Changes in the accumulation of [alpha]-and [beta]-tubulin isotypes during cotton fiber development

Planta (2010) 231:277–291

ORIGINAL ARTICLE

Changes in the accumulation of �- and �-tubulin during bud development in Vitis vinifera L.

Luigi Parrotta · Giampiero Cai · Mauro Cresti

Received: 2 September 2009 / Accepted: 21 October 2009 / Published online: 13 November 2009© Springer-Verlag 2009

Abstract Microtubules play important roles duringgrowth and morphogenesis of plant cells. Multiple isoformsof �- and �-tubulin accumulate in higher plant cells andoriginate either by transcription of diVerent genes or bypost-translational modiWcations. The use of diVerenttubulin isoforms involves the binding of microtubules todiVerent associated proteins and therefore generates micro-tubules with diVerent organizations and functions. Tubulinisoforms are diVerentially expressed in vegetative andreproductive structures according to the developmental pro-gram of plants. In grapevine (Vitis vinifera L.), vegetativeand reproductive structures appear on the same stem, mak-ing this plant species an excellent model to study the accu-mulation of tubulin isoforms. Proteins were extracted fromgrapevine samples (buds, leaves, Xowers and tendrils)using an optimized extraction protocol, separated by two-dimensional electrophoresis and analyzed by immunoblotwith anti-tubulin antibodies. We identiWed eight �-tubulinand seven �-tubulin isoforms with pI around 4.8–5 thatgroup into separate clusters. More acidic �-tubulin isoformswere detected in buds, while more basic �-isoforms wereprevalently found in tendrils and Xowers. Similarly, moreacidic �-tubulin isoforms were used in the bud stage whilea basic �-tubulin isoform was essentially used in leaves andtwo central �-tubulin isoforms were characteristically usedin tendrils and Xowers. Acetylated �-tubulin was notdetected in any sample while tyrosinated �-tubulin wasessentially found in large latent buds and in bursting budsin association with a distinct subset of tubulin isoforms.

The implication of these data on the use of diVerent tubulinisoforms during grapevine development is discussed.

Keywords Electrophoresis · Grapevine · Immunoblot · Protein extraction · Tubulin isoforms

AbbreviationDTT Dithiothreitol

Introduction

The cytoskeleton of plant cells is composed of actin Wla-ments and microtubules that, in combination with associ-ated proteins, provide support for structural stability,anchoring of proteins, assembly of mitotic spindle, mainte-nance of internal order, cytoplasmic streaming, and con-struction of the cell wall (Nick 2007). Microtubules areheterodimeric polymers of �/�-tubulin and are encoded bymultiple genes, whose number is diVerent from species tospecies. Generally, �-tubulins are encoded by a small num-ber of genes (4–6) in contrast to the � subunit (which isencoded by 7–9 genes). For example, the Arabidopsisgenome contains six genes for �-tubulin (Kopczak et al.1992) and nine genes for �-tubulin (Snustad et al. 1992).Tubulin genes are not expressed uniformly in the plantbody but gene variability generates speciWc expression pat-terns in diVerent tissues and throughout developmentalstages. Few examples will clarify this concept. In Arabid-opsis, the �1 tubulin is present only in Xowers, while �2,�3, and �4 are expressed uniformly in roots, leaves andXowers; cells with mitotic and elongation activity accumu-late high levels of the TUA2 �-tubulin gene (Carpenter et al.1993). In rice, TubA1 and TubA2 genes accumulate beforeXowering while TubA3 gene is largely abundant throughout

L. Parrotta · G. Cai (&) · M. CrestiDipartimento Scienze Ambientali, University of Siena, via Mattioli 4, 53100 Siena, Italye-mail: [email protected]

123

278 Planta (2010) 231:277–291

anthesis (Qin et al. 1997). During seed development inHordeum vulgare, a speciWc tubulin gene (HvTUA5) isexpressed only in the embryo (Volodymyr et al. 2007). Incotton, transcripts of �-tubulin genes Tua2/3 and Tua4increased in abundance after anthesis, while levels of Tua1and Tua5 transcripts decrease considerably with the begin-ning of secondary wall synthesis (Whittaker and Triplett1999). During leaf development in barley, TuA2 and TuA4genes are mainly expressed in meristems, while TuA3 geneis expressed during cellular morphogenesis and TuA5/TuA1genes are expressed in post-mitotic cells when transversemicrotubules determine the cell shape (Schröder et al.2001).

Additional information on the diVerential use of tubulinisoforms was also obtained by immunoblot techniques after2D-electrophoresis. In cotton Wbers, nine isoforms of�-tubulin and seven isoforms of �-tubulin were consequentlyidentiWed (Dixon et al. 1994). Tubulin isoforms change rad-ically their expression pattern (some may also disappearduring Wber development) while changes in the structureand organization of microtubules associate with accumula-tion of speciWc �- and �-tubulins and with tyrosination of�-tubulin (Dixon et al. 2000). During maize root develop-ment, four �-tubulin and six �-tubulin isoforms were identi-Wed; the �1 and �4 isoforms predominate in dividing tissues(such as the root apex) while the �2, �3 and �4 isoforms aremore abundant in mature tissues (diVerentiated roots). The�1 and �2 isoforms are largely present in cortical region ofroots, while the �4 and �5 isoforms show an oppositeexpression pattern (Joyce et al. 1992). Similar approachesallowed the identiWcation of four �-tubulin and four �-tubu-lin isoforms in Phaseolus vulgaris (Hussey and Gull 1985)and of four �-tubulin isoforms (Dawson and Lloyd 1985)and six � isoforms (Hussey et al. 1988) in carrot.

Apart from diVerential gene expression, a number ofpost-translational modiWcations (analyzable by immuno-logical approaches) mark subpopulations of microtubules;the term “isotype” will be hereafter used to describe tubu-lins deriving exclusively from the expression of distinctgenes, while the term “isoform” will indicate tubulins thatalso derive from post-translational modiWcations. Post-translational modiWcations probably act individually or incombination at level of single cells to control speciWcmicrotubule functions in particular cell domains. The mostcommon post-translational modiWcations of tubulin aredetyrosination/tyrosination (probably involved in control-ling the binding of plus-end tracking proteins and motorproteins), glutamylation/glycylation (hypothetically involvedin microtubule severing by katanin) and acetylation (foundon stable microtubules in most cells) (Hammond et al.2008). Genomic and proteomic Wndings suggested that thedevelopment of speciWc plant cells and tissues is characterizedby the expression of distinct tubulin genes and consequently

by the use of distinct tubulin isotypes, which are requiredfor the assembly of diVerent microtubule arrays. In additionto gene expression, the local post-translational modiWcationof tubulin is used to modulate the function of distinct micro-tubule arrays by regulating the binding of microtubulesto associated proteins.

Grapevine (Vitis vinifera L.) is a climber plant of Vita-cee that grows using tendrils to attach to physical supports;it produces economically important fruits, the grape, usedfor feeding and wine production. In light of its economicalimportance, the genome of grapevine has been recentlysequenced (Jaillon et al. 2007). Buds of grapevine areplaced laterally with regard to the seedling axis in corre-spondence of each node and can be distinguished as func-tional, dormant, and latent buds (which open after 2 yearsand yield low-productive seedlings) (Mullins et al. 2004).Latent buds produce a lateral shoot that develops into aleaf. This process repeats at least three times; after that, theapical meristem can yield two types of primordia: an addi-tional leaf primordium and an undiVerentiated primordiumthat can develop into either a Xower or a tendril (Carmonaet al. 2008). Therefore, development of grapevine is char-acterized by the presence of reproductive and vegetativemeristematic structures on the same stem unlike Arabidop-sis, in which vegetative meristems are directed to reproduc-tive stage after environmental stimuli (Boss et al. 2003).Although the expression of tubulin genes has been studiedin diVerent plants, a comprehensive model on the use oftubulin isoforms during development of plant organs ismissing. Grapevine is an advantageous plant model inwhich to study the expression of tubulin isoforms in con-nection with development of vegetative (leaves and ten-drils) and reproductive organs (Xowers) starting fromcommon buds. Given their diVerential function and struc-ture, buds, leaves, tendrils and Xowers require the presenceof diVerent microtubule organizations and consequently ofspeciWc tubulin isoforms as well as of diVerent post-transla-tional modiWcations. The aim of the current study is tomonitor changes in the expression of tubulin isoforms dur-ing bud development toward either vegetative or reproduc-tive structures in grapevine. We used a proteomic approachbased on the standardized extraction of proteins from diVer-ent organs followed by protein separation with bidimen-sional electrophoresis and screening with diVerentcommercial antibodies to �- and �-tubulin.

Materials and methods

Antibodies

The list of antibodies to �-tubulin with their epitope posi-tion (as deduced from the datasheets) is reported hereafter.

