Plant Molecular Biology41: 815–824, 1999.© 1999Kluwer Academic Publishers. Printed in the Netherlands.

815

The Arabidopsis thaliana RER1gene family: its potential role in theendoplasmic reticulum localization of membrane proteins

Ken Sato?, Takashi Ueda? and Akihiko Nakano∗Molecular Membrane Biology Laboratory, RIKEN, Wako, Saitama 351-0198, Japan (∗author for correspondence);?these authors contributed equally to this work

Received 4 January 1999; accepted in revised form 15 October 1999

Key words:ER membrane protein, Golgi apparatus, retrieval

Abstract

Many endoplasmic reticulum (ER) proteins are known to be localized to the ER by a mechanism called retrieval,which returns the molecules that are exported from the ER to the Golgi apparatus back to the ER. Signals arerequired to be recognized by this retrieval system. In the work on yeastSaccharomyces cerevisiae, we havedemonstrated that transmembrane domains of a subset of ER membrane proteins including Sec12p, Sec71p andSec63p contain novel ER retrieval signals. For the retrieval of these proteins, a Golgi membrane protein, Rer1p,is essential (Satoet al., Mol. Biol. Cell 6 (1995) 1459–1477; Proc. Natl. Acad. Sci. USA 94 (1997) 9693–9698).To address the role of Rer1p in higher eukaryotes, we searched for homologues of yeastRER1from Arabidopsisthaliana. We identified three cDNAs encodingArabidopsiscounterparts of Rer1p with an amino acid sequenceidentity of 39–46% to yeast Rer1p and namedAtRER1A, AtRER1B, andAtRER1C1. AtRer1Ap and AtRer1Bp arehomologous to each other (85% identity), whereas AtRer1C1p is less similar to AtRer1Ap and AtRer1Bp (about50%). Genomic DNA gel blot analysis indicates that there are several otherAtRER1-related genes, implying thatArabidopsis RER1constitutes a large gene family. The expression of these threeAtRER1genes is ubiquitous invarious tissues but is significantly higher in roots, floral buds and a suspension culture in which secretory activityis probably high. All the threeAtRER1cDNAs complement the yeastrer1 mutant and remedy the defect of Sec12pmislocalization. However, the degree of complementation differs among the three with that ofAtRER1C1beingthe lowest, again suggesting a divergent role of AtRer1C1p.

Introduction

All eukaryotic cells contain a variety of subcellularcompartments with distinct functions. Maintenance oftheir complex structures requires that newly synthe-sized proteins should be correctly delivered to theirdestinations and localized there by specific sortingmechanisms. In the secretory pathway, nascent pro-teins are first targeted to the endoplasmic reticulum(ER) and then transported via the Golgi apparatusto the plasma membrane; each step of such interor-ganellar transport is mediated by vesicles. The ER is

The nucleotide sequence data reported will appear in the EMBL,GenBank and DDBJ Nucleotide Sequence Databases under the ac-cession numbers AB010945 (AtRER1A), AB010946 (AtRER1B) andAB018552 (AtRER1C1).

essential for synthesis, processing and sorting of pro-teins and lipids. To ensure that the ER functions befulfilled properly, ER proteins must be distinguishedfrom other proteins and strictly localized to the ER.To date, several signals have been identified for thelocalization of proteins in the ER. Many ER lume-nal proteins have conserved motifs at their C-termini:KDEL in mammalian cells (Munro and Pelham, 1987;Pelham, 1989) and HDEL in yeast (Pelhamet al.,1988). The receptor for these signals, Erd2p, is presentin the Golgi apparatus and functions to retrieve themistransported proteins and send them back to theER (Lewis and Pelham, 1990; Lewiset al., 1990;Semenzaet al., 1990). Some ER membrane proteinscarry a dilysine motif (KKXX) at their cytoplasmicC-terminal tails, which also acts as a retrieval signal

816

(Letourneuret al., 1994; Cossonet al., 1996). It hasbeen shown that the dilysine motif binds to coatomer(or COPI), which is a complex of coat proteins thatassemble on the transport vesicles formed from theGolgi apparatus (Cosson and Letourneur, 1994). Mu-tations in coatomer subunits cause mislocalization ofproteins carrying the KKXX signal (Letourneuret al.,1994; Cossonet al., 1996). Thus it is believed thatthe proteins harboring the KKXX signal are retrievedfrom the Golgi to the ER via the COPI-coated vesi-cles. We have recently identified the third type of ERlocalization signal in Sec12p and several other ERmembrane proteins (Satoet al., 1996, 1997).

