Plant Molecular Biology 33: 625–633, 1997. 625c 1997 Kluwer Academic Publishers. Printed in Belgium.

Differential accumulation of two glycine-rich proteins duringcold-acclimation alfalfa

Jean-Marc Ferullo, Louis-P. Vezina�, Jimmy Rail, Serge Laberge, Paul Nadeau andYves CastonguayAgriculture et Agroalimentaire Canada, Station de Recherches sur les sols et les grandes cultures, 2560 BdHochelaga, Ste-Foy, Quebec, Canada G1V 2J3 (�author for correspondence)

Received 21 December 1995; accepted in revised form 28 October 1996

Key words: alfalfa, cold acclimation, frost tolerance, glycine-rich proteins

Abstract

Two mRNAs, MsaCiA and MsaCiB, encoding for proteins harboring glycine-rich motifs, accumulate in alfalfaduring cold acclimation. Fusion polypeptides containing the amino acid sequences deduced from these mRNAswere produced in Escherichia coli and used to raise antibodies. Each antibody cross-reacted specifically with solublepolypeptides, MSACIA-32 and MSACIB, respectively. These polypeptides were detectable only in crowns of cold-acclimated plants, even though MsaCiA mRNA accumulated in both crows and leaves during cold acclimation. Theanalysis of parietal proteins showed that several MSACIA-related proteins, with a molecular mass of 32, 41 and68 kDa, did accumulate in leaf cell walls and one of 59 kDa crown cell walls. This diversity is most probably dueto a tissue-specific maturation of MSACIA. A discrepancy was found between the time-course of accumulation ofMSACIB and the one of the corresponding transcript. These results indicate that timing and localization of MSACIAand MSACIB expression are different, and suggest that this differential expression involves both transcriptionaland post-transcriptional events. Comparisons made among six cultivars of contrasting freezing tolerance suggestthat low tolerance could be explained by failure to accumulate proteins like MSACIA and MSACIB at a sufficientlevel.

Introduction

When they are exposed to low temperatures, plantsexhibit physiological adaptations that result in anincreased freezing tolerance [17]. This process,referred to as cold acclimation, is characterized bybiochemical changes in plant tissues, including alter-ations in carbohydrate, protein and lipid composi-tion [11]. Simultaneously, a qualitative modificationof gene expression also takes place during acclima-tion, leading to the accumulation of newly synthes-ized mRNAs [16]. Many of these mRNAs encodeglycine-rich proteins (GRPs) [1, 4, 13, 20, 21, 28, 29],which can be subdivided into several classes, on thebasis of their amino acid sequence. Some of the GRPsare thought to play a structural role in the cell walls[24], others are characterized by the presence of putat-ive RNA-binding regions [7, 25], while the group of

dehydrins, homologuous to stress and/or development-ally regulated proteins, are thought to protect plasmamembranes during stresses that involve an increase inosmotic pressure [6].

In alfalfa (Medicago sativa L. cv. Apica), MsaCiAand MsaCiB cDNAs, encoding for GRPs, have recentlybeen isolated by differential hybridization of a �gt10cDNA library prepared from cold-acclimated plants[15, 19]. The polypeptides deduced from MsaCiA andMsaCiB sequences share common characteristics, forexample a high content of glycine and histidine, theabsence of proline, tryptophane and cysteine, and thepresence of repeated motifs. However, differences intheir primary structure strongly suggest that they playdistinct roles. The putative protein encoded by MsaCiAis a modular polypeptide of 204 amino acid residues,consisting in a series of repeated tri- to hexapeptidesof the G2–4Y0–1N0–1H0–1 type, bordered at the NH2

626

terminus by a hydrophobic peptide potentially involvedin extracellular targeting, interrupted in the middleand bordered at the COOH terminus by 2 sequencescontaining an AVQTE motif [15]. Its amino-acidsequence is typical of plant cell wall glycine-rich pro-teins [14]. In the polypeptide encoded by MsaCib, theglycine residues are more equally distributed amongits sequence of 136 amino-acid residues although oneGEQHGxyGzG motif is repeated 4 times (x, y and z areF, H, V or none), and a putative nuclear targeting signalis present at the COOH terminus [19]. This sequenceis not typical of any known class of glycine-rich pro-teins described up to now, but shares homology withcDNA clones isolated from cold-acclimated Poncirustrifoliata [3].

