Identification of proteins involved in aggregationof human dermal papilla cells by proteomics

Xia Rushan a, Hao Fei a,*, Mou Zhirong b, Wu Yu-zhang b

Journal of Dermatological Science (2007) 48, 189—197

www.intl.elsevierhealth.com/journals/jods

aDepartment of Dermatology, Southwest Hospital, Third Military Medical University,Gaotanyan Road, Chongqing 400038, Chinab Institute of Immunology of PLA, Third Military Medical University, Gaotanyan Road,Chongqing 400038, China

Received 9 February 2007; received in revised form 10 June 2007; accepted 20 June 2007

KEYWORDSComparativeproteome;Dermal papilla cells;Heat shock protein;Mitochondrialribosomal protein S7

Summary

Background: The dermal papilla is a major component of hair, which signals thefollicular epithelial cells to prolong the hair growth process. To date, little is knownabout the significance of the specific protein(s) express in the dermal papilla cells(DPC) with regard to their aggregative behaviour.Objectives: To identify proteins involved in aggregative behaviour of DPC, we com-paratively analyzed the proteome of cells with and without aggregative behaviour.Methods: A series of methods were used, including two-dimensional gel electrophor-esis (2-DE), PDQuest software analysis of 2-DE gels, peptide mass fingerprinting basedon matrix-assisted laser desorption/ionisation-time of flight-mass spectrometry(MALDI-TOF-MS), and NCBInr database searching, to separate and identify differen-tially expressed proteins. Western blotting and reverse transcriptase polymerase chainreaction (RT-PCR) were used to validate the differentially expressed proteins.Results: Image analysis revealed that averages of 618� 22 and 568� 47 protein spotswere detected in passages 3 and 10 DPC, respectively. Twenty-four differential proteinspots were measured with MALDI-TOF-MS. A total of 17 spots yielded good spectra, and15 spots matched with known proteins after database searching. Western blottingconfirmed that heat shocking protein 70 was up-regulated in passage 3 DPC. Over-expression of mitochondrial ribosomal protein S7 was confirmed by RT-PCR, indicatingthat they are involved in aggregation of DPC through some signaling pathway.Conclusions: The clues provided by the comparative proteome strategy utilized herewill shed light on molecular mechanisms of DPC in aggregative behaviour.# 2007 Japanese Society for Investigative Dermatology. Published by Elsevier IrelandLtd. All rights reserved.

Abbreviations: DPC, dermal papilla cells; HSP70, heat shock protein 70; MRPS7, mitochondrial ribosomal protein S7.* Corresponding author. Fax: +86 23 68462522.E-mail addresses: xrs4366@sina.com.cn (X. Rushan), haofei@mail.tmmu.com.cn (H. Fei).

0923-1811/$30.00 # 2007 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.jdermsci.2007.06.013

190 X. Rushan et al.

1. Introduction

The dermal papilla cells (DPC) are specializedstroma cells which are believed to be the sourceof the dermis-derived signaling molecules involvedin hair follicle development and in postnatal haircycling [1]. The tendency of aggregation is one ofthe significant properties of the DPC, and this prop-erty is associated with their biologic function anddifferentiating state [2—6]. Early passage dermalpapilla cells have aggregative behavior and caninduce hair growth in vivo, but, upon further cultur-ing, this property is lost (Fig. 1). Generally speaking,DPC lose the aggregative behavior and the ability ofregulating and controlling the hair cycle in thepassage range 6—15 [7]. However, the cellularmechanisms underlying the aggregative behaviorof DPC are less known. In order to study the aggre-gative behavior of DPC, a variety of techniques havebeen used to identify genes related to the aggrega-tive property of DPC [8,9]. These methods wereused to analyze mRNA expression levels of aggrega-tion-related genes. In protein level, many growthfactors produced by DPC were found to have thefunction of modulating the proliferation of follicularepithelium, such as insulin-like growth factor-I,bone morphogenetic proteins, transforming growthfactor-beta1 and so on. Most of these factors act as acytokine network controlling follicle developmentand hair cycling [10—13]. Our recent studies had

Fig. 1 Cell morphology of DPC in passages 3 and 10. (A) Paswithout aggregative behavior. (C) Passage 10 DPC regain agg

proved that the DPC with aggregative behaviorcould produce soluble molecules which can helpthe DPC losing aggregative behavior regain to aggre-gative behavior [14]. These results suggest that DPCwith aggregative behavior could release soluble,functional factors which stimulate the growth ofDPC and maintain the aggregative behavior ofDPC. Therefore, intensive screening to search forcandidate proteins in low passage DPC is needed toidentify proteins that play a role in the aggregationof DPC.