Sec12p of the yeastSaccharomyces cerevisiaeis atype-II ER membrane protein (Nakanoet al., 1988),which functions as a guanine-nucleotide exchangefactor toward a small GTPase Sar1p and stimulatesbudding of transport vesicles from the ER (Nakanoand Muramatsu, 1989; Barlowe and Schekman, 1993).Sec12p does not possess either HDEL or KKXX ERlocalization signals (Nakanoet al., 1988). To eluci-date how Sec12p is localized to the ER, we isolatedand characterized two recessive mutantsrer1 andrer2(RERstands for retrieval to the ER or retention in theER), which have a defect in the proper localizationof Sec12p (Nishikawa and Nakano, 1993). TheRER1gene codes for a Golgi membrane protein (Rer1p)with four transmembrane domains (TMDs) (Boehmet al., 1994; Satoet al., 1995). By extensive biochem-ical experiments, we have demonstrated that Rer1p isinvolved in the retrieval of Sec12p to the ER (Satoet al., 1995, 1996). Rer1p is required for the cor-rect localization of not only Sec12p but also severalother membrane proteins including Sed4p, Sec63p andSec71p (Satoet al., 1996, 1997). Because the retrievalof these proteins depends on COPI, the role of Rer1pis probably to package the mislocalized ER membraneproteins like Sec12p into the retrograde-directed COPIvesicles. We have further shown, at least for the casesof Sec12p (Satoet al., 1996) and Sec71p (K. Sato, M.Sato, and A. Nakano, unpublished), that the Rer1p-dependent ER retrieval signal exists in their TMDs,even though they have opposite topologies.

In higher plants, various types of ER membranedomains have been morphologically defined, implyingthat more complex mechanisms are necessary for ERprotein localization (Staehelin, 1997). The KDEL orsimilar signals have been found at the C-termini ofseveral plant ER lumenal proteins. Reporter proteinscarrying a H/K/RDEL tetrapeptide at their C-terminiare efficiently localized to the ER (Pelham, 1989; De-

neckeet al., 1992; Vitaleet al., 1993; Haseloffet al.,1997; Staehelin, 1997). AnArabidopsis ERD2ho-mologue has also been identified (Leeet al., 1993;Bar-Peledet al., 1995; Boevinket al., 1998), whichcomplements the lethal phenotype of the yeasterd2deletion mutant. These studies suggest that plants em-ploy a similar system as yeast and animals for thelocalization of ER lumenal proteins. However, mech-anisms for the ER localization of membrane proteinsin plants remain largely unknown. A functional ho-mologue of Sec12p was identified fromArabidopsisthaliana(d’Enfertet al., 1992) and has been shown toreside in the ER (Bar-Peled and Raikhel, 1997), butthe mechanism of its localization has not been studiedyet. No plant ER proteins with the KKXX signal havebeen reported.

As a first step toward understanding the molecu-lar mechanisms for ER membrane protein localizationin plants, we decided to identify and characterizeRer1p, a key player in the retrieval of ER membraneproteins, fromArabidopsis. We found that there areat least three cDNAs encoding Rer1p homologuesin A. thalianaand namedAtRER1A, AtRER1B, andAtRER1C1. With a functional assay to assess the re-trieval of Sec12p, which we developed in yeast, weshow in this paper that all the three cDNAs comple-ment the yeastrer1 mutant, although the degree ofcomplementation differs among these three.

Materials and methods

Plasmid construction

Four cDNA clones of A. thaliana (L.) Heynh,81H10T7, 143C4T7, 136H19T7 and 190G22T7,were kindly provided from the Arabidopsis Bi-ological Resource Center of Ohio State Univer-sity. DNA nucleotide sequences of all cDNAswere determined by the dideoxy method with aDNA sequencer (Model 373A; Applied Biosys-tems, Japan). The open reading frames (ORFs) in81H10T7, 143C4T7 and 136H19T7 were synthesizedby PCR with the following pairs of primers: ATRA1,5′-GAAGATCTATGGACGAAAGCGGAGGT-3′ andATRA2, 5′-CGGGATCCTCAATCCGCACGTGAGCC-3′; ATRB1, 5′-GAAGATCTATGGAAGGAAGCGGTGGTGAC-3′ and ATRB2, 5′-CGGGATCCTCAATCAGCACGAGAGCC-3′; and ATRC-B1, 5′-GAAGATCTATGGAATCCGCAGCAACGGCTG-3′ andATRC-B2, 5′-GAAGATCTTCATTCACTGCTCTCT

817

GTTGG-3′. The DNA fragments were digested withBamHI/BglII or BglII only and placed at theBglIIsite between theTDH3 (GAP: glyceraldehyde-3-phosphate dehydrogenase) promoter and theCMK1(calmodulin kinase) terminator on a single-copy plas-mid pTU1 with the URA3 marker (Uedaet al.,1996). For construction of HA-tagged versions ofAtRer1 proteins, the DNA fragment encoding a hu-man influenza hemagglutinin (HA) epitope [MYPY-DVPDYA] was inserted into theBglII site of pTU1,and the resulting plasmid was named pSKY6. TheScRER1 or AtRER1ORFs were placed just after theHA fragment of pSKY6. The HA-tagged proteinswere detected by immunoblotting with the 16B12mouse monoclonal anti-HA antibody (Berkeley Anti-body Company, Richmond, CA).

DNA manipulations including restriction enzymedigestions, ligations, plasmid isolation, andEs-cherichia colitransformation were carried out by stan-dard methods. Yeast transformation was performedby a lithium thiocyanate method (Keszenman-Pereyraand Hieda, 1988). DNA fragments were purified fromagarose gel pieces with the DNA PREP Kit (AsahiGlass Co., Tokyo).

