Analytical Biochemistry 421 (2012) 330–332

Contents lists available at SciVerse ScienceDirect

Analytical Biochemistry

journal homepage: www.elsevier .com/locate /yabio

Notes & Tips

Proteins isolated with TRIzol are compatible with two-dimensionalelectrophoresis and mass spectrometry analyses

Clifford Young 1, Penelope Truman ⇑Institute of Environmental Science and Research, Kenepuru Science Centre, Porirua, New Zealand

a r t i c l e i n f o a b s t r a c t

Article history:Received 1 August 2011Received in revised form 16 October 2011Accepted 27 October 2011Available online 4 November 2011

Keywords:2-DEMass spectrometryPeptide mass fingerprintProtein extractionProteomicsTRIzol

0003-2697/$ - see front matter � 2011 Elsevier Inc. Adoi:10.1016/j.ab.2011.10.045

⇑ Corresponding author. Fax: +64 4 914 0770.E-mail address: penelope.truman@esr.cri.nz (P. Tru

1 Current address: Novo Nordisk Foundation Center fHealth Sciences, University of Copenhagen, DK-2200 C

2 Abbreviations used: 2-DE, two-dimensional gel elecpH gradient; MALDI–TOF MS, matrix-assisted laser dflight mass spectrometry.

TRIzol is used for RNA isolation but also permits protein recovery. We investigated whether proteins pre-pared with TRIzol were suitable for two-dimensional gel electrophoresis (2-DE) and matrix-assisted laserdesorption/ionization mass spectrometry. Proteins from TRIzol-treated SH-SY5Y cells produced 2-DE spotpatterns similar to those from an equivalent untreated sample. Subsequent identification of TRIzol-trea-ted proteins using peptide mass fingerprinting was successful. TRIzol exposure altered neither the massof myoglobin extracted from sodium dodecyl sulfate (SDS) gels nor the masses of myoglobin peptidesproduced by in-gel trypsin digestion. These findings suggest that proteins isolated with TRIzol remainamenable to proteomic analyses.

� 2011 Elsevier Inc. All rights reserved.

TRIzol is a phenol and guanidine isothiocyanate solution used toobtain RNA from biological samples. After nucleic acid extraction,protein can be precipitated from the phenol and ethanol superna-tant [1]. Because the methodology allows the stepwise isolation ofDNA, RNA and protein, this reagent is useful when genomic, tran-scriptomic and proteomic information is desired from a singlesample [2]. However, we were concerned that TRIzol treatmentmight produce amino acid modifications to undermine both thequality of the protein isolation and the utility of proteomic meth-odologies for protein analysis. Therefore, the suitability of TRIzol-exposed samples for proteomic analyses was evaluated by inspect-ing the two-dimensional gel electrophoresis (2-DE)2 quality fromthese samples and whether subsequent protein identification bypeptide mass fingerprinting was affected.

The effect of TRIzol on protein from human neuroblastoma cellswas assessed first. Approximately 107 SH-SY5Y cells (CRL-2266,American Type Culture Collection, Manassas, VA, USA) were pro-cessed with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) accord-ing to the manufacturer’s instructions, except that the protein wasprecipitated with ice-cold acetone instead of propan-2-ol. After

ll rights reserved.

man).or Protein Research, Faculty ofopenhagen N, Denmark.trophoresis; IPG, immobilizedesorption/ionization time-of-

centrifugation at 12,000g for 10 min and removal of the superna-tant, the pellet was washed five times with 300 mM guanidinehydrochloride in 95% ethanol and twice in ice-cold acetone priorto centrifugation. The dried protein pellet was mixed with 200 llof solubilization solution (8 M urea, 3 M thiourea, 1% dithiothreitol,10 mM tris(2-carboxyethyl)phosphine, 4% 3-[(3-cholamidopro-pyl)dimethylammonio]-1-propanesulfonate and 1% pH 3–10ampholytes) for 2 h before centrifugation at 21,000g for 10 min.An equivalent number of control SH-SY5Y cells were directly solu-bilized and centrifuged in a similar manner for comparison pur-poses. Protein concentrations were determined by Bradford assaybefore performing 2-DE analysis (see Supplementary material).