Advances in the studies on proteomics have madeit possible to compare the total proteins of cellsunder different conditions on a large scale [15,16].The proteomic strategy based on 2-DE and MALDI-TOF-MS has been applied in a variety of studies.

In this study, by using 2-DE and MALDI-TOF massspectrometry, we examined the protein expressionprofile difference in passages 3 and 10 DPC andanalyzed proteins involved in the aggregative beha-vior globally.

2. Materials and methods

2.1. Chemicals and other materials

Bis, Tris, SDS, glycine, TEMED, ammonium persulfate(APS), glycerol, ultra pure urea, bromophenol blue,acrylamide, DTT, CHAPS, agarose and Immobiline

sage 3 DPC with aggregative behavior. (B) Passage 10 DPCregative behavior.

Identification of proteins involved in aggregation of human dermal papilla cells 191

DryStrips (pH 3—10) were purchased from AmershamBiosciences (Uppsala, Sweden). Iodoacetamide(IAA), Brilliant Blue G-250 and Collagenase D (ultrapure grade) came from Sigma (St. Louis, USA).Protease inhibitor cocktail and dispase were pur-chased from Roche (Switzerland). Goat Anti-humanHSP70 mAb and peroxidase-conjugated goat anti-rabbit immunoglobulin were obtained from SantaCruz Biotechnology (Santa Cruz, CA, USA). Otherreagents not mentioned above were domesticmade.

2.2. Dermal papilla separation andcultures

The dermal papilla is derived from human (sixmales, three females, mean age 28 years) non-balding scalp-skin which were obtained from plasticsurgery with the informed consent of donors andapproval of Southwest hospital ethical committee.Dermal papilla were isolated from scalp tissues bytwo-step enzyme method set up previously in ourlaboratory [17]. Firstly, the scalp-skin of nine volun-teers was sterilized and digested in 0.5% (w/v)dispase for 12—16 h at 4 8C and in 0.2% (w/v) col-lagenase D for 6 h at 37 8C sequentially. The digestedtissue was then centrifuged at 550—850 � g for 3—5 min. The dermal papilla was sedimentated at thebottom of tube as a clump of cells, whereas othercells floated in the supernatant. After centrifugedfor several times, the human dermal papilla wasseparated from other types of cells and was culturedin DMEM medium supplemented with 10% foetal calfserum subsequently. DPC were mixed and incubatedat 37 8C in a humidified atmosphere of 95% air and 5%CO2. The passages 3 and 10 DPC were harvested atexponential growth. The cell pellet was washedthree times with 0.1 M PBS, centrifuged and frozeat �70 8C.

2.3. Extraction of water soluble proteins

The passages 3 and 10 DPC pellets (>108 cells) werere-suspended in 100 ml of 40 mM Tris buffer (pH 7.4)and performed by supersonic wave at 40 W for1.5 min. Cell debris were removed by centrifugationat 12,000 � g for 60 min at 4 8C. The supernatantwas accurately sample normalized by Bradford’sprotein assay and used for 2-DE.

2.4. 2-DE

The same amount of 1 mg proteins were loaded onto13 cm, pH 3—10 IPG strips. IPG strips were rehy-drated in rehydration solution (8 M urea, 1% (w/v)CHAPS, 0.2% (w/v) DTT, 0.5% (v/v) pharmalyte

3—10, 0.002% (w/v) bromophenol blue) over night.IEF was conducted by the IPGPhor II system (Amer-sham-Pharmacia Biotech) for 1 h 0 V, 12 h 30 V, 2 h60 V, 1 h 500 V, 1 h 1000 V and at 8000 V untilapproximately 40,000 V h were reached. Focusedstrips were immediately equilibrated for 2 �15 min with buffer (50 mmol/l Tris—HCl, pH 6.8,6 mol/l urea, 30% glycerol, 2% SDS and a trace ofbromophenol blue). DTT (2%, w/v) was added in thefirst step, and iodoacetamide (2.5%, w/v) in thesecond step. The strips were placed on a 1.5-mmthick, 12.5% polyacrylamide gel and sealed with0.1% (w/v) agarose in SDS-electrophoresis buffercontaining 0.01% (w/v) bromophenol blue. The elec-trophoresis was run for 30 min at 15 W per gelfollowed by a further run at 25 W per gel until thebromophenol blue band reached the bottom of thegel. Gels were fixed and stained over night withcolloidal Coomassie blue and destained with deio-nised water until the background was significantlyreduced. The gels were scanned (Gel Doc 2000; Bio-Rad, Hercules, CA, USA), and the images wereprocessed with PDQuest software (Ver 7.0; Bio-Rad).

