Analytical Biochemistry 451 (2014) 28–30

Contents lists available at ScienceDirect

Analytical Biochemistry

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

Notes & Tips

Preparation of nitrophorin 7(D1–3) from Rhodnius prolixus withoutstart–methionine using recombinant expression in Escherichia coli

http://dx.doi.org/10.1016/j.ab.2014.01.0060003-2697/� 2014 Elsevier Inc. All rights reserved.

⇑ Corresponding author at: Max-Planck-Institut für Chemische Energiekonver-sion, Stiftstrasse 34-36, D-45470 Mülheim an der Ruhr, Germany. Fax: +49 208 3063951.

E-mail address: markus.knipp@cec.mpg.de (M. Knipp).1 Current address: Funktionelle Genomforschung der Mikroorganismen, Heinrich-

Heine-Universität Düsseldorf, D-40225 Düsseldorf, Germany.2 Abbreviations used: NP, nitrophorin; NO, nitric oxide; cDNA, complementary DNA;

MALDI–TOF MS, matrix-assisted laser desorption/ionization time-of-flight massspectrometry; PEI, polyethylenimine; SEC, size exclusion chromatography; Mops,3-morpholinopropane-1-sulfonic acid.

3 By analogy to NP4, the presence of Ala1 in NP2(D1A) leads to posttrantruncation of Met0 [11].

Carmen Risse a,1, Johanna J. Taing a, Markus Knipp a,b,⇑a Max-Planck-Institut für Chemische Energiekonversion, D-45470 Mülheim an der Ruhr, Germanyb Fakultät für Chemie und Biochemie, Ruhr-Universität Bochum, D-44801 Bochum, Germany

a r t i c l e i n f o

Article history:Received 28 November 2013Received in revised form 4 January 2014Accepted 13 January 2014Available online 23 January 2014

Keywords:Chelating Sepharose chromatographyNitrophorinPeriplasmRecombinant expressionStart–methionine

a b s t r a c t

The heterologous recombinant expression of proteins in Escherichia coli without start–methionine is acommon problem. The nitrophorin 7 heme properties and function strongly depend on the accurateN-terminal amino acid sequence. Leading protein expression into the periplasm by fusion with the leaderpeptide pelB yields functional protein; however, the folded protein sticks to the cell debris. Therefore, theperiplasmic fraction was dissolved in guanidinium chloride and folded by a drop-in method. Separationfrom impurities including residual pelB–nitrophorin 7 required establishing an unconventionalchromatographic technique using calcium-loaded Chelating Sepharose as cation exchanger and elutionby a linear CaCl2 gradient.

� 2014 Elsevier Inc. All rights reserved.

2

The class of proteins termed nitrophorins (NPs) comprise agroup of secreted ferriheme proteins that are used by the blood-feeding insect Rhodnius prolixus for the transportation of nitric oxide(NO). This small gaseous molecule is bound to the ferriheme iron andreleased on administration of the saliva to the victim’s tissue [1]. An-other recently reported feature of the NPs is their ability to produceNO from NO2

– (nitrite disproportionation reaction) [2–4], leading toclassification as nitrite dismutase (EC 1.7.6.1). From the insectsalivary glands, at first four NPs, designated NP1–4, were isolatedand later recombinantly expressed [5]. Recently, another NP, NP7,appeared from a complementary DNA (cDNA) library having signif-icantly distinct properties [4,6]. The most subtle difference of NP7compared with NP1–4 is its ability to bind to negatively chargedphospholipid membrane surfaces (Kd = 4.8 nM) [6–8].