Yeast strain and culture conditions

The S. cerevisiae rer1-2mutant cells (SKY5-1D:MATα rer1-2 mfα1::ADE2 mfα2::LEU2 bar1::HIS3ura3 leu2 trp1 his3 lys2 ade2) harboring a plasmidpYO324/SEC12-MFα1 were used for the complemen-tation test ofAtRER1(Satoet al., 1995). The cellswere grown in MCD (0.67% yeast nitrogen base with-out amino acids (Difco), 0.5% casamino acids and 2%glucose) medium supplemented appropriately.

Halo assay

The halo assay was performed with a tester yeaststrain BC180 (MATa sst2-2 ura3 leu2 his3 ade2),which is supersensitive toα-factor as described before(Nishikawa and Nakano, 1993). Briefly, the cells ofBC180/pRS316 (ca. 5× 105/plate) were spread onMCD plates buffered at pH 3.5 (Chan and Otte, 1982).Cells to be tested forα-factor production were spot-ted with sterile toothpicks onto these plates and thenincubated at 30◦C for 48 h. Four independent trans-formants were examined for each of the HA-taggedAtRer1 proteins to quantify the amount ofα-factorsecreted.

DNA gel blot analysis

Genomic DNA was prepared fromA. thaliana, theColumbia ecotype, and digested withHindIII andXhoI. Digested DNA was electrophoresed in a 0.7%agarose gel. DNA was denatured and transferred ontoa nylon membrane filter (Hybond-N+, Amersham) ac-cording to the manufacturer’s instructions. Full-lengthORF sequences ofAtRER1A, AtRER1BandAtRER1Cwere used as probes. All probes were labeled with32P-dCTP by using a Random Primer DNA Label-ing Kit (Takara Shuzo, Tokyo, Japan). The filter washybridized at 42◦C for 16 h as described previously(Uedaet al., 1996). The filter was washed twice with2× SSPE (1× SSPE is 0.18 M NaCl, 0.01 M sodiumphosphate, 1 mM Na2EDTA, pH 7.7) and 0.1% SDSat room temperature for 5 min, and twice at 42◦C for30 min (low-stringency conditions). After exposure toan imaging plate for an appropriate duration, the samefilter was washed twice with 0.1× SSPE and 0.1%SDS at 65◦C for 30 min (high-stringency conditions).For visualization of autoradiograms, an imaging platescanner BAS1000 (Fuji Film Co., Tokyo, Japan) wasused.

RNA gel blot analysis

Total RNA was isolated from a suspension culture,roots, rosettes, stems, young siliques, and floral budsof A. thaliana(ecotype Columbia) by a phenol-SDSmethod (Palmiter, 1974). The same probes as in theDNA gel blot analysis were used. The procedures ofelectrophoresis and hybridization were as describedpreviously (Uedaet al., 1996). The filter was washedtwice with 2× SSPE at room temperature for 5 min,twice with 2× SSPE and 0.1% SDS at 65◦C for30 min and twice with 0.1× SSPE and 0.1% SDS at65 ◦C for 30 min. Autoradiograms were visualizedwith the imaging plate scanner BAS1000.

Results

Identification of theA. thaliana RER1gene family

Animal genes that are highly homologous to yeastRER1were recently identified (from man (Füllekruget al., 1997), mouse and nematode), indicating thatthe ER-retrieval system involving Rer1p has been wellconserved during evolution. In the expressed sequencetag (EST) database ofA. thaliana, we found that two

818

Figure 1. A. Amino acid sequence comparison of threeArabidopsisRer1p homologues, AtRer1A, AtRer1B, and AtRer1C1, and yeast Rer1p(accession number D28552). The similar and identical amino acid residues conserved among the four proteins are shaded and boxed, re-spectively. B. The putative splicing acceptor site ofAtRER1C. The arrows indicate the splicing positions that could give rise to two cDNAs,AtRER1C1(C1) andAtRER1C2(C2). An asterisk indicates the Gln-200 residue of AtRer1C1p. C. Molecular phylogenetic analysis of theRer1 family. In addition to the amino acid sequences of theArabidopsisand budding yeast Rer1 proteins, Rer1p homologues ofHomo sapiens(AJ001421),Caenorhabditis elegans(P52879) andSchizosaccharomyces pombe(Z70043) are included. The clustering was performed with theprogram CLUSTAL of Gene Works 2.5.1 (IntelliGenetics) (Higginset al., 1992).

819

cDNA clones, 81H10T7 and 143C4T7, contain se-quences homologous to yeastRER1. To understandhow Rer1p functions in higher plants, we decidedto study these cDNAs and obtained them from theArabidopsis Biological Resource Center. Sequenceanalysis revealed that these clones contained distinctcDNAs. The 81H10T7 clone contained an ORF en-coding a protein of 191 amino acid residues, whereasthe 143C4T7 clone harbored another ORF which en-codes a protein of 195 amino acid residues (Fig-ure 1A). These cDNAs were namedAtRER1AandAtRER1B, and their complete nucleotide and aminoacid sequences have been registered in the DDBJ,EMBL and GenBank databases.