123

Planta (2010) 231:277–291 279

B-5-1-2 (mouse monoclonal, Sigma-Aldrich, St. Louis,MO, USA, diluted 1:5,000) was obtained after immuniza-tion with Sarkosyl-resistant Wlaments from sea urchin andrecognizes an epitope located at the C-terminal end of�-tubulin in a variety of organisms. Yol 1/34 (ChemiconInternational, Temecula, CA, USA, rat monoclonal, diluted1:1,000) is directed to the amino acid sequence 414–422 ofmicrotubules in most eukaryotic cells including mammals,yeast, slime molds and allium. TU-01 (Zymed Laborato-ries, San Francisco, CA, USA, mouse monoclonal, diluted1:500) is directed to a structural domain at the N-terminalof �-tubulin (peptide 65-79). aN-18 (Santa Cruz Biotech-nology, Santa Cruz, CA, USA, goat polyclonal, diluted1:1,000) is an aYnity puriWed goat polyclonal antibodyraised against a peptide mapping near the N-terminus ofArabidopsis �-tubulin. TUB-1A2 (Sigma-Aldrich, mousemonoclonal, diluted 1:1,000) is directed against the tubulinC-terminal tyrosine. 6-11B-1 (Sigma-Aldrich, mousemonoclonal, diluted 1:2,000) recognizes an epitope locatedon the �3 isoform of Chlamydomonas axonemal tubulinwithin four residues of the acetylated Lys40. Since thesequence detected in Chlamydomonas is found in most of�-tubulins, the antibody can potentially detect acetylated�-tubulins from many organisms (protista, plants, inverte-brates, and vertebrates); however, the cross-reacting regionof 6-11B-1 is sometimes divergent and the epitope may beabsent or masked. DM-1A (Sigma-Aldrich, mouse mono-clonal, diluted 1:1,000) is raised against chick brain tubulinand is directed to amino acid 426–430 in the C-terminaldomain, a sequence highly conserved in plant tubulins.ATN-02 (Cytoskeleton, Denver, CO, USA, sheep poly-clonal, diluted 1:1,000) is directed to both �- and �-tubulin.

The following anti �-tubulin antibodies were used. TU-13 (Santa Cruz Biotechnology, mouse monoclonal, diluted1:100) was raised against �-tubulin from brain tissue of pigorigin. The aC-18 antibody (Santa Cruz Biotechnology,goat polyclonal, diluted 1:250) cross-reacted with an epi-tope mapping near the C-terminus of Arabidopsis �-tubu-lin. The aE-14 antibody (Santa Cruz Biotechnology, goatpolyclonal, diluted 1:500) recognized an epitope mappingnear the N-terminus of Arabidopsis �-tubulin. The aF-14antibody (Santa Cruz Biotechnology goat polyclonal,diluted 1:250) is directed to an epitope located within theinternal region of Arabidopsis �-tubulin. The aN-181 anti-body (Santa Cruz Biotechnology, goat polyclonal, diluted1:500) cross-reacted with an epitope mapping at the N-ter-minus of Arabidopsis �-tubulin. D66 (Sigma-Aldrich,mouse monoclonal, diluted 1:250) recognized an epitope(amino acid 427–432) localized at the C-terminal region of

sea urchin �2-tubulin, a sequence highly conserved in Ara-bidopsis �-tubulins. The epitopes recognized by T3526(Sigma-Aldrich, rabbit anti-tubulin antibody, diluted 1:500)are not speciWed. TU-06 (Monosan, Uden, The Nether-lands, mouse monoclonal, diluted 1:300) is raised against�-subunits of pig brain tubulin and is recommended fordetection of a phylogenetically conserved N-terminal struc-tural domain (amino acids 81–95) of �-tubulin of mouse,rat, human, pig and yeast. The peptide sequence is partiallyconserved in Arabidopsis. Tub2.1 (Santa Cruz Biotechnol-ogy, mouse monoclonal, diluted 1:450) was raised againstpuriWed brain �-tubulin of rat origin.

Secondary HRP-conjugated antibodies were as fol-lows: anti-mouse and anti-rabbit from GE HealthCare(Milano, Italy), both diluted 1:5,000; anti-rat from Chem-icon International diluted 1:5,000; anti-sheep from Cyto-skeleton diluted 1:10,000; anti-goat from Santa CruzBiotechnology (sc-2020) diluted 1:5,000. In addition, asecondary antibody conjugated to Alexa Fluor 488 (fromInvitrogen, Carlsbad, CA, USA) was used in Xuorescentblot techniques.

Reagents

Unless diVerently mentioned, all chemicals used in thiswork were purchased from Sigma-Aldrich.

Plant culture and sampling

Plants used in this work [Vitis vinifera (L.) (Sangiovesecultivar), tobacco (Nicotiana tabacum L.) and Arabidop-sis thaliana (L.) Heynh.] were from the Seed and PlantCollection of the Botanical Garden at Siena Universityand grown in the Botanical Garden of Siena University.Tissues and organs from grapevine were collected inspring (from April to May), immediately frozen in liquidnitrogen and stored at ¡80°C until use. The followingsamples of grapevine were collected: small latent buds(·2 mm in height), large latent buds (>3 mm), burstingbuds, leaves, Xowers, and tendrils. Samples of tobaccoand Arabidopsis were collected, weighted, and immedi-ately processed for electrophoresis using the “phenol”method (described below). Extracts of tobacco pollentubes were prepared as described in Persia et al. (2008).Protein extract of bovine brain (used as control) wasprovided by Cytoskeleton Inc.

Protein extraction

In order to obtain protein samples qualitatively and quanti-tatively suitable for electrophoretic analysis, plant materialswere processed with diVerent protocols as follows. Allextractions were performed at least three times.

1 This antibody is diVerent from the aN-18 listed in the �-tubulinsection. The two antibodies have the same name but diVerent codenumbers and cross-react with diVerent proteins.

123

280 Planta (2010) 231:277–291

1. “Tris” method. Samples were powdered in liquid nitro-gen using pestle and mortar. The powder was added tolysis buVer [40 mM Tris, 4% polyvinylpolypyrrolidone(PVPP), 50 mM dithiothreitol (DTT), and plant prote-ase inhibitors], vortexed, or sonicated, and incubatedon ice for 15 min to extract proteins. Samples werecentrifuged at 60,000g for 30 min at 4°C. The pelletwas discarded and the supernatant was diluted inthe rehydration/solubilization buVer (8 M urea, 2 Mthiourea, 2% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 40 mM Tris, traces ofbromophenol blue). For monodimensional analysis,samples were added to Laemmli sample buVer (Laemmli1970).

2. “CHAPS” method. This protocol is similar to the“Tris” method, except for the presence of CHAPS (4%Wnal concentration) in the lysis buVer. After the vortex/extraction/centrifugation steps, the supernatant wasadded to the rehydration/solubilization buVer.

3. “Kit” method. This protocol makes use of two com-mercial kits provided by Bio-Rad (Hercules, CA,USA): the ReadyPrep Protein Extraction Kit for TotalProtein (which contains the detergent ASB-14) and theReadyPrep 2-D CleanUp Kit (which removes deter-gents, salts, lipids, and phenols from the protein sam-ple). The two protocols were carried out following themanufacturer’s instruction. The Wnal supernatantalready contained a buVer suitable for 2-D analysis.

4. “TCA” method. This protocol was derived fromSaravanan and Rose (2004) and was carried out exactlyas described in the “Tris” method, but the Wnal superna-tant was treated with 10% TCA in acetone to precipitateand clean the protein sample. After incubation on ice forat least 2 h, samples were centrifuged at 60,000g for30 min at 4°C. The supernatant was discarded and thepellet was washed with acetone, centrifuged again, andresuspended in the rehydration/solubilization buVer.

5. “Phenol” method. The procedure was derived fromWang et al. (2006) and consisted in the rapid freezingand powdering of samples with pestle and mortar. Thepowder was transferred into centrifuge tubes and sup-plemented with 10% TCA, 50 mM DTT in cold ace-tone. Samples were vortexed and centrifuged at16,000g for 3 min at 4°C; the supernatant was dis-carded while the pellet was washed with 80% cold ace-tone plus 50 mM DTT. After centrifugation at 16,000gfor 3 min (4°C), the supernatant was discarded and thepellet was air-dried for 30 min. The pellet was supple-mented with phenol (pH 8) and SDS-dense buVer (2%SDS, 0.1 M Tris–HCl pH 8, 5% �-mercaptoethanol,30% sucrose) in 1:1 ratio. Samples were mixed by vor-tex and incubated for 10 min. After centrifugation at16,000g for 3 min, the upper phenolic phase was trans-

ferred into a new tube, supplemented with 0.1 Mammonium acetate in 80% methanol, mixed by vortexand incubated at ¡20°C for at least 2 h. After centrifu-gation at 16,000g for 5 min at 4°C, the supernatant wasdiscarded and the pellet was washed with 100% metha-nol and then with 80% acetone containing 50 mMDTT. The Wnal pellet was air-dried, then dissolved inthe rehydration/solubilization buVer.

