1

2

3Q1

4

5

6

7

8

1011

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47484950

J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

ScienceDirectwww.e l sev i e r . com/ loca te / j p ro t

JPROT-01898; No of Pages 10

From hundreds to thousands: Widening the normalhuman Urinome (1)

OO

F

Laura Santuccia,1, Giovanni Candianoa,1, Andrea Petrettob, Maurizio Bruschia,Chiara Lavarellob, Elvira Ingleseb, Pier Giorgio Righettic, Gian Marco Ghiggeria,⁎aNephrology, Dialysis, Transplantation Unit and Laboratory on Pathophysiology of Uremia, Istituto Giannina Gaslini, 16148 Genova, ItalybLaboratory of Mass Spectrometry, Core Facility, Istituto Giannina Gaslini, 16148 Genova, ItalycPolitecnico di Milano, Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Via Mancinelli 7, 20131 Milano, Italy

R

A R T I C L E I N F O

UNCO

R

DOI of original article: http://dx.doi.org/10⁎ Corresponding author at: Division of Nephro

Gaslini, Largo G. Gaslini 5, 16148 Genova, ItaE-mail address: GMarcoGhiggeri@ospedal

1 These two authors have equally contribu

http://dx.doi.org/10.1016/j.jprot.2014.07.0211874-3919/© 2014 Elsevier B.V. All rights rese

Please cite this article as: Santucci L, et alhttp://dx.doi.org/10.1016/j.jprot.2014.07.02

A B S T R A C T

Q3

Article history:Received 20 March 2014Accepted 15 July 2014

Q2

RECTED PIt is currently unknown howmany proteins can be detected in urine. Improving the analytical

approachwould increase their number and potentially strengthen their predictive potential indiseases.We developed a combination of analytical procedures for maximizing sensitivity andreproducibility of normal human urinary proteome analysis based on ultracentrifugation,vesicles separation, combinatorial peptide ligand libraries (CPLLs) and solvent removal ofpigments. Proteins were identified by an Orbitrap Velos Mass Spectrometry.Overall, 3429 proteinswere characterized:most components (1615) were contained in vesicleswhile the remaining 1794 were equally distributed among CPLL and butanol insolublefractions. Several proteins were detected exclusively in one of the phases of the procedure,suggesting that each step is crucial in the fractionation strategy. Many (1724) proteins aredescribed here whose presence in urine has never been reported and represents a potentialsource of information considering that urine is the unique site of excretion of products ofinteraction of metabolic processes. Improving the characterization of normal urinaryproteome would also represent the basis for the analysis of urine biomarkers in humandiseases.

Biological significanceSub-fractionating normal urine by successive steps (vesicle separation, CPLL and solventtreatments) allowed the identification of 3429 proteins, a relevant part (1724) being detectedfor the first time in urine. Several proteins of new description have been implicatedin physiology pathways and in pathologies thus representing a potential source of newinformation on both metabolic processes and diseases.

© 2014 Elsevier B.V. All rights reserved.

Keywords:Urinary proteomeVesiclesCombinatorial peptideligand librariesMass spectrometry

.1016/j.dib.2014.08.006.logy, Dialysis and Transplantation and Laboratory on Pathophysiology of Uremia, Istituto G.ly. Tel.: +39 010 380742; fax: +39 010 395214.e-gaslini.ge.it (G.M. Ghiggeri).ted to the work and share first authorship.

rved.

, From hundreds to thousands: Widening the normal human Urinome (1), J Prot (2014),1

5152

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

2 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

1. Introduction

T

111

112

113

114

115

116

117

118

119Q4

120

121

122123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143Q5

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

UNCO

RREC

The composition of biological fluids represents a potentiallyrelevant source of information for research in human diseasesparticularly in the area of biomarker development. Urine is anideal fluid for such investigations, since it is the physiologic siteof excretion ofmostmolecules deriving from biological process.Products of cell lesions or cell transformation could be alsodetected provided that they are freely filtered through theglomerular sieves. Urine also drains fluids from the urinarytract and is for this reason particularly enriched in proteinsderiving from the kidney, bladder and prostate. Finally, urinecan be easily collected, does not require invasive proceduresand can be handily manipulated and stored [1–7]. Improvingour knowledge on urine protein composition is crucial to anyevolution in the field of mechanisms, early detection andmonitoring of diseases, in particular in those conditions affect-ing the urinary tract.

In recent years, technology advancements have markedlymodified the number of proteins that can be detected in urineandprotocols havebeen developed for the analysis of intra- andinter-individual variability, e.g. age, diet, and time of collection.[8–12]. In fact, proteomic research on urine has a long historicalrecord. In 1979, theAnderson's group published the first studiesby two-dimensional electrophoresis (2-DE) on normal urine[13,14] in which they identified only the major components,in general the same ones present in serum plus some lowmolecular mass peptides. Up to the end of the century nomoreimprovements were obtained [15] when it became clear thatmore than one technique should be utilized for improvingsensitivity. Thus several groups analyzed the “normal humanurinary proteome” by combining liquid chromatography withtandemmass spectrometry (LC–MS) [16,17] and/or utilizing twomethods of sample fractionation, e.g. acetone precipitation andultracentrifugation. [18–21]. These new approaches notablyimproved sensitivity, so that in 2011 the total normal urinaryproteome listed about 2000 proteins [22–30].

One of the general problems encountered with normal urineanalysis is that low expression components are practicallyundetectable since fifty spots represent 99% of totality thusmasking the low-abundance species. Actually, the same dispro-portion between high and low expression proteins also exists inserum and with respect to the composition in major protein,i.e. albumin, transferrin, immunoglobulins, etc., urine is similarto serum. It is for this reason that combinatorial peptide ligandlibraries (CPLLs) [31–33] were developed to increase sensitivitytowards low expression peptides (for reviews, see [34–37]).

