Sequence-specific interaction between S1F, a spinach nuclear factor, and a negative cis-element conserved in plastid-related genes.

The nuclear gene rps1 coding for the spinach plastid ribosomal protein CS1 exhibits both a constitutive and leaf-specific expression pattern. In contrast to other chloroplast-related genes like rbcS and cab, the leaf induction of rps1 expression is light-independent. These unique features of rps1 expression provide good models to study the mechanisms regulating plastid development and differentiation in higher plants. We report on the identification of a spinach leaf nuclear factor, designated S1F, interacting with the rps1 promoter. The S1F binding site is conserved in the promoter region of many plastid-related genes, including rbcS, cab, and rpl21. A binding activity similar to S1F was detected in nuclear extract from dark-grown de-differentiated soybean suspension cells. Through site-specific mutagenesis and transient expression in soybean cell protoplasts, we show that the S1F binding site is a negative element down-regulating the promoter activity of rps1. A ligated tetramer of S1F site was able to repress activity of the cauliflower mosaic virus 35 S promoter extending the negative function of the S1F binding site on promoter activity.

The nuclear gene rpsl coding for the spinach plastid ribosomal protein C S 1 exhibits both a constitutive and leaf-specific expression pattern. In contrast to other chloroplast-related genes like rbcS and cab, the leaf induction of rpsl expression is light-independent. These unique features of rpsl expression provide good models to study the mechanisms regulating plastid development and differentiation in higher plants. We report on the identification of a spinach leaf nuclear factor, designated SlF, interacting with the rpsl promoter. The S1F binding site is conserved in the promoter region of many plastid-related genes, including rbcS, cab, and rp121. A binding activity similar to S1F was detected in nuclear extract from dark-grown dedifferentiated soybean suspension cells. Through sitespecific mutagenesis and transient expression in soybean cell protoplasts, we show that the S1F binding site is a negative element down-regulating the promoter activity of rpsl. A ligated tetramer of S1F site was able to repress activity of the cauliflower mosaic virus 35 S promoter extending the negative function of the S1F binding site on promoter activity.
Sequence-specific DNA binding transcription factors play an important role in regulating developmental and differential processes in eukaryotes. Although the mechanism is not fully understood, it seems that these factors mediate their effects by directly or indirectly interacting with the general transcription apparatus (for review, see Refs. 1 and 2). In higher plants, differentiation of plastids depends essentially on plant cell types in which they reside and environmental conditions such as light and nutrition (for review, see Ref. 3). Sequencespecific transcription factors regulating the expression of nuclear genes coding for plastid proteins would play an important role in responding to these internal and external conditions. Recently, a number of cis-acting elements and trans-acting factors have been identified through studies on photosynthesis-related genes like rbcS and cub whose expression appears limited to chloroplast-containing cells (for review, see Ref. 4). There have been a number of reports demonstrating the existence of both positive and negative cis-* The costs of publication of this article were defrayed in part by the payment 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.
T elements in these promoters (5)(6)(7)(8)(9). Most of the studies, however, have been focused on the light responsiveness and/or organ specificity of these elements. Little is known about the regulatory elements mediating expression of plastid-related genes in a constitutive and/or developmental manner.
We have previously shown that the nuclear rpsl gene coding for the spinach plastid ribosomal protein CS1 is constitutively expressed in most tissues and at most development stages (34). The constitutively expressed rpsl is, however, differentially and light-independently up-regulated in tissues such as leaves that contain photosynthetic competent cells (34). Unlike multifamily genes such as rbcS, in which each member has been shown to be expressed differentially with respect to the level of expression and developmental and/or organ specificity (10, 11), the differential expression of rspl in leaves is achieved through employing an additional transcription start site, from which transcription is much more active than from the constitutive one.' This kind of differential regulation through alternative usage of two initiation sites has been observed in another plastid ribosomal protein coding gene rpl21.l The mechanisms regulating this kind of differential expression are unknown.
With the aim of elucidating the cis-acting elements and trans-acting factors that would be involved in the overlapped differential and constitutive regulation of rpsl gene expression, we have searched for the presence of cis-elements and their binding proteins in the upstream region between nucleotides +23 and -400 of spinach rpsl promoter through in vitro DNA-protein interaction and in vivo transient expression assays. In this work we have identified a spinach nuclear factor, designated SlF, binding to the rpsl promoter. The S1F binding site is conserved in the promoter region of many other plastid-related genes, including rpl21. Through transient transfection assays in soybean protoplasts in which a binding activity similar to S1F was detected, we showed that the S1F site acted as a strong negative element in the native rpsl promoter context. Furthermore, a tetramer of the S1F site was able to repress the CaMV2 35 S promoter by about 3 times. The possible function of the S1F site on the transcription of rbcS and cub genes is discussed.