In order to verify whether these proteins are actu-ally synthesized in alfalfa tissues and to monitor theiraccumulation during cold acclimation, polyclonal anti-bodies were raised against polypeptides containingthe entire putative coding sequences of MsaCiA andMsaCiB. The accumulation of MSACIA and MSA-CIB proteins in crowns and leaves was comparedto the levels of their respective mRNAs. To assessthe potential relationship between the accumulationof MSACIA/MSACIB proteins and freezing toler-ance, six alfalfa cultivars that exhibit contrastingwinter-hardiness were tested for their content in Msa-CiA/MsaCiB transcripts and proteins after cold accli-mation.

Materials and methods

Production of antibodies

The coding sequences of MsaCiA and Msa-CiB cDNAs were first amplified using the poly-merase chain reaction (PCR). Primers for PCR weredeveloped as follows: primers 50-GATCCCCGGGAT-GGATTCGAAAAAGGC-30 and 50-GATCGAATTCT-CAGTTTTGAGTGTTGT-30 anneal to the Msa-CiA cDNA template at nucleotides 33 and469, respectively, and introduce respectively aSmaI site and an EcoRI site; primers 50-GAT-CCCCGGGATGGCAGGAATCATGAA-30 and 50-GACTGAATTCCTAATCACTGTCACTGC-30 annealto MsaCiB cDNA template at nucleotides 53 and 447,respectively, and introduce respectively a SmaI siteand an EcoRI site. These restriction sites have beenchosen because they are absent from MsaCiA andMsaCiB cDNA coding sequences. The PCR products

were digested with Sma I and EcoRI, ligated into thepUC18 vector, and transferred into Escherichia coliDH5 �. Recombinants were selected on Luria-Bertani(X-gal/ampicillin/IPTG) medium. The integrity of thecloned fragments was determined by DNA sequenceanalysis (T7-Sequenase kit, Pharmacia). Fragmentswere subcloned into the pMAL-c2 (New England Bio-labs) expression vector between sites XmnI (com-patible with SmaI) and EcoRI, and transferred toDH5�. The transformants were selected on Luria-Bertani medium by replica plating with and withoutX-Gal/IPTG.

The fusion polypeptides MBP/MSACIA andMBP/MSACIB were obtained by overexpression ofthese plasmids in DH5 �. Sterile rich medium(1% (w/v) bacto-tryptone, 0.5% (w/v) yeast extract,1% (w/v) NaCl, 20 mM D-sucrose), containing100 �g/l ampicillin, was inoculated with 1/100 volof fresh overnight culture, and cultivated at 37 �Cunder vigourous shaking for 3 h. Isopropyl �-D-thiogalactopyranoside was added to the culture at afinal concentration of 0.3 mM and cultivation was pro-longed for 3 additional hours. Bacteria were harvestedby centrifugation (4000g, 20 min, 4 �C), resuspen-ded in 1/10 vol of column buffer (100 mM Tris-HCl,50 mM NaCl, pH 7.4) and frozen overnight at�20 �C.The thawed suspension was sonicated at 0 �C. Duringsonication, cell breakage was monitored by measur-ing protein release in the medium with a commercialBradford assay kit (Bio-Rad). The cell debris werepelleted by centrifugation (9000 � g, 30 min, 4 �C).The clarified supernatant was mixed with 1/100 vol ofamylose matrix (New England Biolabs) and incubatedfor 1 h at 4 �C under moderate stirring. The matrix wascollected on a sintered glass filter and washed with 10vol of column buffer (10 mM Tris-Cl pH 8.0, 25 mMNaCl, 0.1% 2-mercaptoethanol). The fusion proteinwas eluted from the matrix with 1 vol of column buf-fer containing 20 mM maltose. Rabbits were immun-ized against fusion proteins by intramuscular injectionof 0.5 ml of complete Freund’s adjuvant containing100 �g of protein, followed by a second injection, 21days later, of 0.5 ml of incomplete Freund adjuvantcontaining 50 �g of protein. The serum was harvested45 days after the first injection and kept aliquoted at�80 �C.

627

Plant material and cold acclimation, overacclimationand deacclimation treatments

Alfalfa plants (Medicago sativa L.) were grown fromseeds in 15 cm diameter plastic pots (15 plantsper pot) filled with soil, watered daily and fertil-ized weekly with a nutritive solution (20:20:20 g/lN/P/K) under the following environmental conditions:21 �C/17 �C (day/night) temperature, 16 h photoperi-od and 350 �mol m�2 s�1 photosynthetic photon fluxdensity. After 5 weeks of growth, plants were cold-acclimated at 2 �C for 2 weeks under a 8 h photoperiodand 150 �mol m�2 s�1 photosynthetic photon fluxdensity. A group of cold-acclimated plants was trans-ferred to initial growth conditions (deacclimation treat-ment), while another group was subsequently incub-ated at�2 �C during 15 days under the light conditions.This treatment is herein reffered to as overacclimationtreatment, since this proved to further stimulate frosttolerance in alfalfa cv. Apica [4].