Comparison of the 2-DE gel images from control and TRIzol-treated SH-SY5Y proteins revealed similar spot patterns (Fig. 1).Although there were discernible differences in the intensity ofsome spots between treatments, the protein extraction by TRIzolwas predominantly unbiased. Nonetheless, the wash procedureused for the removal of phenol has the potential to result in theselective loss of some proteins.

The representative TRIzol gel displayed vertical streaks origi-nating from the alkaline region of the pH 3–10 immobilized pHgradient (IPG) strip. When a similar TRIzol-treated sample was fo-cused on a pH 4–7 IPG strip (see Supplementary Fig. 1 in Supple-mentary material), only resolved spots and negligible verticalstreaking were observed. These results suggest that 2-DE of SH-SY5Y cells is largely unaffected by TRIzol exposure and some arti-facts can be avoided if IPG strips with different pH ranges are used.The utility of TRIzol protein extraction for 2-DE separations has

100

75

50

37

25

15

+ - -100

75

50

37

25

15

+

Control TRIzol

kDakDa

pIpI

Fig.1. 2-DE of control (left) and TRIzol-treated (right) SH-SY5Y proteins (pH 3–10).

Notes & tips / Anal. Biochem. 421 (2012) 330–332 331

been described previously [3,4], with several TRIzol samples bene-fiting from the removal of contaminants that typically interferewith 2-DE [5,6].

Seven spots from a replicate 2-DE gel of TRIzol-treated SH-SY5Ycells were excised for protein identification (SupplementaryFig. 2A). Briefly, each gel spot was destained and subjected to tryp-sin digestion as described previously [7]. After the extracted pep-tides were analyzed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI–TOF MS),peptide mass lists were submitted to database searches for proteinidentification (see Supplementary material). All seven spots wereconfidently identified by peptide mass fingerprinting (Supplemen-tary Fig. 2B), demonstrating the compatibility of TRIzol-extractedproteins with subsequent analysis by MALDI–TOF MS. Identifica-tion of TRIzol-isolated proteins by peptide mass fingerprintinghas been demonstrated by several groups [2,8].

Although the identification of several TRIzol-exposed SH-SY5Yproteins was successful, there was a possibility that the non-match-ing peptide masses were adducts produced from TRIzol incubation.An investigation was initiated on a well-characterized protein (horsemyoglobin) to ascertain whether the TRIzol methodology inducesmass modifications, with myoglobin masses from TRIzol and controltreatments determined by MALDI–TOF MS (see Supplementarymaterial and Ref. [9]). The observed average masses were not statis-tically significantly different between the treatments, with repre-sentative spectra shown in Fig. 2. A mean average mass of16,994 Da for control horse skeletal myoglobin (95% confidenceinterval: 16,985.0–17,003.1 Da) was calculated from quadruplicatespectra, whereas 16,994.6 Da was obtained from TRIzol-treatedsamples (95% confidence interval: 16,986.3–17,002.8 Da). Thesemass measurements were significantly different (P < 0.001, two-tailed t test) from the molecular mass of 16,951.5 Da obtained froma previous myoglobin study [10] that reported the average mass ofhorse myoglobin without the N-terminal methionine because ofcotranslational cleavage. When cotranslational N-terminal acetyla-tion is also considered [11], the recalculated theoretical mass(16,993.5 Da) is in close proximity to our myoglobin mass measure-ments. The remaining three peaks in each MALDI–TOF spectrum (inorder of descending m/z) correspond well with the respective dou-bly, triply, and quadruply charged masses of extensively cotransla-tionally modified myoglobin. We found no evidence to suggestthat TRIzol causes unusual or considerable mass modifications inmyoglobin, although we cannot exclude mass changes that occurat a low abundance or those that involve small masses.