6. “TCA2” method. Samples were powdered after freez-ing in liquid nitrogen. The powder was resuspended inthe TCA-extraction buVer (10% TCA, 1% PVPP, 2%�-mercaptoethanol in acetone) followed by incubationat ¡20°C overnight. As reported (Saravanan and Rose2004), most proteins precipitated after 30 min. Afterincubation, samples were centrifuged at 10,000g for15 min at 4°C. The supernatant was smoothly removedand discarded. The pellet was washed three times withcold acetone and centrifuged at 10,000g for 5 min(4°C) each time. The supernatant was discarded whilethe pellet was air-dried for 5–10 min and resuspendedin an appropriate volume of rehydration/solubilizationbuVer.

7. “Urea” method. This protocol was derived from Cas-tro et al. (2005) and consisted in the extraction of sam-ples with a lysis buVer containing 7 M urea, 2 Mthiourea, 4% (w/v) CHAPS, 3% (w/v) SDS and 60 mMDTT. Proteins were precipitated with 10 volumes of20% TCA, 0.2% DTT in cold acetone at ¡20°C for1 h. After centrifugation (16,000g, 30 min, 4°C), thepellet was resuspended in the rehydration/solubiliza-tion buVer for 2-D analysis.

Protein concentration

The concentration of proteins in Wnal samples was calcu-lated using the commercial 2-D Quant Kit (GE Health-Care). The assay was carried out according to themanufacture’s instruction using BSA as standard. Eachsample and standard was analyzed at least three times usinga Shimadzu UV-160 spectrophotometer adjusted at 480 nm.

Monodimensional electrophoresis (SDS-PAGE)

Separation of proteins by 1-D electrophoresis was carriedout on polyacrylamide gels as described by Laemmli(1970). Gels (10%) were run in a Mini PROTEAN II cell(Bio-Rad) using 0.75 mm spacers; the stacking gels con-tained 10 wells or 1 large well for slab analysis.

Bidimensional electrophoresis

Protein separation in the Wrst dimension (IEF) was achievedusing Immobiline Dry-Strip (GE HealthCare), 7 or 11

123

Planta (2010) 231:277–291 281

cm-long, with 4–7 pH gradient. Strips were hydrated in therehydration/solubilization buVer (40 mM Tris, 8 M urea, 2M thiourea, 2% CHAPS, traces of bromophenol blue) con-taining 18 mM DTT, 20 �l/ml IPG BuVer (pH 4-7) and theprotein samples. The protein concentration was adjusted to1.5 mg/ml, corresponding to 190 �g of proteins for the7-cm strips and 300 �g of proteins for the 11-cm strip.After hydration, strips were subjected to the Wrst separationusing the Multiphor II (GE HealthCare); run was done at300 V (1 min), from 300 to 3,500 V (1 h and 30 min), 3500V for 3 h and 30 min. After the IEF separation, strips werestored at ¡80°C. Prior to the second dimension, strips wereequilibrated for 15 min in the equilibration buVer (50 mMTris–HCl pH 8.8, 6 M urea, 30% glycerol, 2% SDS, tracesof bromophenol blue, 10 mg/ml DTT). For preliminary sep-arations, 7-cm strips were applied to self-made mini-gels(10%) and run in a Mini PROTEAN II cell (Bio-Rad) using1 mm spacers. Finer separations were obtained by applying11-cm strips to pre-cast gels (Criterion XT Bis–Tris PrecastGel, 10%, 11 cm, Bio-Rad). Gels were run using the Crite-rion Cell (Bio-Rad) at 200 V constant for 1 h and thenstained with Bio-Safe Coomassie (Bio-Rad) or silver (Sil-ver Staining Kit Protein, GE HealthCare). Gel images werecaptured with the Fluor-S Multi-Imager (Bio-Rad). Alter-natively, proteins were blotted onto nitrocellulose mem-branes. At least three replicates of each sample wereanalyzed.

Protein blotting and immunostaining

Blot of proteins onto nitrocellulose membranes (from Bio-Rad) was carried out using either the Mini Trans-Blot Cell(Bio-Rad) for samples separated on 7-cm strips or the Cri-terion Blotter (Bio-Rad) for samples separated on 11-cmstrips. The blot conditions were as described by Towbinet al. (1979). Protein blot was carried out at 100 V (con-stant), 350 mA, 200 W for 45 min. The quality of blots wasevaluated by staining membranes with Ponceau S or bychecking the correct blotting of pre-stained molecular massstandards of the Precision Series (Bio-Rad). Membraneswere blocked overnight in the blocking solution [5% ECLblocking agent (GE HealthCare), 0.1% Tween-20 in TBS(20 mM Tris pH 7.5, 150 mM NaCl)]. After a brief wash inTBS, membranes were incubated for 1 h with the primaryantibodies diluted in TBS-T (TBS containing 0.1% Tween-20). After three washes in TBS, membranes were incubatedwith the secondary antibodies diluted in TBS-T, then exten-sively washed in TBS. The immunological reaction wasvisualized with the ECL Plus Western Blotting DetectionReagents (GE HealthCare) for 5 min. The chemilumines-cent signal was captured with the Fluor-S Multi-Imager(Bio-Rad) using the software Quantity One. In “slab-blot”assays, proteins were loaded in the single large well of the

stacking gel, separated by SDS-PAGE and blotted ontomembranes, which were assembled in the MultiscreenApparatus (Bio-Rad) to obtain 20 identical lanes, separatelyanalyzable.

For determining the pI of tubulin after 2-D separation,the total protein spots in the membrane were labeled withDeep Purple (GE HealthCare) according to the manufac-turer’s instructions. Membranes were then blocked asoutlined above and incubated with a primary antibody to�-tubulin (B-5-1-2). After washes, membranes were incu-bated with a secondary antibody conjugated to Alexa Fluor488. Signals were detected using the Quantity One softwareby selecting the 610 BP Wlter for Deep Purple and the 520LP Wlter for Alexa Fluor 488. Images were then superim-posed and pseudo-colored using the Multi Channel Viewerof Quantity One. For comparing the position of tubulinwith that of molecular standards, extracted proteins weremixed with 2D standards (Bio-Rad) and were separated by2D-electrophoresis. After blotting and staining with DeepPurple, protein spots in the membrane were compared withstandard proteins and the pI of tubulin (revealed by theAlexa Fluor 488 signal) was determined consequentlyusing the MrpI Tool of PDQuest.

Analysis of gels and blots

Images of gels and blots were captured using the Fluor-SMulti-Imager and analyzed by Quantity One (for 1D-gels)and PDQuest (for 2D-gels) (software from Bio-Rad). Theexposure time was 1–10 min for blots developed withchemiluminescence, 0.5 s for pre-stained molecular massstandards, 5–7 s for gels stained with Coomassie/silver.Analysis of protein spots in gels and blots was done usingthe Spot Detection Wizard of PDQuest. First, we selectedthe faintest spot in the scan (thus setting the sensitivity andminimum peak value parameters); then, we selected thesmallest spot in the scan (to set the size scale parameter)and the largest spot (setting the radius of the backgroundsubtraction rolling ball and the streak removal rolling).Three independent gels of each sample were used for spotcomparison. Further analysis of spots was done using theSpot and Matchset tools.

All blots were developed using identical conditions,from the ECL incubation time to the exposure time. Allimages were processed correspondingly using the Auto-scale command to improve the quality of gels/blots and theBackground Subtraction command to remove the back-ground noise. The relative intensity of single spots was cal-culated with the Volume Tool of Quantity One software(Bio-Rad); three independent blots were used. Alterna-tively, spots were also analyzed with the ImageJ software(available at http://www.rsbweb.nih.gov/ij/) using the Mea-sure command.

123

282 Planta (2010) 231:277–291

Results

Optimization of the extraction protocol

In order to optimize protein extraction from grapevine interms of quality, quantity, and reproducibility, we testeddiVerent protocols. The “urea” method did not provide anadequate extraction level because only few protein bandswere visible in 1-D gels (Fig. 1a, lane 3), among which aprominent band at 50 kD (presumably the ribulose 1,5-bis-phosphate carboxylase/oxygenase enzyme). The “TCA”protocol was not equally satisfactory because only fewscratches were detected in the gel (lane 5). The “kit”method made use of two speciWc commercial kits (the Wrstfor protein extraction, the second for sample cleaning) andallowed obtaining a good-quality sample (lane 4) but stillcontaminated by interfering substances. The “Tris” methodused a basic soluble buVer and was the less eVective proto-col (lane 8). The “CHAPS” method was substantially simi-lar to the previous one except for the addition of CHAPS,which increased the extractability of proteins (lane 9); how-ever, the abundant 50-kD band was absent. The qualitativeyield of “TCA-2” method was good (lane 10), at least com-parable to the “phenol” method. Although the TCA-2 pro-tocol is technical simpler, it was not convincing in terms ofreproducibility and quantitative yield. The “phenol”method was the best-tested protocol (lane 6); several bands

were observed, the most abundant at 50 kD. The sampleappeared deWcient of interferences and immunoblot with apolyclonal anti-tubulin showed positive cross-reactivitywith a 50-kD band (Fig. 1b, lane 6). Immunoblot analysiswas also positive for the “kit” (lane 4) and TCA-2 (lane 10)methods but cross-reactivity was generally lower. The“phenol” method was therefore chosen for the extraction ofproteins from diVerent grapevine samples.