Plasmid Constructions, Site-directed Mutagenesis, and Polymerase
Chain Reactions (PCR)-The plasmid pHE was constructed as follows. The rpsl promoter region from +23 to -400 was generated by

23515
inserted into the EcoRI and SmaI sites of pTZ19U (Bio-Rad). For plasmids pHECAT and pHEM4CAT, the HindIII-BamHI fragment from the plasmid pCaMVCN (Pharmacia LKB Biotechnology Inc.), containing the structural gene of chloramphenicol acetyltransferase and the nos transcription terminator, was ligated into BamHI-and HindIII-digested pHE or pHEM4 (see below). For plasmid p2/3-35SCAT, a ligated tetramer of the synthesized double-stranded oligonucleotide (named S l ) was inserted into the XbaI site upstream to the 35 S promoter of plasmid pCaMVCN.

SEQUENCE 1
For clustered point mutagenesis in the promoter region of rpsl (MI, M4, and M7, see Fig. 2B), the plasmid pHE was converted into single strands, annealed respectively with three oligonucleotides bearing nucleotide substitutions (see Fig. 2B) and synthesized with Klenow enzyme as described (12). The mutants were identified by sequencing (13). Amplification of the region between nucleotides -176 and -257 of wild type and mutant rpsl promoters was done by PCR (13) with the following oligonucleotides as primers.
Preparation of Nuclear Extracts-Nuclear extracts from mature spinach leaves were prepared essentially as described by Green et al. (14). Spinach leaves bought from a local market were washed with cold water and homogenized at 4 "C in buffer I (1 M hexylene glycol, 10 mM PIPES/KOH, pH 7.0, 10 mM MgC12, 0.5% Triton X-100, 5 mM 2-mercaptoethanol, and 0.8 mM PMSF) at 4 "C. Homogenates were filtered through 1000-pm and then 80-pm nylon meshes, and nuclei were pelleted at 3000 X g for 10 min. Pellets were then resqspended with buffer I1 (0.5 M hexylene glycol, 10 mM PIPES/ KOH, pH 7.0, 10 mM MgCl,, 5 mM 2-mercaptoethanol, and 0.8 mM PMSF) plus 0.5% Triton X-100. Nuclei were centrifuged again at 3000 X g for 10 min and washed with buffer I1 without Triton. After centrifugation at 3000 X g for 10 min, nuclei were then lysed with buffer 111 (110 mM KCl, 15 mM HEPES/KOH, pH 7.5,5 mM MgCl,, and 1 mM DTT), and chromatin was precipitated by addition of 0.1 volume of 4 M ammonium sulfate at 4 "C for 30 min and eliminated by centrifugation at 40,000 X g for 90 min. The nuclear proteins in the supernatants were precipitated by addition of ammonium sulfate to a final concentration of 0.3 g/ml. After incubation at 4 "C for 30 min, precipitated proteins were collected by centrifugation at 10,000 X g for 15 min. Pellets were resuspended with a mimimal volume of buffer IV (40 mM KC1,25 mM HEPES/KOH, pH 7.5,O.l mM EDTA, 10% glycerol, and 1 mM DTT) and dialyzed against buffer IV for 3 h with two changes of buffer.
DNase Protection and Gel Shift Assays-For DNase protection (17), the EcoRI-BamHI fragment containing the nucleotide sequence from -400 to +23 relative to cap site +I of the rpsl promoter was labeled with Klenow fragment of DNA polymerase I and with ["' PI dATP for the 3' end of the top strand or with [32P]dGTP to label the lower strand, and incubated with DNase I in the presence or absence of spinach leaf nuclear extracts. The digested products were then electrophoresed on a 5% sequencing gel, together with a G+A scission reaction (18) of the same DNA, as marker. For gel shift analysis, the XbaI-digested DNA fragments from PCR amplification or synthesized double-stranded oligonucleotides SI (boldface letters):