Crowns and leaves of at least 10 randomly chosenplants were sampled at intervals throughout the growthand acclimation period, ground into a fine powder inliquid nitrogen with a mortar and a pestle, and kept at�80 �C until use.

Analysis of RNA

Total RNA was extracted according to de Vries et al.[8] with slight modifications. Powdered material (5 g)was mixed with 10 ml of a pre-heated (90 �C 1:1 (v/v)mixture of phenol containing 0.1% (w/v) 8-hydroxy-quinoline (pre-equilibrated with a solution of 200 mMTris-HCl pH 8.0, 100 mM LiCl, 5 mM EDTA) and100 mM Tris-HCl pH 9.0 containing 1% (w/v) SDS,100 mM LiCl, 10 mM EDTA. The sample was stirredfor 15 min at room temperature. A 5 ml portion ofchloroform was added and the sample was stirred foran additional 15 min. Extracts were centrifuged at20 000� g (25 �C, 30 min). The aqueous phase wasremoved, reextracted with chloroform as describedabove, and brought to 2 M LiCl. RNA was allowedto precipitate for 16 h at 4 �C. After centrifugationat 12 000 � g (4 �C 30 min), the RNA pellet waswashed with 2 M LiCl and twice with 80% (v/v) eth-anol, vacuum-dried, dissolved in TE (10 mM Tris-HClpH 7.4, 1 mM EDTA), and stored at �80 �C. RNAcontent was estimated by UV absorption at 260 nm.

For northern blot analysis, 15 �g of total RNA wasseparated by electrophoresis on denaturing agarosegels [9], vacuum-transferred onto Hybond-N mem-

branes (Amersham), and hybridized overnight at 68 �Cin 2� SSC containing 0.25% (w/v) skimmed powderedmilk. The gel-purified cDNA inserts of MsaCiA andMsaCiB (50 ng) were radiolabelled by random priming(Pharmacia oligolabelling kit) using [��32]P-dCTP(3000 Ci/mmol), and used as probes for hybridiza-tion. Membranes were washed 5 times with 2� SSCand autoradiographed in �80 �C on Kodak X-o-MatXAR film and intensifying screens.

Analysis of proteins

Powdered material (0.5 g) was homogenized in 4 mlof protein extraction buffer (50 mM Tris-HCl pH 8.0containing 0.04% (v/v) 2-mercapto-ethanol and 6 mMPMSF), with an Ultra-turrax homogenizer for 30 s atmaximum speed. The homogenate was filtered through2 layers of Miracloth and centrifuged at 20 000� g for30 min. One volume of 2� loading buffer (Tris-HCl100 mM pH 6.8 containing 20% (v/v) glycerol 0.1%SDS and 0.01% (w/v) bromophenol blue) was addedto 100 �l of supernatant. The sample was heated at100 �C for 2 min and centrifuged 2 min at 10 000� g.A 10 �l portion was loaded on a 12.5% acrylamide gelfor SDS-PAGE separation (Mini-Protean II apparatus,Bio-Rad). Gels were then stained with Coomassie blueor electrotransferred onto nitrocellulose membranes.For immunodetection, membranes were blocked for1 h in M-PBS (137 mM NaCl, 3 mM KCl, 8 mMNa2HPO4, 1.5 mM KH2PO4 pH 7.4, 2.5% (w/v)skimmed powdered milk), incubated 1 h in M-PBScontaining the rabbit antiserum at a 10�3 dilution,washed 5 times with T-PBS (M-PBS without milk andwith 0.05% (v/v) Tween 20), incubated 1 h in M-PBScontaining 125I-labelled donkey anti-rabbit antibody(Amersham, 590 kBq/�g) at a 10�3 dilution, washed 5times with T-PBS and autoradiographied at�80 �C onKodak X-o-Mat XAR film with intensifying screens.