Using salivary gland homogenates, NP7 could be isolated fromthe bulk through binding to liposomes and subsequent matrix-assisted laser desorption/ionization time-of-flight mass spectrom-etry (MALDI–TOF MS), where it was demonstrated that the

N-terminus is extended by the residues Leu1–Pro–Gly3 comparedwith NP1–4 [8]. However, extensive biophysical characterization ofNP7 compared with a truncated version, NP7(D1–3), showed thatthe presence of Leu1–Pro–Gly3 is crucial for NP7 functionality.However, recombinant protein expression in Escherichia coli typi-cally requires an initial Met residue so that NP1–3 and NP7 containa non-native Met0. Some amino acids in position 1, such as Ala,lead to posttranslational truncation of Met0 [9,10], as is the casefor NP4 [11]. X-ray crystallography revealed that its N-terminusforms salt bridges with Asp129 and Glu32, suggesting that itmay have an important structural function [12]. Spectroscopiccomparison of Met0–NP2 with NP2 without Met0 and NP2(D1A)3

reveals that Met0 can have a remarkable influence on the hemeproperties [11,13]. In the case of NP7, the meaning of the extendedN-terminus is not clear [14]. Therefore, it was intended to produceNP7 and N-terminal deletion mutants of NP7 without Met0 for fu-ture structural and functional characterization.

The omission of Met0 in heterologous expression is a generalproblem. In principle, N-terminal fusion with a tag sequence con-taining a peptidase cleavage site provides a solution. Some pepti-dases, such as thrombin, cut their substrate peptide C-terminal oftheir recognition site so that on scission no N-terminal leftoversare present. However, NPs contain Cys close to the N-terminus(NP1–4: Cys2; NP7: Cys5) that forms an intramolecular disulfide

slational

Notes & Tips / Anal. Biochem. 451 (2014) 28–30 29

bond [15]. Hence, the protein structure is so constrained that thecleavage efficiency is void. Some chemical methods are availableto cleave Met0, but yields are typically low and side reactionsare frequent [16]. Therefore, another method for the generationof Met0-depleted NP7 and mutants was developed.

The cDNA coding for NP7 was inserted into pET-26b(+) (Nova-gen), which provides a pelB leader sequence for expression intothe periplasmic space [17]. pelB is cut by E. coli’s signal peptidasesuch that the choice of the residue in position 1 is left to the exper-imenter. For the cloning, the pET-26b(+) restriction site MscI wasused. An StuI site was added to the 5’-end of the NP7 cDNA, andthe two blunt ends where ligated. Details of the cloning strategyare displayed in Supplementary material. Ligation of the 3’-endwas performed through BamHI already present in the templateplasmid [6]. The resulting plasmid is termed ppelB–NP7Kan. Fromthis plasmid, N-terminal sequence deletion variants were derivedthrough standard polymerase chain reaction (PCR)-based muta-genesis, yielding ppelB–NP7(D1)Kan, ppelB–NP7(D1–2)Kan, andppelB–NP7(D1–3)Kan. The plasmid production and the subsequentgene sequencing were carried out by GenScript Inc. (Piscataway,NJ) according to our instructions.

The expression of functional NP2 without Met0 by pelB fusionwas already successfully performed with NP2 [13]. By analogy,for NP7 the expression culture medium was supplemented with5-aminolevulinic acid as a precursor for heme biosynthesis toallow assembly of the protein with its cofactor. In addition, theperiplasm established the two disulfide bonds of the protein.4

The brown color of the cells demonstrated that NP7 was yielded inits assembled form in the periplasm. However, on isolation of theperiplasm according to the method reported for NP2 [13], only verylow yields of NP7 were obtained. MALDI–TOF MS showed that thenative NP7 sequence was obtained. The absorbance spectra showeda Soret maximum at 404 nm like in Met0–NP7. However, the browncolor of the cell debris indicated that most of the NP7 coprecipitatedwith the cell debris during centrifugation. Because of its positivelycharged surface, we tried to separate NP7 from cell debris by theaddition of (i) positively charged competitors such as divalent metalions, L-arginine, and polyethylenimine (PEI), (ii) negatively chargedcompetitors such as L-glutamate and heparin, and (iii) amphiphilicdetergents such as cholate and Chaps (3-[(3-cholamidopropyl)dim-ethylammonio]-1-propanesulfonate). However, although yield in-creased slightly, in particular with L-arginine and PEI, the resultswere not satisfactory (<0.2 mg/L expression culture).