Screening of �- and �-tubulin antibodies

Cross-reactivity of grapevine tubulin was tested using sev-eral commercial antibodies from diVerent manufactures,which were reported to cross-react with �- and �-tubulinfrom diVerent sources. The position of corresponding epi-topes (when available) is schematically reported in Fig. 2a,b, otherwise the epitope mapping is approximately esti-mated. In our hands, all antibodies cross-reacted with braintubulin (data not shown). The �-tubulin antibodies werealso screened against protein extracts from tobacco leaves,used as a plant control because tubulin isoforms have beenlargely characterized in this plant species (Smertenko et al.1997b). In addition, tobacco extracts were prepared exactlyas grapevine extracts in order to exclude that the absence ofcross-reactivity was dependent on the extraction method.Only three antibodies to �-tubulin provided adequate cross-reactivity: B-5-1-2, aN-18 and the polyclonal ATN-02. Thethree antibodies labeled speciWcally a 50-kD band (Fig. 2c,arrow). The TU-01 (not shown) did not cross-react withtobacco extracts.

The �-tubulin antibodies were Wrst screened againsttobacco leaf extracts (Fig. 2d, sample L) and bovine braintubulin (sample T). As result, only the aE-14 practicallycross-reacted with both samples while most of antibodiesrecognized bovine brain tubulin but showed imprecise sig-nals when probed on tobacco leaf extracts. Consequently,the aE-14 antibody seemed the appropriate choice to moni-tor �-tubulin isoforms. The aE-14 antibody also cross-reacted with Arabidopsis extracts (not shown).

The �- and �-tubulin antibodies were tested against pro-tein extracts from grapevine leaves prepared according to the“phenol” method (Fig. 2e). Only one antibody to �-tubulin(B-5-1-2) was found to cross-react speciWcally with a 50-kDband (Wrst two lanes of the immunoblot). The other antibod-ies gave negative or indecipherable results. Comparably,slab-blot with �-tubulin antibodies on grapevine leaf extractsshowed the cross-reactivity of aE-14 only (Fig. 2f, Wrst fourlanes). This Wnding highlights the low cross-reactivity ofantibodies against grapevine �- and �-tubulin. Apparently,�- and �-tubulin migrated at the same level in the mono-dimensional gel, indicating that the diVerence in molecularweight is not signiWcant. Alternatively, the percentage ofacrylamide used in the mono-dimensional gel (10%) was

Fig. 1 Extraction of proteins from grapevine leaves. a Electrophore-sis of proteins extracted with diVerent methods. Lane 1 and 7 molecu-lar mass markers. Lane 2 bovine brain tubulin (10 �g). Grapevine leafproteins (10 �g) were extracted with the “urea” (lane 3), “kit” (lane 4),“TCA” (lane 5), “phenol” (lane 6), “Tris” (lane 8), “CHAPS” (lane 9)and “TCA-2” (lane 10) methods. Lanes 1–6 and 7–10 are from twoseparate gels. b Detection of tubulin in samples by immunoblot withthe polyclonal ATN-02 antibody. The blots of lanes 2–6 are from thesame membrane

123

Planta (2010) 231:277–291 283

not probably appropriate for a clear separation of �- and�-tubulin of grapevine. Given that B-5-1-2 and aE-14 antibod-ies were speciWcally raised against each tubulin subunit, it isunlikely that they cross-reacted with the unrelated subunit.

2-D electrophoresis of samples from grapevine

After completing the screening of antibodies, weapproached the standardization of 2-D electrophoresis and

Fig. 2 Immunoblot assay with diVerent antibodies to �- and �-tubulin.a Position of epitopes recognized by antibodies to �-tubulin; the nameof antibody is in bold when the exact mapping of epitopes is available.Black boxes in the �-tubulin polypeptidic chain indicate �-strandswhile white boxes indicate �-helices. b Position of epitopes recognizedby antibodies to �-tubulin; the name of antibody is in bold when theepitope mapping is exactly known. Black boxes in the �-tubulin poly-peptide indicate �-strands while white boxes indicate �-helices. Thesecondary structure of � and � tubulin was obtained by submitting theamino acid sequence of �1 and �1 tubulin of Arabidopsis to thePredictProtein service for sequence analysis, structure and functionprediction (http://www.predictprotein.org/). Images are simply used toshow the epitope position of antibodies with regard to the �1 and �1tubulin of Arabidopsis and are not intended to provide information ontubulin secondary structure. c Slab-gel immunoblot with �-tubulinantibodies on protein extracts from tobacco leaves (100 �g). The codeof antibodies and their working dilution is shown on the top. Only

antibodies B-5-1-2, aN-18 and ATN-02 cross-reacted with a 50-kDband (arrow). d Immunoblot with �-tubulin antibodies on tobacco leafextract (L) and bovine brain tubulin (T). The antibody code is shownon the top; all blots were overexposed in order to detect faint signals.Only the aE-14 antibody cross-reacted with both samples. e Slab-blotwith �-tubulin antibodies on protein extracts from mature leaves ofgrapevine. The previously tested TU-01 antibody is not shown but itdid not provide cross-reactivity. About 100 �g of proteins were loaded.The code and dilution of antibodies is shown on the top. Only the anti-body B-5-1-2 showed a signiWcant cross-reactivity against a 50-kDband. f Slab-blot with �-tubulin antibodies on protein extracts fromgrapevine leaves. The code and dilution of antibodies is shown againon the top. Only the antibody aE-14 showed a signiWcant cross-reactiv-ity. B-5-1-2 was used as control. The Tub2.1 antibody, which gave noreactivity, is not shown. Only the blot segment around 50 kD is shownin panels c–f

123

284 Planta (2010) 231:277–291

immunoblot using 10% Criterion XT gels (Bio-Rad). When300 �g of proteins were loaded, a typical Coomassie-stained gel contained around 400 spots (as determined bythe software PDQuest) (Fig. 3). The most of protein spotsranged between 100 and 20 kD. 2D-gels showed that spotswere resolved with much higher quality compared with 1-Delectrophoresis and consequently validated the analysis oftubulin spots by immunoblot.

After protein spot detection by PDQuest, spots were com-pared across gels by creating a matchset and selecting the“small bud” gel as the template for the matchset standard(Fig. 3, some landmark spots are indicated by arrows). Sam-ple gels were subsequently compared using the scatter plotanalysis of PDQuest, which indicated their relatedness. Weselected a 2-fold-factor range for comparing diVerent sam-ples. As shown by the correlation coeYcient (r), the similar-ity level of samples was not high. The highest correlation

coeYcient was observed when comparing small latent budsto leaves (r = 0.61), while the lowest coeYcient was foundwhen comparing bursting buds to Xowers (r = 0.10). Thecomparison between Wnal developmental points (leaves vs.tendrils, leaves vs. Xowers, tendrils vs. Xowers) alwaysreturned low values of correlation coeYcient, suggestingthat the three ending points had diVerent protein patterns.The comparison of blot results (see below) with analysis ofthe correlation coeYcient (as an index of protein variability)indicated that tubulin spots did not change considerably theirexpression pattern compared with modiWcations in the num-ber and intensity of the other protein spots.

