

- 2"- TAA

I . ~ . . . . -, ,... TAG

I

TAG

uxz3-

d 8 s pMBS O L

% 19 23 42 44 18 30 27 32 48 41 43 49 26 33 52

kDa

- 43

- 29

-18

- 14

- 6

18228 Novel and Divergent Chloroplast Ribosomal Proteins

2 4 6 8 10 12 14 16 Fraction number

6 8 lo 12 14 16 17 CI M r

kDa

MBS44 *- - 29

- 18 - 14

- 6

5 6 7 8 9 Fraction number

- 10 11 12 13 14 15 16 17 CI M

kDa

- 29

c M B S 3 0

- 14

- 6

- 29

I ~ M B S I S L I S

- 14

- 6

40% 4 10% 40% 4 sucrose gradient

10% sucrose gradient

FIG. 4. Incorporation of chloroplast L21 protein and its derivatives in E. coli ribosomes and polysomes. The experimental procedure used is described in the legend to Fig. 2. A, sucrose gradient profile showing the separation of ribosomal subunits, ribosomes, and polysomes; B, incorporation of mature L21 protein with its NTE and CTE (pMBS44); C, incorporation of L21 moiety essentially without the NTE (pMBS30); D, incorporation of L21 with a truncated NTE (pMBS18). The numbers above the electrophoresis lunes are those of the sucrose gradient fractions (see the gradient profile). The fraction number of the 70 S peak is underlined. CZ, cell lysate; Mr, prestained protein markers. The bund at -14 kDa (fractions 15-17, C) arises from a nonspecific reaction of antiserum with lysozyme.

structurally and functionally important residues left un- changed. From this point of view and based on the earlier finding that chloroplast RPs are incorporated into E. coli ribosome (34-36), we used L21 protein and its deletion derivatives to develop a model system for studying the minimal structural requirements for RP-ribosome interaction. The absence of im- munocross-reaction between chloroplast L21 and E. coli L21 by the serum we used (Ref. 19 and this study) made it suitable for tests under in vivo physiological conditions.

Several L21 constructs representing the main parts of the protein were examined for incorporation in E. coli ribosome in vivo. The analysis showed that L21 derivatives covering the homologous region (pMBS30) or together with the NTE (pMBS18 and -44) were specifically incorporated in both ribosomes and polysomes (Fig. 4). Approximately 94,64, and 84% of the expressed protein was found in ribosomal particles with pMBS30, pMBS18, and pMBS44, respectively. This roughly corresponds to 4040% of all of the ribosomes in the cell extracts, based on A260nm quantitation of ribosomes and a semi- quantitative immunoassay of the protein. The COOH-terminal part of L21 (pMBS33) is incorporated in the E. coli ribosome (70% of the expressed protein) but not into polysomes (Fig. 5A). However, the NH,-terminal protein fragment (pMBS42) is mainly found in the supernatant fraction of the gradient (only 20% in ribosomes). I t showed some tailing into the subunit region of the gradient, indicating a weak affinity for ribosomes (Fig. 5B) .

A particularly interesting finding was the incorporation into the E. coli ribosome of the chloroplast-specific NTE in L21. This 67-residue peptide has no homology to any known RPs (or to any sequence in data bases). From its length and amino acid sequence, the L21-NTE has the potential to form a new domain

in the structure of chloroplast L21. Expression of L21-NTE (pMBS49) led to the accumulation of a relatively stable product in both cell extracts and in ribosomes (Fig. 50). Its stability to cellular proteases (in contrast to the fate of other fragments of the same size, e.g. pMBS26 and -52) (Fig. 3B) indicates a compact structure, and thus it could be a distinct protein domain. Its incorporation into E. coli ribosome (35% of the expressed protein) can be taken as evidence for a relatively weak but specific ribosomal interaction. This conclusion is further supported by the data with pMBS48 that additionally include half the transit peptide (30% of the expressed protein incorporated in ribosomes) (Fig. 5C).

