phylogenetic analyses and expression studies reveal two · phylogenetic analyses and expression...

Phylogenetic Analyses and Expression Studies Reveal TwoDistinct Groups of Calreticulin Isoforms in Higher Plants1

Staffan Persson*, Magnus Rosenquist, Karin Svensson, Rafaelo Galvao, Wendy F. Boss, andMarianne Sommarin

Department of Plant Biochemistry, Lund University, 22100 Lund, Sweden (S.P., K.S., M.S.); Department ofPathology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215 (M.R.); andBotany Department, North Carolina State University, Raleigh, NC 27695 (R.G., W.F.B.)

Calreticulin (CRT) is a multifunctional protein mainly localized to the endoplasmic reticulum in eukaryotic cells. Here, wepresent the first analysis, to our knowledge, of evolutionary diversity and expression profiling among different plant CRTisoforms. Phylogenetic studies and expression analysis show that higher plants contain two distinct groups of CRTs: aCRT1/CRT2 group and a CRT3 group. To corroborate the existence of these isoform groups, we cloned a putative CRT3ortholog from Brassica rapa. The CRT3 gene appears to be most closely related to the ancestral CRT gene in higher plants.Distinct tissue-dependent expression patterns and stress-related regulation were observed for the isoform groups. Further-more, analysis of posttranslational modifications revealed differences in the glycosylation status among members within theCRT1/CRT2 isoform group. Based on evolutionary relationship, a new nomenclature for plant CRTs is suggested. Thepresence of two distinct CRT isoform groups, with distinct expression patterns and posttranslational modifications, supportsfunctional specificity among plant CRTs and could account for the multiple functional roles assigned to CRTs.

Calreticulin (CRT) is a highly conserved proteinmainly localized to the endoplasmic reticulum (ER)in plants and to the ER/sarcoplasmic reticulum inmammals (for review, see Crofts and Denecke, 1998;Michalak et al., 1999; Hadlington and Denecke, 2000;Johnson et al., 2001). CRT is a multifunctional pro-tein, suggested to be involved in over 40 intra- andextracellular processes in mammalian cells. How-ever, the main focus has been on its role in calciumsignaling (Camacho and Lechleiter, 1995; Nakamuraet al., 2001; Arnaudeau et al., 2002) and as a chaper-one (Hebert et al., 1996; Saito et al., 1999; Nakamuraet al., 2001). CRT comprises three major subdomains:a highly conserved N domain, a high-affinity butlow-capacity Ca2�-binding P domain, and a low-affinity but high-capacity Ca2�-binding C domainending with an ER retention signal (Michalak et al.,1999).

Although the role of CRTs as chaperone-like pro-teins and in calcium signaling is well established inmammals, the functions of CRT have been elusive inplants until recently. Plant CRTs have been shown tobind calcium with similar characteristics as their

mammalian homologs (Chen et al., 1994; Hassan etal., 1995; Navazio et al., 1995; Coughlan et al., 1997; Liand Komatsu, 2000) and recently also to havecalcium-storing functions in the ER of plant cells(Persson et al., 2001; Wyatt et al., 2002). In contrast tomost animal CRTs, glycosylation of CRTs is generallyobserved in plants (Navazio et al., 1995, 2002; Pagnyet al., 2000). Plant CRTs are up-regulated in responseto a variety of stress-mediated stimuli, e.g. pathogen-related signaling molecules (Denecke et al., 1995;Jaubert et al., 2002) and gravistimulation (Heilmannet al., 2001), and are highly expressed during mitosis(Denecke et al., 1995), embryogenesis (Borisjuk et al.,1998), and in floral tissues (Chen et al., 1994; Deneckeet al., 1995; Nelson et al., 1997). In addition, CRTpreferentially localizes to plasmodesmata in maize(Zea mays) root tips and is suggested to be involved inregulation of the closure of plasmodesmata (Baluskaet al., 1999).

The dogma for CRT in human (Homo sapiens) andmouse (Mus musculus) has been: one gene, onemRNA, and one protein. However, recently, an ad-ditional isoform was identified (Persson et al., 2002b).The sequence of the newly discovered CRT isoformdiffered significantly from the previously establishedisoform but still contained typical CRT features. Phy-logenetic analysis revealed that the early duplicationof the CRT genes in mammals generated two distinctCRT groups. In plants, two or more isoforms exists inseveral species, e.g. Arabidopsis, maize, and barley(Hordeum vulgare; GenBank accession no. 190454;Chen et al., 1994; Kwiatkowski et al., 1995; Nelson etal., 1997). With the exception of the Arabidopsis iso-forms, the reported isoforms all have a high sequence

1 This work was supported by the Conselho Nacional de Des-envolvimento CientR fico e Tecnologico, Ministerio da Ciencia eTecnologica (Brazil; fellowship to R.G.), in part by The SwedishResearch Council (grant to M.S.), and in part by the NationalAeronautics and Space Administration and the National ScienceFoundation (grant to W.F.B.).

* Corresponding author; e-mail [email protected];fax 46 – 46 –2224116.

Article, publication date, and citation information can be foundat www.plantphysiol.org/cgi/doi/10.1104/pp.103.024943.

Plant Physiology, November 2003, Vol. 133, pp. 1–12, www.plantphysiol.org © 2003 American Society of Plant Biologists 1 of 12

_________________________________________________________________________________________________________

This article is published in Plant Physiology Online, Plant Physiology Preview Section, which publishes manuscripts accepted forpublication after they have been edited and the authors have corrected proofs, but before the final, complete issue is publishedonline. Early posting of articles reduces normal time to publication by several weeks._________________________________________________________________________________________________________

www.plantphysiol.orgon March 6, 2020 - Published by Downloaded from Copyright © 2003 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

similarity, implying a recent duplication of the CRTgene in these species. Of the three described CRTsequences in Arabidopsis, the CRT1 and CRT2 se-quences share higher sequence homology to eachother than compared with the third isoform, CRT3(Nelson et al., 1997). In an evolutionary context, thisimplies that two duplications of the CRT gene inArabidopsis took place at different times.

Here, we report that both monocotyledons andeudicotyledons contain two distinct groups of CRTs.The early duplication of the CRT gene in plants isstrikingly similar to the duplication of the CRT genein mammals (Persson et al., 2002b). The intron/exonorganization of CRT genes encoding different iso-forms reinforces the prediction of a common ancestryfor the CRT gene. To verify the existence of the twoisoform groups, we cloned a CRT3 ortholog in Bras-sica rapa based on the known Arabidopsis sequencecharacteristics. In addition, analyzes of tissue-dependent and stress-related expressions and post-translational modifications of the different isoformswere carried out to evaluate in silica predictions andto test the proposed evolutionary model.

RESULTS

Phylogenetic Analysis of CRT Amino AcidSequences in Plants

The Arabidopsis genome harbors three expressedCRT genes (Nelson et al., 1997). In addition, the Ara-bidopsis genome also contains a putative CRT pseu-dogene (locus At1g56390), consisting of four poten-tial exons, corresponding to exons 1, 2, 3, and 6 ofCRT1 (data not shown).

To get a more complete picture of the number ofCRT isoforms identified in plants, we performed astandard BLASTP analysis at the National Center forBiotechnology Information (NCBI) using CRT pro-tein sequences corresponding to the Arabidopsis iso-forms. We found 18 unique protein sequences anno-tated as CRT (data not shown). From these, a rootedphylogenetic tree was created (Fig. 1). In both mono-cotyledons and eudicotyledons, there appears to beat least two CRT isoforms with high sequence iden-tity, e.g. CRT1 and CRT2 in maize, Arabidopsis, andbarley (Fig. 1).