Immunoblot with �- and �-tubulin antibodies

The purpose of this study was to characterize the �- and�-tubulin isoforms expressed during development of buds

Fig. 3 Analysis by 2-D electrophoresis of protein samples fromgrapevine. Proteins were extracted from small, large, and burstingbuds, and from leaves, tendrils and Xowers, separated by isoelectrofo-cusing on 11-cm strips (pH 4-7) and then on 10% Criterion XT gels.Gels were stained with Bio-Safe Coomassie. Most of protein spots

showed a molecular mass ranging from 20 to 100 kD. Molecular massstandards are indicated on the top left image but refer to all otherimages as well. Arrows indicate landmark proteins, while the white cir-cle indicates a large 50-kD protein spot present in leaves (presumablyribulose-1,5-bisphosphate carboxylase-oxygenase)

123

Planta (2010) 231:277–291 285

to leaves/tendrils/Xowers in grapevine. As shown by blotanalysis, the �- and �-tubulin subunits showed a pI around4.8–5, with an average molecular mass of approximately 50kD (Fig. 4a). The pI of grapevine tubulins was estimated bymatching the immunolabeled tubulin spots to 2D-image ofXower proteins visualized in the same membrane with DeepPurple. The �-tubulin isoforms migrated faster in the SDSsecond dimension and were slightly less acidic in the IEFWrst dimension than the �-tubulin isoforms (Fig. 4b). The �-and �-tubulin isoforms focused into distinct single clustersin the IEF dimension. We identiWed at least eight isoformsof �-tubulin in the examined samples (Fig. 5a), which havebeen numbered according to their position. The �-tubulinisoforms accumulated diVerentially throughout develop-mental stages. Isoforms 3–7 were present during all devel-opmental stages although with dissimilar proportions. Incontrast, isoforms 1 and 2 were hardly detectable in smalllatent buds but were present in all other stages; isoform 8was found in early stages (from small latent buds to burst-ing buds), sometimes in the leaf stage as well but not in ten-drils and Xowers. Isoforms 4-5 were the most abundant ininitial stages (from small to bursting buds), while isoform 5decreased signiWcantly in leaves. In contrast, tendrils andXowers showed a comparable proportion of isoforms; morebasic isoforms (from 4 to 7) were abundantly used in ten-drils and Xowers while more acidic isoforms (from 1 to 3)were less used. When the initial stages were compared,small latent buds were characterized by a limited number ofisoforms (mainly 3–5); other isoforms were missing (suchas 1–2) or expressed at low levels (such as 6–8). The rangeof isoforms increased during the transition from small tolarge/bursting buds, with all isoforms being represented.Later developmental stages were marked by a dramatic

change in isoform composition. The leaf stage was charac-terized by the prevalent use of acidic isoforms (2–4) withother isoforms present at lower level. Tendrils and Xowersshowed a peculiar isoform composition, with basic iso-forms (5–7) being the most abundant while acidic isoforms(1–3) were detectable at lower level. As a general trend,bud development was characterized by the progressiveappearance of all �-tubulin isoforms while the Wnal devel-opmental stages were characterized by the common pres-ence of tubulin isoforms 1–7 with acidic isoforms beingused prevalently in leaves and more basic isoforms abun-dantly used in tendrils and Xowers. The most acidic iso-forms (1–3) are never used in larger proportion (except forleaves), while the central and most basic isoforms are fre-quently used.

Fig. 4 IdentiWcation of tubulin position in the protein spot pattern.a Matching of �-tubulin with protein spots in the Xower sample. The�-tubulin (red, indicated by arrow) was labeled with the B-5-1-2 anti-body and with an Alexa Fluor 488-labeled secondary antibody and wasthen superimposed to the blot membrane stained with Deep Purple(blue). The approximate pI of tubulin was deduced by comparing theposition of �-tubulin with that of other protein spots. b Composite blotof �- and �-tubulin isoforms of grapevine. The double blot indicatesthe relative position of � and �-tubulin isoforms in Xowers; the twospot clusters have approximately the same pI but � isoforms move fast-er than � isoforms

Fig. 5 Blots with �-tubulin antibodies on small buds, large buds,bursting buds, leaves, tendrils and Xowers from grapevine. a Immuno-blot analysis. Position of corresponding spots is indicated by numbersin all samples. b Analysis of spot volume made with ImageJ softwareon the blot images. Intensity of spots is reported in Y-axis as IntegratedDensity (that is the sum of the pixel values in the image or selectionand is equivalent to the product of Area and Mean Gray Value). Val-ues are average of three independent measurements with their standarddeviation. Results obtained with Quantity One are equivalent

123

286 Planta (2010) 231:277–291

Seven �-tubulin isoforms were found in the samplesexamined (Fig. 6) and, like �-tubulin isoforms, wereindicated according to their position from acidic to basic.Comparison of �-tubulin isoforms from small latent budsto mature organs indicated that �-isoforms were alsodiVerentially expressed during development of buds toleaves/tendrils/Xowers. A faint signal of isoform 1 wasconsistently found in large latent buds but was rarelydetected in other samples. The more basic isoform 7 wasalso hardly detectable; since isoform 7 was constantlyfound in leaves, it represented a speciWc marker for thisorgan. Isoforms 2 and 3 were consistently found in theinitial developmental stage (small latent buds); giventheir abundance in this organ, isoforms 2–3 were considered

highly speciWc for the initial stage of bud development.The transition from small to large latent buds wasmarked by the progressive decrease of isoform 2 and bythe concomitant appearance of isoforms 4–5, which rep-resented the most abundant isoforms in other samples.Isoform 3 progressively decreased from small to burstingbuds; it was consistently present in leaves but hardly intendrils and Xowers. Isoform 4 was found in all samples,slightly in small and bursting buds but more abundantlyin large latent buds, leaves, tendrils, and Xowers. Givenits high abundance in tendrils and Xowers, isoform 4 wasconsidered a valuable marker for these two organs. Likeisoform 4, isoform 5 was weakly detected in small latentbuds but was more abundant in all other samples; inter-estingly, isoform 5 was highly expressed in bursting budsbut decreased in terminal developmental stages (leaves,tendrils and Xowers). In a more striking contrast, isoform6 was used at very low levels in all organs examined;although present in initial stages (small and large latentbuds), it was not found as buds opened; although it wasrarely present in leaves, it was never detected in tendrilsand Xowers; consequently, isoform 6 marked the prelimi-nary stages of bud development. Isoform 7 was exclu-sively found (although at low levels) in large latent budsand leaves, suggesting that it was a speciWc marker forthese developmental stages. Therefore, isoforms 4 and 5were the only ones found in all samples. Small latentbuds were characterized by a unique isoform composi-tion, whereas large latent buds and leaves were similar inthe pattern of �-tubulins, with central tubulin isoformsbeing mostly represented. Conversely, bursting buds,tendrils and Xowers used a distinct pattern of �-tubulins,in which isoforms 4 and 5 were preferentially presentalthough with a diVerent relative concentration. Burstingbuds were consistently marked by isoform 5 while ten-drils and Xowers used prevalently isoform 4 in greaterproportions. Overexposure of the �-tubulin immunoblotof tendrils sometimes showed a very faint signal of iso-form 3, while no additional isoform other than 4 and 5were detected in Xowers, suggesting that the presence ofadditional tubulin isoforms in tendrils and Xowers isextremely rare.

Analysis of acetylated and tyrosinated tubulin

We also checked for the presence of two common post-translational modiWcations of �-tubulin (acetylation andtyrosination) in all samples. The 6-11B-1 antibody(directed to acetylated �-tubulin) did not cross-react withany protein in grapevine samples but it did with extractsfrom both calf brain and tobacco pollen tubes (Fig. 7a). Onthe contrary, the Tub-1A2 antibody to tyrosinated �-tubulincross-reacted with a 50-kD band in large buds and (to a

Fig. 6 Blots with �-tubulin antibodies on small buds, large buds,bursting buds, leaves, tendrils and Xowers from grapevine. a Immuno-blot analysis. Position of corresponding spots is indicated by numbersin all samples. b Analysis of spot volume made with ImageJ softwareon the blot images. Intensity of spots is reported in Y-axis as IntegratedDensity (that is the sum of the pixel values in the image or selection andis equivalent to the product of Area and Mean Gray Value). Values areaverage of three independent measurements with their standard devia-tion. Measuring spot volumes with Quantity One returned comparableresults

123

Planta (2010) 231:277–291 287

lesser extent) in bursting buds (Fig. 7b). The antibody alsocross-reacted (as expected) against tyrosinated tubulin ofcalf brain and tobacco pollen tubes. When probed by 2-Dimmunoblot, tyrosinated �-tubulin was detected in largelatent buds as a group of more basic spots (Fig. 7c, num-bered in Roman numbers from I to V). Apparently, thetyrosinated �-tubulin spots did not match with spots previ-ously identiWed by B-5-1-2. Similarly, tyrosinated �-tubulinwas found to a lesser extent in bursting buds (Fig. 7d), butonly isoforms I, II, and IV were detected. Again, tyrosi-nated tubulin spots in bursting buds did not match with�-tubulin detected by B-5-1-2 but they matched with tyrosi-nated spots found in large latent buds. The high backgroundlevel of Fig. 7d is due to the low abundance of tyrosinatedtubulin in bursting buds and to the automatic regulation ofgrayscale by Quantity One.