Additional proteolytic fragments (over those originally found in cell extracts) could be detected on ribosomes with pMBS44 (Fig. 4B), which carries the whole NTE. This could indicate that this extension is exposed on the surface of the ribosome and is readily accessible to proteases in a manner shown earlier with L13 (34). Multiple degradation products were also observed on ribosomes with products of pMBS19 and -43 (data not shown).

The results demonstrate that as low as 30% sequence identity in chloroplast L21 is sufficient for it to compete with E. coli L21 and assemble itself in E. coli ribosome under physiological conditions. The whole protein molecule is involved in this interaction (as expected for a structural RP), the COOH-terminal part apparently being rather important. Since most RPs are involved, with increasing evidence (37, 38), in creation and maintenance of rRNA conformation and ribosomal functional centers, they probably are bound in the ribosome via multiple contact sites. Hence, the principles of their interaction with ribosomes (and functional roles) would be difficult to study by simple site-directed mutagenesis. We suggest that a more


Fraction number Fraction number

FIG. 5. Investigation of the regions of chloroplast L21 responsible for interacting with E. coli ribosome. A, COOH-terminal part of L21 (pMBS33); B, NH,-terminal part of L21 (pMBS42); C, NH,-terminal part of L21, including a segment of transit peptide (pMBS48); D, the NH,-terminal extension of L21 (pMBS49). The numbers above the lunes are of the sucrose gradient fractions; the number of the 70 S peak is underlined. CL, cell lysate, My, prestained protein markers.

pMBS 44

@ 6 8 U 12 14 16 17 CI, 6 8 lo 12 14 16 17 CL M rr - 29

"" 0 MBS33+

MUS48+

40?4.( "10% 40% 4 sucrose graclicnt

lOV0 S I I C ~ U S C gradient

pMBS 44

- 18

I l4 e MUS42

1 6

"1 8

-14

tMus49

"

FIG. 6. Inhibitory effect of chloroplast L21 expression on E. coli cell growth. E. coli XL1 cells were transformed with the different L21 constructs shown, and colonies were streaked on LB plates with or without IF'TG and incubated at 37 "C. Plasmid pMBS44, which inhibited E. coli growth, contains the complete mature L21 with both extensions. Other constructs were the following: pMBS30, homologous region of L21; pMBS33, COOH-terminal part of L21;

A&7- pMBS42 and -49, NH,-terminal parts of L21. They showed no noticeable effect on cell growth.

pMRS30

+IPTG -1PTG

promising approach would be to use heterologous proteins and their engineered fragments (as demonstrated in this work). First, the minimal requirements for binding is roughly esti- mated; then, the fine resolution of binding sites is achieved by further dissection of the fragments in combination with site- directed mutagenesis of conserved regions. This approach would be especially valuable in conjunction with E. coli strains lacking one of the RPs; more than 15 such lacking mutants are available (39).

Effect of Chloroplast L21 Expression on E. coli Growth: Evo- lution of Chloroplast RP Extensions-What are the conse- quences of assembling a chloroplast RP in E. coli ribosome either as an extra component or as substitute for an E. coli homologue? Previously, the functional complementation of rpsL mutation (S12, streptomycin resistance) in E. coli with the chloroplast rpsl2 gene was experimentally shown (36). Since S12 is one of the most conserved RPs, this result was not surprising. I t would be more interesting if a highly divergent protein like L21 could similarly complement, but no L21 mutants of E. coli have been reported (39). However, the observed ability of chloroplast L21 protein to interact with E. coli ribo-

somes in vivo offered at least the possibility to examine the physiological consequence of this interaction.