To address these issues, gel pieces containing intact myoglobinfrom TRIzol and control treatments were subjected to peptide massfingerprint analysis (see Supplementary material). The sequence

coverage from TRIzol-treated myoglobin samples was similar tothat obtained in the controls, which ranged from 79% to 84% be-tween all technical replicates (Supplementary Table 1A). Impor-tantly, the vast majority of the submitted masses were declaredby database searches to match myoglobin tryptic peptides (Supple-mentary Table 2). These observations suggest that TRIzol treat-ment does not hinder the identification of myoglobin by peptidemass fingerprinting.

On closer inspection of the myoglobin peptide matches, massesobtained from control and TRIzol peptide mass fingerprint spectrawere adjudged by database searches to match the peptideGLSDGEWQQVLNVWGK (without the N-terminal methionine). Arecent study identified the N-terminal acetylated peptide Ac-GLSDGEWQQVLNVWGK from horse myoglobin [12], but massesthat could correspond to this peptide were not detected, raisingdoubt of the existence of an N-terminally acetylated myoglobin.However, evidence of an N-terminal acetylated peptide was froma mass matching the trypsin miscleavage product Ac-GLSDGEWQQVLNVWGKVEADIAGHGQEVLIR, with one of its twotryptophans oxidized (theoretical MH+ = 3461.745). Although thispeptide was identified from only one TRIzol sample, this matchprovides evidence of the extensively cotranslationally modifiedform of myoglobin and corroborates our previous myoglobin mea-surements. We hypothesize that this long peptide was not rou-tinely selected by database searches because of its low intensityor poor peak shape.

Because it possesses a similar mass to acetylation, a single carba-mylation (43 Da increase) was considered as a possible explanationfor the observed mass of myoglobin. When both N-terminal and ly-sine carbamylation were also selected as variable modifications, themyoglobin sequence coverage obtained in all samples generally in-creased to between 81% and 90% (Supplementary Table 1B). Despitethe additional settings, the number of peptide hits remained thesame as before. Several peptide matches from both control andTRIzol spectra (Supplementary Fig. 3B) were adjudged to representthe doubly carbamylated HLKTEAEMKASEDLK (theoreticalMH+ = 1815.891), its related oxidation product (MH+ = 1831.886)and the doubly carbamylated FDKFKHLKTEAEMK with anoxidized residue (MH+ = 1853.922). However, these masses in theoriginal searches accounted for the myoglobin peptidesGLSDGEWQQVLNVWGK (theoretical MH+ = 1815.902), one of itsassociated tryptophan oxidation products (MH+ = 1831.897) andGHHEAELKPLAQSHATK (MH+ = 1853.962), respectively (Supple-mentary Fig. 3A). Because the unmatched peptide masses from theoriginal searches were not responsible for the carbamylated pep-tides, we consider the new carbamylation matches to be false posi-tives. Despite myoglobin containing many potential carbamylation

0

8493

566216993

4247

31000

Intensity

4000 20000Mass (m/z)

0

8493

5661

169944243

20000

Intensity

4000 Mass (m/z) 20000

Control

TRIzol

Fig.2. MALDI–TOF spectra of control (top) and TRIzol-treated (bottom) myoglobin.

332 Notes & Tips / Anal. Biochem. 421 (2012) 330–332

sites due to its high lysine content, there was no evidence for thewidespread presence of this modification in control or TRIzol sam-ples prepared in urea-containing solutions or for an increase in car-bamylation caused by TRIzol. These suggestions are substantiated bythe lower Mascot scores obtained from the carbamylation searches,where they also failed to surpass the 1% significance level (Supple-mentary Table 1). Importantly, the large acetylated peptide was re-ported with search settings that included carbamylation. Becausemultiple carbamylations would have increased the mass of intactmyoglobin beyond the single carbamylation we were looking for,the combination of N-terminal methionine cleavage and acetylationremains the most likely reason for the mass difference between theobserved and theoretical myoglobin masses.