We further realized that the thirdA. thalianahomo-logue of RER1was present on chromosome II fromthe sequence information available from the genomeproject. In the database, thisRER1homologue wasregistered to be encoding a protein of 211 aminoacid residues (GenBank AC002391: locus 2642434).With primers designed from this genomic sequence,we tried to amplify the ORF from the cDNA li-brary by PCR, but never succeeded. By looking intothe cDNA database again, we discovered that cDNAclones 136H19T7 and 190G22T7 were likely to con-tain the same sequence as the above genomic locus.Determination of the nucleotide sequence of the clone136H19T7 revealed a new ORF (AtRER1C1) encodinga protein of 212 amino acids, which is nearly identicalto the predicted protein from the genomic sequenceexcept for its carboxyl-terminal sequence (Figure 1A).This difference can be explained by the fact that thegenomic sequence of this gene has one more potentialintron. Because the C-terminal part encoded by thecDNA sequence can be well aligned with yeast Rer1p,AtRer1Ap and AtRer1Bp, we believe that this is thetrue ORF of this gene and named itAtRER1C1.

On the other hand, the cDNA clone 190G22T7contained a DNA fragment encoding a C-terminal halfdomain, which is identical to AtRer1C1p except thatit lacks the Gln-200 residue. The 3′-noncoding se-quences of these two cDNAs were completely thesame as that of the genomic locus 2642434, suggest-ing that they were derived from alternative splicing. Infact, the putative splicing acceptor site in the genomicsequence contains ...AGCAGTAT, in which either oftwo AGs may act as the acceptor signal (Figure 1B).If the downstream AG was utilized, one glutamineresidue would be omitted, which exactly correspondsto Gln-200 of AtRer1C1p. We tentatively named theincomplete cDNA 190G22T7AtRER1C2.

Figure 2. Hydropathy profile of the predicted AtRer1 proteins. Thealgorithm of Kyte and Doolittle (1982) was used with a windowof 11 amino acid residues. Hydrophobic portions are above thehorizontal line.

The gene products ofAtRER1A, B, andC1 show46%, 44%, and 39% identity with the yeast Rer1protein, respectively. AtRer1Ap is closely relatedto AtRer1Bp (85%), while AtRer1C1p shares rel-atively low homology with AtRer1Ap (53%) andAtRer1Bp (51%). The result of phylogenic analysisof theseAtRER1cDNAs together with otherRER1homologues in animals and fungi is shown in Fig-ure 1C. Hydropathic analysis (Kyte and Doolittle,1982) predicts four membrane-spanning domains ineach AtRer1 protein (Figure 2), suggesting that theyare integral membrane proteins like yeast Rer1p (Satoet al., 1995).

Characterization of theAtRER1gene family inArabidopsis

DNA gel blot analysis was performed on theAtRER1genes. Genomic DNA ofA. thaliana (ecotypeColumbia) was digested withHindIII and XhoI, sepa-rated on an agarose gel, transferred to a nylon mem-brane, and hybridized with the full-length ORFs ofAtRER1. NeitherAtRER1Anor AtRER1BcDNA con-tain any restriction sites forHindIII or XhoI in theirORF, andAtRER1Ccontains oneHindIII site.

As shown in Figure 3A, the full-lengthAtRER1AORF probe hybridized with many fragments under

820

Figure 3. DNA gel blot analysis of theAtRER1genes. GenomicDNA was prepared fromA. thaliana (Columbia ecotype), and di-gested withHindIII and XhoI. Digested DNA was subjected togel-blot hybridization. Full-length ORF sequences of threeAtRER1genes were used as probes. A.AtRER1A, low-stringency con-ditions. B. AtRER1A, high-stringency conditions. C.AtRER1B,high-stringency conditions. D.AtRER1C, high-stringency condi-tions. Numbers at the left indicate molecular length markers inkilobases. H,HindIII; X, XhoI.

low-stringency conditions. With a high-stringencywash, allAtRER1genes hybridized with only one ortwo bands (Figure 3B, C, D). Some bands detectedunder the low-stringency wash did not correspondto any of those after the high-stringency wash, sug-gesting that there are several otherAtRER1-relatedgenes in theA. thalianagenome, i.e. they constitutea large gene family. Database searching also revealsthat mammals have manyRER1homologues. Consid-ering the fact that the yeastS. cerevisiaehas only oneRER1gene, it is obvious that higher eukaryotes havedeveloped divergence of this gene family.

Next, we examined the expression ofAtRER1genes in various tissues ofA. thaliana. As shown inFigure 4, all the threeAtRER1genes are ubiquitouslyexpressed in theArabidopsistissues we examined.The expression ofAtRER1Ais high in roots and flo-ral buds, lower in the suspension culture, rosettes,stems and seedlings, and even lower in cotyledonsand siliques. The expression pattern ofAtRER1BandAtRER1Cis similar to that ofAtRER1A, except thattheir transcription is higher in the suspension culture.