2.5. In-gel tryptic digest and massspectrometry

Spots were excised from the stained gel. Destainedwith 25 mM ammonium bicarbonate/50% acetoni-trile (Sigma) and dried with a SpeedVac plus SC110 (Savant Holbook, HY, USA). The gel was rehy-drated in trypsin solution (Promega, Madison, WI,USA). After incubation overnight at 37 8C, peptideswere first eluted with 5% TFA in 40 8C for 1 h, then2.5% TFA/50% acetonitrile at 30 8C for 1 h andremoval of acetonitrile by centrifugation in avacuum centrifuge. The peptides were concen-trated by using pipette tips C18 (Millipore, Bedford,MA). Analyses were performed primarily usingMALDI-TOF mass spectrometer (Burker company,German). Peptide mixtures were analyzed using asaturated solution of a-cyano-4-hydroxycinnamicacid (Sigma) in acetone containing 1% TFA. Peptideswere selected in the mass range of 800—4000 Da.The peptide sequence was determined with Mascotsoftware. Sequence homology was analyzed usingMascot program and the NCBI BLAST online searchservice. Peptide masses were assumed to be mono-isotopic masses, and cystines to be iodoacetamided.The peptide mass tolerance was set to 100 ppm, andthe maximum of missed cleavage sites was set to 1.Species was set to homo sapiens (human). Positiveidentification was achieved only when a 100 ppmmass accuracy met with a significant probabilityPROWL software score, and nearly all dominantsignals of the spectrum were assigned to the

192 X. Rushan et al.

identified protein. Candidates were further evalu-ated by comparison with their calculated massand pI, using the experimental values obtained from2-DE.

2.6. Western blot analysis

The indicated amounts of protein extracts obtainedfrom DPC were separated on a 12.5% (w/v) SDS-polyacrylamide gel. Then, proteins were trans-ferred to PVDF membrane (Roche). After blockingwith TBS-Tween 20 (TBST) containing 5% skim milk,the membranes were incubated with goat anti-human HSP70 mAb diluted 1:500 in TBST for 1 h,followed by peroxidase-conjugated goat anti-rabbitimmunoglobulin diluted 1:5000 in TBST for 1 h.Finally, membranes were washed three times withTBT, and blots were developed by the enhancedchemiluminescence (ECL, Roche). As control forequal protein loading, blots were re-stained usingmonoclonal mouse anti-glyceraldehyde-3-phos-phate dehydrogease (GAPDH) (kangchen Biotech,China). The intensity of each band was measuredby Quantity one (Ver 4.5, Bio-Rad).

2.7. RT-PCR experiment

Total RNA was extracted from passages 3 and 10DPC using Tripure according to the manufacturer’sinstruction. Reverse transcription with oligo(dT)priming was performed to generate cDNAs from2 mg total RNA using Superscript II following theinstruction provided by the manufacturer. DNAamplification was carried out with Taq DNA poly-merase (Takara, Japan) using the following pri-mers: b-actin (618 bp), 50-CGG GAC CTG ACT GACTAC CTC-30 and 50-CAA GAA AGG GTG TAA CGCAAC-30; MRPS7 (287 bp), 50-AAG CCA GTG GAG GAGCTA A-30 and 50-GCT TGA TGG AAG ATG GTG TA-30.PCR conditions were 94 8C for 5 min and then26 cycles of 94 8C for 30 s, 44 8C for 30 s and72 8C for 60 s, followed by incubation at 72 8Cfor 5 min for MRPS7 amplification. Amplified frag-ments were separated by electrophoresis on 1%agarose gels and visualized by ethidium bromidestaining. The intensity of each band was measuredby Quantity one (Ver 4.5, Bio-Rad), and intensityof MRPS7 were corrected by the intensity of b-actin.

2.8. Statistical analysis

Value is expressed as mean � standard deviation(S.D.). Student’s t-test and the x2-test were usedfor statistical analyses. Differences were consideredsignificant at P < 0.05.

3. Results

3.1. Cell morphology in passages 3 and 10DPC

The DPC in passage 3 demonstrated a distinctivesingle cell morphology and formed cell aggregatesat confluence which were believed as aggregativebehavior (Fig. 1A). However, passage 10 DPC wereradically different from the cell morphology of pas-sage 3 DPC and were believed to be lost this prop-erty (Fig. 1B). DPC losing aggregation regainaggregative behaviour when cultured supernantantof low passage DPC were added to them (Fig. 1C).