Because the folding of Met0–NP7 and Met0–NP7(D1–3) frominclusion bodies is a well-established method [7,18], it wasdecided to unfold the protein from the periplasmic fraction sepa-rating protein from debris. The method is described for the prepa-ration of NP7(D1–3), but it should work for all four constructsmentioned above. Expression in BL21(DE3) cells was performedin 1-L cultures in LB medium supplemented with 34 lg/L kanamy-cin and 0.005% Antifoam 204 (Sigma–Aldrich). Cells were grown atfirst at 37 �C. When OD600 � 0.8 was reached, expression was in-duced with 0.1 mM isopropyl-b-D-thiogalactopyranoside (IPTG)and growth was continued for 21 h at approximately 21 �C. Typi-cally, approximately 4 g of cells was obtained from a 1-L expres-sion culture. The cell pellet was resuspended in 50 ml of 30 mMMops (3-morpholinopropane-1-sulfonic acid)/NaOH (pH 8.0) and20% sucrose. Then 1 mM Na2EDTA was added, and the suspensionwas stirred for 10 min at room temperature. On centrifugation(10 min, 8800g, 4 �C), the cell pellet was collected and suspendedin 25 ml of ice-cold 5 mM MgSO4 to release the periplasmic pro-teins. After stirring for 10 min on ice, the solution was cleared by

4 Cytosolic expression does not yield folded protein because the low reductionpotential does not allow the formation of disulfide bonds.

centrifugation (10 min, 8800g, 4 �C) before the addition of 3.5 Mguanidinium chloride (GdmCl), 2 mM tris(2-carboxyethyl)phos-phonine hydrochloride (TCEP�HCl), and 10 mM Mops/NaOH (pH7.0) to the supernatant. The solution was left overnight at 4 �C tomaximize protein solubilization and then was centrifuged(40 min, 100,000g, 4 �C). The supernatant was used to establishthe protein fold by the drop-in method followed by extensivedialysis, heme insertion, and size exclusion chromatography(SEC) as described previously [7]. At this stage, the pooled fractionsfrom SEC were subjected to MALDI–TOF MS. Besides someimpurities, two major signals [M + H]+ were obtained, reflectingNP7(D1–3) without Met0 (observed: 20,586 ± 20 Da; calculated:20,570 Da) and the uncleaved fusion pelB–NP7(D1–3) (observed:22,802 ± 20 Da; calculated: 22,799 Da) where the signal intensityof pelB–NP7(D1–3) is approximately half of the signal intensityof the cleaved protein (Fig. 1A).

The separation of these two forms requires an additional chro-matographic step. Because of the high surface basicity of NP7, cat-ion exchange chromatography was considered. However, attemptsto purify NP7 by common cation exchange media resulted in loss ofprotein due to precipitation on the resin. We ascribe this to thegeneral tendency of NP7 to precipitate at the high salt concentra-tion typically required for protein elution. This instability also ex-cludes hydrophobic interaction chromatography.

Instead, an unprecedented method was developed where pro-tein was bound to a 5-ml HiTrap Chelating Sepharose (GE Health-care) loaded with Ca2+. Elution was performed with a lineargradient of 0 to 40 mM CaCl2 (Fig. 1B), keeping the ionic strengthlow compared with common elution with NaCl or other salts dur-ing conventional cation exchange chromatography. Fractions werecollected and analyzed by MALDI–TOF MS, showing that someweakly binding impurities were eluted in a pre-peak (fractions10–14) that contained only negligible amounts of NP7(D1–3) andpelB–NP7(D1–3). In the second peak (fractions 15–30), in thebeginning (fraction 20) mostly NP7(D1–3) was obtained ([M +H]+: 20,588 ± 20 Da; [M + 2H]2+: 10,296 ± 10 Da) plus a little bitof pelB–NP7(D1–3) ([M + H]+: 22,799 ± 20 Da) together with someminor impurity. Fraction 22 contained highly enriched NP7(D1–3)([M + H]+: 20,584 ± 20 Da; [M + 2H]2+: 10,924 ± 10 Da) besides alittle bit of pelB–NP7(D1–3) ([M + H]+: 22,791 ± 20 Da). Finally,in the right flank (fraction 24) of the peak, both forms were con-tained in low quantities. Consequently, fractions 21 to 23 werepooled so that approximately 1 mg of largely enriched (70–80%purity, judged by sodium dodecyl sulfate–polyacrylamide gel elec-trophoresis [SDS–PAGE]) NP7(D1–3) was obtained from 1 L ofexpression culture, which represents a more than 5-fold increasein yield compared with the isolation of soluble protein from theperiplasm and is sufficient for the attempted structural and func-tional characterizations of the protein. The protein shows the samefeatures in the absorbance spectrum as the protein prepared frominclusion bodies [7,18], demonstrating the proper protein fold,heme insertion, and disulfide integrity. The protein is able to bindthe various low-molecular-weight heme ligands such as NO andimidazole, as does the protein obtained from inclusion bodies [7].However, further purification of the protein is required, and thiscan be performed by SEC as described for the Met0-containingform [7]. Furthermore, the detailed functional, structural, and spec-troscopic characterization of the protein is beyond the scope of thisarticle and is planned for investigation in future experiments.