Figure 4. Expression ofAtRER1homologues inArabidopsistis-sues. Total RNA was extracted from a suspension culture, roots,rosettes, stems, young siliques, and floral buds ofA. thaliana(Columbia ecotype) and subjected to gel blot hybridization with thesame probes as used in Figure 3.

AtRER1cDNAs complement the defect of the yeastrer1mutant in correct ER localization of yeast Sec12protein

In order to assess the function of theseAtRER1genes,we examined whether their cDNAs can complementthe yeastrer1 mutant. In our previous studies, we de-veloped a convenient assay to detect mislocalizationof an ER membrane protein, Sec12p, in yeast cells,and utilized it for isolation ofrer mutants and theircharacterization (Nishikawa and Nakano, 1993; Satoet al., 1995, 1996, 1997). The basic scheme of thisassay is summarized in Figure 5. The fusion proteinof Sec12p with a precursor of yeastα-factor (Sec12-Mfα1p) is used as the marker of this assay.α-factoris secreted fromα-mating type yeast cells and trig-gers mating witha-mating type cells. When the newlysynthesizedα-factor precursor (Mfα1p) is transportedfrom the ER to thetrans compartment of the Golgiapparatus, it is processed by three kinds of proteases,Kex2p, Kex1p and Ste13p, and the maturated peptidepheromone (α-factor) is secreted. In wild-type cells,Sec12-Mfα1p is recycling between the ER and thecis-Golgi (black arrows). If Sec12-Mfα1p is mislocal-ized to thetrans-Golgi compartment in therer mutant,three proteases act on the Mfα1p moiety (Fulleret al.1988) and produce the matureα-factor, which is thensecreted to the medium (gray arrows). The amountof the matureα-factor produced can be quantitatively

821

Figure 5. A cartoon to explain the biosynthetic pathway of theSec12-Mfα1 fusion protein in the wild type and therer1 mutantyeast cells. Sec12-Mfα1p is a fusion protein of Sec12p and a pre-cursor of yeastα-mating factor and is used as a marker of Sec12plocalization. If this fusion protein is mislocalized to thetrans-Golgiregion in the mutant, the Kex2 protease acts on the Mfα1p moi-ety and produces the matureα-factor. The secretion of the matureα-factor from the mutant cells is quantitated by the halo assay. Thelower three arrows indicate the route that Sec12p never encountersin the wild-type cells.

measured by the halo assay on agar plates. The ma-ture α-factor arrests the growth ofMATa yeast cells(a mutant supersensitive toα-factor) at the G1 phase,and thus causes the formation of a halo (growth inhi-bition zone). As shown in Figure 6, therer1-2 mutantexpressing Sec12-Mfα1p secretes a large amount ofmatureα-factor and makes a large halo (see Vec inFigure 6). This halo+ phenotype of therer1-2 mutantis completely complemented by the wild-type yeastRER1gene (RER1).

To test complementation of the yeastrer1 mu-tant by theArabidopsis RER1cDNAs, the ORFs ofAtRER1clones were inserted between the yeastTDH3promoter and theCMK1 terminator, and introducedinto therer1-2mutant expressing the Sec12-Mfα1p ona 2α-based multicopy plasmid. The results of the haloassay are shown in Figure 6. AllArabidopsis RER1cDNAs could more or less complement therer1-2mu-tant, indicating that theArabidopsisRer1 proteins canat least partially replace the function of yeast Rer1p.We also constructed HA-epitope-tagged versions ofthe AtRer1 proteins. The HA epitope is composed of9 amino acid residues from human influenza hemag-glutinin and HA-tagged proteins can be analyzed withthe 16B12 mouse monoclonal anti-HA antibody. HA-tagged AtRer1Ap, AtRer1Bp and AtRer1C1p were

Figure 6. Complementation test of the yeastrer1-2 mutant withArabidopsis thaliana RER1homologues. Therer1-2 mutant strain(SKY5-1D) expressing the Sec12p-Mfα1p fusion protein was trans-formed with a vector only (Vec) or plasmids harboring yeastRER1(RER1), AtRER1A, AtRER1B, or AtRER1C1. The transformantswere examined forα-factor secretion by the halo assay. The plateswere incubated at 30◦C for two days.

Figure 7. Immunoblotting analysis of the HA-tagged AtRer1 pro-teins in therer1 mutant. Therer1-2 mutant (SKY5-1D) harboringpSKY6 (lane 1), pSKY6/RER1(lane 2), pSKY6/AtRER1A(lane 3),pSKY6/AtRER1B(lane 4) or pSKY6/AtRER1C1(lane 5) weregrown at 30◦C in MCD medium. Extracts were prepared by ag-itation with glass beads in the SDS gel sampling buffer. Proteins(50 µg) were separated on an SDS gel and the HA-tagged pro-teins were detected by immunoblotting with the 16B12 monoclonalantibody.