Separation of exosomes has been the second evolution[38–44]. Exosomes are extracellular membrane-bound nano-vesicles secreted by diverse cell types along the urinary tract[45,46]. They are expected to play a role in immune systemmodulation and in the transport of proteins, mRNAs andmicroRNA, overall influencing angiogenesis, cell proliferationand tumor cell invasion [47,48]. Vesicles are ubiquitous andurine is very enriched of them [49–57]. However, their purifica-tion isnot an easy task, since, in urine, they interactwith solubleproteins such as of uromodulin and, in spite of minimizingaggregation phenomena by treatment with reducing agents,they appear to be frequently contaminated [41,58–61]. Overall,

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

CPLL treatment and purified vesicle collection, coupled withmodern mass spectrometry, substantially increase our analyt-ical view on the global urinary proteome and should representthe basis to develop common laboratory strategies.

In this study, we merged different methods of proteinseparation to optimize the efficiency of low-abundance proteindiscovery. The objective was to bring the field close to a finalsurvey, and put the basis for planning studies in humandiseases. The result is the development of a sequence oftechniques that altogether allow the detection of 3500 singleproteins in normal urine and constitutes a good starting pointfor the objectives above.

ED P

RO

OF

2. Materials and methods

2.1. Materials

ProteoMiner™, urea, thiourea, 3-[3-cholamidopropyl di-methylammonio]-1-propansulfonate (CHAPS), dithioerythritol(DTE), Tris(hydroxymethyl)aminomethane (Tris), ethylendiami-no tetracetic acid (EDTA) and sodiumdodecyl sulfate (SDS) werepurchased by Sigma-Aldrich (St Louis, Mo, USA). Completeprotease inhibitor cocktail tablets were from Roche Diagnostics,(Basel, CH). Sequencing grade bovine trypsinwas from Promega(Madison,WI, USA). All other chemicals of analytical gradewerefrom J.T. Baker (Deventer, Holland).

2.2. Sample collection and urine preparation

Second morning urine samples (approximately 160 mL) fromhealthy volunteers (12 individuals, age 35–45 years, 6 malesand 6 females) were collected after informed consent andwere immediately added with tablets of protease inhibitor,chilled on ice and centrifuged at 4 °C for 10 min at 1000 ×g toeliminate cell debris in accord to Standard Protocols (seeEuroKUP-European Kidney and urine proteomics at http://www.eurokup.org). At this point urine was frozen and main-tained at −80 until further processing. Second step wascentrifugation at 17,000 ×g in a JA-20 rotor (Beckman Avanti™J-25) for 30 min at 16 °C to remove the urinary microscopicparticle sediments. Then the collected urine was pooled (twodifferent pools, each consisting of three males and threefemales) and analyzed in technical triplicates to confirm thereproducibility of the experiments. An aliquot of the total17,000 ×g urinary supernatant, after Bradford protein assay,was dialyzed three times against 25 mM sodium phosphatepH 7.4 in 3500 MWCO Spectra/Por® cellulose membranes at4 °C. After lyophilization this material was stored at −80 °Cuntil use.

2.3. Urinary vesicles isolation

The 17,000 ×g urinary supernatant (80 mL) was ultra-centrifuged in 8 tubes at 48,000 rpm in a Ti 90 rotor (BeckmanOptima™ l-90K) for 75 min at 18 °C. The ultracentrifugationstep was repeated by adding the same volume used beforeuntil vesicles were isolated; DTT (200 mg/mL) was added to48,000 rpm pellets and incubated for 2 min at 95 °C to avoidthe formation of high molecular weight complexes of vesicles

sands: Widening the normal human Urinome (1), J Prot (2014),

T

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

3J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

UNCO

RREC

with uromodulin. The pellets were pooled together andwashed three times with Tris–HCl 65 mM pH 8.8 and centri-fuged a 14,000 rpm for 10 min at 4 °C and stored at −80 °Cuntil mass spectrometry analysis.

2.4. Butanol precipitation

The 48,000 rpm supernatant was dialyzed versus water;aliquots of 50 mL were added with 100 μL of acetic acid(about pH 3–4) and after a few minutes with 10 mL ofn-butanol; this mixture was rotated in agitation for 90 minand centrifuged at 4,000 rpm for 10 min at 18 °C. Threedifferent phases were obtained: a first phase consistedin a protein pellet, a second was the supernatant (forProteoMiner™, see below) and a third phase consisted inpigments (that were discarted), yielding two fractions: CPLL-beads and unbound or flow-through.

2.5. Combinatorial solid-phase ligand library(ProteoMiner™)

The phase deriving from butanol extraction was lyophilizedand loaded onto a column of 150 μL peptide library beads(ProteoMiner™) equilibrated in 25 mM phosphate buffer,pH 7.4 as already described by Candiano et al. [33]. Sodiumazide and EDTA were added to samples. The eluate in Tris–SDS (CPLL-beads) and the unbound fraction were preserved at−80 °C until analysis by mass spectrometry.

2.6. Mass spectrometry

2.6.1. ApparatusA linear Trap Quadropole (LTQ) Orbitrap Velos Pro MassSpectrometry was used for the analysis of all urinary sub-fractions.