For cell-wall protein extraction, 0.5 g of poweredmaterial was suspended into 5 ml of protein extractionbuffer, and centrifuged 10 min at 20 000�g. The super-natant was discarded and the pellet re-extracted 6 timesas above. The protein content in the last supernatant,measured with Bio-Rad protein assay, was < 5% ofthe protein content in the first supernatant. The pel-let was suspended in 5 ml of a CaCl2 solution (mix-ture 3:2 of protein extraction buffer and 0.5 M CaCl2),incubated under moderate shaking for 45 min at roomtemperature and centrifuged 15 min at 20 000�g. Thesupernatant was isolated and the pellet was extracted asecond time with the CaCl2 solution. The two super-

628

natant were pooled and 4 vol of acetone at �20 �Cwere added. Proteins were precipitated overnight at�20 �C, and pelleted by centrifugation at 20 000� g

for 20 min. The pellet was washed with 80% acetone,vacuum-dried, and resuspended in 500 �L of proteinextraction buffer. Samples for SDS-PAGE were pre-pared as above.

Results

Preparation of fusion polypeptides

The heterologous expression of the pMAL-MsaCiAconstruct in E. coli yielded ca. 8 mg of a 65 kDapolypeptide per litre of media. Affinity purification onamylose matrix yielded a protein mixture containing atleast 95% of the 65 kDa polypeptide (Fig. 1). Attemptsto separate the 65 kDa polypeptide from the minorpolypeptides by a second affinity chromatography orby ion-exchange chromatography were unsuccessful.

With the pMAL-MsaCiB construct, a 60 kDa poly-peptide accumulated in E. coli at a concentration of ca.0.2 mg/l (Fig. 1). This polypeptide was efficiently puri-fied by affinity chromatographyon the amylose matrix,with no detectable contaminants.

In both cases, eluates from the amylose matrix wereused without further purification to raise antisera. In thefollowing text, these antisera are designated as ‘anti-MBP/MSACIA’ and ‘anti-MBP/MSACIB’, respect-ively.

Accumulation of MsaCiA and MsaCiB transcripts inplant tissues and identification of the correspondingprotein

Northern blot analyses showed that the MsaCiA tran-script accumulated at a high level in both crowns andleaves during acclimation (Fig. 2). MsaCiB accumu-lated only in crowns, but was present at a low basallevel in acclimated and nonacclimated leaves. Anti-MBP/MSACIA and anti-MBP/MSACIB antisera wereused for immunodetection of protein from leaves andcrowns from control plants, and from plants acclimated2 �C for 15 days. Since antisera were raised againstfusion polypeptides that contained the MBP moiety,the results were compared with immunodetections per-formed with a commercial anti-MBP antiserum. Itwas shown that the three antibodies gave equal cross-reactivity with MBP, and that the anti-MBP antiserum

did not cross-react with any protein from the plantextracts (Fig. 2).

Immunodetection with anti-MBP/MSACIA re-vealed a broad, intense band of ca. 32 kDa, thatwas detectable only in soluble extracts from cold-acclimated crowns. Herein after MSACIA-32 desig-nates this soluble protein of 32 kDa. The antibodyrevealed also a diffuse signal (multiple band poorlyresolved) present in extracts from cold-acclimatedleaves and crowns. No signal was visible in extractsfrom unacclimated materials. To understand the natureof the multiple bands and why only a faint, diffusesignal was detected in leaves while the MsaCiA tran-script accumulated as much as in crowns, we tested thehypothesis that MSACIA might undergo differentialmaturation that could lead to an increase in molecularmass and to the insolubilization of the protein into thecell wall. Cell wall proteins were isolated from unac-climated and acclimated materials. Immunodetectionwith anti-MBP/MSACIA revealed 3 proteins from leafcell-walls (one broad band at ca. 32 kDa, and two sharpbands at 41 kDa and 68 kDa) and 1 protein in crown cellwalls (59 kDa) that cross-reacted with anti-MSACIAand that were induced by cold acclimation (Fig. 3).

Anti-MBP/MSACIB revealed a sharp protein bandof 23 kDa that was detectable only in crown extractsand highly enhanced after cold acclimation. In theabsence of MSACIB in leaves, the kinetics of MSA-CIA/B accumulation during cold acclimation wereperformed on crowns only. Moreover, since pariet-al MSACIA-related proteins represented at most 1/40of MSACIA-32 (based on immunodetection signals),only MSACIA-32 was considered for kinetics experi-ments.