The topology of the phylogenetic tree reveals anearly duplication event in the species Arabidopsis,from which the CRT1/CRT2 and the CRT3 isoformsderive, perhaps predating the evolutionary split ofplants into dicots and monocots (Soltis et al., 1999).This early divergence of CRTs in Arabidopsis advo-cates the existence of orthologous isoforms (CRT3s)in other plant species. Therefore, a standard BLASTNwith Arabidopsis CRT3 was performed at the NCBIagainst expressed sequence tags (ESTs) and full-length cDNAs from various plant species. Two puta-tive full-length mRNAs (GenBank accession nos.AY105822 and AP003316), predicted from genomic

sequences, were obtained from maize and rice (Oryzasativa). The sequences were denoted maize CRT3 andrice CRT3, respectively.

To investigate sequence homology to Crts from thetwo isoform groups, the maize CRT3 was translatedinto an amino acid sequence and aligned with Ara-bidopsis CRT1, CRT2, and CRT3 and maize CRT1and CRT2 (Fig. 2A). The maize CRT3 sequence con-sists of 415 amino acids and shows 70% identity tothe Arabidopsis CRT3 isoform but only 57% and 55%identity to the maize CRT1 and CRT2 isoforms, re-spectively (data not shown). Several of the typicalCRT sequence features, conserved among CRT pro-teins from different kingdoms (for review, see Micha-lak et al., 1999) were conserved in the maize CRT3sequence. These include: three Cys residues impor-tant for the correct folding of CRTs (Matsuoka et al.,1994), a potential ER signal sequence located in the Nterminus, two triplets of conserved regions in the Pdomain, and a typical ER retention motif (HDEL) inthe far end of the C terminus (Fig. 2A).

Figure 1. Phylogenetic analysis of CRT protein sequences in plants.A rooted phylogenetic tree with topology representative for plantCRTs, generated with the Chlamydomonas reinhardtii CRT as out-group. A heuristic search using the maximum parsimony method wasdone on the alignment of 18 unique plant CRT protein sequences.Two major groups are prominent: monocotyledons and eudicotyle-dons. These groups are presented in different shades of gray.

Persson et al.

2 of 12 Plant Physiol. Vol. 133, 2003 www.plantphysiol.orgon March 6, 2020 - Published by Downloaded from Copyright © 2003 American Society of Plant Biologists. All rights reserved.


The two triplets of conserved regions in the P do-main, denoted I and II, are well conserved in maizeCRT3 (Fig. 2A). Comparison of the 18 aligned plantCRT sequences gave the repeats in region I the con-sensus sequence of PXXIXDPXXKKPEXWDD and inregion II the consensus sequence of GXWXAXXIXN-PXYK (data not shown). In animal CRTs, the repeat Iand II consensus sequences are PXXIXDPDAXKPED-WDE and GXWXPPXIXNPXYX, respectively (Micha-lak et al., 1999). Thus, the two triplet repeats areconserved but not identical in vertebrates and plants.

To verify expression of the putative maize and riceCRT3 genes, a BLASTN search was performed atNCBI against ESTs from rice and maize, respectively.We obtained six ESTs from maize, and one EST fromrice with E values below 2 e-64 (score � 200), con-firming that the CRT3 gene is transcribed in maizeand rice (Table I).

To corroborate that the putative maize and riceCRT3 isoforms are orthologs to the ArabidopsisCRT3 isoform, an exhaustive phylogenetic analysis ofArabidopsis, maize, rice, and C. reinhardtii CRT pro-teins was performed. A tree was constructed usingCRT from C. reinhardtii as outgroup (Fig. 2B). Twodistinct clades of CRTs, supported by high bootstrapvalues, were evident: a CRT1/CRT2 isoform groupand a CRT3 isoform group. To emphasize the exis-tence of the two clusters, we are using the labelCRT1/CRT2 isoform group and CRT3 isoform groupfor CRTs belonging to the respective isoform cluster.The isoform-specific clades were similar when ananalogous analysis was performed using correspond-ing CRT nucleotide sequences (data not shown).

CRT Gene Maps

Genomic CRT sequences from the different iso-forms in both monocotyledons and eudicotyledons

Table I. EST analysis of CRTs from Arabidopsis, maize, and rice

Isoform Arabidopsis Maize Rice

CRT1 65 81a 43a

CRT2 56 – –CRT3 12 6 1

a Due to sequence similarities between the isoforms CRT1 andCRT2 in monocotyledons, ESTs correlating to respective isoformscould not be distinguished.

the sequence alignment corresponds to two triplets of conservedregions in the P domain of the proteins. The black line (ER-R)overlaying the immediate C terminus corresponds to an ER retentionsignal. The approximate position of the three domains (N, P, and C)are indicated. B, Rooted phylogenetic tree based on the proteinalignment, including CRT1/2, and CRT3 protein sequences from rice(GenBank accession nos. BAA88900, and BAC06263, respectively)and the CRT protein sequence from C. reinhardtii (GenBank acces-sion no. CAB54526), the latter used as outgroup. Two distinct clus-ters can be observed: CRT1 and CRT2 isoforms versus CRT3 iso-forms. Bootstrap values are indicated on respective branch.

Figure 2. Sequence comparison of CRT isoforms from Arabidopsisand maize. Comparison of the amino acid sequences from Arabi-dopsis CRT1, CRT2, and CRT3 with maize CRT1, CRT2, and CRT3(GenBank accession nos. AAC49695, AAK74014, and AAC49697with CAA86728, AAF01470, and AY105822, respectively) was madeusing a ClustalW analysis algorithm. A, Vertical alignments betweenthe sequences for identical and similar amino acids are highlighted indifferent shades of gray. The black line (SS) overlaying the immediateN terminus of the CRT isoforms corresponds to a putative ER signalsequence segment. The black arrows indicate the positions of threehighly conserved Cys residues. The black lines (I and II) overlaying

Two Isoform Groups of Plant Calreticulins

Plant Physiol. Vol. 133, 2003 3 of 12 www.plantphysiol.orgon March 6, 2020 - Published by Downloaded from Copyright © 2003 American Society of Plant Biologists. All rights reserved.


were analyzed to obtain the exon/intron organiza-tions for the CRT genes in higher plants. In Arabi-dopsis, individual CRT exon lengths were generallyconserved between the genes (Fig. 3).

Analogous information from monocotyledonswere obtained using BLASTN searches of the ricegenome (for draft descriptions, see Goff et al., 2002;Yu et al., 2002) at the NCBI with cDNAs correspond-ing to the CRT1/CRT2 isoform group (GenBank ac-cession no. AB021259) and the CRT3 isoform group.A 100% sequence identity was obtained when align-ing the CRT3 cDNA with the corresponding genomicclone (GenBank accession no. AAAA01001172).However, when aligning the CRT1/CRT2 cDNA withthe reverse complement of the only genomic se-quence hit obtained (GenBank accession no.AAAA01007283), we discovered that the hit onlyshowed 98% sequence identity with the query entryand, thus, harbored a novel CRT isoform. Thegenomic sequence corresponding to the rice CRTused in the initial BLASTN search is currently notavailable. Hence, it appears that the evolution of CRTisoforms in rice is similar to the isoform evolution inmaize and barley, yielding two closely related para-logs for the CRT1/CRT2 isoform group and a moredistantly related CRT3 isoform (Fig. 1).