Discussion

Extraction of proteins and cross-reactivity of grapevine tubulin

Unlike animal cells, extraction of proteins from plant tis-sues is not technically simple. Physical barriers (cell wall)and interfering molecules (polyphenols, terpenes, organicacid, pectins, waxes, etc.) hinder the extractability and anal-ysis of proteins by 2-D electrophoresis. Since grapevineorgans and tissues can be considered recalcitrant to proteinextraction, the proper analysis of tubulin isoforms requiredthe optimization of an extraction protocol capable of yield-ing clean and reproducible amounts of protein. Therefore,the immunoblot results need to be supported by adequateextraction protocols that warrant be introducing and brieXydiscussing. In order to optimize the extraction of proteinsfrom grapevine tissues, we tested seven diVerent protocols,selected either from the literature or from the self-made oravailable as commercial kits. One method (the so-called“phenol”) was successful for all organs and yielded cleanprotein samples coupled with high reproducibility and tech-nical simplicity. The method is based on the removal ofcontaminants by washing with organic solvents and on theextraction of proteins with phenol (Wang et al. 2006) andworks eYciently with diVerent recalcitrant plants (Wanget al. 2003).

We tested several commercial antibodies to both N- andC-terminal epitopes of �- and �-tubulin in order to identifythe most of tubulin isoforms. In view of the poor level ofcross-reactivity, the screening of anti-tubulin antibodies isan integral part of this work and is basic for the immunoblotresults. Only one antibody to �-tubulin (B-5-1-2) was foundto cross-react with protein extracts from grapevine leaves.This antibody recognizes one epitope in the carboxyl-termi-nal segment of �-tubulin in evolutionary divergent organ-isms, such as ox, tobacco and Arabidopsis. B-5-1-2 wassuccessfully used to localize all tubulin isoforms in Arabid-opsis (Kopczak et al. 1992) and provided signiWcant resultsin rye roots (Kerr and Carter 1990). Another antibody usedto identify �-tubulin isoforms in diVerent plants is YOL1/34 (Dixon et al. 1994) but it was ineVective in grapevine.The TU-01 antibody identiWed �-tubulin in tobacco cells(Smertenko et al. 1997a) but it was unsatisfactory ingrapevine. The DM1A antibody recognizes a conservedregion at the C-terminus of �-tubulin, but it worked weaklyin grapevine.

Like �-tubulin, commercial antibodies to �-tubulinshowed weak cross-reactivity with grapevine samples. Theonly exception was the aE-14 antibody from Santa Cruz,which cross-reacts with an epitope mapping near the N-ter-minus of �-tubulin of Arabidopsis thaliana. Compared withthe C-terminal domain, which is the most variable region of

Fig. 7 Analysis of acetylated and tyrosinated �-tubulin. a Immunoblotwith 6-11B-1 antibody to acetylated �-tubulin on proteins extractedfrom calf brain (lane 1), tobacco pollen tubes (lane 2), small buds (lane3), large buds (lane 4), bursting buds (lane 5), leaves (lane 6), Xowers(lane 7), tendrils (lane 8) and separated by 1D-electrophoresis. Signalwas detected in extracts of calf brain and tobacco pollen tubes (used ascontrol). No signal was detected in grapevine samples. b Immunoblotwith Tub-1A2 antibody to tyrosinated �-tubulin on the same samples.Signal was detected in calf brain (lane 1) and more faintly in pollentubes (lane 2), used again as control. In grapevine, cross-reactivity wasfound essentially in large buds and to a lesser extent in bursting buds.Only the blot portion around 50-kD is shown. c 2-D immunoblot ofproteins from large latent buds probed with antibody to tyrosinated�-tubulin. Spots are numbered in Roman from I to V. Presumptiveposition of tubulin spots recognized by B-5-1-2 is indicated by the dot-ted box. d 2-D immunoblot of proteins from bursting buds probed withantibody to tyrosinated �-tubulin. Signal is weak like in 1-D blots.Spots are numbered again in Roman according to their equivalent inlarge latent buds. Presumptive position of tubulin spots recognized byB-5-1-2 is indicated by the dotted box

123

288 Planta (2010) 231:277–291

tubulin, the N-terminal domain is much less variable.Unfortunately, the exact epitope recognized by aE-14 is notknown and the manufacture did not provide information.Nevertheless, aE-14 also cross-reacted with tobacco tubulinand Arabidopsis tubulin (the latter is not shown). Other�-tubulin antibodies showed very faint or no cross-reactivity.The DM1B antibody, successfully used in other plant spe-cies (Dixon et al. 1994), was not tested because unavailableat the time of experiments (it now available in the SantaCruz catalog; http://www.scbt.com/datasheet-58880-beta-tubulin-dm1b-antibody.html). The TUB 2.1 antibody cross-reacted with tubulin of rye roots (Kerr and Carter 1990) butit was not functioning in our case; in addition, it recognizesonly a subset of �-tubulin isoforms. The TU-06 antibodyworked greatly against brain tubulin and labeled �-tubulinsfrom tobacco cells but it did not work against grapevineextract (Smertenko et al. 1997b). �-tubulin isoforms ofArabidopsis thaliana were successfully identiWed using the2-10-B6 antibody (Snustad et al. 1992) but the probe is notcommercially available (at least to our knowledge).

Tubulin isoforms in buds, tendrils, Xowers, and leaves

It was reported that the expression level of cytoskeletal pro-teins does not change considerably among diVerent grapecultivars and is not aVected by water deWcit or salt stress(Vincent et al. 2007); consequently, diVerences in �- and�-tubulin composition depends on the diVerential use of iso-forms rather than on responses to environmental changes.Another source of variability in tubulin composition couldbe cold treatment. For example, speciWc �- and �-tubulinisoforms are overexpressed in rye roots after 4 days of coldacclimation (Kerr and Carter 1990). We are conWdent that acold response is not accountable of changes in the composi-tion of tubulin isoforms in grapevine because we alwayscollected samples in April/May, when temperatures neverdecreased signiWcantly. We also believe that variability oftubulin isoforms is not dependent on protein degradationcaused by extraction because we never observed immuno-blot signals with lower molecular mass and other works didnot report degradation of �-tubulin in protein samplesextracted with the same method (Wang et al. 2003, 2004).

We found that �- and �-tubulin isoforms grouped intosingle clusters. The presence of a single �-tubulin cluster ingrapevine is diVerent from what observed in cotton Wbers(Dixon et al. 1994) where �-tubulin isoforms aggregated intwo distinct clusters but it is similar to what observed inArabidopsis (Kopczak et al. 1992) and rye roots (Kerr andCarter 1990). The situation in tobacco is intermediate(Smertenko et al. 1997b). The separation of tubulin iso-forms was obtained using a pH gradient of 4–7, which isthe most preferable pH gradient as shown by the severalpapers on 2-D analysis of animal and plant tubulin. Since

grapevine tubulin spots were detected in single clusterswith pI of 4.8–5, IEF could be also done with pH gradientshypothetically providing more resolution. We found thatstrips with pH gradients of 4.5–5.5 (supposed more ade-quate) are only available as 7-cm long and are not compati-ble with our system. Although separation of tubulinisoforms with pH gradient of 4.5–5.4 was reported (Towbinet al. 2001), spots still clustered together. We think that ourresults are comparable with those of other manuscripts inwhich single tubulin spots in the clusters were hardly dis-tinguishable. Because of clustering, the identiWcation,quantitation and comparison of tubulin spots were done bydistinct software, PDQuest (Bio-Rad) and ImageJ (freelyavailable). Analysis was done on three sample replicatesand by two operators independently. As result, PDQuestand ImageJ reported analogous data on the number and rel-ative quantity of tubulin spots.

The identiWcation of eight �-tubulin and seven �-tubulinisoforms is in line with results from other plant species (seereferences in the “Introduction”). On the basis of consensussequences of either �- or �-tubulin from Arabidopsis (madewith Clustal W), a preliminary screening using the BLAT andBLAST tools for database searching in grapevine 8X genome(Bioinformatics section of Vigna web site, www.vitisgenome.it/en/index.php5) indicated the presence of 6 putativesequences of �-tubulin and 8 putative sequences of �-tubulin.

The pattern of �- and �-tubulin isoforms changed duringbud development toward vegetative and reproductive struc-tures but �-isoforms changed less signiWcantly than �-iso-forms. Small buds showed a peculiar pattern of � and �isoforms consisting of a restricted number of � and use ofvery acidic �-isoforms. The transition from buds to leavesis characterized by the increased of the more acidic �-tubu-lin isoforms compared with the more basic ones. Expres-sion of leaf-speciWc tubulin genes is not surprising becausemature leaves show a less mitotic activity compared toyounger stages, grow to a lesser extent, and then need aspeciWc microtubule composition to maintain the cell struc-ture. Our observations are in agreement with current mod-els suggesting that changes in the expression of tubulinisotypes generate microtubule arrays with distinct organiza-tion and function that control the establishment of plant cellshapes, probably modulating the binding of microtubules todiVerent associated proteins. For example, genes coding forspeciWc �-tubulins (HVATUB2 and HVATUB4) areexpressed in barley meristems while the isotype HVATUB3is expressed during leaf morphogenesis and the two iso-types HVATUB1 and HVATUB5 are exclusively expressedin post-mitotic cells when microtubules determine the cellshape in the mesophyll (Schröder et al. 2001). In cottonWbers, a larger number of �-tubulins were identiWed after 10days from anthesis than after 20 days (when growth slowsdown) (Dixon et al. 1994), suggesting that dividing tissues

123

Planta (2010) 231:277–291 289

require diVerent tubulin isoforms. In soybean, cells show-ing rapid growth and division have a larger number andhigher levels of tubulin isoforms (Han et al. 1991). SpeciWctubulin genes (such as �1 and �4) are also predominantlyexpressed in rapidly dividing tissues of maize whereasother isotypes (such as �2, �3 and �4) predominate inmature tissues (root and pollen) (Joyce et al. 1992).