Kinetics of accumulation of MSACIA-32 and MSACIBin crowns during acclimation, overacclimation anddeacclimation

The accumulation of MSACIA-32 and MSACIB incrowns during cold acclimation were monitored byimmunodetection and compared to the accumulationtime-course of their respective mRNAs during coldacclimation and deacclimation treatments (Fig. 4).Proteins were extracted from crowns at days 2, 4,8 and 15 of the acclimation period, at day 15 ofthe subsequent overacclimation period, and at days1, 2 and 5 of the deacclimation period. SDS-PAGEanalysis of protein extracts revealed no significantvariation in total protein content between samples(Fig. 4). Immunodetections with anti-MBP/MSACIA

629

Figure 1. SDS-PAGE of proteins from E. coli containing pMAL-c2/MsaCiA (A) and pMAL-c2/MsaCiB (B). NI, before induction; I, 3 h afterinduction with IPTG; P, purified; PC, purified and concentrated 10-fold.

Figure 2. A. Northern blots of leaf and crown total RNA extracts probed with MsaCiA and MsaCiB. B. Immunodetections of proteins separatedby SDS-PAGE with anti-MBP, anti-MBP/MSACIA and anti-MBP/MSACIB, respectively. MBP, maltose-binding protein; LC, leaf, control;LA, leaf, cold-acclimated; CC, crown, control; CA, crown, cold-acclimated.

Figure 3. Analysis of cell wall proteins. CaCl2 extraction and immunodetection were performed as described in Materials and methods, LC,leaf, control; LA, leaf, cold-acclimated; CC, crown, control; CA: crown, cold-acclimated.

630

and anti-MBP/MSACIB revealed that the levels ofboth MSACIA-32 and MSACIB increased progress-ively between day 2 and day 15 of the acclimationperiod and were unchanged after the overacclimationtreatment. The accumulation time-course of MsaCiAwas similar to the one of MSACIA/B, whereas MsaCiBaccumulated rapidly within the first 4 days of acclima-tion, progressively decreased to initial level betweenday 4 and day 15, and accumulated again during theoveracclimation treatment.

During the deacclimation period, the levels of Msa-CiA/B transcripts decreased rapidly within 1 day, asdid the level of MSACIA-32. The level of MSACIBdecreased only slightly within the 5 days of deac-climation, as did the diffuse signal revealed by anti-MBP/MSACIA.

Comparison between cultivars

The levels of MSACIA-32, MSACIB and corres-ponding transcripts were assessed in cold-acclimatedcrowns of 6 alfalfa cultivars that exhibit contrastingwinterhardiness: CUF 101 and Moapa 69 (sensitive),Iroquois and Apica (hardy), and spredor and Rambler(very hardy) [5] (Fig. 5). No significant variation wasfound in the total protein and RNA contents betweenall cultivars (data not shown). The lowest levels ofboth transcripts and proteins for MsaCiA and MsaCiBwere obtained with the cold-sensitive cultivars CUF101 and Moapa 69. In the other cultivars, the levels ofMsaCiA/MsaCIB transcripts and proteins were higherexcept for MsaCiA transcript in Spredor. In all cases,the maximum level was observed in Apica.

Discussion

The production of fusion polypeptides with the pMALexpression vector has allowed us to raise 2 antiserathat specifically cross-reacted with alfalfa proteins thataccumulate upon cold acclimation. It is worth noticingthat the expression of MBP/MSACIA yielded about40-fold more heterologous protein than the expres-sion of MBP-MSACIB. Moreover, the induction ofthe expression of MBP/MSACIB by IPTG was leth-al for the bacteria. This indicates that although bothpolypeptides share common characteristics, they havedifferent biochemical properties, at least in E. coli.Although the antigene used for the preparation ofanti-MBP/MSACIA yielded multiple bands on elec-trophoresis in denaturating conditions, the resulting

antiserum was specific for proteins that accumulatein alfalfa during cold acclimation. The impossibilityto separate MBP/MSACIA from contaminants by ion-exchange chromatography (Fig. 1) suggests that thesepeptides most probably arise from a degradation ofMBP/MSACIA upon sample preparation for electro-phoresis.

The theorical molecular masses of MSACIA andMSACIB proteins, deduced from their correspondingnucleotide sequence, are 21.8 and 14.5 kDa, respect-ively, whereas the respective molecular masses of theproteins detected with anti-MBP/MSACIA and anti-MBP/MSACIB in alfalfa extracts are 32 (major band)and 23 kDa, respectively. Such discrepancies concern-ing the molecular masses of proteins derived from cold-induced genes have been reported by several authors[10, 13, 18, 21, 22, 27]. This suggests that these pro-teins might undergo post-translational modifications.However, the amino-acid sequences of MSACIA andMSACIB contain no potential sites of N-glycosylation,and MSACIB possesses no apparent signal for endo-plasmic reticulum targeting. Moreover, the MSACIBmoiety resulting from the cleavage of MBP/MSACIBwith factor Xa had the same apparent size as the plantprotein (data not shown). An alternative explanationof this discrepancy between theoretical and apparentMWs has been proposed, invoking the high hydropili-city of GRPs which would result in strong avidity forSDS, and thus in a increase in molecular weight onSDS-PAGE [13, 27].