well expressed in therer1-2 mutant and detected asbands of 22.8, 22.5, and 27.5 kDa respectively (Fig-ure 7). These HA-tagged AtRer1 proteins showedsimilar complementation activities to the untaggedproteins. We quantitated the halo size of four indepen-dent transformants for each of the HA-tagged AtRer1proteins. The average amounts of secretedα-factorfrom therer1-2mutant expressingHA-AtRER1A, HA-AtRER1Bor HA-AtRER1C1on a single-copy plasmidwere calculated to be 34.9±3.9%, 20.9±3.5%, and60.5±7.1% of the vector control (100%), respectively,while theHA-ScRER1could complement completely(<9±0.7%). Considering the fact that the expressionlevel of HA-AtRer1Cp was the highest among the

822

Figure 8. Comparison of amino acid sequences in the TMDs ofyeast Sec12p and Sed4p andArabidopsisAtSec12p.

three AtRer1 products, these observations indicate thatAtRER1Bcomplements the yeastrer1-2 mutant bestand AtRER1Cis the weakest. This may imply thatAtRer1C1p has the most divergent role among thethree.

Discussion

In this paper, we have described the identificationand characterization ofA. thaliana RER1cDNAs.The presence of EST clones and the result of DNAgel blot analysis indicate that there are manyAra-bidopsisgenes that are homologous to yeastRER1.We obtained three cDNAs,AtRER1A, AtRER1BandAtRER1C1, which complemented the defect of theyeastrer1 mutant in correct ER localization of Sec12p,although the degree of complementation differed.Thus, theRER1gene family has developed diversityduring evolution but still conserves its function.

Yeast Rer1p is located in the Golgi apparatus andits function is to recognize ER membrane proteins thatare exported from the ER and retrieve them back tothe ER. An intriguing question withArabidopsis RER1would then be whether they are differentiated in theirroles of such Golgi-to-ER protein retrieval. Staehelin(1997) proposes that plant ER has many discrete func-tional domains. If this is true, the vesicle recyclingbetween the ER and the Golgi may also have manydifferent routes.

The difference in the complementation abilities ofAtRER1cDNAs is interesting in terms of their pos-sible differentiation. AtRer1Cp is the weakest amongthe three. In fact, the scores of amino acid iden-tity between yeast Rer1p and AtRer1Ap, AtRer1Bpand AtRer1Cp are 46%, 44% and 39%, respectively.Phylogenic analysis indicates thatAtRER1Cis mostdistantly located among the three cDNAs. These ob-servations may imply that AtRer1Cp has evolved amore specialized role.

AtRer1Ap and AtRer1Bp are 85% identical. How-ever, the complementation of the yeastrer1-2 mutantby AtRer1Ap is weaker than that by AtRer1Bp. Thisdifference may provide a hint to understand whichamino acid residues in Rer1p are important for the

recognition and retrieval of Sec12p in yeasts andplants. We have demonstrated that the TMDs of yeastSec12p and Sed4p function as the Rer1-dependentretrieval signal (Satoet al., 1996). Comparison be-tween these two TMDs and the TMD ofArabidopsisSec12p is shown in Figure 8. We have also shown thatthe asparagine and glutamine residues in the TMD ofyeast Sec12p are important for the Rer1p-dependentER localization (Satoet al., 1996). In the TMD ofAtSec12p, however, the corresponding residues arechanged to leucine and tyrosine, respectively. We donot have evidence at the moment that AtSec12p isretrieved by AtRer1p inArabidopsiscells, but thisdifference in the TMD sequences may well affect therecognition mechanism in retrieval. Mutational analy-sis of yeast Rer1p and AtRer1p would be useful toelucidate the molecular basis of such recognition. Itshould also be mentioned that the C-terminal half ofRer1p is very well conserved between yeast andAra-bidopsisproteins, suggesting an important role for thisdomain.

As for AtRER1C, we found two very similarcDNAs, which may be produced by alternative splic-ing of a common gene. This implies that the expres-sion ofAtRER1Cis uniquely regulated, whose preciserole in higher plants remains to be revealed. In contrastto Rer1p, theArabidopsisgenome appears to haveonly one gene for Erd2p, the receptor in the Golgiapparatus for the ER proteins harboring KDEL/HDELsignals (Leeet al., 1993; Bar-Peledet al., 1995).Both RER1andERD2are single genes in yeast; thelatter is essential for growth but the former is not.Why higher plants have developed different diversi-fication for these two Golgi-to-ER retrieval systems isan interesting question from an evolutionary point ofview.

The three AtRER1 genes are ubiquitously ex-pressed in allArabidopsistissues we examined. Theirexpression is high in roots, floral buds, and suspen-sion culture in which the secretory activity is probablyhigh. Similar tendencies of expression have also beenobserved with other secretory genes ofArabidopsis,such asAtSEC12, AtSAR1, AtGDI1, AtGDI2 andRMA1 (Ueda et al., 1996; Bar-Peled and Raikhel,1997; Matsuda and Nakano, 1998; Uedaet al., 1998).The expression ofAtRER1Aseems to be increasedby NaCl treatment. A 3- to 4-fold induction of theAtRER1AmRNA level in seedlings by 30 min NaCltreatment has been reported (Pihet al., 1997). Thephysiological meaning of such a stress response ofAtRer1Ap is unclear at the moment though.