2.6.2. Sample preparation and procedureSamples for mass spectrometry were solubilized in 0.1 mL of 4%SDS, 50 mM DTT, and 0.1 M Tris/HCl, pH 7.6, at 90 °C for 5 minand briefly sonicated and were processed by the FASP procedureusing 30 kVivacom filtrationdevices (Sartorius) [62]. Briefly, 50 μLaliquots were mixed with 0.2 mL of 8 M urea in 0.1 M Tris/HCl,pH 8.5 (UA solution), loaded into the filtration devices andcentrifuged at 14,000 g for 15 min. The concentrates werediluted in the deviceswith 0.2 mLofUA solution and centrifugedagain. After centrifugation, the concentrates were mixed with0.1 mL of 50 mM iodoacetamide in UA solution and incubatedin darkness at room temperature (RT) for 30 min followed bycentrifugation for 15 min. Then, the concentrate was dilutedwith 0.2 mL of 8 M urea in 0.1 M Tris/HCl, pH 8.5 (UB solution),and concentrated again. This step was repeated twice. Thesamples were then diluted with 120 μL of 40 mM NaHCO3

containing 1 μg of trypsin. Following anovernight digestion, thepeptides were collected by centrifugation of the filter units for20 min and the filter device was rinsed with 50 mL 0.5 M NaCland centrifuged in order to suppress hydrophobic interactions.

2.6.3. NanoLC setupThe sample was first loaded from the sample loop onto atrapping column (Acclaim PepMap C18; 2 cm × 100 μm × 5 μm,

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

ED P

RO

OF

100 Å) using the loading solvent at a flow rate of 5 μL/min for5 min. The trapping column was then switched in-line with theseparation column and the peptides eluted with increasingorganic solvent at a flow rate of 300 nL/min. The peptideseparations were carried out at 30 °C in a 75-μm ID × 15 cm2 μm, 100 Å C18 column mounted in a thermostated columncompartment. The peptides were separated and eluted with alinear gradient of 5–35% solution B (99.9% ACN, 0.1% formic acid)in 150 min.

2.6.4. Mass spectrometer setupThe mass spectrometer LTQ-Orbitrap Velos Pro was operatedin the positive ionization mode. Single MS survey scans wereperformed in the Orbitrap, recording a mass window between350 and 1650 m/z using a maximal ion injection time of250 ms. The resolution was set to 60,000 and the automaticgain control was set to 1,000,000 ions. The lock mass optionwas enabled allowing the internal recalibration of spectrarecorded in the Orbitrap by polydimethylcyclosiloxane back-ground ions (protonated (Si(CH3)2O)6; m/z 445.120025). Theexperiments were done in data-dependent acquisition modewith alternating MS and MS/MS experiments. The minimumMS signal for triggering MS/MS was set to 500 ions, with themost prominent ion signal selected for MS/MS using anisolation window of 2 Da. The m/z values of signals alreadyselected for MS/MS were put on an exclusion list for 60 s usingan exclusion window size of ±10 ppm. In all cases, onemicro-scan was recorded. CID was done with a target valueof 5000 ions in the linear ion trap, a maximal ion injectiontime of 150 ms, normalized collision energy of 35%, aQ-value of 0.25 and an activation time of 10 ms. A maximumof 20 MS/MS experiments were triggered per MS scan.

The standard filter of 10 ppm was chosen on the basis ofcurrent accepted criteria for deep and unbiased proteomecoverage (see for example [63]). As reported below and inaccord with the literature [24], we did not restrict our searchto the “two-peptide” rule and adopted as well the singlepeptide identification approach, as suggested by Gupta andPevzner [64]. These authors confirmed that the “two-peptide”rule reduces the number of protein identifications in thetarget databasemore significantly than in the decoy databaseand results in increased false discovery rates, compared tothe case when single-hit proteins are not discarded.

2.6.5. Data analysisRaw MS files were processed with the Thermo ScientificProteome Discoverer software version 1.3. Peak list files weresearched by the MASCOT and SEQUEST search engine againstUniprot human database (release 2012_07) containing bothforward and reversed protein sequences. The resulting peptidehits were filtered for a maximum 1% FDR using Percolator, thepeptide mass deviation was set to 10 ppm and a minimum ofsix 6 amino acids per identified peptide were required. TheDatabase search parameters were mass tolerance precursor20 ppm [65], mass tolerance fragment CID 0.8 Da with dynamicmodification of deamidation (N, Q), oxidation (M) and staticmodification of alkylation with IAM (C). For all searches theoption trypsinwith twomissed cleavageswas selected. Proteinswere grouped by applying the maximum parsimony rule. Thecomputational methods used to determinate the absolute

sands: Widening the normal human Urinome (1), J Prot (2014),

T

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332Q6

333

334

4 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

C

protein quantification was average of the three most abundantpeptides.

2.6.6. Bioinformatic analysisLabel free experiments were analyzed with the freely availablePerseus software, which includes all necessary functionalities(http://www.perseusframework.org/Perseus_1.4.1.3.zip). Proteingroups were filtered to require in at least one experi-mental group three valid values. Label free intensities werelogarithmized and missing values were imputed with randomnumbers from a normal distribution,whosemean and standarddeviation were chosen to best simulate low abundance valuesclose to noise level. Single peptide identificationwas accepted incase the same peptide could be identified 4 times on twobiological replicates performed in triplicate. A modified ANOVAandT-testwith permutation based FDR statistics (6)was appliedto filter the first results. We performed 250 permutationsand required an FDR of 0.05 and 0.01 for Anova and T-testrespectively. The parameter S0 was empirically optimizedto separate outliers from the background distribution. Hierar-chical clustering of resulting proteins was performed onlogarithmized intensities after z-score normalization of thedata, using Euclidean distances. The protein categorical anno-tation was supplied by Gene Ontology biological process,molecular function, and cellular component. All annotationwas extracted from the UniProt database. The “annotationmatrix algorithm”, implemented in Perseus software, and theHeatmap were used to describe the differential enrichment ofspecific Gene Ontology biological process terms in the samples.The software filters protein annotation terms by testing thedifference among means for any protein annotation from theoverall ratio distribution for all cells of the expression matrix.The statistical test is a two-dimensional non-parametric

UNCO

RRE

Fig. 1 – A: Flowchart of the procedure used for analysis of the urisupernatant was treated with a butanol precipitation and with PCandiano et al. [33]. The 5 fractions thus obtained (untreated, veprocessed by mass spectrometry analysis. B: Heat map of Pearsonthe label-free replicates (identified with “1 or 2 or 3” and “a or b” rthe five different fractions.