The obtained genomic clones from rice were usedto generate individual exon lengths for the CRT iso-forms. Although only one genomic clone correspond-ing to the CRT1/CRT2 isoform group was found, a100% sequence identity in the splice regions to bothCRT1 and CRT2 in maize gave the exon lengths forboth isoforms (Fig. 3). Although exons encoding theC terminus vary considerably for different isoformsand species, both in lengths and splice codons, allexons encoding the N and P domains of the proteinsare conserved, with splice sites located in the samecorresponding regions (Fig. 3).

Evolution of the CRT Gene in Higher Plants

In silica mapping of the Arabidopsis CRTs revealsthat the CRT2 and CRT3 genes are located closelytogether on chromosome 1 at loci At1g09210 (Gen-Bank accession no. AY045656) and At1g08450 (Gen-Bank accession no. U66345), respectively. The CRT1gene is also located on chromosome 1 at locusAt1g56340 (GenBank accession no. U66343). To in-vestigate potential relationships between major du-plication events in the Arabidopsis genome and theevolution of the CRT gene, we examined if any of theCRT loci were situated in known duplicated genomicsegments. Although the CRT1 and CRT2 genes werefound in a region that was duplicated approximately50 million years ago (Vision et al., 2000; blocks 8a and8b in Fig. 1), the CRT3 gene locus is located in aregion without any major duplication activity re-ported (data not shown).

The evolutionary split between the monocotyle-dons maize and rice has been estimated to be 52 � 15

million years ago (Bremer, 2002). When comparingCRT1/CRT2 sequences for these species, an identityof approximately 85% is observed. Because the se-quence identity between CRT1 and CRT2 in Arabi-dopsis is 83%, the predicted time of the genomicduplication of the Arabidopsis CRT locus appearsprobable.

A close examination of the exon/intron patterns ofCRTs in different species revealed an apparent pat-tern of evolution. Overall, the sizes of exons, includ-ing exon fusion products, are conserved among iso-forms and species investigated, with the exception ofexons 1 and 11 to 13 (exon nos. for the CRT3 isoforms;Fig. 3). The CRT3 gene has 14 exons in Arabidopsis,rice, and maize, with exon sizes highly conservedexcept for exon 1 and 12 (Fig. 3). A comparison ofgenes from the CRT1/CRT2 isoform groups amongspecies revealed that CRT1 and CRT2 in both maizeand rice contain 14 exons similar to the CRT3 genes,whereas CRT1 and CRT2 in Arabidopsis only contain12 and 13 exons, respectively. This result is predictedfrom exon fusions of exons 4 to 6 in ArabidopsisCRT3, generating larger exons: exon 4 in CRT1 andexon 5 in CRT2 (Fig. 3). Thus, the conservationamong the CRT genes is highest for the CRT3 iso-forms in the investigated species, whereas the CRT1/CRT2 isoforms show evolutionary deviations amongmonocotyledons and eudicotyledons.

Cloning and Expression of CRT Orthologs in B. rapa

To obtain additional information regarding or-thologous CRT1/2 and CRT3 isoforms, a standardBLASTN search was performed against ESTs fromvarious plant species at the NCBI. Although severalplant species were found to have ESTs correspondingto either putative CRT1/2 or CRT3 isoforms (score �200, respectively), only B. rapa contained ESTs corre-lating to both isoform groups (data not shown). TheESTs corresponding to the CRT1/2 and CRT3 isoformgroups from B. rapa were aligned, and the overlap-ping sequences were used to generate specific probesfor the putative CRT1/2 and CRT3 isoforms. Bothprobes recognized a band at an approximate size of1.4 kb of the total RNA from B. rapa leaves, indicatingthat the two isoform groups are present in B. rapa(data not shown).

Overlapping ESTs for CRT3 from B. rapa were alsoused to generate sequence specific primers againstthe 5� end of the putative CRT3 isoform, whereas anoligo(dT15) primer was used for extension from the 3�end. A 1,300-bp nucleotide sequence was obtained.Of these, the first 1146 were sequenced (GenBankaccession no. AY336743), and an open reading frameencoding 381 amino acids was generated (Fig. 4). Wewere unable to obtain the sequence for the farC-terminal end, most likely due to secondary struc-tures in the nucleotide sequence (Technical Support,MWG Biotech, Ebersberg, Germany). The deducedamino acid sequence shows high homology to the

Persson et al.



Figure 3. Gene maps of CRT genes in Arabidop-sis and rice. Black boxes, Exons; white boxes,introns. Exon and intron sizes are indicated innumber of base pairs within each box. The genenames and corresponding species are listed tothe left. The vertical arrows to the right indicatethe potential direction of evolution. Genes bestrepresenting the most ancestral CRT genes,CRT3s, are within a gray bracket, and the direc-tion of gene divergence is indicated with adja-cent arrows. Conservation of exon size betweengenes is indicated by shaded areas.




CRT3 isoforms in Arabidopsis and maize (91%, and72% identity, respectively) but only 58%, and 57%identity to the Arabidopsis CRT1, and CRT2 iso-forms, respectively. Furthermore, the B. rapa CRT3isoform contained the typical CRT features indicatedin Figure 4. The sequence was aligned with CRT1,CRT2, and CRT3 from Arabidopsis and CRT from C.reinhardtii, and an exhaustive phylogenetic analysiswas performed. Using the CRT from C. reinhardtii asoutgroup, two distinct clusters were obtained withthe putative B. rapa CRT3 sequence closely clusteredwith the Arabidopsis CRT3 (data not shown). Thesedata strengthen the hypothesis advocating two dis-tinct CRT isoform groups in both mono- andeudicotyledons.

Tissue-Dependent Expression of CRT Isoforms inArabidopsis and Maize

Northern-blot analyzes of various tissue types fromboth Arabidopsis and maize were performed to de-termine where the CRT1, CRT2, and CRT3 isoformsare expressed (Fig. 5, A–F). cDNAs corresponding toCRT1, CRT2, and CRT3 isoforms and CRT1/CRT2isoforms were used as probes for Arabidopsis andmaize, respectively. For the maize CRT3 isoform, a101-nucleotide probe corresponding to the 3�-UTRwas generated, and cross-reactivity among theprobes was checked. None of the probes showed anycross-reactivity within respective species (Fig. 5, Aand D).

The Arabidopsis CRT1 and CRT2 isoforms weremainly expressed in leaves, roots, and flowers, with alower expression in the inflorescence stem (Fig. 5, Band C). On the other hand, the Arabidopsis CRT3isoform was predominantly expressed in leaves androots and was only detected at very low levels in theinflorescence stem (Fig. 5, B and C). A similar expres-sion pattern was observed in maize, where the CRT1/CRT2 isoforms were present in all investigated tis-sues (data not shown), and the CRT3 isoform wasmost abundant in leaves and roots (Fig. 5E). Becausethe northern-blot analysis only revealed relative ex-pression levels within each isoform group, we per-formed an EST analysis for the Arabidopsis, maize,and rice CRT cDNAs at the NCBI. Substantially moreESTs corresponding to the CRT1/CRT2 isoform groupthan to the CRT3 isoform group were obtained, sug-gesting that CRT1 and CRT2 are expressed in higherabundance in higher plants (Table I).