823

Yeast Rer1p is predominantly localized to thecis-Golgi region in the steady state, while the humancounterpart was observed in the ER-Golgi interme-diate compartment and in the Golgi apparatus (Satoet al., 1995; Boehmet al., 1997; Füllekruget al.,1997). The localization of AtRer1p inArabidopsisis another important question to be addressed. Bio-chemical and cytological analysis of AtRer1p will notonly reveal the mechanism of ER membrane proteinlocalization but also provide useful tools to study theendomembrane system in higher plants.

Acknowledgements

We are grateful to the Arabidopsis Biological Re-source Center for supplying cDNAs containingAtRER1homologues. This work was supported bya Special Coordination Fund for promoting Scienceand Technology from the Science and TechnologyAgency of Japan, by a Grant-in-Aid for ScientificResearch on Priority Areas and a Grant-in-Aid for En-couragement of Young Scientists from the Ministry ofEducation, Science, Sports and Culture of Japan, andby a President’s Special Research Grant from RIKEN.

References

Bar-Peled, M. and Raikhel, N.V. 1997. Characterization ofAtSEC12andAtSAR1. Plant Physiol. 114: 315–324.

Bar-Peled, M., Conceição, A.S., Frigerio, L. and Raikhel, N.V.1995. Expression and regulation ofaERD2, a gene encodingthe KDEL receptor homolog in plants, and other genes encodingproteins involved in ER-Golgi vesicular trafficking. Plant Cell 7:667–676.

Barlowe, C. and Schekman, R. 1993.SEC12encodes a guanine nu-cleotide exchange factor essential for transport vesicle buddingfrom the ER. Nature 365: 347–349.

Boehm, J., Ulrich, H.D., Ossig, R. and Schmitt, H.D. 1994. Kex2-dependent invertase secretion as a tool to study the targetingof transmembrane proteins which are involved in ER→Golgitransport in yeast. EMBO J. 13: 3696–3710.

Boehm, J., Letourneur, F., Ballensiefen, W., Ossipov, D., Démol-lière, C. and Schmitt, H.D. 1997. Sec12p requires Rer1p forsorting to coatomer (COPI)-coated vesicles and retrieval to theER. J. Cell Sci. 110: 991–1003.

Boevink, P., Oparka, K., Santa Cruz, S., Martin, B., Betteridge, A.and Hawes, C. 1998. Stacks on tracks: the plant Golgi apparatustraffics on an actin/ER network. Plant J. 15: 441–447.

Chan, R.K. and Otte, C.A. 1982. Isolation and genetic analysis ofSaccharomyces cerevisiaemutants supersensitive to G1 arrest bya factor andα factor pheromones. Mol. Cell. Biol. 2: 11–20.

Cosson, P. and Letourneur, F. 1994. Coatomer interaction with di-lysine endoplasmic reticulum motifs. Science 263: 1629–1631.

Cosson, P., Démollière, C., Hennecke, S., Duden, R. and Le-tourneur, F. 1996.δ- andζ -COP, two coatomer subunits homolo-gous to clathrin-associated proteins, are involved in ER retrieval.EMBO J. 15: 1792–1798.

Denecke, J., De Rycke, R. and Botterman, J. 1992. Plant and mam-malian sorting signals for protein retention in the endoplasmicreticulum contain a conserved epitope. EMBO J. 11: 2345–2355.

d’Enfert, C., Gensse, M. and Gaillardin, C. 1992. Fission yeast and aplant have functional homologues of the Sar1 and Sec12 proteinsinvolved in ER to Golgi traffic in budding yeast. EMBO J. 11:4205–4211.

Füllekrug, J., Boehm, J., Röttger, S., Nilsson, T., Mieskes, G.and Schmitt, H.D. 1997. Human Rer1 is localized to the Golgiapparatus and complements the deletion of the homologousRer1 protein ofSaccharomyces cerevisiae. Eur. J. Cell. Biol 74:31–40.

Fuller, R.S., Sterne, R.E. and Thorner, J. 1988. Enzymes requiredfor yeast prohormone processing. Annu. Rev. Physiol. 50: 345–362.

Haseloff, J., Siemering, K., Prasher, D. and Hodge, S. 1997. Re-moval of a cryptic intron and subcellular localization of greenfluorescent protein are required to mark transgenic Arabidopsisplants brightly. Proc. Natl. Acad. Sci. USA. 94: 2122–2127.

Higgins, D.G., Bleasby, A.J. and Fuchs, R. 1992. CLUSTAL V: im-proved software for multiple sequence alignment. Comput. Appl.Biosci. 8: 189–191.

Jackson, M.R., Nillson, T. and Peterson, P.A. 1993. Identification ofa consensus motif for retention of transmembrane proteins in theendoplasmic reticulum. EMBO J. 9: 3153–3162.

Jackson, M.R., Nillson, T. and Peterson, P.A. 1993. Retrieval oftransmembrane proteins to the endoplasmic reticulum. J. CellBiol. 121: 317–333.