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

Mann–Whitney test, with a Benjamini–Hochberg multiplehypothesis testing correction that was controlled by using afalse discovery rate threshold of 0.05. All categories that survivedthe test for at least one of the samples were converted to onerow of the annotation matrix. The values for each row arethe differences between groups of the Mann–Whitney testmentioned above. Fisher's exact test was performed witha false discovery rate value of 0.02. See data available atChorusProject.org under project name “From hundreds tothousands: widening the normal human Urinome”.

ED P

RO

OF3. Results

3.1. Protein composition of un-fractionated normal urine

Without fractionation, urine is, on average, found to contain1176 unique proteins. This number has been obtained byanalyzing with mass spectrometry the second morning urinesamples from 12 young normal volunteers (6 females, 6males,age 35–45 years.) pooled in two samples that were evaluatedthree times each. We calculated the coefficient of variation(CV = standard deviation/mean) across the three replicatesof the same sample for each protein, identified to providea direct quantitative measure of technical variability, andbetween the several samples for biological variability [11]. Themedian CV value of the six runs (technical variability) rangedfrom 0.012 to 0.047 while that of the biological variabilityranged between 0.11 and 0.15.

The list of single proteins detected in un-fractioned urine isgiven in Table 1 in Data in Brief article. Technical variabilityamong six determinations was negligible in terms of presenceor absence, relevant if concentration was considered.

nary proteome. After obtaining vesicles from the pellet, theroteoMiner followed by Tris & SDS–DTE elution as persicles, precipitate, CPLL-beads eluate and unbound) werecorrelation coefficients showing good reproducibility amongespectively for technical triplicates of biological duplicates) of

sands: Widening the normal human Urinome (1), J Prot (2014),

T335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

Fig. 2 – Venn Diagram of the 3429 non-redundant proteinsdetected in a normal urinary proteome with a peptide ID of1% FDR, at a peptide mass deviation of 10 ppm and aminimum of six 6 amino acids per identified peptide. Thetotal proteins identified in each fraction was: 1176(untreated), 1615 (vesicles), 993 (precipitate), 1488(CPLL-beads) and 1345 (unbound).

5J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

3.2. Normal urine composition step by step

Fractionating urine by a multistep procedure increased thenumber of detectable proteins to an overall of 3429, differentlydistributed among the successive steps of the sequence. Thismeans that each fractionation step increases the concentra-tion of single proteins that could not otherwise be detected in

UNCO

RREC

Fig. 3 – Unsupervised hierarchical clustering analysis of protein expand filtering for three valid values for each technical replicate ofshows the segregation of the each urinary fraction in five clustersanalysis, of the same fractions, represents in each row the clusteurine fractions with six values of technical replicates are displayproteins in the different fractions.

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

ED P

RO

OF

unfractionated samples. When the results were matchedwith the Arabidopsis database (most frequently utilizing onepeptide and in a few three peptides as matching system) thenumber of identified proteins corresponded to 77 differentaccessions, represented in most cases by interspecies pre-served molecules with basic fundamental functions for celllife (e.g. heat shock proteins, histone chaperone proteins,peroxidase, RNA polymerase). Comparing peptides identifiedby using the human database (n 14,023) with Arabidopsisdatabase (n 92) supported the accuracy of the results reportedin our work.

This improvement was obtained by using about 150 mL ofurine and utilizing the multi-step flow chart illustrated inFig. 1A, that includes: 1) ultracentrifugation for obtainingvesicles, 2) extraction with butanol of supernatant derivingfrom ultracentrifugation to eliminate pigments and 3) CPLL toenhance the signal of low-expression components. Variabilityin purification efficiency was evaluated by repeating eachsingle passage three times on the same pool of urine and wasfound very low. The heatmap (Fig. 1B) analysis of compositionof each purification step (untreated, vesicles, precipitateafter butanol extraction, bound and unbound material toCPLL-beads) shows a good reproducibility among the repli-cates of five sample preparation methods. In this case, thecolor map highlights the groups of samples that have similarcorrelations on a scale of colors from black (maximun) towhite (minimum) where the strict replication is confirmed bydrawing a theoretical diagonal (in gray in Fig. 1B).

Proteins that were purified uniquely in a step of theprocedure were numerous (overall 1678), as shown by the

ression profiles of the five urine fractions after normalizationbiological duplicates. A) Principal component analysis (PCA), according to the different sample preparations. B) Heat mapring of protein expression profiles, while in each column theed. The strongest colors indicate the relative abundances of

sands: Widening the normal human Urinome (1), J Prot (2014),

371

372

373Q7

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

6 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

number in bold in the Venn diagram of Fig. 2 summarizingthese aspects; details on the identity of single proteins arelisted in Table 1 in the Data in Brief article. This figure alsoreports that most proteins were contained in vesicles (1615),followed byCPLL (1488) and by other precipitated andunboundfractions (993 and 1345 proteins respectively). Fig. 3B displaysa heat map analysis relative to the concentration of eachsingle protein in different purification phases: this image (inwhich black means high levels and white minimum) clearlyshows clusters of proteins in vesicles, CPLL, etc., withminimalnumbers of proteins shared by the different procedures. Thisfinding suggests a good purification efficacy overall. The sameconcept is illustrated in Fig. 3A that shows, with the principalcomponent analysis (PCA), clusters of proteins in differentfractions of the method. Also a good reproducibility was

UNCO

RRECT

Fig. 4 – Volcano plot used graphically to represent the results of tonly species that exceeded the threshold for significance in the Tthe − log 10 of the P-value. In all panels the proteins of the untrea(panel b), unbound (panel c) and vesicles (panel d) are shown.