3.2. Two-dimensional electrophoresis andimage analysis of human DPC

To study the different proteins involved in aggrega-tion of DPC, total proteins of DPC were subjected to2-DE analysis. Typical Coomassie blue-stained gelsare shown in Fig. 2. Several hundred spots wereclearly identifiable in each gel. Image analysis of 2-Dgels revealed that averages of 618 � 22 and568 � 47 protein spots were detected in two groups,respectively, and the majority of these protein spotswere matched. About 300 protein spots werematched between passages 3 and 10 DPC, and thecorrelation coefficient was 0.75 by correlation ana-lysis of gels. The unmatched spots representedthose related to the aggregation of DPC as new orabsent proteins, which were the main alteration ofproteome. We select 24 protein spots having thedifference beyond five times for further study.

3.3. Identification of spots by MALDI-TOFMS

The spots of proteins were spread at the side of the2-DE map corresponding to a 4.0—9.0 pH range.However, to ensure an exact identification, an‘‘in-gel digestion’’ of manually excised spots wascarried out using trypsin with high specificity, andthe peptide mixtures deriving from tryptic hydro-lysis were analyzed by MALDI-TOF MS. Out of a totalof 24 spots excised from the gels, 17 spots yieldednearly perfect MALDI spectra. Fifteen spots werepreliminarily identified by PMF, these being the mostrepresented, as reported in Table 1. Fig. 3 shows thetypical PMF of spot 4, whose peptides matchedhuman transgelin. Mascot Search Results are shownin Fig. 4. The detailed matching result is listed inTable 2. The identified proteins are involved in thecellular processes of intercellular junctions, anti-apoptosis, metabolism, signal transduction and pro-tein post-translational processes. One spots with

Identification of proteins involved in aggregation of human dermal papilla cells 193

Fig. 2 2-DE patterns of cytoplasmic protein in passage 3 DPC (A) and passage 10 DPC (B). Protein (1.0 mg) was loadedand separated in IPG strips (pI range of 3—10), and the gels were stained with Brilliant Blue G-250. Spot numbers refer toproteins summarized in Table 1.

good MALDI spectra was returned with inconclusivematch results, suggesting that it might be unknownproteins.

3.4. Immunoblot confirmation of the up-regulation of HSP70 in passage 3 DPC

To confirm and extend the 2-DE results, HSP70expression in passage 3 was compared with thatof passage 10 DPC by Western blot analysis(Fig. 5). Equal protein loading was confirmed by

Table 1 Differential displayed protein spots preliminarily

Spot ID NCBInr code Peptidesmatched

TheoreticpI/Mr (kDa)

Sequencecoverage

1 gij45934285 6/39 6.04/17,298 51%2 gij51479152 8/36 6.60/15,820 69%

3 gij7705738 6/20 10/28,258 23%3-1 gij48255907 22/43 8.88/22,522 68%4 gij56967119 18/55 8.32/36,631 51%

5 gij913159 12/25 7.42/21,027 67%6 gij50345982 12/61 9.42/44,476 38%

10 gij67464392 20/50 8.22/60,277 49%

12 gij32891807 11/34 7.13/22,219 65%16 gij13921685 4/19 10.27/4,392 65%17 gij31417921 9/12 8.02/50,335 35%18 gij62897129 12/26 5.28/71,083 23%19-1 gij55960375 21/42 8.41/24,609 91%20-1 gij50845388 26/48 8.53/40,671 64%21 gij238427 9/44 8.63/30,737 32%

", Up-regulated in passage 3 DPC; –—, expressed in passage 10 DPCa (+) Expressed in passage 3 DPC.

parallel GAPDH immunoblotting. HSP70 was foundsignificantly up-regulated in passage 3 DPC. Thisresult confirmed our 2-DE data.

3.5. Semi-quantity RT-PCR

A specific antibody against MRPS7 is unavailable. Totest if the reduced content of MRPS7 in passage 3DPC reflected a decrease in transcription, compara-tive MRPS7 RT-PCR of passages 3 and 10 DPC wereperformed. Primers of MRPS7 were located within

identified by PMF

Score Protein name Alterationa

66 Globoside synthase mutant +100 ATP synthase, H+ transporting,

mitochondrial F0 complex,subunitd isoform b

+

69 Mitochondrial ribosomal protein S7 "168 Transgelin "139 Chain B, human annexin A2 in the

presence of calcium ions+

111 Neuropolypeptide h3 +93 ATP synthase, H+ transporting,

mitochondrial F1 complex, alphasubunit isoform b

+

87 Chain A, human muscle pyruvatekinase (Pkm2)

+

68 Biliverdin reductase B +69 Unnamed protein product +90 TKT protein +

108 Heat shock 70 kDa protein "230 Transgelin 2 "236 Annexin A2 isoform1 "80 Porin 31HM –—

.