Keszenman-Pereyra, D. and Hieda, K. 1988. A colony procedurefor transformation ofSaccharomyces cerevisiae.Curr. Genet. 13:21–23.

Kyte, J. and Doolittle, R.F. 1982. A simple method for displayingthe hydropathic character of a protein. J. Mol. Biol. 157: 105–132.

Lee, H., Gal, S., Newman, T.C. and Raikhel, N.V. 1993. The Ara-bidopsis endoplasmic reticulum retention receptor functions inyeast. Proc. Natl. Acad. Sci. USA 90: 11433–11437.

Letourneur, F., Gaynor, E.C., Hennecke, S., Démollière, C.,Duden, R., Emr, S.D., Riezman, H, and Cosson, P. 1994.Coatomer is essential for retrieval of dilysine-tagged proteins tothe endoplasmic reticulum. Cell 79: 1199–1207.

Lewis, M.J. and Pelham, H.R.B. 1990. A human homologue of theyeast HDEL receptor. Nature 348: 162–163.

Lewis, M.J., Sweet, D.J. and Pelham, H.R.B. 1990. TheERD2genedetermines the specificity of the luminal ER protein retentionsystem. Cell 61: 1359–1363.

Matsuda, N. and Nakano, A. 1998.RMA1, anArabidopsis thalianagene whose cDNA suppresses the yeastsec15mutation, encodesa novel protein with a RING finger motif and a membrane anchor.Plant Cell Physiol. 39: 545–554.

Munro, S. and Pelham, H.R.B. 1987. A C-terminal signal preventssecretion of luminal ER proteins. Cell 48: 899–907.

Nakano, A. and Muramatsu, M. 1989. A novel GTP-binding protein,SAR1, is involved in transport from the endoplasmic reticulum tothe Golgi apparatus. J. Cell. Biol. 109: 2677–2691.

Nakano, A., Brada, D. and Schekman, R. 1988. A membraneglycoprotein, Sec12p, required for protein transport from theendoplasmic reticulum to the Golgi apparatus in yeast. J. CellBiol. 107: 851–863.

824

Nishikawa, S. and Nakano, A. 1993. Identification of a gene re-quired for membrane protein retention in the early secretorypathway. Proc. Natl. Acad. Sci. USA 90: 8179–8183.

Palmiter, R. 1974. Magnesium precipitation of ribonucleoproteincomplexes. Expedient techniques for the isolation of under-graded polysomes and messenger ribonucleic acid. Biochemistry13: 3606–3615.

Pelham, H.R.B. 1989. Control of protein exit from the endoplasmicreticulum. Annu. Rev. Cell Biol. 5: 1–23.

Pelham, H.R.B., Hardwick, K.G. and Lewis, M.J. 1988. Sorting ofsoluble ER proteins in yeast. EMBO J. 7: 1757–1762.

Pih, K.T., Jang, H.J., Kang, S.G., Piao, H.L. and Hwang, I.1997. Isolation of molecular markers for salt stress responses inArabidopsis thaliana. Mol. Cell 7: 567–571.

Sato, K., Nishikawa, S. and Nakano, A. 1995. Membrane proteinretrieval from the Golgi apparatus to the endoplasmic reticulum(ER): characterization of theRER1gene product as a componentinvolved in ER localization of Sec12p. Mol. Biol. Cell 6: 1459–1477.

Sato, M., Sato, K. and Nakano, A. 1996. Endoplasmic reticulumlocalization of Sec12p is achieved by two mechanisms: Rer1p-dependent retrieval that requires the transmembrane domainand Rer1p-independent retention that involves the cytoplasmicdomain. J. Cell Biol. 134: 279–293.

Sato, K., Sato, M. and Nakano, A. 1997. Rer1p as common ma-chinery for the endoplasmic reticulum localization of membraneproteins. Proc. Natl. Acad. Sci. USA 94: 9693–9698.

Semenza, J.C., Hardwick, K.G., Dean, N. and Pelham, H.R.B. 1990.ERD2, a yeast gene required for the receptor-mediated retrievalof luminal ER proteins from the secretory pathway. Cell 61:1349–1357.

Staehelin, L.A. 1997. The plant ER: a dynamic organelle composedof a large number of discrete functional domains. Plant J. 11:1151–1165.

Ueda, T., Matsuda, N., Anai, T., Tsukaya, H., Uchimiya, H.and Nakano, A. 1996. An Arabidopsis gene isolated by anovel method for detecting genetic interaction in yeast encodesthe GDP dissociation inhibitor of Ara4 GTPase. Plant Cell 8:2079–2091.

Ueda, T., Yoshizumi, T., Anai, T., Matsui, M., Uchimiya, H. andNakano, A. 1998. AtGDI2, a novel Arabidopsis gene encoding aRab GDP dissociation inhibitor. Gene 206: 137–143.

Vitale, A., Ceriotti, A. and Denecke, J. 1993. The role of theendoplasmic reticulum in protein synthesis, modification andintracellular transport. J. Exp. Bot. 44: 1417–1444.