Stress Induction of the CRT Genes in Arabidopsis

To obtain information about the regulation of thedifferent CRT isoforms, we used Arabidopsis cellsuspension cultures. All three CRT isoforms wereexpressed in the suspension cells (Fig. 6A). We alsoinvestigated the expression of the different isoformsduring different phases of the growth period. Al-though both CRT1 and CRT2 showed a high expres-sion during the 3 first d, corresponding to a rapidphase of growth, the CRT3 expression was moreevenly distributed over the growth period examined(Fig. 6B). Because high initial expression of the CRTisoforms would diminish putative up-regulations ofthe genes in response to certain stresses, we chose toperform stress experiments on 4-d-old cell cultures.We used salt, tunicamycin (an inhibitor of N-linkedglycosylation processes), and ABA as stress media-tors and monitored changes in the expression forCRT1, CRT2, and CRT3 after 30-min and 4- and 12-htreatments (Fig. 6C). CRT3 showed a fast response tosalt and tunicamycin, with a severalfold increase inexpression for both treatments (30 min in Fig. 6C). Incontrast, both CRT1 and CRT2 showed no major in-crease in expression in response to 30-min treat-ments. However, after 4 h of stress exposure, both theCRT1 and CRT2 expression increased severalfold inresponse to tunicamycin (4 h in Fig. 6C). The induc-tion of CRT1 and CRT2 was maintained and furtherincreased after 12 h of tunicamycin treatment (12 h inFig. 6C). On the other hand, the increased CRT3expression observed at 30 min was no longer evident.

To investigate if a similar induction of the CRTgenes occurs in whole plants, we performed stressexperiments with Arabidopsis plants grown on liq-uid medium. In addition to the treatments describedabove, plants were subjected to drought and EGTAtreatments. CRT1, CRT2, and CRT3 transcripts allincreased in whole plants after 2 h of tunicamycin

Figure 4. Cloning of CRT3 from B. rapa. Nucleotide sequence anddeduced amino acid sequence of the CRT3 isoform (GenBank ac-cession no. AY336743) in B. rapa. Several CRT characteristics areindicated in accordance with Figure 2. The transparent boxes indi-cate the three conserved Cys residues.

Persson et al.



treatment (Fig. 7). Furthermore, the CRT3 expressionwas increased in response to salt stress, similar towhat was observed in the cell cultures (compare Figs.6C and 7). Hence, in addition to differences in tissue-dependent expression, differences are seen in stress-induced expression among the Arabidopsis CRTisoforms.

Variation in Glycosylation Status among CRTIsoforms in Arabidopsis

Potential differences in posttranslational modifica-tions among CRT isoforms were analyzed using theMacVector 7.0 software (Oxford Molecular GroupPlc, Oxford). The two isoform groups differ in thenumber of negatively charged amino acids in theC-terminal region (data not shown). In addition, wefound three potential glycosylation sites in the CRT1sequence but only one in the CRT2 and CRT3 se-quences, respectively (Fig. 8A). Putative differencesin the number of attached glycans were also sug-

gested by western blots, where three bands (a–c inFig. 8B), corresponding to CRTs, were obtained. Toconfirm that the size differences of the bands weredue to attached glycans, an Arabidopsis homogenatewas treated with the glycosidase PNGase F, whichremoves N-linked glycans. Figure 8C shows that theupper band disappears after a brief PNGase F treat-ment, indicating that glycans attached to this CRTwas easily accessible for the glycosidase. Further-more, one band showed remarkable resistance to theglycosidase and disappeared only after prolongedtreatment (band termed b in Fig. 8C). Thus, the dif-ferent CRT isoforms harbor differences in attachedN-linked glycans, potentially in numbers of attachedglycans, which show variations in resistance towardglycosidase PNGase F.

DISCUSSION

CRTs have been implicated in a variety of cellularprocesses, spanning from a mediator of cellular ad-

Figure 5. Expression of CRT isoforms in Arabi-dopsis and maize. A and D, Examination ofcross-hybridization of the Arabidopsis CRT1,CRT2, and CRT3 (A) and maize CRT1/2 andCRT3 (D) isoform probes. One hundred nano-grams of each probe was applied to the mem-brane and subsequently probed with radiola-beled CRT1, CRT2, and CRT3 probescorresponding to respective species. The cross-hybridization was performed in parallel withnorthern-blot hybridizations. B and E, Northern-blot analyses of total RNA (Arabidopsis, 10 �glane�1; and maize, 14 �g lane�1) from varioustissues in Arabidopsis (B) and maize (E). Mem-branes were probed with radiolabeled CRT1,CRT2, and CRT3 probes (Arabidopsis) and aCRT3 probe (maize) within respective species.Radiolabeled rRNA was used as a control. Ara-bidopsis experiments were performed indepen-dently three times and gave similar expressionprofiles. Maize experiments were performedonce to confirm Arabidopsis patterns. C, Visual-ization of relative expression of CRT isoforms invarious tissues for Arabidopsis. Asterisk, MaizeCRT3 probe corresponds to a 101-nucleotide 3�-untranslated region (UTR) segment.




hesion in the extracellular matrix to regulation ofcalcium signaling and protein folding in the ER lu-men (Johnson et al., 2001). The functional diversity ofthe protein has lead to a search for additional CRTisoforms, resulting in the discovery of a second iso-form (Crt2), present in several mammalian species(Persson et al., 2002b). To further corroborate diver-sity among plant CRT proteins, we report here theexistence of two distinct CRT isoform groups amonghigher plant species.

Several investigations have established that plantscontain two or more CRT isoforms (Chen et al., 1994;Kwiatkowski et al., 1995; Nelson et al., 1997). Thepresent nomenclature for plant CRTs suggests thatduplication events resulted in two or several or-

thologs. However, it now appears that plant CRT1and CRT2 isoforms rather represent paralogous iso-forms within respective species, whereas the CRT3isoform appears to have orthologs (Figs. 1 and 2B).Therefore, we have used the label CRT1/CRT2 iso-form group and CRT3 isoform group for CRTs be-longing to the respective isoform cluster. To avoidfuture misunderstandings regarding functional as-pects of CRT isoforms, we suggest a reevaluation ofthe CRT nomenclature within the Viridiplantae king-dom. Proposed names should be in accordance withcurrent ontology, i.e. CRT1a and CRT1b for theCRT1/CRT2 isoforms and CRT3 remaining as CRT3.

Alignment of the putative maize CRT3 isoformwith other plant CRT isoforms revealed that the se-quence contains several features typical for CRT pro-teins, e.g. an ER signal sequence in the N terminus,three Cys residues important for the proper foldingof the protein, the three tandem repeats in the Pdomain, and the ER retrieval signal in the C terminus

Figure 7. Stress induction of CRT expression in Arabidopsis plants.Northern-blot analyses of total RNA (10 �g lane�1) from Arabidopsisplants grown on liquid medium. Membranes were probed with ra-diolabeled CRT1, CRT2, and CRT3 probes. Radiolabeled rRNA wasused as a control. Plants were treated with 150 mM NaCl, 15 �gmL�1 tunicamycin, 10 mM EGTA, and 100 �M ABA or exposed todrought stress. Plants were harvested after 2-h treatments. The ABAtreatment was performed as a separate experiment. Experiments wereperformed twice and gave similar expression patterns.

Figure 6. Analysis of CRT expression in Arabidopsis suspension cellcultures. Northern-blot analyses of total RNA (5 �g lane�1) fromArabidopsis cell suspension cultures. Membranes were probed withradiolabeled CRT1, CRT2, and CRT3 probes. Radiolabeled rRNA wasused as a control. A, Untreated material probed with the differentisoform probes. B, Upper, Fresh weight for the Arabidopsis cellsuspension cultures at different days in culture. Lower, Relativeexpression of the isoforms corresponding to different days in culture.C, Arabidopsis suspension culture cells treated with either 150 mM

NaCl, 15 �g mL�1 tunicamycin, or 100 �M abscisic acid (ABA). Cellswere either treated for 30 min, 4 h, or 12 h. Radiolabeled rRNA wasused as a control. Experiments were performed twice and gavesimilar expression patterns.