The principle of this novel technique can be rationalized as de-picted in Fig 1C. The resin is first charged with Ca2+ coordinated bythe iminodiacetic acid moieties. When the protein is loaded, NP7and Ca2+ compete for the resin carboxylates. Impurities are essen-tially washed out before elution of NP7 occurs at approximately20 mM CaCl2. The protein binds to the resin also without Ca2+, dem-onstrating that binding to the resin is not Ca2+ dependent. However,

Fig.1. (A) MALDI–TOF MS spectrum of the fractions of NP7(D1–3) and pelB–NP7(D1–3) obtained from SEC. (B) The material from panel A was first dialyzed against 10 mMMops/NaOH (pH 6.8) and 2% glycerol and then was applied to Ca2+-loaded Chelating Sepharose. Elution was performed with a linear gradient of 0 to 40 mM CaCl2. Here, 10 llof peak fractions was rendered salt free using ZipTipC18 (Millipore) and analyzed by MALDI–TOF MS using 3-(4-hydroxy-3,5-dimethoxyphenyl)prop-2-enoic acid as matrix.(C) Principle of the purification of NP7(D1–3) by Chelating Sepharose as explained in the text.

30 Notes & Tips / Anal. Biochem. 451 (2014) 28–30

without Ca2+, dramatic losses of protein occur, indicating that theunhindered binding of NP7(D1–3) results in irreversible aggrega-tion on the column resin. Elution from the resin was also possiblewith a gradient of L-glutamate that supposedly competes with theresin for the positive charges of NP7. However, the best yields wereobtained using Ca2+ pre-charging and Ca2+ gradient elution.

In summary, the peculiar influence of the initial Met0 added toNP7 by recombinant expression requires its removal. Periplasmicexpression truncates Met0, but sufficient soluble protein cannotbe obtained due to its being attached to cell debris. Therefore,the protein was solubilized by a chaotrope for the separation fromcell debris and was folded thereafter. The sensitivity of NP7 againsthigh ionic strength required the development of a novel chromato-graphic purification strategy using Ca2+-charged Chelating Sephar-ose. This method may be considered for the purification of otherproblematic cases of protein production.

Acknowledgments

The authors are grateful for the technical assistance of NorbertDickmann. This work is supported by the Max Planck Society andby the Cluster of Excellence RESOLV (EXC 1069) funded by theDeutsche Forschungsgemeinschaft.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.ab.2014.01.006.

References

[1] F.A. Walker, Nitric oxide interaction with insect nitrophorins and thoughts onthe electron configuration of the {FeNO}6 complex, J. Inorg. Biochem. 99 (2005)216–236.

[2] C. He, M. Knipp, Formation of nitric oxide from nitrite by the ferriheme bprotein nitrophorin 7, J. Am. Chem. Soc. 131 (2009) 12042–12043.

[3] C. He, H. Ogata, M. Knipp, Formation of the complex of nitrite with theferriheme b b-barrel proteins nitrophorin 4 and nitrophorin 7, Biochemistry 49(2010) 5841–5851.

[4] M. Knipp, C. He, Nitrophorins: nitrite disproportionation reaction and othernovel functionalities of insect heme-based nitric oxide transport proteins,IUBMB Life 63 (2011) 304–312.