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

F

confirmed by the near overlap of the six samples (tripledtechnical, duplicate biological). In other words, the multistepprocedure described here augments the possibility of isolatingsingle proteins from urine with a good degree of purity (seesection below).

How much every single step improves the purificationefficacy in respect to un-fractioned urine and which proteinstake major benefit from this approach is shown in Fig. 4 and inSupplementary Table 1. In the former case, by exploiting aVolcanoplot, the level of proteinswhich exceeded the thresholdfor significance in the T-test was compared to un-fractionedurine. The statistically significant under-expressed (on the left)and over-expressed (on the right) proteins are plotted in blackvs. gray. The proteins considered in all these plots were 1526and were differently distributed among the different phases of

ED P

RO

O

he T-tests for differential expression of proteins. It includes-test. The log 2 fold change for each protein is plotted againstted fraction (control) versus CPLL-beads (panel a), precipitate

sands: Widening the normal human Urinome (1), J Prot (2014),

OO

F

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416417

418

419

420

421

422

423

424

425

426

427

428

Fig. 6 – Venn diagram comparing our study with four otherrelevant reports on urinary proteome. A total of 2883proteins of new description have been identified(see Supplementary Table 2).

7J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

the purification procedure. For example, when consideringCPLL (shown in a), at least 323 proteins were concentrated by afactor of 4 or more and the same was for other conditions ofpurification (Fig. 4b, c, d; the identification of over-expressedproteins in the Volcano plot is reported in SupplementaryTable 1).

3.3. Functional clusters and single protein, where and how

Fig. 5 shows that, when considering the function of any singleprotein purified in different steps, there was a sort of clusterseparation mainly related to the presence of nuclear proteinsin vesicles. In fact, vesicles contain proteins deriving from thenucleus andmitochondria that are, in general, nucleotide andATP binding families (in red in Fig. 5) while, vice versa,receptor, cell adhesion, signal, trans-membrane and secretedproteins are maximally contained in the fraction derivingfrom CPLL capture (Fig. 5).

Remarkably, in the data presented in SupplementaryTable 2 there are 1724 proteins that have never been detectedin urine before, which, on this basis, can be defined as‘proteins of new description’ in this biological fluid. Detectionof new proteins in urine means that other recent studies onthe normal urine proteome failed to detect their presence.However, as shown by the Venn diagram presented in Fig. 6,

UNCO

RRECT

429

430

431

432

433

Fig. 5 – Gene Ontology analysis of the five urine fractions,obtained via the scheme of Fig. 1A, and displaying theclassification of biological processes of proteins exhibitingdifferent physic-chemical proprieties.

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

ED P

R

other relevant studies on urine proteome [23,24,30,66] coulddetect urine proteins of new description on their own. Overall,considering our and other studies, 2883 new proteins havebeen described (Supplementary Table 2 identifies all singlenew urinary proteins). A detailed discussion consideringevery component, its significance and possible use as abiomarker is far beyond the scope of the present workand will be presented separately for different organs andapparatuses.

434435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

4. Discussion

The main objective of our study was to widen the knowledgeon protein composition of normal urine while maintaining ahigh reproducibility. The strategy was to utilize moderntechniques of analysis (mass spectrometry of last generation,CPLLs) and traditional approaches (ultracentrifugation, sol-vent extraction, etc.) while combining them in a virtuoussuccession of steps altogether allowing for maximal efficien-cy. We thus identified 3429 non-redundant proteins in normalurine with a net increment of 2253 proteins not detectable inun-fractioned urine. One basic methodological aspect relatedto the sensitivity of the method needs to be discussed here. Infact, as reported above and in accord with the literature [64],we adopted as standard a cut off 10 ppm. The same limit iswidely utilized in the literature and represents, in our opinion,the gold standard of the technique (see for example [63]).Studies on urine proteome almost uniquely utilized the filterof 10 ppm and the most up-dated databases [23,24,30,66]utilized this condition (Fig. 6). Actually, when adopting morestringent conditions of analysis (i.e. 3 ppm) on the same urine,only 3071 proteins and 13,123 peptides could be detected witha decrease of 10% and 6% respectively (Supplementary Table3). However, we chose to present the results generatedfollowing similar parameters as described in current litera-ture; comparison to results obtained with less or higherstringent conditions will also be feasible when other data-bases will appear. Many of the newly described proteins that

sands: Widening the normal human Urinome (1), J Prot (2014),

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495496

497

498499

500

501

502503

504

505

506

507

508

509

510

511Q8

51 2

513514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

8 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

could be visualized were much more concentrated in one ormore steps of the procedure. It is of interest that functionalclusters of proteins were purified by selected procedures (e.g.nuclear proteins in vesicles) suggesting that analysis and/orpurification strategies are simplified.

