RESEARCH

A Vector Design that Allows Fast and Convenient Productionof Differently Tagged Proteins

Omar Scapolan • Andrea N. Mazzarello • Maria Bono •

Marzia Occhino • Vasily Ogryzko • Marco Bestagno •

Paolo Scartezzini • Silvia Bruno • Franco Fais • Fabio Ghiotto

� Springer Science+Business Media, LLC 2011

Abstract Recombinant-tagged proteins have a wide-

spread use in experimental research as well as in clinical

diagnostic and therapeutic approaches. Well-stocked sets

of differently tagged variants of a same protein would be of

great help. However, the construction of differently tagging

vectors is a demanding task since cloning procedures need

several tailored DNA inserts. In this study, we describe a

novel vector system that allows a cost- and time-effective

production of differently tagged variants of a same protein

by using the same DNA fragment and a set of vectors each

carrying a different tag. The design of these expression

vectors is based on an intronic region that becomes func-

tional upon cloning the insert sequence, splicing of which

attaches a certain tag to the protein termini. This strategy

allows for the cloning of the fragment that codes for the

protein of interest, without any further modification, into

different vectors, previously built and ready-to-use, each

carrying a tag that will be joined to the protein. Proof of

principle for our expression system, presented here, is

shown through the production of a functional anti-GD2 Fab

fragment tagged with biotin or polyhistidine, or a combi-

nation of both, followed by the demonstration of the

functional competencies of both the protein and the tags.

Keywords Expression vector � Protein tag � Recombinant

antibody � Fab fragment � Immunotherapy

Introduction

Recombinant fusion proteins are nowadays a fundamental

tool in several fields of experimental research and clinical

applications, e.g., for protein functional studies or patient-

specific antibody-based therapies.

Tagging a protein allows the analysis of different

aspects of its function, especially if specific antibodies are

not available. Depending on the tag design, it is possible to

localize, label with a secondary reagent, or purify the given

protein. For example, a protein can be tagged with GFP

family proteins, polyhistidine tags, and antigenic epitopes

for which specific monoclonal antibodies are available

(e.g., c-myc, V5) or peptides bound to one biotin molecule

[1, 2].

In situ recombinant antibody-based approaches are

becoming promising therapies in several malignancies.

However, their tumor-targeting efficacy has often remained

to be improved, especially in the case of solid tumors.

Modification of their format, i.e., conversion of monova-

lent recombinant antibodies (r-Abs) into multivalent for-

mats increases their functional affinity. Oligomerization

domains from non-Ig proteins have been proposed to

Omar Scapolan and Andrea N. Mazzarello contributed equally to the

work.

O. Scapolan � A. N. Mazzarello � M. Bono � M. Occhino �S. Bruno � F. Fais (&) � F. Ghiotto

Department of Experimental Medicine, University of Genoa, Via

De Toni 14, 16132 Genoa, Italy

e-mail: franco.fais@unige.it

V. Ogryzko

Universite Paris-Sud 11, CNRS, Interactions Moleculaires et

Cancer, Institut de Cancerologie Gustave-Roussy, UMR 8126,

Villejuif, France

M. Bestagno

International Centre for Genetic Engineering and Biotechnology,

AREA Science Park Padriciano 99, 34012 Trieste, Italy

P. Scartezzini

Department of Gerontology, E.O. Ospedali Galliera, Genoa,

Italy

Mol Biotechnol

DOI 10.1007/s12033-011-9469-4

generate multimeric r-Abs. As an example, fusion proteins

of single chain Fv (scFv) with the NC1 trimerization region

of collagen XVIII resulted in trimeric antibodies [3], and

fusion of scFv with the multimerization domain of p53

generated tetrameric antibodies [4]. Furthermore, antitu-

mor effector molecules (e.g., TNF-alpha or IL-2) fused to

recombinant antibodies that bind specific tumor antigens

have great promise in tumor immunotherapy.

The examples described above represent only a few of

the numerous applications of recombinant-tagged fusion

proteins. To date, multiple tagging strategies are not

straightforward since the preparation of differently tagged

proteins requires the use of different expression vectors,

each one with particular cloning requirements, thus making

the cloning a time-consuming task. It would be very

helpful if we have the possibility to produce differently

tagged variants of the same protein without undertaking

onerous processes. Fast and easy production of differently

tagged variants of a same protein should become routine in

research and clinical laboratories, which often need to

switch within a short time among different fusion variants.

In this study, we describe a novel expression vector

system that allows a cost- and time-effective production of

any recombinant protein with any different tag at its C

terminal. An experimental validation of the feasibility and

the functionality of this system are presented.