[5] W.R. Montfort, A. Weichsel, J.F. Andersen, Nitrophorins and relatedantihemostatic lipocalins from Rhodnius prolixus and other blood-suckingarthropods, Biochim. Biophys. Acta 1482 (2000) 110–118.

[6] J.F. Andersen, N.P. Gudderra, I.M.B. Francischetti, J.G. Valenzuela, J.M.C. Ribeiro,Recognition of anionic phospholipid membranes by an antihemostatic proteinfrom a blood-feeding insect, Biochemistry 43 (2004) 6987–6994.

[7] M. Knipp, H. Zhang, R.E. Berry, F.A. Walker, Overexpression in Escherichia coliand functional reconstitution of the liposome binding ferriheme proteinnitrophorin 7 from the blood sucking bug Rhodnius prolixus, Protein Expr. Purif.54 (2007) 183–191.

[8] M. Knipp, R.P.P. Soares, M.H. Pereira, Identification of the native N-terminus ofthe membrane attaching ferriheme protein nitrophorin 7 from Rhodniusprolixus, Anal. Biochem. 424 (2012) 79–81.

[9] P.-H. Hirel, J.-M. Schmitter, P. Dessen, G. Fayat, S. Blanquet, Extent of N-terminal methionine excision from Escherichia coli proteins is governed by theside-chain length of the penultimate amino acid, Proc. Natl. Acad. Sci. USA 86(1989) 8247–8251.

[10] F. Frottin, A. Martinez, P. Peynot, S. Mitra, R.C. Holz, C. Giglione, T. Meinnel, Theproteomics of N-terminal methionine cleavage, Mol. Cell. Proteomics 5 (2006)2336–2349.

[11] R.E. Berry, T.K. Shokhireva, I. Filippov, M.N. Shokhirev, H. Zhang, F.A. Walker,The effect of the N-terminus on heme cavity structure, ligand equilibrium, andrate constants and reduction potentials of nitrophorin 2 from Rhodniusprolixus, Biochemistry 46 (2007) 6830–6843.

[12] E.M. Maes, A. Weichsel, J.F. Andersen, D. Shepley, W.R. Montfort, Role ofbinding site loops in controlling nitric oxide release: structure and kinetics ofmutant forms of nitrophorin 4, Biochemistry 43 (2004) 6679–6690.

[13] R.E. Berry, D. Muthu, T.K. Shokhireva, S.A. Garrett, H. Zhang, F.A. Walker, NativeN-terminus nitrophorin 2 from the kissing bug: similarities to and differencesfrom NP2(D1A), Chem. Biodivers. 9 (2012) 1739–1755.

[14] A. Oliveria, A. Allegri, A. Bidon-Chanal, M. Knipp, A.E. Roitberg, S. Abbruzzetti,C. Viappiani, F.J. Luque, Kinetics and computational studies of ligand migrationin nitrophorin 7 and its 1–3 mutant, Biochim. Biophys. Acta 2013 (1834)1711–1721.

[15] M. Knipp, J.J. Taing, C. He, Reduction of the lipocalin type heme containingprotein nitrophorin: sensitivity of the fold-stabilizing cysteine disulfidestoward routine heme–iron reduction, J. Inorg. Biochem. 105 (2011) 1405–1412.

[16] M. Suenaga, H. Ohmae, N. Okutani, T. Kurokawa, T. Asano, T. Yamada, O.Nishimura, M. Fujino, Further studies on the chemical cleavage of an N-terminal extra methionine from recombinant methionylated proteins, J. Chem.Soc. Perkin Trans. 1 (1999) 3727–3733.

[17] F.J.M. Mergulhão, D.K. Summers, G.A. Monteiro, Recombinant protein secretionin Escherichia coli, Biotechnol. Adv. 23 (2005) 177–202.

[18] M. Knipp, F. Yang, R.E. Berry, H. Zhang, M.N. Shokhirev, F.A. Walker,Spectroscopic and functional characterization of nitrophorin 7 from theblood-feeding insect Rhodnius prolixus reveals an important role of itsisoform-specific N-terminus for proper protein function, Biochemistry 46(2007) 13254–13268.