5'CT-243AGAATTGTTACCATGAAACA-224 (corresponding to
the spinach rpsl S1F site) and 3A (boldface letters): GATCC-313AATTTTCATGGATACAATT-zg5 (corresponding to the pea rbcS-SA promoter region), were labeled by [32P]dCTP or ["PI dGTP and Klenow enzyme, incubated with spinach or soybean nuclear extract and electrophoresed on a nondenaturing polyacrylamide gel in 0.5 X TBE, as described (19). In some experiments, unlabeled competitor DNA at 50-fold molar excess was included in the incubation mixture. Transient Transfection of Soybean Protoplasts and Chloramphenicol Acetyltransferase (CAT) Assays-Protoplasts from soybean suspension cells were prepared by incubating cells at exponential growth phase with 0.2% cellulase and 0.02% pectolyase in buffer CPWl3 (16) at 28 "C for 3 h. After filtering through 80-pm nylon mesh, and washing with buffer CPW13, about 1,000,000 protoplasts were transfected with 20 pg of plasmid DNA by the calcium-polyethylene glycol method (20). The transfected protoplasts were incubated at 28 "C in the dark for 24 h and then extracted for CAT assays as described (21).
Southwestern Blot and UV Cross-linking Analysis-As described in Ref. 22, about 20 pg of nuclear protein were separated by 12% SDS-polyacrylamide gel electrophoresis and transferred onto nitrocellulose filter as described (13). The filter was prehybridized in a solution containing 5% nonfat dry milk and 10 mM HEPES, pH 8.0, for 30 min at room temperature and then transferred to binding buffer containing 10 mM HEPES, pH 8.0,50 mM NaC1,lO mM MgC12, 0.1 mM EDTA, 1 mM DTT, 0.25% nonfat milk, and 5' end-labeled DNA (5 X lo5 cpm/ml).
UV cross-linking analyses were done as described (23). The synthetic 22-mer double-stranded oligonucleotide (SI) was end-labeled by Klenow filling. After DNA-protein binding reaction performed as described for gel shift assays, the binding mixture was irradiated with UV light at 254 nm at a distance of 4 cm for 20 min. The cross-linked protein-DNA complexes were separated on a 12% SDS-polyacrylamide gel and revealed by autoradiography.

Identification of Three cis-Elements in the SI Promoter-
We have previously observed that nuclear factors interact with the upstream region from nucleotide +23 to -400 of the spinach rpsl promoter (34). To define protein factors and their DNA binding sequences, we analyzed this upstream DNA fragment by DNase I protection assay. As shown in Fig.   1, three sites of protection by nuclear extracts from mature spinach leaves were observed on both strands (Fig. L4). Examination of these sites revealed three sequences as indicated in Fig. 1B.
Site 2, which is slightly protected from DNase I digestion by nuclear extract, contains a GT-rich sequence ("""GGTGTATGTGAAAA""O). It is not clear at this stage whether this element is related to the previously characterized GT-1 sites, which are probably involved in the light responsiveness of many light-regulated genes (24-26). Interestingly, site 3 (-'"AAAGTTAGTTAAAAGAT-" reverse) is highly similar to one of the binding sites (AAGTTAAAAA) of chalcone synthase silencer binding protein SBF-1, which has been shown to be related to GT-1 (27).
Our attention was turned to site 1, which appears to be a new plant &-element, as judged by its nucleotide sequence: -243AGAATTGTTACCATGAAA-226. To define the nucleotides essential for binding activity, we introduced three sets of site-specific mutations in this element by using oligonucleotide-directed mutagenesis (Fig. 2B). DNA fragments spanning from -176 t o -257 were prepared by PCR amplification from wild type and mutated clones and used as probes in gel shift assays. As shown in Fig. 2 A , when incubated with spinach leaf nuclear extract, the ""P-labeled wild type fragment was nearly completely shifted.
The DNA fragments bearing nucleotide substitutions on both sides of the element (MI and M7) maintained their binding activity as the wild A B type, demonstrating that these nucleotides are not important for binding. In contrast, when nucleotides -233, -234, -235, and -237 were changed (M4), binding activity of the D N A fragment was almost completely lost. At least one of the mutated base pairs must be required for binding activity.