194 X. Rushan et al.

Fig. 3 Typical MALDI-TOF-MS spectrum of spot 3-1 from the 2-DE map. The MS spectrum of the peptide mixture wasobtained from a typical in-gel digestion of the 2-DE separated protein. Numbers refer to peptides are listed in Table 2 andmatched peptides with transgelin are listed in Fig. 4.

exon 159 (forward) and exon 445 (reverse), respec-tively, generating a PCR product of 287 bp. In linewith the 2-DE findings, the extent of MRPS7 expres-sion exhibited considerable decrease in passage 10DPC (Fig. 6).

4. Discussion

In the present study, we report for the first time theapplication of a proteomic strategy for the compara-tive analysis of DPC with and without aggregativecharacters. Our approach was based on separation oftheproteinsby2-DE, computational imageanalysis ofthe resulting proteomemaps, and protein identifica-tion using mass spectrometry. Usually, large scalecomparative differential gene expression studies

Fig. 4 Mascot search results of spot 4, which is matched witNumber of mass values searched is 43 and number of mass v

can be addressed in terms of mRNA (transcriptome)or protein (proteome) levels. Transcriptome technol-ogy offers some technical advantages over proteo-mics, such as the possibility to work with minimalquantities of biological samples and well establishedplatforms towork in a high-throughputmode [18,19].Nevertheless, studies in yeast and mammalian cellshave shown that mRNA abundance is not directlycorrelated to protein levels [20,21]. This lack ofcorrelation makes us study the aggregative behaviorof DPC at the post-transcriptional level.

As a method to screen proteins on a large-scale,proteomics has been driven forward by the advent ofthe genome era, and it has advantages in analyzingtotal or specific proteins from cells. In this study, wepreliminarily identified 15 proteins, including trans-gelin, MRPS7, HSP70 and so on. Except for Porin 31

h transgelin. Peptides from spot 4 are shown in bold red.alues matched is 22. Sequence coverage is 68%.

Identification of proteins involved in aggregation of human dermal papilla cells 195

Table 2 The detailed result of database searching of spot 4 (listed in Table 1 whichmatched the record gij48255907 inthe NCBInr database.

TheoreticalMW (Da)

Peptidemass M—H+ (Da)

DMass (comparedwith theoretical) (Da)

Start—end Peptide

853.38 853.38 0.01 5—12 GPSYGMSR964.50 964.49 0.01 100—108 AAEDYGVIK989.58 989.57 0.01 50—57 LGFQVWLK993.46 993.44 0.02 147—154 GDPNWFMK

1009.49 1009.43 0.06 147—154 GDPNWFMK oxidation (M)1081.45 1081.46 �0.00 22—29 YDEELEER1203.64 1203.60 0.01 49—57 QMEQVAQFLK Pyro-glu (N-term Q)1209.55 1209.55 �0.00 21—29 KYDEELEER1220.64 1220.62 0.01 90—99 QMEQVAQFLK1236.64 1236.62 0.03 90—99 QMEQVAQFLK oxidation (M)1243.65 1243.63 0.03 79—89 VPENPPSMVFK1294.63 1294.60 0.03 162—172 EFTESQLQEGK1515.84 1515.83 0.01 65—78 LVNSLYPDGSKPVK1529.74 1529.71 0.03 109—121 TDMFQTVDLFEGK1545.76 1545.70 0.05 109—121 TDMFQTVDLFEGK oxidation (M)2108.11 2108.08 0.03 30—47 LVEWIIVQCGPDVGRPDR2317.10 2317.07 0.02 109—128 TDMFQTVDLFEGKDMAAVQR oxidation (M)2741.32 2741.45 �0.13 65—89 LVNSLYPDGSKPVKVPENPPSMVFK

MH, 14 identified proteins were specific or over-expresed in DPC with aggregative behaviour. Theaggregative behaviour of DPC is manifested as aresult of increased production of proteins specieswhich are in line with the Stenn’s study [22]. Theseproteins are related to various cellular responses,such as intercellular junctions, anti-apoptosis, cel-lular metabolism, signaling transduction and so on.They are novel signaling molecules or targets withno previously known function in aggregative beha-vior of DPC.