Besides the general considerations above, it is noteworthythat many of the proteins here described have never beenreported in urine before and have been defined here as ‘ofnew description’ on this basis. Several of them are of interestto clinical chemistry as biomarkers of relevant human dis-eases as the list is quite long and involves pathologies ofdifferent organs and systems. A detailed description herewould appear too long and out of the scope of the presentinvestigation. Reports focused on specialized journals wouldbetter suit the objective to attract clinician interest and givethem the opportunity to look at urine for diagnosis of selecteddiseases.

T531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

EC

5. Conclusions

By coupling Orbitrap Mass spectrometry with a few simplefractionation steps almost 3500 proteins could be detected innormal urine, 50% of which are of new description. Assem-bling databases dedicated to different medical specialitieswould constitute the basis for looking at human pathologies.Given the number of proteins detected for the first timein urine, the wealth of information gathered here could formthe basis to consider studies in human pathologies. Diseasesaffecting the kidney and the urinary tract are of particularinterest considering that tissue proteins (for malformation),mediators of inflammation and renal autoantibodies (forglomerulonephritis) are preferentially excreted into urine.

Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.jprot.2014.07.021.

R 552

553

554

555

556

557

RConflict of interest

All authors declare no conflicts of interest.

O 558

559

560

561

562

563

564

NCTransparency document

The Transparency document associated with this article canbe found, in the online version.

565

566

567

568

569

570

571

572

573

574

575

576

577

578

U

Acknowledgfments

The Giannina Gaslini Institute provided financial and logisticsupport to the study. This work was also supported bythe Italian Ministry of Health ‘Ricerca Corrente’ and fromcontributions derived from ‘Cinque per mille dell'IRPEF’. Wealso acknowledge contributions from the Renal Child Foun-dation, Fondazione La Nuova Speranza (‘Progetto integrato perla definizione dei meccanismi implicati nella glomerulo sclerosifocale’) and Italian Society of Nephrology (Progetto Ricercando).

579

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

ED P

RO

OF

R E F E R E N C E S

[1] Decramer S, Gonzalez de Peredo A, Breuil B, Mischak H,Monsarrat B, Bascands JL, et al. Urine in clinical proteomics.Mol Cell Proteomics 2008;7:1850–62.

[2] Lee RS, Monigatti F, Briscoe AC,Waldon Z, FreemanMR, SteenH. Optimizing sample handling for urinary proteomics.J Proteome Res 2008;7:4022–30.

[3] Thongboonkerd V. Practical points in urinary proteomics.J Proteome Res 2007;6:3881–90.

[4] Thongboonkerd V, Chutipongtanate S, Kanlaya R. Systematicevaluation of sample preparation methods for gel-basedhuman urinary proteomics: quantity, quality, and variability.J Proteome Res 2006;5:183–91.

[5] Thongboonkerd V, Saetun P. Bacterial overgrowth affectsurinaryproteomeanalysis: recommendation for centrifugation,temperature, duration, and the use of preservatives duringsample collection. J Proteome Res 2007;6:4173–81.

[6] Vlahou A, Schanstra J, Frokiaer J, El Nahas M, Spasovski G,Mischak H, et al. Establishment of a European Network forurine and kidney proteomics. J Proteomics 2008;71:490–2.

[7] Zurbig P, Renfrow MB, Schiffer E, Novak J, Walden M, WittkeS, et al. Biomarker discovery by CE-MS enables sequenceanalysis via MS/MS with platform-independent separation.Electrophoresis 2006;27:2111–25.

[8] Candiano G, Santucci L, Petretto A, Bruschi M, Dimuccio V,Urbani A, et al. 2D-electrophoresis and the urine proteomemap: where do we stand? J Proteomics 2010;73:829–44.

[9] Court M, Selevsek N, Matondo M, Allory Y, Garin J, MasselonCD, et al. Toward a standardized urine proteome analysismethodology. Proteomics 2011;11:1160–71.

[10] Molina L, Salvetat N, Ameur RB, Peres S, Sommerer N, JarrayaF, et al. Analysis of the variability of human normal urine by2D-GE reveals a “public” and a “private” proteome. J Proteomics2011;75:70–80.

[11] Nagaraj N, Mann M. Quantitative analysis of the intra- andinter-individual variability of the normal urinary proteome.J Proteome Res 2011;10:637–45.

[12] Mischak H, KolchW, Aivaliotis M, Bouyssie D, Court M, DihaziH, et al. Comprehensive human urine standards forcomparability and standardization in clinical proteomeanalysis. Proteomics Clin Appl 2010;4:464–78.

[13] Anderson NG, Anderson NL, Tollaksen SL. Proteins of humanurine. I. Concentration and analysis by two-dimensionalelectrophoresis. Clin Chem 1979;25:1199–210.

[14] Anderson NG, Anderson NL, Tollaksen SL, Hahn H, Giere F,Edwards J. Analytical techniques for cell fractions. XXV.Concentration and two-dimensional electrophoretic analysisof human urinary proteins. Anal Biochem 1979;95:48–61.

[15] Edwards JJ, Tollaksen SL, Anderson NG. Proteins of humanurine. III. Identification and two-dimensional electrophoreticmap positions of some major urinary proteins. Clin Chem1982;28:941–8.

[16] Spahr CS, Davis MT, McGinley MD, Robinson JH, Bures EJ,Beierle J, et al. Towards defining the urinary proteome usingliquid chromatography–tandemmass spectrometry. I. Profilingan unfractionated tryptic digest. Proteomics 2001;1:93–107.

[17] Bueler MR, Wiederkehr F, Vonderschmitt DJ. Electrophoretic,chromatographic and immunological studies of humanurinary proteins. Electrophoresis 1995;16:124–34.

[18] Thongboonkerd V, McLeish KR, Arthur JM, Klein JB. Proteomicanalysis of normal human urinary proteins isolated byacetone precipitation or ultracentrifugation. Kidney Int 2002;62:1461–9.