Five L21 constructs (pMBS30, -33, -42, -44, and -49) were chosen for this experiment. They represent the central homologous core of the protein with or without the extensions, or mainly the NH,- and COOH-terminal extensions. The results of the experiment (Fig. 6) showed that the expression of the full- length mature L21 protein carrying the NTE (pMBS44) se- verely inhibited E. coli growth, whereas the expression of any of the other constructs was almost neutral. This suggests that the presence of the NTE on the L21 molecule alters some basic character of the protein and makes its presence on the E. coli ribosome toxic to the cell.

The function of the L21 protein is not understood a t present. It is not known to be a part of any of the functional centers of the ribosome, but it interacts as an assembly protein (40) with two regions of 23 S rRNA (41). I t participates in building the structure of the functional 50 S subunit (42). Evidently, the presence of the 67-residue extension on the NH, terminus of the higher plant chloroplast L21 protein does not disturb the latter's binding to E. coli ribosomes. A similar observation was

18230 Novel and Divergent Chloroplast Ribosomal Proteins

N E

Psrpl [ L21 c

kDa

.43

- 29

- 18

- 14

- 6

FIG. 7. Screening of chloroplast ribosomes from maize (Zm, Zea nays ) , spinach (So, S. oleracea), and C. reinhardtii (Cr) for the presence of Psrp-1 and L21. The material was electrophoresed in SDS and Western blotted and immunostained with pool 49 antiserum. The positions of Psrp-1 and L21 are indicated. The cross-reacting bund of Psrp-1 in C. reinhurdtii is above the L21 bund; it is barely visible in this photograph.

made earlier with chloroplast L13 carrying a 43-residue NTE (35). This property of chloroplast RPs carrying additional long sequences that interact with E. coli ribosomes could give some insight into their evolution. The role of the long NTEs could be to help create the specific chloroplast ribosomal structure in higher plants, quite similar (l), yet different (2), from that of its prokaryotic ancestor(s). They could have emerged as new structural modules, added to preexisting prokaryotic ancestral structural cores, and coevolved with the latter to adapt the core to the evolving chloroplast. The cases of chloroplast RPs L21 and especially S18 (encoded in the chloroplast DNA), with its variably repeated heptapeptide motif (43), may illustrate this process. Interestingly, chloroplast L21 of M. polymorpha, which is encoded the organelle DNA, does not contain an NTE (as first pointed out in Ref. 7), indicating that most NTEs may have appeared after the evolution of the common ancestor of land plants.

Screening of Ribosomes from Different Organisms for Pres- ence of Psrp-1-A protein specifically cross-reacting with antiserum to Psrp-1 was found in the chloroplast ribosomes of spinach, maize, and C. reinhardtii (Fig. 7). No such protein was detected in cytoplasmic ribosomes (maize), E. coli ribosomes/ cell extracts, or in cell extracts of a cyanobacterium, Synecho- cystis PCC 6803 (data not shown). The oxygenic cyanobacterial group is thought to belong to the ancient endosymbiotic pro- genitors of chloroplasts (44). Thus, it appears that Psrp-1 is a ubiquitous chloroplast-specific RP present in mono- and dicot plants and algae. L21 homologues were detected in chloroplast ribosomes of maize and Chlamydomonas but with significantly low intensity of cross-reaction (Fig. 7).

The absence of Psrp-1 cross-reacting material in plant cytosolic and Synechocystis ribosomes makes it difficult to speculate on the evolutionary origin of this novel protein. As pointed out (2), it could be a ribosomal component that was present in the ancestral prokaryotic chloroplast progenitor (but now maintained only in chloroplasts) or a protein recruited into ribosomes at the early stage of photosynthetic eukaryotic evolution. A discernible identity of parts of the Psrp-1 sequence to certain bacterial open reading frames in data bases has been suggested (i.e. upstream open reading frames of diu gene in Bacillus subtilis and pheA in E. coli; downstream of rpoN in Azotobacter vinelandii, Klebsiella pneumoniae, Pseudomonas

putida, and Rhizobium me l l i l~ t i ) .~ The recently initiated work on sequence analysis of Synechocystis RP genes (29)' and completion of the sequencing projects on eukaryotic cytosolic RPs of yeast and mammals (e.g. Ref. 45) should shed further light on this problem.