Immunodetection of plant extracts with anti-MBP/MSACIA reveals a major, 32 kDa polypeptide(MSACIA-32), plus a diffuse signal that seems to cor-repond to multiple bands of higher molecular mass.The nature of these bands is unknown. They may cor-respond either to other genes from MsaCiA family orto post-translational modifications of MSACIA. Thesecond hypothesis seems more probable since (1) allcDNAs isolated with MsaCiA as probe were of equalor slightly shorter size than MsaCiA, and (2) amplific-ation of genomic DNA with primers that annealed withMsaCiA cDNA extremities gave only one product ofthe same size as MsaCiA (data not shown).

During cold acclimation, MSACIA accumulatesonly in crowns, while the level of the correspondingtranscript increases in both leaves and crowns. Thisobservation could be related to a lower rate of messen-ger translation and/or a higher rate of protein turnoverin leaves. Nevertheless, the presence of a faint, diffusesignal in western blot analyses from cold acclimatedleaves (Fig. 2) led us to raise the hypothesis that MSA-

631

Figure 4. Northern and western blots of RNA and protein extracts from alfalfa crowns. A0, A2, ..., A15, after 0, 2, ..., 15 days at 2 �C; F15,after 15 days at 2 �C and 15 days at �2 �C; D1, D2, D5, after 15 days at 2 �C and 1, 2 and 5 days at 21/17 �C; MsaCiA, anti-MBP/MSACIAused for immunodetection and MsaCiA cDNA probe for northern blot; MsaCiB, anti-MBP/MSACIB used for immunodetection and MsaCiBcDNA probe for northern blot.

CIA might be differentially processed in leaves andin crowns. The band at 32 kDa could correspond toan immature form of the protein, which, in its matureform, would no longer be separated as a single band onSDS-PAGE. Since MSACIA is probably an extracel-lular GRP, it may become insoluble following cross-linking with other cell wall components [14]. There-fore, the absence of induction of MSACIA-32 in leafextracts might result from a tissue-specific maturation.The analysis of cell wall proteins supports this hypo-thesis, since proteins immunologically related to MSA-CIA were found in walls of both acclimated leaves andcrowns (Fig. 3). The diversity of the proteins present incell walls suggests that this tissue specificity of matura-tion would be not only quantitative,but also qualitative.These hypotheses will be tested by further character-ization of the multiple bands.

During acclimation and overacclimation treat-ments, MsaCiA and MsaCiB transcripts show dif-

ferential kinetics of accumulation in alfalfa crowns:the accumulation of MsaCiA transcript is sustainedthroughout acclimation, whereas the level of MsaCiBtranscript increases during the first 4 days of accli-mation and then decreases to the initial level. Sub-sequent overacclimation does not affect the amountof MsaCiA transcript, while MsaCiB transcript re-accumulates to the maximum observed after 4 daysof acclimation. Such a differential response indicatesthat the regulation of the expression of MsaCiA andMsaCiB genes by low temperatures follows distinctelicitation pathways. The existence of several, inde-pendent mechanisms of induction of gene expressionby cold has been already proposed by other authors toexplain differential kinetics of induction or differentialresponses to ABA or drought [18, 22, 23, 26]. Whilethe accumulation profile of MSACIA is closely relatedto that of the corresponding transcript, MSACIB levelevolves much slower compare to the rapid increase and

632

Figure 5. Northern and western blots of RNA and protein extracts from acclimated crowns from alfalfa cultivars of constrasting winter-hardiness.CUF, Cuf 101; Moa, Moapa 69; Iro, Iroquois; Api, Apica; Spr, Spredor; Ram, Rambler; MsaCiA, anti-MBP/MSACIA used for immunodetectionand MsaCiA cDNA probe for northern blot; MsaCiB, anti-MBP/MSACIB used for immunodetection and MsaCiB cDNA probe for northernblot.

decrease in the corresponding transcript. Such differ-ences in turnover rates for low temperature-inducedproteins have been observed with BN28 in Brassicanapus [2], compare to CAP85 [12, 21] and CAP160[12] in spinach. The level of BN28 decreases rapidlyduring deacclimation, while the others exhibit a slowturnover, as does MSACIB. It has been suggested thatshow turnover of some cold-induced proteins wouldenable plants to maintain frost tolerance during shortperiods of warmer temperatures [21].