Among 14 preliminarily identified proteins, trans-gelin was overexpressed in DPC with aggregation.Transgelin is an actin-binding protein of an unknownfunction cross-linking actin filaments. This protein istransformation sensitive and gels actin, so Shapland

Fig. 5 HSP70 immunoblotting of passages 3 and 10 DPC.Western blot analysis revealedmarkedly overexpression inpassage 3 DPC (P3) when compared with that in passage 10DPC (P10).

et al. named it transgelin [23]. Transgelin is found infibroblasts and smooth muscle cells. It is also pre-sent in normal mesenchymal cells, secondary cul-tures of mouse and rat embryo fibroblasts. But theyare absent in many apparently normal fibroblast celllines. This protein is involved in various types of cellmobility and exocytosis, and can be regulated bycalcium [24]. It has been shown that down-regula-tion of transgelin may be an important early event intumour progression and is discussed as a diagnosticmarker for breast and colon cancer development[25]. In our study, we find that transgelin is up-expressed in passage 3 DPC and down-regulated inpassage 10 DPC which may reflect cell structural

Fig. 6 Comparative MRPS7 RT-PCR of passages 3 and 10DPC. In line with the results at the 2-DE level, the RT-PCRresults revealed marked down-regulation of MRPS7 inpassage 10 DPC.

196 X. Rushan et al.

changes from aggregation to non-aggregation.Whether it is a marker for the aggregation of DPCneed to be further investigated.

A significant reduction in the levels of the porin 31HM has been observed in DPC with aggregation. Porin31 HM is also named voltage-dependent anion-selec-tive channel protein 1 (VDAC-1) or outer mitochon-drial membrane protein porin 1. It is a pore-formingprotein discovered 26 years ago in the mitochondrialouter membrane [26]. It has recently been describedas being aNADH:ferricyanide reductase in theplasmamembrane where it establishes a novel level ofapoptosis regulation putatively via its redox activity.It regulates the cell growth and death [27]. Porinproteins are involved in the regulation of cellularmetabolism with a higher level of productionreported in hypoxic neuronal cells [28]. In this study,Porin 31 HM expression was decreased in passage 3DPC. Here, we propose that the down-regulation ofporin 31 HM reflects the increase of cellular meta-bolism in DPC with aggregative behavior.

Using a proteomics strategy, we detected twointeresting proteins involved in aggregative growthof DPC. One protein was heat shock proteins 70(HSP70). Heat shock proteins (HSPs) have a physio-logic function in unstressed cells, which is believedto include a role as a ‘‘molecular chaperone’’. Theparticipation of molecular chaperones in the pro-cess of senescence and in the mechanisms of age-related diseases is currently under investigation inmany laboratories [29]. Chaperone functionsmediated by HSP family constitute a fundamentalmechanism that governs the life span of organisms.Among different classes of chaperones, HSP70 arenow the major candidates in the gene-longevityassociation studies [30]. A significant associationof one HSP70 haplotype gene with male longevitywas observed [31]. Mortalin is the chaperone ofmitochondrial HSP70 protein which is a heat-unin-ducible stress protein involved in immortalizationand tumorigenesis. There are two mortalin alleles,mot-1 and mot-2, in mouse. Whereas an overexpres-sion of mot-1-induced senescence in NIH 3T3 cells,overexpression of mot-2 promoted their malignantproperties. So mortalin protein have differentialaging phenotypes [32]. In this work, it was up-regulated in DPC with aggregative behavior. Therelationship between HSP70 and aggregation ofDPC has not been determined before, so it is worthyof further research.

Another protein identified was human MRPS7.MRPS7 is also namedmitochondrial ribosomal proteinS7 or 30S ribosomal protein S7 homolog [33]. It is a 28-kDa protein with a pI of 10. MRPS7 is located at thehead of the small (30S) subunit of the ribosome andfaces into the decoding centre. It is one of the

primary 16S rRNA-binding proteins responsible forinitiating theassembly of theheadof the30S subunit.MRPS7 has been shown to be the major proteincomponent to cross-link with tRNA molecules boundat both theaminoacyl-tRNA (A) andpeptidyl-tRNA (P)sitesof the ribosome.MRPsare thought tobe involvedin the maintenance of the mitochondrial DNA [34].MRPS7 clearly plays an important role in ribosomefunction. In this study, MRPS7 was over-expressed inDPC with aggregative behavior which maybe relatedto synthesize more proteins to maintain the aggre-gative behavior of DPC.