Acknowledgment-We thank Klaus von Knoblauch for skillful tech- nical assistance.

REFERENCES 1. Boynton, J. E., Gillham, N. W., and Lambowitz, A. M. (1980) in Ribosomes:

Structure, Function, and Genetics (Chamblis, G., Craven, G. R., Davies, J., Davis, K., Kahan, L., and Nomura, M., eds) pp. 903-950, University Park Press, Baltimore

2. Subramanian, A. R. (1993) ?Fends Biochem. Sci. 18,177-180 3. Capel, M. S., and Bourque, D. P. (1982) J. Biol. Chem. 257,7746-7755 4. D0rne.A. M.. Eneas-Filho, J., Heizmann, P., and Mache. R. (1984) Mol. & Gen.

5. Schmidt, R. J., Myers, A. M., Gillham, N. W., and Boynton, J. E. (1984) Mol. Genet. 193, 129-134

6. Subramanian, A. R., Stahl, D.. and Prombona, A. (1990) in The Molecular Biol. Evol. 1, 317-334

Biology ofPlastids (Bogorad, L., and Vasil, I. K., eds) pp. 191-215, Academic Press, New York

12790-12795

(1990) Mol. & Gen. Genet. 223, 167

10. Gantt, J. S. (1988) Cum Genet. 14, 519-528 Biochem. Mol. Biol. Int. 29, 25-31

11. Bisanz-Seyer, C., and Mache, R. (1992) Plant Mol. Biol. 18,337344 12. Choli, T., Franceschi, F., Yonath, A,, and Wittmann-Liebold, B. (1993) Biol.

13. Sugiura, M., Torazawa, K., and Wakasugi, T. (1990) in The Danslational Chem. Hoppe-Seyler 374,377483

Apparatus of Photosynthetic Organelles (Mache, R., Stutz, E., and Subra- manian, A. R., eds) pp. 59-69, Springer-Verlag, Berlin

14. Bourque, D. P., Elhag, G., Bonham-Smith, P., Thomas, F., McCreery, T., and Glinsman-Gibson, B. (1990) in The Danslational Apparatus of Photosyn- thetic Organelles (Mache, R., Stutz, E., and Subramanian, A. R., eds) pp. 85-93, Springer-Verlag, Berlin

15. Lagrange, T., Carol, P., Bisanz-Sayer, C., and Mache, R. (1990) in The ?Fans- lational Apparatus ofPhotosynthetic Organelles (Mache, R., Stutz, E.. and

16. Subramanian, A. R., Smooker, P. M., and Giese, K. (1990) in The Ribosome: Subramanian, A. R., eds) pp. 107-115, Springer-Verlag, Berlin

Structure, Function, and Evolution (Hill, W. E., Dahlberg,A., Garrett, R.A., Moore, P. B., Schlessinger, D., and Warner J., eds) pp. 655-663, American

17. Kossel, H. (1990) in The Danslational AppQrQtUS of Photosynthetic Organelles Society for Microbiology, Washington, D. C.

Verlag, Berlin (Mache, R., Stutz, E.. and Subramanian, A. R., eds) pp. 1-17, Springer-

18. Gillham, N. W., Hams, E. H., Randolph-Anderson, B. L., Boynton, J. E.,

lational Apparatus of Photosynthetic Organelles (Mache, R., Stutz, E., and Hauser, C. R., McElwain, K. B., and Newman, S. M. (1990) in The ?Fans-

Subramanian, A. R., eds) pp. 127-144, Springer-Verlag, Berlin 19. Smooker, P. M., Kruft, V., and Subramanian, A. R. (1990) J. Biol. Chem. 265,