Overall, no systematic differences between TRIzol-treated andcontrol myoglobin tryptic peptides were detected. We concludethat the TRIzol procedure is compatible with peptide mass finger-printing for the identification of proteins.

In summary, we have shown in this report that the isolation ofproteins with TRIzol does not appear to introduce any artifacts ormass modifications that interfere with intact protein analysis orpeptide mass fingerprinting. With only minor changes to the origi-nal protocol, this method allows the isolation of complex proteinmixtures for further analysis by standard proteomic methods. Thisis an especially useful application of the reagent when systemsbiology approaches are desired from limited amounts of startingmaterial.

Acknowledgments

We thank Daniel Kay for technical assistance. We are grateful tothe Centre for Biodiscovery (Victoria University of Wellington) foraccess to its mass spectrometer and the advice from T. William Jor-dan and Pisana Rawson. We thank Brian Chait (Rockefeller Univer-sity) for his assistance in the molecular mass determination ofintact myoglobin. This project was funded by the New ZealandMinistry of Research, Science and Technology Capability Fund.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ab.2011.10.045.

References

[1] P. Chomczynski, A reagent for the single-step simultaneous isolation of RNA,DNA and proteins from cell and tissue samples, BioTechniques 15 (1993) 532–537.

[2] J. Xiong, Q. Yang, J. Kang, Y. Sun, T. Zhang, G. Margaret, W. Ding, Simultaneousisolation of DNA, RNA, and protein from Medicago truncatula L., Electrophoresis32 (2011) 321–330.

[3] R.H. Butt, T.A. Pfeifer, A. Delaney, T.A. Grigliatti, W.G. Tetzlaff, J.R. Coorssen,Enabling coupled quantitative genomics and proteomics analyses from ratspinal cord samples, Mol. Cell. Proteomics 6 (2007) 1574–1588.

[4] R. Sudha, N. Kawachi, P. Du, E. Nieves, T.J. Belbin, A. Negassa, R.H. Angeletti,M.B. Prystowsky, Global proteomic analysis distinguishes biologic differencesin head and neck squamous carcinoma, Lab. Invest. 87 (2007) 755–766.

[5] P.A. Kirkland, J. Busby, S. Stevens Jr., J.A. Maupin-Furlow, Trizol-based methodfor sample preparation and isoelectric focusing of halophilic proteins, Anal.Biochem. 351 (2006) 254–259.

[6] F.W.-F. Lee, S.C.-L. Lo, The use of Trizol reagent (phenol/guanidineisothiocyanate) for producing high quality two-dimensional gelelectrophoretograms (2-DE) of dinoflagellates, J. Microbiol. Methods 73(2008) 26–32.

[7] M. Learmonth, A. Aitken, Protein identification by in-gel digestion and massspectrometric analysis, in: J.M. Walker (Ed.), The Proteomics ProtocolsHandbook, Humana, Totowa, NJ, 2005, pp. 311–314.

[8] F.W.-F. Lee, D. Morse, S.C.-L. Lo, Identification of two plastid proteins in thedinoflagellate Alexandrium affine that are substantially down-regulated bynitrogen-depletion, J. Proteome Res. 8 (2009) 5080–5092.

[9] S.L. Cohen, B.T. Chait, Mass spectrometry of whole proteins eluted from sodiumdodecyl sulfate–polyacrylamide gel electrophoresis gels, Anal. Biochem. 247(1997) 257–267.

[10] J. Zaia, R.S. Annan, K. Biemann, The correct molecular weight of myoglobin, acommon calibrant for mass spectrometry, Rapid Commun. Mass Spectrom. 6(1992) 32–36.

[11] B. Polevoda, F. Sherman, Na-terminal acetylation of eukaryotic proteins, J. Biol.Chem. 275 (2000) 36479–36482.

[12] T. Kishimoto, J. Kondo, T. Takai-Igarashi, H. Tanaka, Accurate mass comparisoncoupled with two endopeptidases enables identification of protein termini,Proteomics 11 (2011) 485–489.