In addition, seven metabolic enzymes, neuropo-lypeptide h3 and annexin A2 are overexpressed inDPC with aggregation. Most of metabolic enzymesare associated with the glycolytic pathway and thetricarboxylic acid (TCA) cycle. Neuropolypeptide h3was previously present in epithelial, muscular tis-sue, nervous tissue and testis [35,36]. We found thatneuropolypeptide h3 was present and overex-pressed in DPC with aggregation. Annexin A2 maybe involved in the regulation of Ca2+-dependentexocytosis and cell—cell adhesion mechanism[37]. But, their roles in aggregative behavior ofDPC are unclear and further research is needed todraw a conclusion.

In brief, we compared the proteomic profile ofcytoplasmic proteins in DPC with and without aggre-gative behavior. This strategy provided an efficientresolution to analyze aggregation-related proteinsdirectly at the protein level. Further studies will beperformed to determine the mechanism by whichthese proteins play a critical role in the aggregationof early passage DPC. This proteome analysis maycontribute to the elucidation of molecular mechan-ism of aggregative behavior in DPC.

Acknowledgments

This research was supported by the China Postdoc-toral Science Foundation (No. 2005038477). XiaRushan and Mou Zhirong contributed equally tothe study. Part of our studies was conducted inthe Institute of Immunology, the Third Military Med-ical University, Chongqing, China, so we thank Prof.Yu-Zhang Wu, Doctor Yu-Jun He and Wan-Ling Li.

References

[1] Hardy MH. The secret life of the hair follicle. Trends Genet1992;8:55—61.

[2] Jahoda CA, Oliver RF, Reynolds AJ, et al. Trans-species hairgrowth induction by human hair follicle dermal papillae. ExpDermatol 2001;10:229—37.

Identification of proteins involved in aggregation of human dermal papilla cells 197

[3] Inamatsu M, Matsuzaki T, Iwanari H, Yoshizato K. Establish-ment of rat dermal papilla cell lines that sustain the potencyto induce hair follicles from afollicular skin. J Invest Derma-tol 1998;111:767—75.

[4] Robinson M, Reynolds AJ, Gharzi A, Jahoda CA. In vivoinduction of hair growth by dermal cells isolated from hairfollicles after extended organ culture. J Invest Dermatol2001;117:596—604.

[5] Almond-Roesler B, Schon M, SchonMP, et al. Cultured dermalpapilla cells of the rat vibrissa follicle. Proliferative activity,adhesion properties and reorganization of the extracellularmatrix in vitro.. Arch Dermatol Res 1997;289:698—704.

[6] Bratka-Robia CB, Mitteregger G, Aichinger A, et al. Primarycell culture and morphological characterization of caninedermal papilla cells and dermal fibroblasts. Vet Dermatol2002;13:1—62.

[7] Horne KA, Jahoda CA, Oliver RF. Whisker growth induced byimplantation of cultured vibrissa dermal papilla cells in theadult rat. J Embryol Exp Morphol 1986;97:111—24.

[8] Zhiqiang S, Jiwen W, Fei H, et al. Identification of differ-entially expressed genes HSPC016 in dermal papilla cellswith aggregative behaviour. Arch Dermatol Res 2005;297:114—20.

[9] Sleeman MA, Murison JG, Strachan L, Kumble K, Glenn MP,McGrath A, et al. Gene expression in rat dermal papilla cells:analysis of 2529 ESTs. Genomics 2000;69:214—24.

[10] McElwee KJ, Hoffmann R. Growth factors in early hairfollicle morphogenesis. Eur J Dermatol 2000;10:341—50.

[11] Itami S. Pathomechanism of androgenetic alopecia and newtreatment. Nippon Ronen Igakkai Zasshi 2004;41:598—600.

[12] Inui S, Fukuzato Y, Nakajima T, et al. Identification ofandrogen-inducible TGF-beta1 derived from dermal papillacells as a key mediator in androgenetic alopecia. J InvestigDermatol Symp Proc 2003;8:69—71.

[13] O’Shaughnessy RF, Christiano AM, Jahoda CA. The role ofBMP signalling in the control of ID3 expression in the hairfollicle. Exp Dermatol 2004;13:621—9.

[14] Luo Y, Hao F, Zhong BY, et al. The biological activities ofconditioned medium derived from human dermal papillacells cultured in vitro. Chin J Dermatol 2004;37:648—50(in Chinese).