16699-16703 20. Martin, W., Lagrange, T., Li, Y. F., Bisanz-Seyer, C., and Mache, R. (1990) Cum

Genet. 18,553-556 21. Lagrange, T., Franzetti, B.,Axelos, M., Mache. R., and Lerbs-Mache, S. (1993)

Mol. Cell. Biol. 13, 2614-2622 22. Umezono, IC, and Ozeki, H. (1987) ?Fends Genet. 3,281-287 23. Giese, K., and Subramanian, A. R. (1989) Biochemistry 28,3525-3529 24. Sambrook, J., Fritsch. E. F., and Maniatis, T. (1989) Molecular Cloning: A

Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

25. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. (1991) Current Protocols in Molecular Biology, John Wiley & Sons, Inc., New York

7. Johnson, C. H., KruR, V., and Subramanian, A. R. (1990) J. Biol. Chem. 265,

8. Zhou, D. X., and Mache, R. (1989) Mol. Gen. Genet. 219,204-208; Correction

9. Schmidt, J., Srinivasa, B., Weglohner, W., and Subramanian, A. R. (1993)

26. Zell, R., and Fritz, H. J. (1987) EMBO J. 6,1809-1815 27. Miller, J. H. (1972) Experiments in Molecular Genetics pp. 3136, Cold Spring

28. Laemmli, U. K. (1970) Nature 227,680-685 Harbor Laboratory, Cold Spring Harbor, NY

29. Schmidt, J., Bubunenko, M., and Subramanian, A. R. (1993) J. Bid. Chem.

30. Bartsch, M., Kimura, M., and Subramanian,A. R. (1982) Proc. Natl. Acad. Sci.

31. Subramanian, A. R., and Davis, B. D. (1973) J. Mol. Bid . 74.45-56 32. Meadows, J. W., and Robinson, C. (1991) Plant Mol. Bid. 17,1241-1243 33. Subramanian, A. R. (1983) Prog. Nucleic Acid Res. Mol. Biol. 28, 101-142 34. Weglohner, W., Schmidt, J., Giese, K., and Subramanian, A. R. (1993) The

Trnnslational Apparatus (Nierhaus, K, Franceschi, F., Subramanian, A. R., Erdmann, V., and Wittmann-Liebold, B., eds) pp. 701-712, Plenum Press, New York

268,27447-27457

U. S. A. 79, 6871-6875

35. Giese, K., and Subramanian, A. R. (1991) FEBS Lett. 288, 72-76 36. Liu, X. Q.. Gillham, N. W., and Boynton, J. E. (1989) J. Biol. Chem. 264,

16100-16108

B. Baum, personal communication. M. Sugiura, personal communication.

Novel and Divergent Chloroplast Ribosomal Proteins 18231 37. Stern, S., Powers, T., Changchien, L., and Noller, H. F. (1989) Science 244, 42. Vasiliev, V. D., Serdyuk, I. N., Gudkov, A. T., and Spirin, A. S. (1986) in

38. Brimacombe, R. (1991) Biochimie 73,927-936 Structure, Function, and Genetics of Ribosomes (Hardesty, B., and Kramer,

39. Dabbs, E. R. (1991) Biochimie 73, 639445 G., eds) pp. 128-142, Springer-Verlag, Berlin

40. Nierhaus, K. H. (1991) Biochimie 73, 739-755 41. Osswald, M., Greuer, B., and Brimacombe, R. (1990) Nucleic Acid Res. 18, I . K., eds) pp. 303-330, Academic Press, New York

783-790

43. Weglobner, W., and Subramanian, A. R. (1991) FEBS Lett. 279, 193-197 44. Gray, M. W. (1990) in The Molecular Biology ofk'lastids (Bogorad, L., and Vasil,

67554760 45. Wool, I. G., Chan,Y. L., Gluck,A., and Suzuki, K. (1991)Biochimie 73,861-870

Documents

Recognition of Novel and Divergent Higher Plant Chloroplast