One of the most remarkable observations is that grape-vine leaves had a distinct repertoire of �-tubulin isoformswhile tendrils and Xowers showed an equivalent pattern ofboth � and � isoforms. In the literature, there are no data ontubulin expression in tendrils. On the other hand, Xowersare often characterized by speciWc �/�-tubulin isotypes. Themost striking example is the expression of speciWc tubulinisoforms in the pollen of various plant species. In Arabid-opsis, the TUB9 gene is expressed mainly in Xowers withhigh expression levels in pollen, growing pollen tubes, andovules (Cheng et al. 2001). In sunXower, the �(pi)-tubulinis expressed primarily in the male gametophyte (Evrardet al. 2002). Similarly, the OsTUB8 gene of rice is primar-ily expressed in anthers and in mature pollen, suggesting arole in the formation of microtubules during anther and pol-len development or throughout pollen tube growth (Yoshik-awa et al. 2003). The PwTUA1 gene is speciWcallyexpressed in the pollen of Picea wilsonii according to eachstage of pollen tube growth and is likely involved in vesicletransport (Yu et al. 2009). Use of speciWc �-tubulin genesduring Xower and pollen development was also reported,suggesting that microtubules of reproductive tissues requirea speciWc composition of � and � tubulin isoforms (Chenget al. 2001). Tubulin genes are also speciWc for vegetativetissues and are not expressed in pollen, such as the �5-tubu-lin gene found only in the vegetative tissues of mature car-rot plants (Hussey et al. 1988). In Arabidopsis, the TUA6gene is critical for root development and transgenic plantswith reduced expression of TUA6 showed no irregularitiesin the aerial parts (Bao et al. 2001).

When compared to leaves, both tendrils and Xowers useprevalently more basic �-tubulin isoforms, with large abun-dance of isoforms 6 and 7 (hardly detectable during buddevelopment). Moreover, tendrils and Xowers use the samepattern of �-tubulins (isoforms 4 and 5). The use of analo-gous tubulin isoforms in very diVerent organs (such as ten-drils and Xowers) is not atypical. For example, two MADS-box genes (VFUL-L and VAP1) are expressed in lateralmeristems of grapevine giving rise to either inXorescencesor tendrils and expression of neither VFUL-L nor VAP1 wasfound in vegetative organs, suggesting a putative commonorigin between tendrils and Xowers (Calonje et al. 2004). Inaddition, tendrils from in vitro grown shoot tips of grapevinecan develop into inXorescences after addition of benzylade-nine (BA) or 6-(benzylamino)-9-(2-tetrahydropyranyl)-9H-purine (PBA) (Srinivasan and Mullins 1978). Therefore, a

comparable expression pattern of �- and �-tubulin isoformsbetween tendrils and Xowers is not surprising and is likelyto support a common developmental plan.

Our results indicate that �-tubulin isoforms mark thedevelopment of grapevine organs more critically than � iso-forms do. The importance of using speciWc �-tubulin iso-forms during plant development is conWrmed by evidencethat the orientation of cellulose microWbrils in secondarycell walls of Eucalyptus grandis is correlated with expres-sion of the �-tubulin gene EgrTUB1, suggesting that pro-duction and organization of cell walls is strictly associatedwith speciWc �-tubulin isoforms (Spokevicius et al. 2007).

In grapevine, post-translational modiWcations of tubulinare typical of speciWc developmental stages because tyrosi-nated tubulin was found in large and in bursting buds, whileacetylated tubulin was not identiWed. In plant and animalcells, tubulins can be modiWed post-translationally by addi-tion or removal of chemical groups in order to generatesubsets of speciWc microtubules (Hammond et al. 2008).Missing of acetylated tubulin in grapevine is not surprisingbecause such post-translational modiWcation could berestricted to speciWc cell types or be even absent, like indeveloping cotton Wbers (Dixon et al. 2000). For example,acetylated �-tubulin was detected in maize leaves but not inroots, pollen, and anthers (Wang et al. 2004). Since acety-lated microtubules were shown to interact preferentiallywith motor proteins (kinesin-1 and dynein) (Fukushimaet al. 2009), acetylated tubulin was expectedly identiWed inassociation with microtubules of the tobacco generative cell(Astrom 1992) and at the poles of mitotic spindle intobacco (Smertenko et al. 1997b).

Detyrosination is a second post-translational modiWca-tion of plant cell tubulin. In cotton Wbers, diVerent tubulinisoforms are tyrosinated according to diVerential develop-mental stages (Dixon et al. 2000). Tyrosine residues oftubulin are also phosphorylated as an additional mechanismof post-translational modiWcation (Blume et al. 2008). Thedetyrosinated form of tubulin also interacts speciWcallywith kinesin-1 (Liao and Gundersen 1998) while tyrosi-nated microtubules interact preferentially with CLIP170(Hammond et al. 2008), suggesting that such post-transla-tional modiWcation aVects the interplay between microtu-bules and associated proteins. Earlier studies indicated thatdetyrosinated microtubules are less dynamic than tyrosi-nated ones (Kreis 1987); if this model is applied to grape-vine, buds presumably contain highly dynamicmicrotubules, in line with the high dynamics of cells withinbuds. Since such modiWcation is missing in the Wnal devel-opmental stages of grapevine (leaves, tendrils and Xowers),we hypothesize that tyrosinated tubulin is required for thedevelopmental program of buds. We cannot explain whyTUB-1A2 cross-reacted with tubulin isoforms not detectedby B-5-1-2. We are conWdent that B-5-1-2 did not recognize

123

290 Planta (2010) 231:277–291

spots other than those described in this manuscript. A simi-lar result was also described in maize leaves, where TUB-1A2 detected tubulin spots not identiWed by A-11126, anantibody directed (unlike B-5-1-2) to the N-terminaldomain of �-tubulin (Wang et al. 2004). Similarly, the anti-tyrosinated �-tubulin antibody YL 1/2 cross-reacted withone tubulin isoform not recognized by the anti-tubulin YOL1/34 in cotton Wbers (Dixon et al. 2000).

In summary, our data indicate that the developmentalpathway of grapevine buds to leaves/tendrils/Xowers ischaracterized by the use of distinct tubulin isoforms. DiVer-ences are more evident for �- than for �-isoforms. Data inthis manuscript are in line with results in the literature indi-cating that speciWc tubulin isoforms are preferentiallyexpressed in diVerent plant tissues and are required for theassembly of microtubule arrays with distinct functions.Given the developmental program of meristematic struc-tures (buds) toward either structural (tendril) or energetic(leaf) or reproductive organs (Xower), grapevine is a goodplant model to study the accumulation of tubulin isoformsin association with the structural and functional diVerencesof microtubule arrays.

Acknowledgments We sincerely thank Prof. Mario Pezzotti(Dipartimento di Scienze, Tecnologie e Mercati della Vite e del Vino,University of Verona, Italy), Prof. Enrico Pè (Scuola SuperioreSant’Anna, Pisa, Italy) and Prof. Luca Bini (Dipartimento di BiologiaMolecolare, University of Siena, Italy) for helpful suggestions and crit-icisms. This work was funded in the framework of the Vigna Project(http://www.vitisgenome.it/en/index.php5).

ConXict of interest statement All authors declare that no Wnancial/commercial conXicts of interest exist.