Persson et al.



(Fig. 2A; Michalak et al., 1999). Although the signalsequence for the maize CRT3 isoform is distinctlydifferent compared with the other aligned isoforms(Fig. 2A), it does contain typical features for ER lo-calization, i.e. positively charged amino acid(s) inclose proximity of the N terminus, an aliphatic/hy-drophobic stretch downstream of the positivelycharged amino acid(s), and a few polar amino acidstogether with an Ala/Leu immediately upstream ofthe cleavage site (von Heijne, 1985), suggestingproper ER targeting.

Examining the genomic organization of the CRTgenes in Arabidopsis and rice showed that the struc-ture of the gene is highly conserved (Fig. 3). TheCRT3 genes in both Arabidopsis and rice have 14exons, with high similarity in exon sizes. In contrast,the CRT1 and CRT2 in Arabidopsis consist of 12 and13 exons, respectively, whereas the CRT1/CRT2 iso-forms in rice consist of 14 exons. The best represen-tation of an ancestral CRT gene, therefore, is pro-vided by the CRT3 gene in higher plants. From thegenomic sequences, it is also evident that the regionscorresponding to the N and P domains of the proteinshow a high degree of conservation. In contrast, ex-ons corresponding to the C domain are less wellconserved. The rate of conservation among the exonsis also reflected in the amino acid sequences, wherethe C domain is less conserved than the other do-mains. Apparently, the selection pressure is higherfor the N and P domains, possibly due to structuralor functional importance, compared with the Cterminus.

It is believed that the main Ca2�-binding capacityof CRT proteins is given by the number of negativelycharged amino acids in their C-terminal region(Baksh and Michalak, 1991). The differences in sizeand sequence among exons corresponding to the Cterminus of the CRTs, therefore, imply differences in

the efficiency of Ca2� binding. When comparing thenumber of negatively charged amino acids in the Cdomain for the different isoforms, it is apparent thatthe CRT3 isoforms contain less acidic residues thanboth the CRT1 and CRT2 isoforms in all investigatedspecies (37%, 35%, and 26% for Arabidopsis CRT1,CRT2, and CRT3, respectively). If the negativelycharged residues truly correspond to the amount ofCa2� that CRT can withhold, the CRT3 isoformsshould have less overall effect on the ER Ca2� levels.In addition, there are more ESTs corresponding to theisoforms in the CRT1/CRT2 isoform group. This sup-ports CRT1 and CRT2 as being the major isoforms,possibly due to an enhanced Ca2�-binding efficiency,and may indicate a less dominant role for the CRT3isoform in Ca2� homeostasis.

Implications of the C domain sequence and lengthvariations might also lie in its sensitivity to proteo-lytic activity. Earlier reports have shown that the Cdomain is sensitive to degradation, which might af-fect the subcellular location and functionality ofCRTs (Corbett et al., 2000; Persson et al., 2002a).Therefore, the differences observed in the domaincould affect both the stability of the protein, possiblyfunctioning as a turnover mechanism, or as a switchfor other subcellular localizations and interactingcomponents of CRTs.

A close examination reveals that the last exon, con-taining 12 coding nucleotides, corresponds to the ERretrieval signal, important for the retention/retrievalof resident ER proteins (Gomord et al., 1999). Becausethe CRT protein also is suggested to be involved inprocesses occurring outside the ER, mechanisms toescape the ER retrieval machinery have been sug-gested (Baldan et al., 1996; Eggleton and Llewellyn,1999). Also, as mentioned above, the sensitivity of theC domain to proteolytic activity could alter the local-ization of CRTs (Corbett et al., 2000). Therefore, it is

Figure 8. Differences in N-linked glycosylationstatus among different CRT isoforms in Arabi-dopsis. A, Analysis of CRT protein sequencespredicted three potential glycosylation sites inCRT1 and one in CRT2 and CRT3, respectively.Analysis was performed using MacVector 7.0Software. B, Arabidopsis homogenate was ana-lyzed by 10% (w/v) SDS-PAGE (15 �g proteinlane�1), blotted, and immunostained with poly-clonal antibodies against maize CRT (1:5,000[v/v]). Bands (a–c) next to lanes correspond toCRTs with differences in glycosylation status. C,Arabidopsis homogenate treated withN-glycosidase F (PNGase F) for 5 to 60 minunder native conditions. The panel to the rightshows lane probed with polyclonal antibodiesagainst Arabidopsis calnexin (CNX; 1:1,000[v/v]).




tempting to speculate whether an alternative splicevariant, lacking the HDEL signal, exists.

Here, we show that CRT1 and CRT2 were mostabundant in floral, root, and leaf tissues, with a lowerexpression in stem tissues for both Arabidopsis andmaize (Fig. 5B; data not shown). In contrast, CRT3from Arabidopsis and maize showed highest expres-sion in leaves and roots (Fig. 5, B and E). The higherrelative expression of maize CRT3 in roots might bebecause the plants were soil grown, whereas theArabidopsis plants were grown on liquid medium(compare Fig. 5, B with E). Earlier reports haveshown that CRT, although present in various tissuesin curled-leaf tobacco (Nicotiana plumbaginifolia;Borisjuk et al., 1998), Arabidopsis (Nelson et al.,1997), tobacco (Nicotiana tabacum; Denecke et al.,1995), and barley (Chen et al., 1994), was most abun-dant in floral tissues. Although none of the latterinvestigations have performed expression studies forthe full set of CRT isoforms, the overall expressionpatterns reported here are consistent with thesereports.

Members of the different isoform groups responddifferently to applied external stimuli. Although theCRT3 was induced already after 30 min in responseto salt or tunicamycin treatments, the CRT1 andCRT2 isoforms showed a slower induction in Arabi-dopsis. The faster response of the CRT3 isoformcould be due to an overall low expression level of thisisoform, implied by the small number of reportedESTs, and, therefore, lead to a compensatory up-regulation of CRT3. Another plausible explanationwould be that the different CRTs participate in dif-ferent regulatory pathways and would support func-tional diversity among the CRT isoforms. A several-fold induction of the CRT2 gene was reportedrecently in response to both tunicamycin and dithio-threitol in Arabidopsis (Martinez and Chrispeels,2003), supporting the up-regulation reported here.

Examining the amino acid sequences for potentialposttranslational modifications revealed that themembers from the two isoform groups in Arabidop-sis contain different numbers of negatively chargedamino acids and that there might be a difference innumbers of attached glycans in the CRTs. Here, weshow that Arabidopsis CRTs contain differences inattached N-linked glycan moieties, potentially due todifferences in numbers of N-linked glycans (Fig. 8).Both plant and mammalian CRTs can be glycosylated(Jethmalani et al., 1994; Navazio et al., 2002). Al-though the function of glycosylation of CRTs remainselusive, a potential role could be to mediate a redis-tribution of CRTs to other cellular compartments(Jethmalani et al., 1997). Furthermore, the structuralcomposition of the attached glycans also correspondsto which compartments CRT has been translocatedthrough, i.e. the complexity of the glycan is enhancedwhen modified by enzymes associated with the Golgiapparatus (Crofts et al., 1999; Pagny et al., 2000;

Navazio et al., 2002). Therefore, the structures of theglycan moieties have been used to monitor if CRTcan escape out of the ER in plants (Navazio et al.,2002). Because the glycosidase used in this study,PNGase F, removes N-linked glycans that lack a core�-(1,3) Fuc residue, it seems likely that the investi-gated CRTs did not translocate beyond the ER orcis-Golgi compartments (Pagny et al., 2000; Navazioet al., 2002).