[15] Swinbanks D. Government backs proteome proposal. Nature1995;378:653.

[16] Righetti PG, Castagna A. Recent trends in proteome analysis.Adv Chromatogr 2003;42:269—321.

[17] Yang WB, Hao F, Song ZQ, et al. Apoptosis of the dermalpapilla cells of hair follicle associated with the expression ofgene HSPCO16 in vitro. Exp Dermatol 2005;14:209—14.

[18] Nguyen C, Rocha D, Granjeaud S, Baldit M, Bernard K,Naquet P, et al. Differential gene expression in the murinethymus assayed by quantitative hybridization of arrayedcDNA clones. Genomics 1995;29:207—16.

[19] Velculescu VE, Zhang L, Vogelstein B, Kinzler KW. Serialanalysis of gene expression. Science 1995;270:484—7.

[20] Gygi SP, Rochon Y, Franza BR, Aebersold R. Correlationbetween protein and mRNA abundance in yeast. Mol CellBiol 1999;19:1720—30.

[21] Anderson L, Seilhamer J. A comparison of selected mRNAand protein abundances in human liver. Electrophoresis1997;18:533—7.

[22] Stenn KS, Combates NJ, Eilertsen KJ, et al. Hair folliclegrowth controls. Dermatol Clin 1996;14:543—58.

[23] Shapland C, Hsuan JJ, Totty NF, Lawson D. Purification andproperties of transgelin: a transformation and shape changesensitive actin-gelling protein. J Cell Biol 1993;121:1065—73.

[24] Lawson D, Harrison M, Shapland C. Fibroblast transgelin andsmooth muscle SM22alpha are the same protein, the expres-sion of which is down-regulated in many cell lines. Cell MotilCytoskeleton 1997;38:250—7.

[25] Sitek B, Luttges J, Marcus K, et al. Application of fluores-cence difference gel electrophoresis saturation labelling forthe analysis of microdissected precursor lesions of pancrea-tic ductal adenocarcinoma. Proteomics 2005;5:2665—79.

[26] Lawen A, Ly JD, Lane DJ, et al. Voltage-dependent anion-selective channel 1 (VDAC1)–—a mitochondrial protein,rediscovered as a novel enzyme in the plasma membrane.Int J Biochem Cell Biol 2005;37:277—82.

[27] Baker MA, Lane DJ, Ly JD, et al. VDAC1 is a transplasmamembrane NADH-ferricyanide reductase. J Biol Chem2004;279:4811—9.

[28] Shinohara Y, Ishida T, Hino M, et al. Characterization of porinisoforms expressed in tumor cells. Eur J Biochem 2000;267:6067—73.

[29] Brocchieri L, Conway de Macario E, Macario AJ. Chapero-nomics, a new tool to study ageing and associated diseases.Mech Ageing Dev 2007;128(1):125—36.

[30] Singh R, Kolvraa S, Rattan SI. Genetics of human longevitywith emphasis on the relevance of HSP70 as candidategenes. Front Biosci May 2007;12:4504—13.

[31] Singh R, Kolvraa S, Bross P, Christensen K, et al. Heat-shockprotein 70 genes and human longevity: a view from Den-mark. Ann NY Acad Sci May 2006;1067:301—8.

[32] Deocaris CC, Yamasaki K, Kaul SC, Wadhwa R. Structural andfunctional differences between mouse mot-1 and mot-2proteins that differ in two amino acids. Ann NY Acad Sci2006;1067:220—3.

[33] Cavdar Koc E, Blackburn K, Burkhart W, Spremulli LL. Iden-tification of a mammalian mitochondrial homolog of ribo-somal protein S7. Biochem Biophys Res Commun 1999;266:141—6.

[34] Myers AM, Pape LK, Tzagoloff A. Mitochondrial proteinsynthesis is required for maintenance of intact mitochon-drial genomes in Saccharomyces cerevisiae. EMBO J 1985;4:2087—92.

[35] Bollengier F, Mahler A. Localization of the novel neuropo-lypeptide h3 in subsets of tissues from different species. JNeurochem 1988;50:1210—4.

[36] Seddiqi N, Bollengier F, Alliel PM, et al. Amino acid sequenceof the Homo sapiens brain 21—23-kDa protein (neuropoly-peptide h3), comparison with its counterparts from Rattusnorvegicus and Bos taurus species, and expression of itsmRNA in different tissues. J Mol Evol 1994;39:655—60.

[37] Siever DA, Erickson HP. Extracellular annexin II. Int J Bio-chem Cell Biol 1997;29:1219—23.