References

Astrom H (1992) Acetylated alpha-tubulin in the pollen tube microtu-bules. Cell Biol Int Rep 16:871–881

Bao Y, Kost B, Chua NH (2001) Reduced expression of alpha-tubulingenes in Arabidopsis thaliana speciWcally aVects root growth andmorphology, root hair development and root gravitropism.Plant J 28:145–157

Blume Y, Yemets A, Sulimenko V, Sulimenko T, Chan J, Lloyd C,Dráber P (2008) Tyrosine phosphorylation of plant tubulin. Planta229:143–150

Boss PK, Buckeridge EJ, Poole A, Thomas MR (2003) New insightsinto grapevine Xowering. Funct Plant Biol 30:593–606

Calonje M, Cubas P, Martinez-Zapater JM, Carmona MJ (2004) Floralmeristem identity genes are expressed during tendril developmentin grapevine. Plant Physiol 135:1491–1501

Carmona MJ, Chaib J, Martinez-Zapater JM, Thomas MR (2008) Amolecular genetic perspective of reproductive development ingrapevine. J Exp Bot 59:2579–2596

Carpenter JL, Kopczak SD, Snustad DP, SilXow CD (1993) Semi-constitutive expression of an Arabidopsis thaliana �-tubulin gene.Plant Mol Biol 21:937–942

Castro AJ, Carapito C, Zorn N, Magne C, Leize E, Van DA, ClementC (2005) Proteomic analysis of grapevine (Vitis vinifera L.) tis-sues subjected to herbicide stress. J Exp Bot 56:2783–2795

Cheng Z, Snustad DP, Carter JV (2001) Temporal and spatial expres-sion patterns of TUB9, a �-tubulin gene of Arabidopsis thaliana.Plant Mol Biol 47:389–398

Dawson PJ, Lloyd CW (1985) IdentiWcation of multiple tubulins intaxol microtubules puriWed from carrot suspension cells. EMBOJ 4:2451–2455

Dixon DC, Seagull RW, Triplett BA (1994) Changes in the accumula-tion of �- and �-tubulin isotypes during cotton Wber development.Plant Physiol 105:1347–1353

Dixon DC, Meredith WR, Triplett BA (2000) An assessment of�-tubulin isotype modiWcation in developing cotton Wber. IntJ Plant Sci 161:63–67

Evrard JL, Nguyen I, Bergdoll M, Mutterer J, Steinmetz A, LambertAM (2002) A novel pollen-speciWc alpha-tubulin in sunXower:structure and characterization. Plant Mol Biol 49:611–620

Fukushima N, Furuta D, Hidaka Y, Moriyama R, Tsujiuchi T (2009)Post-translational modiWcations of tubulin in the nervous system.J Neurochem 109:683–693

Hammond JW, Cai D, Verhey KJ (2008) Tubulin modiWcations andtheir cellular functions. Curr Opin Cell Biol 20:71–76

Han IS, Jongewaard I, Fosket DE (1991) Limited expression of adiverged beta-tubulin gene during soybean (Glycine max [L.]Merr.) development. Plant Mol Biol 16:225–234

Hussey PJ, Gull K (1985) Multiple isotypes of alpha- and beta-tubulinin the plant Phaseolus vulgaris. FEBS Lett 181:113–118

Hussey PJ, Lloyd CW, Gull K (1988) DiVerential and developmentalexpression of beta-tubulins in a higher plant. J Biol Chem263:5474–5479

Jaillon O, Aury JM, Noel B, Policriti A, Clepet C, Casagrande A,Choisne N, Aubourg S, Vitulo N, Jubin C, Vezzi A, Legeai F,Hugueney P, Dasilva C, Horner D, Mica E, Jublot D, Poulain J,Bruyere C, Billault A, Segurens B, Gouyvenoux M, Ugarte E,Cattonaro F, Anthouard V, Vico V, Del FC, Alaux M, Di GG,Dumas V, Felice N, Paillard S, Juman I, Moroldo M, Scalabrin S,Canaguier A, Le CI, Malacrida G, Durand E, Pesole G, Laucou V,Chatelet P, Merdinoglu D, Delledonne M, Pezzotti M, LecharnyA, Scarpelli C, Artiguenave F, Pe ME, Valle G, Morgante M,Caboche M, Adam-Blondon AF, Weissenbach J, Quetier F,Wincker P (2007) The grapevine genome sequence suggestsancestral hexaploidization in major angiosperm phyla. Nature449:463–467

Joyce CM, Villemur R, Snustad DP, SilXow CD (1992) Tubulin geneexpression in maize (Zea mays L.): change in isotype expressionalong the developmental axis of seedling root. J Mol Biol 227:97–107

Kerr GP, Carter JV (1990) Tubulin isotypes in rye roots are altered dur-ing cold acclimation. Plant Physiol 93:83–88

Kopczak SD, Haas NA, Hussey PJ, SilXow CD, Snustad DP (1992)The small genome of Arabidopsis contains at least six expressed�-tubulin genes. Plant Cell 4:539–547

Kreis TE (1987) Microtubules containing detyrosinated tubulin areless dynamic. EMBO J 6:2597–2606

Laemmli UK (1970) Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 227:680–685

Liao G, Gundersen GG (1998) Kinesin is a candidate for cross-bridg-ing microtubules and intermediate Wlaments. Selective binding ofkinesin to detyrosinated tubulin and vimentin. J Biol Chem273:9797–9803

Mullins MG, Bouquet A, Williams LE (eds) (2004) Biology of theGrapevine. Cambridge University Press, Cambridge

Nick P (2007) The plant cytoskeleton—new jobs for a versatile net-work. Protoplasma 230:125–127

Persia D, Cai G, Del Casino C, Faleri C, Willemse MTM, Cresti M(2008) Sucrose synthase is associated with the cell wall of tobac-co pollen tubes. Plant Physiol 147:1603–1618

123

Planta (2010) 231:277–291 291

Qin X, Giani S, Breviario D (1997) Molecular cloning of three rice al-pha-tubulin isotypes: diVerential expression in tissues and duringXower development. Biochim Biophys Acta 1354:19–23

Saravanan RS, Rose JKC (2004) A critical evaluation of sampleextraction techniques for enhanced proteomic analysis of recalci-trant plant tissues. Proteomics 4:2522–2532

Schröder J, Stenger H, Wernicke W (2001) �-Tubulin genes are diVer-entially expressed during leaf cell development in barley (Horde-um vulgare L.). Plant Mol Biol 45:723–730

Smertenko A, Blume Y, Viklicky V, Draber P (1997a) Exposure oftubulin structural domains in Nicotiana tabacum microtubulesprobed by monoclonal antibodies. Eur J Cell Biol 72:104–112

Smertenko A, Blume Y, Viklicky V, Opatrny Z, Draber P (1997b)Post-translational modiWcations and multiple tubulin isoforms inNicotiana tabacum L. cells. Planta 201:349–358

Snustad DP, Haas NA, Kopczak SD, SilXow CD (1992) The smallgenome of Arabidopsis contains at least nine expressed beta-tubulin genes. Plant Cell 4:549–556

Spokevicius AV, Southerton SG, MacMillan CP, Qiu D, Gan S,Tibbits JFG, Moran GF, Bossinger G (2007) Beta-tubulin aVectscellulose microWbril orientation in plant secondary Wbre cellwalls. Plant J 51:717–726

Srinivasan C, Mullins MG (1978) Control of Xowering in the grape-vine (Vitis vinifera L.): formation of inXorescences in vitro by iso-lated tendrils. Plant Physiol 61:127–130

Towbin H, Staehelin T, Gordon J (1979) Electrophoretic transfer of pro-teins from polyacrylamide gels to nitrocellulose sheets: procedureand some applications. Proc Natl Acad Sci USA 76:4350–4354

Towbin H, Özbey O, Zingel O (2001) An immunoblotting method forhigh-resolution isoelectric focusing of protein isoforms on immo-bilized pH gradients. Electrophoresis 22:1887–1893

Vincent D, Ergul A, Bohlman MC, Tattersall EA, Tillett RL, WheatleyMD, Woolsey R, Quilici DR, Joets J, Schlauch K, Schooley DA,Cushman JC, Cramer GR (2007) Proteomic analysis revealsdiVerences between Vitis vinifera L. cv. Chardonnay and cv. Cab-ernet Sauvignon and their responses to water deWcit and salinity.J Exp Bot 58:1873–1892

Volodymyr VR, Nese S, Yaroslav B, Winfriede W (2007) Distincttubulin genes are diVerentially expressed during barley graindevelopment. Physiol Plant 131:571–580

Wang W, Scali M, Vignani R, Spadafora A, Sensi E, Mazzuca S, CrestiM (2003) Protein extraction for two-dimensional electrophoresisfrom olive leaf, a plant tissue containing high levels of interferingcompounds. Electrophoresis 24:2369–2375

Wang W, Vignani R, Scali M, Sensi E, Cresti M (2004) Post-transla-tional modiWcations of �-tubulin in Zea mays L. are highly tissuespeciWc. Planta 218:460–465

Wang W, Vignani R, Scali M, Cresti M (2006) A universal and rapidprotocol for protein extraction from recalcitrant plant tissues forproteomic analysis. Electrophoresis 27:2782–2786

Whittaker DJ, Triplett BA (1999) Gene-speciWc changes in alpha-tubulin transcript accumulation in developing cotton Wbers. PlantPhysiol 121:181–188

Yoshikawa M, Yang G, Kawaguchi K, Komatsu S (2003) Expressionanalyses of beta-tubulin isotype genes in rice. Plant Cell Physiol44:1202–1207

Yu Y, Li Y, Li L, Lin J, Zheng C, Zhang L (2009) Overexpression ofPwTUA1, a pollen-speciWc tubulin gene, increases pollen tubeelongation by altering the distribution of alpha-tubulin and pro-moting vesicle transport. J Exp Bot 60:2737–2749

123