Presence of S l F Site in Other
Plastid-related Genes-Searches in other well documented plant genes allowed us to discover in the promoter region of pea rbcS-3A a sequence homologous to site 1 over a stretch of 18 base pairs with only two nucleotides changed (Fig. 3B). Pea rbcS-SA promoter has been extensively studied before (4). This element had not been, however, identified. To ascertain that site 1 binding factor (SlF) binds indeed to pea rbcS-3A promoter, we performed competition experiments with gel shift assays. In these experiments, we used as probes two 22-mer double-stranded oligonucleotides, one (SI) corresponding to site 1 of rpsl and the other (3A) to the homologous region of pea rbcS-3A. When extracts from leaves were used, a similar shifted band was observed with either probe (Fig. 3A). This shifted band can be competed away by a 50-fold excess of unlabeled probe.
Sequence comparison further revealed that this element is conserved also in the promoter region of other rbcS and cab genes of many dicot plants ( Fig. 3B and Refs. [30][31][32][33]. A highly conserved S1F binding site was found in the upstream region of another plastid ribosomal protein gene, rp121. The consensus sequence deduced from the comparison corresponds well to the core sequence of the S1F site determined in Fig. 2.

Conseruation of a S1F-related Binding Activity in Darkgrown Dedifferentiated Soybean Suspension
Cells-Darkgrown soybean suspension cells were available in our labora-  numbers 1-3. The GA lanes represent chemical cleavage of the same fragment at G and A residues, as size markers. E , the sequence of the rpsl promoter from +23 to -400 is presented, as well as three protected regions (underlined, with numbering as in panel A ) . Two transcription start sites at +1 and -41 are marked by arrows.
between nucleotide -400 and +23 of the spinach rPS1 gene. interesting to know whether these single type plastid-contain- Conservation of S1F site in other plastid-related gene promoters. A, gel shift assays with synthesized doublestranded oligonucleotides. SI represents the double-stranded oligonucleotide to the spinach rpsl S1F site. 3A represents the doublestranded oligonucleotide corresponding to the pea rbcS-3A promoter region. Labeled S1 or 3A were incubated with (+) or without (-) spinach leaf nuclear extracts and electrophoresed. Where indicated, a 50-fold molar excess of unlabeled DNA was added in the incubation mixture for competition. NS represents an unrelated 25-mer doublestranded oligonucleotide. B, comparison of putative S1F sites found in the rbcS, cab, and rp121 genes with the rpsl S1F site. Conserved nucleotides are marked by boldface letters.
ing cells had maintained the S1F-like binding activity. For this purpose, we performed gel shift assays using nuclear extracts from these soybean cells. Incubation of wild type and mutated DNA fragments used in Fig. 2 with soybean nuclear extracts gave rise to a shifted pattern similar to that with spinach leaf extracts (Fig. 4B), implying that S1F activity was preserved in these dedifferentiated cells. The soybean shifted band migrated slightly faster than the spinach leaf band (Fig. 4A). It is not clear whether this difference of mobility was due to differences in size and/or physiological state of the factors from those two materials. These data suggest that S1F is an ubiquitous sequence-specific factor present in different cell types.
SlF Site Down-regulates SI Promoter as Well as CaMV35 S Promoter in Protoplasts of Soybean Suspension Cells-Since S1F site binding activity exists in soybean cells, which are convenient materials for protoplast preparation and transient expression assays, we decided to use these cells to determine the function of the S1F site in the context of the rpsl promoter in vivo. For this purpose, we transfected the soybean protoplasts with, respectively, plasmid constructs containing the bacterial chloramphenicol acetyltransferase gene driven by the upstream region from +23 to -400 of the rpsl promoter or by its mutated version (M4). Activities of these constructs were determined by CAT assays using the protoplast extracts 24 h after transfection. As shown in Fig. 5A, the mutated promoter was 7 times more active than the wild type, demonstrating negative function of the S1F site of rpsl promoter in the soybean protoplasts.
In order to confirm its negative function, we fused a tetramer of S1F site to the upstream region -343 to +8 of the cauliflower mosaic virus 35 S promoter, which acts as a strong constitutive promoter in most organs of plants (28) and in protoplasts of cultured cells. The chimeric and the original (35 S) promoters were used, respectively, to drive CAT gene expression in protoplasts of soybean suspension cell. The activities of both promoters were determined by CAT assays of extracts from transiently transfected protoplasts. As shown in Fig. 5B, the fused S1F sites-35 S promoter was about 3 times less active than the 35 S promoter, indicating that S1F sites were able to repress 35 S promoter in transient expression.
Taken together, these data strongly support the notion that the S1F site is a negative element in soybean suspension cells.
Determination of the Molecular Size of S1F-In the first attempt to characterize SlF, we used Southwestern blotting to determine the molecular mass of S1F. Spinach leaf nuclear proteins were separated by SDS-polyacrylamide gel electro-FIG. 4. Demonstration of S1F binding activity in darkgrown dedifferentiated soybean suspension cell extracts. A, '"P-labeled PCR amplified wild type DNA fragment used as in Fig. 2 was incubated without (-) or with nuclear extracts from spinach leaves ( S L ) or soybean cells (SC) and electrophoresed. €3, gel shift assays with nuclear extracts from soybean cells incubated with the wild type and the mutated probes used in Fig. 2. phoresis and transferred to a nitrocellulose membrane. The membrane was then incubated with the '"P-labeled tetramer of S1F site. As shown in Fig. 6, a band of about 30 kDa was revealed by the radioactive DNA probe. To confirm this observation, UV cross-linking technique was used. After incubation of the leaf nuclear extract with "'P-end-labeled 22base pair synthetic oligonucleotide of the S1F site, the binding complexes were subsequently irradiated with UV light and then separated by SDS-polyacrylamide gel electrophoresis. In these experiments, four radioactive bands were observed in the absence of specific competitors (Fig. 6, lune 3). When 50 times the amount of cold probe was added during the binding reaction, the band migrating at about 37 kDa disappeared (Fig. 6, lane Z ) , indicating that this band was the specific S1F-S1F site complex, and that the others were produced by nonspecific DNA-protein interactions. Taking the DNA weight into account, the size of this band is consistent with the Southwestern result. These data indicated that S1F is a protein of about 30 kDa.