In conclusion, together with the establishment of asecond CRT isoform (Crt2) in animals (Persson et al.,2002b), the data presented here show that two dis-tinct CRT forms are generally present in both animalsand higher plants. In addition, differences in expres-sion patterns, regulation, and posttranslational mod-ifications support multifunctional roles of CRTs.

MATERIALS AND METHODS

Computational Analysis of CRT Protein Sequences

Protein sequences corresponding to different CRT isoforms were ob-tained from the Swissprot and GenBank databases via the NCBI (http://www.ncbi.nlm.nih.gov, using BLASTP; Altschul et al., 1997). Sequenceswere compared within each species to eliminate incorrect or redundantentries. Multiple alignment of 18 CRT protein sequences was performedusing ClustalW, the MacVector 7.0 software package. The alignment wascarried out using the Blossum series matrix, with an open gap penalty of 10and an extend gap penalty of 0.05, followed by manual adjustments. Heu-ristic searches using the maximum parsimony method were performed onthe aligned sequences using the PAUP 4.0b8a software (Sinauer Associates,Inc. Publishers, Sunderland, MA), with tree bisection-reconnection branch-swapping algorithm and gaps treated as missing data. Alleloforms andprotein sequences under 100 amino acids in length were excluded from thephylogenetic analysis. The CRT amino acid sequence from Chlamydomonasreinhardtii (GenBank accession no. CAB54526) was used as the outgroup(phylum Chlorophyta). Support of the phylogeny was estimated usingbootstrap analysis of 100 replicates with heuristic searches.

The Arabidopsis CRT2 and CRT3 sequences were used to obtain corre-sponding Brassica rapa ESTs using a BLASTN analysis (Altschul et al., 1997)via the NCBI. Only ESTs with scores higher than 200 and E values lowerthan 6e-65 were considered ESTs for putative CRT2 and CRT3 isoforms.

Sequence Comparison of CRT Proteins

Protein sequences corresponding to Arabidopsis CRT1, ArabidopsisCRT2, Arabidopsis CRT3, maize (Zea mays) CRT1, maize CRT2, and maizeCRT3 (GenBank accession nos. AAC49695, AAK74014, AAC49697,CAA86728, AAF01470, and translated from nucleotide sequence AY105822,respectively), obtained from the GenBank database via the NCBI, were usedfor sequence analysis. A ClustalW multiple alignment of protein sequenceswas carried out as described above. The two conserved triplet regions,denoted I and II in the alignment (Fig. 2A), were obtained from Michalak etal. (1999) and correspond to PXXIXDPDAXKPEDWDE (three times) andGXWXPPXIXNPXYX (three times).

Exon/Intron Organization

Arabidopsis CRT exons were obtained at the NCBI. BLAST analyzesusing Arabidopsis ESTs against the Arabidopsis genome at the NCBI wereperformed to retrieve genomic clones harboring the corresponding gene.The exon/intron organization was obtained by ClustalW pair-wise align-ment of each CRT mRNA with corresponding genomic clone and subse-quent manual adjustment. The same CRT mRNA sequences were used toconfirm the genomic localization.

A rice (Oryza sativa) cDNA corresponding to CRT3 (GenBank Accessionno. AP003316) was obtained performing BLASTN analysis at NCBI with the

Persson et al.



Arabidopsis CRT3 sequence against the nonredundant database with a ricelimit. The obtained rice CRT1/CRT2 and CRT3 sequences were used toBLAST search the rice genome. Obtained genomic sequences (CRT1/CRT2,GenBank accession no. AAAA01007283; and CRT3, GenBank accession no.AAAA01001172) were aligned against available mRNA sequences and man-ually adjusted to obtain corresponding intron and exon organizations. Genemaps were drawn manually based on obtained exon and intron sizes.

Plant Material

B. rapa subsp. pekinensis plants were grown in a growth chamber with a14-h-light/10-h-dark photoperiod at 22°C and 70% relative humidity until 5weeks after germination. Arabidopsis ecotype Columbia-0 were grown onsoil with a 10-h-light/14-h-dark photoperiod at 22°C and 70% humidityuntil 6 weeks after germination. The Arabidopsis plants were then trans-ferred to a greenhouse with a 16-h-light/8-h-dark photoperiod until flow-ering was obtained. Arabidopsis roots and Arabidopsis plants for stress-related experiments were obtained from Arabidopsis ecotype Columbia-0and grown on Murashige and Skoog liquid medium with a 14-h-light/10-h-dark photoperiod at 22°C and 70% relative humidity until 6 weeks aftergermination. Arabidopsis cell cultures were maintained in 25 mL of liquidculture medium (Gamborg B5 salts, 15 g L�1 Suc, 0.1 mg L�1 2,4-dichlorophenoxyacetic acid, 1 mg L�1 kinetin, and 2 mm KH2PO4 [pH 5.7])at 22°C with gyratory shaking at 150 rpm in continuous light. Cells weresubcultured weekly with a 10% (v/v) inoculum. Maize cv Pioneer 3183plants were grown in soil in a greenhouse supplemented with lights aspreviously described (Perera et al., 1999). Six-week-old plants were used forRNA isolation. For endosperm samples, variety W64A� kernels were har-vested 18 d after pollination during the summer of 2002.

Stress Experiments

Arabidopsis plants, grown on liquid medium, were treated with either150 mm NaCl, 10 mm EGTA, 15 �g mL�1 tunicamycin, and 100 �m ABA orsubjected to drought (plants removed from liquid medium). To avoid re-sponses to fresh medium, respective treatments were added to medium thathad been changed 4 d earlier. Treatments were terminated after 2 h, andplants were harvested into liquid nitrogen and stored at �80°C. Water ormethanol (solvent for tunicamycin and ABA) were used as controltreatments.

For the Arabidopsis suspension cultures, cells were harvested 4 d afterinoculation into new medium and treated with 150 mm NaCl, 15 �g mL�1

tunicamycin, or 100 �m ABA. Flasks contained approximately 8 g of cellsper 50 mL of inoculum when treatments were initiated. Five milliliters wasremoved after 30-min and 4- and 12-h treatments from each flask. Cells wereharvested by centrifugation (250g) for 2 min, immersed into liquid nitrogen,and stored at �80°C.

Cloning and Northern-Blot Analysis

B. rapa leaves were harvested 5 weeks after germination, and total RNAwas prepared using a conventional phenol/chloroform extraction. TotalRNA was used as template for reverse transcriptase-PCR, using a primerdesigned against putative CRT3 ESTs for the 5� end (5�-ATGAGATTAA CC-CAAAACAAGC-3�) and an oligo(dT15) primer (Boehringer MannheimScandinavia AB, Bromma, Sweden) for the 3� end. A QIAGEN OneStepRT-PCR kit (QIAGEN, Merck Eurolab AB, Spånga, Sweden) was used forthe first strand synthesis and subsequent PCR step. Obtained products wereseparated on a 1.5% (w/v) agarose gel. The putative CRT3 fragment wasexcised and sequenced for identification.