Material and Methods

Vectors

pPSI–BiotC Vector

The construction of the pPSI–BiotC vector was achieved

by introducing a fragment coding for a 50 incomplete

intron, a spacer and a biotin acceptor sequence into the

eucaryotic expression vector pIRESpuro3 (Clontech,

Mountain View, CA). The DNA fragment to be cloned into

the pIRESpuro3 vector was built using a sequential

amplification approach that allowed the joining of three

segments: an intron, originally belonging to the human

IgG1 gene, that is incomplete at the 50 end, a sequence

coding for the GGGGSGGGGSGGGGS spacer, and a

sequence coding for the GLNDIFEAQKIEWHE peptide—

a specific substrate for the Escherichia coli BirA enzyme

[5]. This enzyme binds a biotin molecule to the lysine

present in the substrate peptide [6]. In brief, the sequence

encoding the intron and the spacer was amplified using

primers Intron iCPF and Spacer R (Table 1) and a con-

struct previously made in our lab as template [7]. The

sequence coding the GLNDIFEAQKIEWHE peptide was

amplified by primers Spacer F and Tag iCPR (Table 1)

using, as template, a construct previously made in our lab

that carries the peptide. The two fragments were joined by

means of PCR amplification with primers Intron iCPF and

Biot iCPR (Table 1). The PCR product was cloned into the

pIRESpuro3 expression vector cut with EcoRI and BamHI

restriction enzymes, by means of a commercial kit

designed to join DNA fragments that have 15 bases of

homology at their ends (in this case DNA insert and line-

arized vector), namely, the InFusionTM

Advantage PCR

Cloning Kit (Clontech), following the manufacturer’s

instructions.

pPSI–HisTag Vector

The pPSI–HisTag vector was built following the same

strategy as the pPSI–BiotC vector. The fragment to be

cloned into the pIRESpuro3 vector was built amplifying

the intron and the spacer as for pPSI–BiotC and then using

the primers Intron iCPF and HisTag iCPR (Table 1) for

adding six histidines at the 30 of the sequence. Again, the

resulting insert was cloned into the EcoRI/BamHI cut

pIRESpuro3 vector using the InFusionTM

Advantage PCR

Cloning Kit (Clontech).

pPSI–BHC Vector

The pPSI–BHC vector is characterized by the presence of

both the sequence substrate of the BirA enzyme and six

histidines with a spacer sequence (GGGGSGGGGS)

between the two. It was built following the same strategy

as the previous vectors. The fragment to be cloned into the

pIRESpuro3 vector was built amplifying the intron and the

spacer from pPSI–BiotC vector with primers Intron iCPF

and ExtHis R and then using primers Intron iCPF and

BioHisTag iCPR to complete the spacer and to add the

histidines (see Table 1 for primer sequences). The resulting

insert was cloned into the EcoRI/BamHI cut pIRESpuro3

vector using the InFusionTM

Advantage PCR Cloning Kit

(Clontech).

For all PCR reactions, the proofreading PrimeSTAR�

HS DNA polymerase (Takara, Shiga, Japan) was used. For

each cloning procedure, several clones were sequenced

with an automated sequenator (Applied Biosystems, Foster

City, CA), and one clone with the correct sequence was

selected for further cloning experiments.

Constructs

pBiGFP/CD33L–hBirA

The pBiGFP/CD33L–hBirA construct was built starting

from the pBiEGP vector, previously made in our lab, that

carries an IRES sequence followed by the EGFP gene,

Mol Biotechnol

used as reporter gene for expression of the upstream-cloned

gene. This vector carries the resistance for geneticin. The

hBirA gene, a version of the E. coli BirA gene that

underwent codon optimization, was amplified from a

construct previously described [8]. By means of two rounds

of PCR amplifications using, first, primers CD33L/hBirA F

and hBirA BamHI CPR, and then CD33L EcoRI CPF and

hBirA BamHI CPR (Table 1), we inserted, at the 50 of the

gene, the leader sequence of the CD33 protein, thus

allowing hBirA to enter the endoplasmic reticulum. The

resulting fragment was then inserted in the pIRESneo

vector (Clontech). Subsequently, after amplification with

the primers CD33BirA iCPF and CD33BirA iCPR primers

(Table 1), the CD33L/hBirA sequence was subcloned into

the pBiGFP vector using the InFusionTM

Advantage PCR

Cloning Kit (Clontech).

pIREShygro/antiGD2VL

The light chain variable region (VL) of the anti-GD2

monoclonal antibody (mAb) 126 was amplified using the

pcDNA3–mocSIP construct as template [9]. The latter

construct codes for a single chain molecule that includes

the VL leader sequence, VL and VH regions of the 126

mAb. The VL region was amplified together with its leader

sequence. In order to obtain a chimeric human/mouse Fab,

the human constant kappa region was added to the 126

mAb variable region by means of a sequential PCR

approach: the pcDNA3–mocSIP construct was amplified

with primers GD2VK iCPF and VK/CK R, while human B

cell cDNA with primers VK/CK F and IGKC iCPR. The

two fragments were then joined using primers GD2VK

iCPF and IGKC iCPR (Table 1). The resulting chimeric

kappa light chain was cloned into the pIREShygro vector

(Clontech) by means of the InFusionTM

Advantage PCR

Cloning Kit (Clontech).