DISCUSSION
Studies on the expression of nuclear genes coding for plastid ribosomal proteins provide good opportunities in searching for nuclear transcription factors involved in biogenesis and development of plastids in higher plants. In this paper we have described the identification of a spinach leaf nuclear factor S1F that binds to the rpsl promoter. The S1F binding site is conserved in the promoter region of many other plastidrelated genes including; rbcS, cub, and rp121. Presence of the S1F binding site in those functionally different promoters and the preservation of the S1F binding activity in the dedifferentiated soybean cells imply that S1F would be an ubiquitous factor involved in regulation of plastid differentiation and development.
Site-specific mutagenesis and transient expression assays have shown that the S1F site down-regulates the promoter activity of rpsl in the protoplasts of dark-grown dedifferentiated soybean suspension cells. This negative function is confirmed by the observation that a tetramer of S1F site, when linked upstream to the region between -343 and +8, represses the strong CaMV 35 S promoter by about 3-fold.
While it may be different in the plant system, the results of transient expression would, however, reflect at least in part the function of the rpsl S1F site in amyloplast-containing cells such as those in roots, in which transcription of rpsl has been shown to occur only at basal level (34). We believe it is likely that the S1F site functions similarly in the promoter context of the photosynthesis-related genes. Specifically, Ueda et al. (9) have shown the existence of a negative element in the upstream region between -496 and -374 of tomato rbcS-3A in mature leaves. Similarly, a negative responsive element localized between -347 and -100 from the cap site of the pea cab gene AB80 has been characterized by its ability to silence the expression of a constitutive promoter in roots (5). The finding of S1F binding sites in the negative elementcontaining region of both tomato rbcS-3A and pea cub promoters (Fig. 3) suggests that S1F would be engaged in negative regulation of both genes. It is conceivable that these differentially functioning negative responsive element could be generated through unique combinations of a common negative cis-element (e.g. S1F site) with other different cis-elements present in those promoters.
In eukaryotes, many negative transcription factors can also function as activators, depending on the nature of the DNA binding site and interaction with other proteins (reviewed in Ref. 29). Specifically, it has been shown that GT-1 binding sites, which are involved in the up-regulation of many lightresponsive genes, could repress constitutive transcription in the dark when fused to CaMV 35 S promoter (6). For SlF, we do not know presently whether it functions differently in regulating rpsl expression in different tissues in which different rpsl promoter binding activities have been detected (34).:' As described previously, differential expression of rpsl and rp121 is regulated through alternative usage of two start sites.
Since the S1F site is present in both promoters, it will be interesting to know whether S1F plays a role in transcription start site selection. These problems can be solved only through experiments with transgenic plants which are now in progress.