Total RNA was obtained using a conventional chloroform/phenol ex-traction from either 5-week-old leaves from B. rapa or various tissues fromArabidopsis grown for 10 weeks and from Arabidopsis cell cultures. Duringtotal RNA preparation from Arabidopsis flowers, 2 mm dithiothreitol wasincluded during the extraction step. The RNA was separated using anagarose/formaldehyde gel and blotted to a Hybond N� membrane (Amer-sham Pharmacia Biotech, Uppsala). For maize, tissues were excised from 6-to 7-week-old maize plants. The upper portions of the maize plants wereharvested and frozen in liquid N2. Then, roots were removed from the soil,washed in water, excised, and frozen in liquid N2. Samples were stored at�80°C, and total RNA was extracted by using TRIzol Reagent (Invitrogen,

Carlsbad, CA). The RNAs were separated in a formaldehyde-containing gel(1.5% [w/v] agarose, 40 mm triethanolamine, and 2 mm Na2EDTA). TheRNAs were transferred from the gel to a nylon membrane (Osmonics,Minnetonka, MN) and immobilized on the membrane using UV cross-linking (UV Stratalinker, Stratagene, La Jolla, CA). cDNA clones harboringArabidopsis CRT1, CRT2, CRT3, and maize CRT1/2 were used as full-lengthprobes. A 101-nucleotide sequence corresponding to the 3� UTR for maizeCRT3 (corresponding to 5�-CTATAAAAGTCCCCAAATATTGCATTCCTC-AAAAGCATAAGCTGGAAGTTGCTTCGGACATTGTGGGTGCTTTTCAA-TAATAATAATTGATTCGCCTGGTCAGAA-3�) was obtained from MWGBiotech and used as isoform-specific probe. Because the maize CRT1 andCRT2 isoforms are 98% identical on a nucleotide level (UTRs included), wewere unable to generate specific probes for these isoforms. Probes wereradiolabeled using the Rediprime II kit (Amersham Pharmacia Biotech).Cross-hybridizations among probes were investigated by dot-blotting 100ng of either probe, or for the B. rapa control, 100 ng of the cloned CRT3isoform, onto a Hybond-N� membrane. The membranes were subsequentlyhybridized with each probe in parallel with the membranes containing theseparated RNA to be investigated. Hybridization was carried out usingExpressHyb Hybridization Solution (CLONTECH Laboratories, Palo Alto,CA), essentially according to the manufacturer’s protocol. A radiolabeledcDNA corresponding to 26S rRNA was used as a control.

Analysis of N-Linked Glycans

Arabidopsis CRT1, CRT2, and CRT3 protein sequences were analyzedusing the MacVector 7.0 software. Six-week-old greenhouse-grown Arabi-dopsis plants were homogenized using a tight-fitting glass-glass homoge-nizer in 200 mm Suc, 25 mm HEPES-KOH (pH 7.0), 3 mm EGTA, 1 mmMgSO4, 1 mm phenylmethylsulfonyl fluoride, and 1 mm dithiothreitol. Thehomogenate was centrifuged at 1,000g for 10 min at 4°C, and the superna-tant was used for glycosidase analysis, SDS-PAGE, and immunoblotting.Recombinant PNGase F (New England Biolabs, Beverly, MA) was used forthe N-linked glycosidase treatment. Deglycosidation was carried out essen-tially according to the manufacturer’s protocol under native conditions.

SDS-PAGE and Immunoblotting

The Arabidopsis homogenate was solubilized by the addition of one-third of 3.33� sample buffer (250 mm Tris-HCl [pH 6.8], 6% [w/v] SDS, 33%[v/v] glycerol, 15% [v/v] �-mercaptoethanol, and 0.02% [w/v] bromphenolblue). Equal amounts of solubilized proteins were separated on a 10% (w/v)Laemmli SDS-polyacrylamide gel. Gels were either stained with CoomassieBrilliant Blue or transferred to a hydrophobic polyvinylidene fluoride mem-brane (Immobilon-P Transfer membrane, Millipore Corporation, Bedford,MA). Proteins were wet blotted at 100 V for 1 h. After transfer, the blottingmembranes were blocked with 4% (w/v) blocking reagent (Bio-Rad Labo-ratories, Hercules, CA) in Tris-buffered saline with 0.2% (v/v) Tween 20 for1 h at room temperature. The membranes were incubated with either anantiserum against CRT from maize diluted 1:5,000 (v/v) or an antiserumagainst calnexin from Arabidopsis diluted 1:1000 (v/v). Polyclonal antibod-ies were raised against a purified deglycosylated maize CRT, essentiallydescribed by Pagny et al. (2000). Immunodecoration was visualized withchemiluminescent detection of horseradish peroxidase according to the ECLwestern detection reagent protocol (Amersham Pharmacia Biotech).

Distribution of Materials

Upon request, all novel materials described in this publication will bemade available in a timely manner for noncommercial research purposes.

ACKNOWLEDGMENTS

We thank Drs. Donald P. Shepley and Hans Bohnert for supplying uswith the Arabidopsis CRT1 and CRT3 clones. We thank Dr. Neil E. Hoffmanfor supplying the Arabidopsis calnexin antibodies, Dr. Steve Huber forsupplying the maize plants, Drs. Rebecca S. Boston and Jeff Gillikin for themaize endosperm. and Dr. Eric Ruelland for supplying us with Arabidopsissuspension cells. We also thank Mrs. Adine Karlsson for supplying hydro-




ponic Arabidopsis material and Mr. Magnus Alsterfjord and Drs. UrbanJohanson and Jenny Xiang for valuable suggestions.

Received April 7, 2003; returned for revision June 5, 2003; accepted August1, 2003.

LITERATURE CITED

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W,Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new generation ofprotein database search programs. Nucleic Acids Res 25: 3389–3402

Arnaudeau S, Frieden M, Nakamura K, Castelbou C, Michalak M, De-maurex N (2002) Calreticulin differentially modulates calcium uptakeand release in the endoplasmic reticulum and mitochondria. J Biol Chem277: 46696–46705

Baksh S, Michalak M (1991) Expression of calreticulin in Escherichia coli andidentification of its Ca2� binding domains. J Biol Chem 266: 21458–21465

Baldan B, Navazio L, Friso A, Mariani P, Meggio F (1996) Plant calreticulinis specifically and efficiently phosphorylated by protein kinase CK2.Biochem Biophys Res Commun 221: 498–502

Baluska F, Samaj J, Napier R, Volkmann D (1999) Maize calreticulinlocalizes preferentially to plasmodesmata in root apex. Plant J 19: 481–488

Borisjuk N, Sitailo L, Adler K, Malysheva L, Tewes A, Borisjuk L, Man-teuffel R (1998) Calreticulin expression in plant cells: developmentalregulation, tissue specificity and intracellular distribution. Planta 206:504–514

Bremer K (2002) Gondwanan evolution of the grass alliance of families(Poales). Evol Int J Org Evolution 56: 1374–1387

Camacho P, Lechleiter JD (1995) Calreticulin inhibits repetitive intracellularCa2� waves. Cell 82: 765–771

Chen F, Hayes PM, Mulrooney DM, Pan A (1994) Identification andcharacterization of cDNA clones encoding plant calreticulin in barley.Plant Cell 6: 835–843

Corbett EF, Michalak KM, Oikawa K, Johnson S, Campbell ID, EggletonP, Kay C, Michalak M (2000) The conformation of calreticulin is influ-enced by the endoplasmic reticulum luminal environment. J Biol Chem275: 27177–27185