[19] Weissinger EM, Wittke S, Kaiser T, Haller H, Bartel S, Krebs R,et al. Proteomic patterns established with capillaryelectrophoresis and mass spectrometry for diagnosticpurposes. Kidney Int 2004;65:2426–34.

sands: Widening the normal human Urinome (1), J Prot (2014),

T

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

9J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

UNCO

RREC

[20] Schaub S, Wilkins J, Weiler T, Sangster K, Rush D, NickersonP. Urine protein profiling with surface-enhancedlaser-desorption/ionization time-of-flight massspectrometry. Kidney Int 2004;65:323–32.

[21] Oh J, Pyo JH, Jo EH, Hwang SI, Kang SC, Jung JH, et al.Establishment of a near-standard two-dimensional humanurine proteomic map. Proteomics 2004;4:3485–97.

[22] Pieper R, Gatlin CL, McGrath AM, Makusky AJ, Mondal M,Seonarain M, et al. Characterization of the human urinaryproteome: a method for high-resolution display of urinaryproteins on two-dimensional electrophoresis gels with a yieldof nearly 1400 distinct protein spots. Proteomics 2004;4:1159–74.

[23] Adachi J, Kumar C, Zhang Y, Olsen JV, Mann M. The humanurinary proteome contains more than 1500 proteins, includinga large proportion of membrane proteins. Genome Biol 2006;7:R80.

[24] Marimuthu A, O'Meally RN, Chaerkady R, Subbannayya Y,Nanjappa V, Kumar P, et al. A comprehensive map of thehuman urinary proteome. J Proteome Res 2011;10:2734–43.

[25] Ru QC, Katenhusen RA, Zhu LA, Silberman J, Yang S, OrchardTJ, et al. Proteomic profiling of human urine usingmulti-dimensional protein identification technology.J Chromatogr A 2006;1111:166–74.

[26] Soldi M, Sarto C, Valsecchi C, Magni F, Proserpio V, Ticozzi D,et al. Proteome profile of human urine with two-dimensionalliquid phase fractionation. Proteomics 2005;5:2641–7.

[27] Sun W, Li F, Wu S, Wang X, Zheng D, Wang J, et al. Humanurine proteome analysis by three separation approaches.Proteomics 2005;5:4994–5001.

[28] Zerefos PG, Vougas K, Dimitraki P, Kossida S, Petrolekas A,Stravodimos K, et al. Characterization of the human urineproteome by preparative electrophoresis in combination with2-DE. Proteomics 2006;6:4346–55.

[29] Kentsis A, Monigatti F, Dorff K, Campagne F, Bachur R, SteenH. Urine proteomics for profiling of human disease usinghigh accuracy mass spectrometry. Proteomics Clin Appl 2009;3:1052–61.

[30] Zerefos PG, Aivaliotis M, Baumann M, Vlahou A. Analysis ofthe urine proteome via a combination of multi-dimensionalapproaches. Proteomics 2012;12:391–400.

[31] Castagna A, Cecconi D, Sennels L, Rappsilber J, Guerrier L,Fortis F, et al. Exploring the hidden human urinary proteomevia ligand library beads. J Proteome Res 2005;4:1917–30.

[32] Candiano G, Dimuccio V, Bruschi M, Santucci L, Gusmano R,Boschetti E, et al. Combinatorial peptide ligand libraries forurine proteome analysis: investigation of different elutionsystems. Electrophoresis 2009;30:1–7.

[33] Candiano G, Santucci L, Bruschi M, Petretto A, D'Ambrosio C,Scaloni A, et al. “Cheek-to-cheek” urinary proteome profilingvia combinatorial peptide ligand libraries: a novel,unexpected elution system. J Proteomics 2012;75:796–805.

[34] Boschetti E, Righetti PG. The art of observing rare proteinspecies in proteomes with peptide ligand libraries. Proteomics2009;9:1492–510.

[35] Righetti PG, Boschetti E. The ProteoMiner and theFortyNiners: searching for gold nuggets in the proteomicarena. Mass Spectrom Rev 2008;27:596–608.

[36] Righetti PG, Boschetti E, Zanella A, Fasoli E, Citterio A.Plucking, pillaging and plundering proteomes withcombinatorial peptide ligand libraries. J Chromatogr A 2010;1217:893–900.

[37] Santucci L, Candiano G, Bruschi M, D'Ambrosio C, Petretto A,Scaloni A, et al. Combinatorial peptide ligand libraries for theanalysis of low-expression proteins: validation for normalurine and definition of a first protein MAP. Proteomics 2012;12:509–15.

[38] Pisitkun T, Shen RF, Knepper MA. Identification and proteomicprofiling of exosomes in human urine. Proc Natl Acad Sci U S A2004;101:13368–73.

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

ED P

RO

OF

[39] Zhou H, Yuen PS, Pisitkun T, Gonzales PA, Yasuda H, Dear JW,et al. Collection, storage, preservation, and normalization ofhuman urinary exosomes for biomarker discovery. Kidney Int2006;69:1471–6.

[40] Merchant ML, Powell DW,Wilkey DW, Cummins TD, DeegensJK, Rood IM, et al. Microfiltration isolation of human urinaryexosomes for characterization by MS. Proteomics Clin Appl2010;4:84–96.

[41] Rood IM, Deegens JK, Merchant ML, Tamboer WP, Wilkey DW,Wetzels JF, et al. Comparison of threemethods for isolation ofurinary microvesicles to identify biomarkers of nephroticsyndrome. Kidney Int 2010;78:810–6.