The comparison between cultivars has shown thatthe differences observed in the levels of MSACIA/Bare consistent with the level of their correspondingtranscripts, except for MSACIA in Spredor cultivar.The lowest contents in both MsaCiA/B transcripts andcorresponding proteins were observed in CUF 101 andMoapa 69, which are the less hardy cultivars among the6 tested. However, the highest contents are not foundin the hardiest cultivars, i.e. Spredor or Rambler, butin Apica, which is a cultivar of intermediary winter-hardiness.

These data underscore that there is only a partialcorrelation between the steady state levels in MSA-CIA/B and winter-hardiness. While cold-sensitivitymay be explained by incapacity to accumulate suf-ficient amounts of cold-induced compounds such asMSACIA/B, high levels in these compounds do notconfere higher winter-hardiness. However, a recent

field study proved that Apica was the cultivar thatreached the highest level of frost tolerance in the middleof winter [5], but deacclimated more rapidly thanSpredor and Rambler at spring. This result is consist-ent with the fact that the highest levels in MSACIA/Bafter acclimation were found in Apica, and exempli-fy the idea that winter-hardiness cannot be reduced togenuine frost tolerance. The ability to maintain coldacclimation at early spring is probably an importanttrait of winter-hardiness, and one can postulate thatrestoring these traits in cultivars of low tolerance, eitherthrough classical breeding or genetic transformation,could improve their rusticity.

The cellular function of MSACIA and MSACIB hasnot been elucidated yet. The cell wall localization ofMSACIA-related proteins supports the hypothesis thatMSACIA play a structural role as other plant GRPs[14]. MSACIB should have a distinct function, sinceit is a soluble protein, most probably intracellular. Itsaccurate cytolocalization is under investigation. Therole of these proteins in the acquisition of frost toler-ance is being assessed with transgenic plants harboringthe sense and antisense MsaCiA/B gene constructs.

In conclusion, this study shows that MSACIA andMSACIB accumulate during alfalfa cold acclimation,and that this accumulation is differentially regulated intime and in space. Moreover, it seems that MSACIAundergoes complexe, organ-specific maturation. A low

633

steady-state level of these proteins after cold acclima-tion may account for low tolerance of sensitive alfalfacultivars. Although much remains to be done to elucid-ate the cellular functions of MSACIA and MSACIB,they are likely to play a role in frost tolerance mech-anisms. The question of the relationship between theirhigh glycine content and cold adaptation remains oneof the most intriguing aspects of these proteins, espe-cially if we consider that their biochemical functionsare probably distinct.

Acknowledgements

We thank Pierre Lechasseur for his excellent assistancein performing RNA extraction and analysis.

References

1. Arora R, Wisniewski ME: Cold acclimation in geneticallyrelated (sibling) deciduous and evergreen peach (Prunus per-sica L. Batsch). II. A 60-kilodalton bark protein in cold-acclimated tissues of peach is heat stable and related to thedehydrin family of proteins. Plant Physiol 105: 95–101 (1994).

2. Boothe JG, de Beus MD, Johnson-Flanagan AM: Expressionof a low-temperature-induced protein in Brassica napus. PlantPhysiol 108: 795–803 (1995).

3. Cai Q, Guy CL, Moore GA: Characterization of an unusualgroup II LEA gene family in citrus responsive to low tem-perature. GenBank/EMBL submissions L39004 and L39005(1995).

4. Castonguay Y, Nadeau P, Laberge S: Freezing tolerance andalteration of translatable mRNAs in alfalfa (Medicago sativaL.) hardened at subzero temperatures. Plant Cell Physiol 34:31–38 (1993).

5. Castonguay Y, Nadeau P, Lechasseur P, Chouinard L: Dif-ferential accumulation of carbohydrates in alfalfa cultivars ofcontrasting winterhardiness. Crop Sci 35: 509–516 (1995).

6. Close TJ, Fenton RD, Yang A, Asghar R, DeMason DA, CroneDE, Meyer NC, Moonan F: Dehydrin: the protein. In: Close TJ,Bray EA (eds) Plant Responses to Cellular Dehydration DuringEnvironmental Stress, pp. 104–118. Current Topics in PlantPhysiology, vol. 10. American Society for Plant Physiology,Rockville, MD (1993).

7. Cretin C, Puigdomenech P: Glycine-rich RNA-binding pro-teins from Sorghum vulgare. Plant Mol Biol 15: 783–785(1990).

8. De Vries S, Hoge H, Bisseling T: Isolation of total and polyso-mal RNA from plant tissues. In: Gelvin SB, Schilperoort RA(eds) Plant Molecular Biology Manual B6: 1–13 (1988).