Coughlan SJ, Hastings C, Winfrey R Jr (1997) Cloning and characterizationof the calreticulin gene from Ricinus communis L. Plant Mol Biol 34:897–911

Crofts AJ, Denecke J (1998) Calreticulin and calnexin in plants. Trends PlantSci 3: 396–399

Crofts AJ, Leborgne-Castel N, Hillmer S, Robinson DG, Phillipson B,Carlsson LE, Ashford DA, Denecke J (1999) Saturation of the endoplas-mic reticulum retention machinery reveals anterograde bulk flow. PlantCell 11: 2233–2248

Denecke J, Carlsson LE, Vidal S, Hoglund AS, Ek B, van Zeijl MJ,Sinjorgo KM, Palva ET (1995) The tobacco homolog of mammaliancalreticulin is present in protein complexes in vivo. Plant Cell 7: 391–406

Eggleton P, Llewellyn DH (1999) Pathophysiological roles of calreticulin inautoimmune disease. Scand J Immunol 49: 466–473

Goff SA, Ricke D, Lan TH, Presting G, Wang R, Dunn M, Glazebrook J,Sessions A, Oeller P, Varma H et al. (2002) A draft sequence of the ricegenome (Oryza sativa L. ssp. japonica). Science 296: 92–100

Gomord V, Wee E, Faye L (1999) Protein retention and localization in theendoplasmic reticulum and the Golgi apparatus. Biochimie 81: 607–618

Hadlington JL, Denecke J (2000) Sorting of soluble proteins in the secretorypathway of plants. Curr Opin Plant Biol 3: 461–468

Hassan AM, Wesson C, Trumble WR (1995) Calreticulin is the major Ca2�

storage protein in the endoplasmic reticulum of the pea plant (Pisumsativum). Biochem Biophys Res Commun 211: 54–59

Hebert DN, Foellmer B, Helenius A (1996) Calnexin and calreticulin pro-mote folding, delay oligomerization and suppress degradation of influ-enza hemagglutinin in microsomes. EMBO J 15: 2961–2968

Heilmann I, Shin J, Huang J, Perera IY, Davies E (2001) Transient disso-ciation of polyribosomes and concurrent recruitment of calreticulin andcalmodulin transcripts in gravistimulated maize pulvini. Plant Physiol127: 1193–1203

Jaubert S, Ledger TN, Laffaire JB, Piotte C, Abad P, Rosso MN (2002)Direct identification of stylet secreted proteins from root-knot nematodesby a proteomic approach. Mol Biochem Parasitol 121: 205–211

Jethmalani SM, Henle KJ, Gazitt Y, Walker PD, Wang SY (1997) Intracel-lular distribution of heat-induced stress glycoproteins. J Cell Biochem 66:98–111

Jethmalani SM, Henle KJ, Kaushal GP (1994) Heat shock-induced promptglycosylation: identification of P-SG67 as calreticulin. J Biol Chem 269:23603–23609

Johnson S, Michalak M, Opas M, Eggleton P (2001) The ins and outs ofcalreticulin: from the ER lumen to the extracellular space. Trends CellBiol 11: 122–129

Kwiatkowski BA, Zielinska-Kwiatkowska AG, Migdalski A, Klecz-kowski LA, Wasilewska LD (1995) Cloning of two cDNAs encodingcalnexin-like and calreticulin-like proteins from maize (Zea mays) leaves:identification of potential calcium-binding domains. Gene 165: 219–222

Li Z, Komatsu S (2000) Molecular cloning and characterization of calreti-culin, a calcium-binding protein involved in the regeneration of ricecultured suspension cells. Eur J Biochem 267: 737–745

Martinez IM, Chrispeels MJ (2003) Genomic analysis of the unfoldedprotein response in Arabidopsis shows its connection to important cel-lular processes. Plant Cell 15: 561–576

Matsuoka K, Seta K, Yamakawa Y, Okuyama T, Shinoda T, Isobe T (1994)Covalent structure of bovine brain calreticulin. Biochem J 298: 435–442

Michalak M, Corbett EF, Mesaeli N, Nakamura K, Opas M (1999) Calre-ticulin: one protein, one gene, many functions. Biochem J 344: 281–292

Nakamura K, Zuppini A, Arnaudeau S, Lynch J, Ahsan I, Krause R, PappS, De Smedt H, Parys JB, Muller-Esterl W et al. (2001) Functionalspecialization of calreticulin domains. J Cell Biol 154: 961–972

Navazio L, Baldan B, Dainese P, James P, Damiani E, Margreth A, MarianiP (1995) Evidence that spinach leaves express calreticulin but not calse-questrin. Plant Physiol 109: 983–990

Navazio L, Miuzzo M, Royle L, Baldan B, Varotto S, Merry AH, HarveyDJ, Dwek RA, Rudd PM, Mariani P (2002) Monitoring endoplasmicreticulum-to-Golgi traffic of a plant calreticulin by protein glycosylationanalysis. Biochemistry 41: 14141–14149

Nelson D, Glaunsinger B, Bohnert HJ (1997) Abundant accumulation of thecalcium-binding molecular chaperone calreticulin in specific floral tissuesof Arabidopsis thaliana. Plant Physiol 114: 29–37

Pagny S, Cabanes-Macheteau M, Gillikin JW, Leborgne-Castel N, LerougeP, Boston RS, Faye L, Gomord V (2000) Protein recycling from the Golgiapparatus to the endoplasmic reticulum in plants and its minor contri-bution to calreticulin retention. Plant Cell 12: 739–756

Perera IY, Heilmann I, Boss WF (1999) Transient and sustained increases ininositol-1,4,5-trisphosphate precede the differential growth response ingravistimulated maize pulvini. Proc Natl Acad Sci USA 96: 5838–5843

Persson S, Love J, Tsuo PL, Robertson D, Thompson WF, Boss WF (2002a)When a day makes a difference: interpreting data from endoplasmicreticulum-targeted green fluorescent protein fusions in cells grown insuspension culture. Plant Physiol 128: 341–344

Persson S, Rosenquist M, Sommarin M (2002b) Identification of a novelcalreticulin isoform (Crt2) in human and mouse. Gene 297: 151–158

Persson S, Wyatt SE, Love J, Thompson WF, Robertson D, Boss WF (2001)The Ca2� status of the endoplasmic reticulum is altered by induction ofcalreticulin expression in transgenic plants. Plant Physiol 126: 1092–1104

Saito Y, Ihara Y, Leach MR, Cohen-Doyle MF, Williams DB (1999) Calre-ticulin functions in vitro as a molecular chaperone for both glycosylatedand non-glycosylated proteins. EMBO J 18: 6718–6729

Soltis PS, Soltis DE, Chase DW (1999) Angiosperm phylogeny inferredfrom multiple genes as a tool for comparative biology. Nature 402:402–404

Vision TJ, Brown DG, Tanksley SD (2000) The origins of genomic dupli-cations in Arabidopsis. Science 290: 2114–2117

von Heijne G (1985) Signal sequences, The limits of variation. J Mol Biol184: 99–105

Wyatt SE, Tsou PL, Robertson D (2002) Expression of the high capacitycalcium-binding domain of calreticulin increases bioavailable calciumstores in plants. Transgenic Res 11: 1–10

Yu J, Hu S, Wang J, Wong GK, Li S, Liu B, Deng Y, Dai L, Zhou Y, ZhangX et al. (2002) A draft sequence of the rice genome (Oryza sativa L. ssp.indica). Science 296: 79–92

Persson et al.



phylogenetic analyses and expression studies reveal two · phylogenetic analyses and expression...

Documents