[42] Zhang Y, Li Y, Qiu F, Qiu Z. Comprehensive analysis oflow-abundance proteins in human urinary exosomes usingpeptide ligand library technology, peptide OFFGELfractionation and nanoHPLC-chip-MS/MS. Electrophoresis2010;31:3797–807.

[43] Taylor DD, Zacharias W, Gercel-Taylor C. Exosome isolationfor proteomic analyses and RNA profiling. Methods Mol Biol2011;728:235–46.

[44] Wang Z, Hill S, Luther JM, Hachey DL, Schey KL. Proteomicanalysis of urine exosomes by multidimensional proteinidentification technology (MudPIT). Proteomics 2012;12:329–38.

[45] Gruenberg J, Stenmark H. The biogenesis of multivesicularendosomes. Nat Rev Mol Cell Biol 2004;5:317–23.

[46] Thery C, Zitvogel L, Amigorena S. Exosomes: composition,biogenesis and function. Nat Rev Immunol 2002;2:569–79.

[47] Li XB, Zhang ZR, Schluesener HJ, Xu SQ. Role of exosomes inimmune regulation. J Cell Mol Med 2006;10:364–75.

[48] Valadi H, Ekstrom K, Bossios A, Sjostrand M, Lee JJ, Lotvall JO.Exosome-mediated transfer of mRNAs and microRNAs is anovel mechanism of genetic exchange between cells. Nat CellBiol 2007;9:654–9.

[49] Moon PG, You S, Lee JE, Hwang D, Baek MC. Urinary exosomesand proteomics. Mass Spectrom Rev 2011;30:1185–202.

[50] van Balkom BW, Pisitkun T, Verhaar MC, Knepper MA.Exosomes and the kidney: prospects for diagnosis andtherapy of renal diseases. Kidney Int 2011;80:1138–45.

[51] Santucci L, Candiano G, Bruschi M, Bodria M, Murtas C,Petretto A, et al. Urinary proteome in a snapshot: normalurine and glomerulonephritis. J Nephrol 2013;26:610–6.

[52] Alvarez ML, Khosroheidari M, Kanchi Ravi R, Distefano JK.Comparison of protein, microRNA, and mRNA yields usingdifferentmethods of urinary exosome isolation for the discoveryof kidney disease biomarkers. Kidney Int 2012;82:1024–32.

[53] Cheruvanky A, Zhou H, Pisitkun T, Kopp JB, Knepper MA, YuenPS, et al. Rapid isolation of urinary exosomal biomarkers usinga nanomembrane ultrafiltration concentrator. Am J PhysiolRenal Physiol 2007;292:F1657–61.

[54] Moon PG, Lee JE, You S, Kim TK, Cho JH, Kim IS, et al.Proteomic analysis of urinary exosomes from patients ofearly IgA nephropathy and thin basement membranenephropathy. Proteomics 2011;11:2459–75.

[55] Pisitkun T, Johnstone R, Knepper MA. Discovery of urinarybiomarkers. Mol Cell Proteomics 2006;5:1760–71.

[56] Raj DA, Fiume I, Capasso G, Pocsfalvi G. A multiplexquantitative proteomics strategy for protein biomarkerstudies in urinary exosomes. Kidney Int 2012;81:1263–72.

[57] Simpson RJ, Lim JW, Moritz RL, Mathivanan S. Exosomes:proteomic insights and diagnostic potential. Expert RevProteomics 2009;6:267–83.

[58] Fernandez-Llama P, Khositseth S, Gonzales PA, Star RA,Pisitkun T, Knepper MA. Tamm–Horsfall protein and urinaryexosome isolation. Kidney Int 2010;77:736–42.

[59] Musante L, Saraswat M, Duriez E, Byrne B, Ravida A, Domon B,et al. Biochemical and physical characterisation of urinarynanovesicles following CHAPS treatment. PLoS One 2012;7:e37279.

sands: Widening the normal human Urinome (1), J Prot (2014),

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

10 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X

[60] Gonzales PA, Pisitkun T, Hoffert JD, Tchapyjnikov D, Star RA,Kleta R, et al. Large-scale proteomics and phosphoproteomicsof urinary exosomes. J Am Soc Nephrol 2009;20:363–79.

[61] Gonzales PA, Zhou H, Pisitkun T, Wang NS, Star RA, KnepperMA, et al. Isolation and purification of exosomes in urine.Methods Mol Biol 2010;641:89–99.

[62] Wisniewski JR, Zougman A, Nagaraj N, Mann M. Universalsample preparation method for proteome analysis. NatMethods 2009;6:359–62.

[63] Branca RM, Orre LM, Johansson HJ, Granholm V, Huss M,Perez-Bercoff A, et al. HiRIEF LC-MS enables deep proteomecoverage and unbiased proteogenomics. Nat Methods 2014;11:59–62.

UNCO

RRECT

Please cite this article as: Santucci L, et al, From hundreds to thouhttp://dx.doi.org/10.1016/j.jprot.2014.07.021

[64] Gupta N, Pevzner PA. False discovery rates of proteinidentifications: a strike against the two-peptide rule.J Proteome Res 2009;8:4173–81.

[65] Hsieh EJ, HoopmannMR, MacLean B, MacCoss MJ. Comparisonof database search strategies for high precursor mass accuracyMS/MS data. J Proteome Res 2010;9:1138–43.

[66] Li QR, Fan KX, Li RX, Dai J, Wu CC, Zhao SL, et al. Acomprehensive and non-prefractionation on the protein levelapproach for the human urinary proteome: touchingphosphorylation in urine. Rapid Commun Mass Spectrom2010;24:823–32.

ED P

RO

OF

sands: Widening the normal human Urinome (1), J Prot (2014),