9. Fourney RM, Miyakoshi J, Day III RS, Paterson MC: Northernblotting: efficient RNA staining and transfert. Focus 10: 1(1988).

10. Guo W, Ward RW, Tomashow MF: Characterization of acold-regulated wheat gene related to Arabidopsis cor47. PlantPhysiol 100: 915–922 (1992).

11. Guy CL: Cold acclimation and freezing stress tolerance: roleof protein metabolism. Annu Rev Plant Physiol Plant Mol Biol41: 187–223 (1990).

12. Guy C, Haskell D, Neven L, Klein P, Smelser C: Hydration-state-responsive proteins link cold and drought stress in spin-ach. Planta 188: 265–270 (1992).

13. Houde M, Danyluk J, Laliberte JF, Rassart E, Dhindsa RS, Sar-han F: Cloning, characterization, and expression of a cDNAencoding a 50 kDa protein specifically induced by cold accli-mation in wheat. Plant Physiol 99: 1381–1387 (1992).

14. Keller B: Structural cell wall proteins. Plant Physiol 101: 1127–1130 (1993).

15. Laberge S, Castonguay Y, Vezina L-P: New cold- and drought-regulated gene from Medicago sativa. Plant Physiol. 101:1411–1412 (1993).

16. Lee SP, Chen TH-H: Molecular biology of plant cold hardi-ness development. In: Li PH, Christersson L (eds) Plant ColdHardiness, pp. 1–29. CRC Press, Boca Raton, FL (1993).

17. Levitt J: Responses of Plants to Environmental Stresses, Vol I:Chilling, Freezing, and High Temperature Stresses. AcademicPress, New York (1980).

18. Mantyla E, Lang V, Tapio Palva E: Role of abscissic acidin drought-induced freezing tolerance, cold acclimation, andaccumulation of LTI78 and RAB18 proteins in Arabidopsisthaliana. Plant Physiol 107: 141–148 (1995).

19. Monroy FA, Castonguay Y, Laberge S, Sarhan F, Vezina L-P, Dhindsa RS: A new cold-induced alfalfa gene is associatedwith enhanced hardening at subzero temperature. Plant Physiol102: 873–879 (1993).

20. Muthalif MM, Rowland LJ: Identification of dehydrin-like pro-teins responsive to chilling of floral buds of blueberry (Vac-cinium, section Cyanococcus). Plant Physiol 104: 1439–1447(1994).

21. Neven LG, Haskell DW, Hofig A, Li Q-B, Guy CL: Charac-terization of a spinach gene responsive to low temperature andwater stress. Plant Mol Biol 21: 291–305 (1993).

22. Nordin K, Heino P, Tapio-Palva E: Separate signal pathwaysregulate the expression of a low-temperature-induced gene inArabidopsis thaliana (L.) Heynh. Plant Mol Biol 16: 1061–1071 (1991).

23. Nordin K, Vahala T, Tapio Palva E: Differential expressionof two related, low-temperature-induced genes in Arabidopsisthaliana (L.) Heynh. Plant Mol Biol 21: 641–653 (1993).

24. Showalter AM: Structure and function of plant cell wall pro-teins. Plant Cell 5: 9–23 (1993).

25. Sturm A: A wound-inducible glycine-rich protein from Daucuscarota with homology to single-stranded nucleic acid-bindingproteins. Plant Physiol 99: 1689–1692 (1992).

26. Welin B, Ake O, Nylander M, Tapio Palva E: Characterizationand differential expression of dhn/lea/rab-like genes duringcold acclimation and drought stress in Arabidopsis thaliana.Plant Mol Biol 26: 131–144 (1994).

27. Weretilnyk EA, Hanson AD: Molecular cloning of a plantbetaine-aldehyde dehydrogenase, an enzyme implicated inadaptation to salinity and drought. Proc Natl Acad Sci USA87: 2745–2749 (1990).

28. Wolfraim LA, Langis R, Tyson H, Dhindsa RS: cDNAsequence, expression, and transcript stability of a coldacclimation-specific gene, cas18, of alfalfa (Medicago falc-ata) cells. Plant Physiol. 101: 1275–1282 (1993).

29. Wolfraim LA, Dhindsa RS: Cloning and sequencing of thecDNA for cas17, a cold acclimation-specific gene of alfalfa.Plant Physiol 103: 667–668 (1993).