pPSI–BiotC/antiGD2VH, pPSI–HisTag/antiGD2VH,

and pPSI–BHC/antiGD2VH

In this case as well, the heavy chain variable region (VH) of

the 126 mAb was amplified using the pcDNA3–mocSIP

construct as template. However, since the VH region in the

pcDNA3–mocSIP construct lacks the leader sequence, we

inserted the b2-microglobulin leader sequence at the 50 of

the VH segment. To complete the chimeric Fab we added

the CH1 domain of the human IgM constant region at the 30.In brief, three rounds of PCR amplification with primers

b2 mL/GD2VH F and VH/CH R, b2 mL/joining F and

VH/CH R, GD2VH iCPF, and VH/CH R, respectively,

Table 1 Primers used for the construction of vectors and constructs

The 15 base pairs of the primer that were identical with the 15 base pairs of each extremity of the cut vector, as requested by the InFusionTM

Advantage PCR Cloning Kit (Clontech), are underlined. The six base pairs of the primer IgMCH1 iCPR used to reconstitute the 50 boundary

of the intron are shaded in light gray

Mol Biotechnol

were performed to build the b2-microglobulin leader

sequence/VH fragment using the pcDNA3–mocSIP con-

struct as template. To build the IgM CH1 fragment a single

round of PCR amplification with primers VH/CH F and

IgM CH1 iCPR was performed, using human B cell cDNA

as template. Finally, the two fragments were joined by PCR

amplification with primers GD2VH iCPF and IgM CH1

iCPR to build the complete b2-microglobulin leader

sequence/VH/IgMCH1 fragment (see Table 1 for primer

details). The PCR product was subsequently cloned into the

pPSI–BiotC, pPSI–HisTag and pPSI–BHC vectors by

means of the InFusionTM

Advantage PCR Cloning Kit

(Clontech). For all PCR reactions the proofreading

PrimeSTAR� HS DNA polymerase (Takara) was used. For

every cloning procedure, several clones were sequenced

with an automated sequenator (Applied Biosystems), and

one clone with the correct sequence was selected for cell

transfection.

Cell Line Cultures, Transfections, and Supernatant

Production

HEK293, HeLa, and IMR32 cell lines were cultured in

DMEM GlutamaxTM

medium (Invitrogen, Milan, Italy)

supplemented with penicillin, streptomycin, and 10% of

fetal calf serum (Invitrogen). Transfections and co-trans-

fections of the HEK293 cell line were performed using

Lipofectamine 2000 (Invitrogen) following the manufac-

turer’s instructions. Transfected cells were selected with

Geneticin (Invitrogen), Hygromycin (Invitrogen), and

Puromycin (Invitrogen) as requested, at the concentra-

tion of 300 lg/mL for Geneticin, and Hygromycin, and

0.5 lg/mL for Puromycin. For the production of the bio-

tinylated recombinant anti-GD2 Fab fragment, stably

transfected HEK293 cells were cultured in OptimemTM

medium (Invitrogen) without adding fetal calf serum for

72–96 h. For the production of the biotinylated recombi-

nant antiGD2 Fab fragment, biotin was added to the

medium at the concentration of 0.2 mg/L. For staining

purpose, supernatants were concentrated 35–40 times using

Amicon Ultra 30 K concentrators (Millipore, Billerica,

MA). For the production of the biotinylated/histidine-tag-

ged recombinant anti-GD2 Fab fragment, stably transfected

HEK293 cells were cultured in DMEM GlutamaxTM

med-

ium (Invitrogen) supplemented with penicillin, streptomy-

cin, 10% of fetal calf serum (Invitrogen), and 0.2 mg/L

biotin for 72–96 h.

Biochemical Analysis

Polyhistidine-tagged and biotinylated/polyhistidine-tagged

recombinant anti-GD2 Fab fragments were purified by

adding 10 and 100 mL, respectively, of culture medium to

nickel-charged resin columns (Ni–NTA agarose, Qiagen,

Milan, Italy). Resin was washed and the bound protein

eluted according to the manufacturer’s instruction. We

obtained 15 lg of recombinant protein for each milliliter of

culture supernatant. The purified protein was dialyzed and

concentrated ten times in PBS with Amicon Ultra 30 K

concentrators (Millipore). The polyhistidine-tagged

recombinant anti-GD2 Fab fragment was detected by

sodium dodecyl sulfate polyacrylamide gel electrophoresis

(SDS-PAGE) after loading 26 lg of protein on 12% gel

followed by Coomassie blue staining. The biotinylated/

polyhistidine-tagged recombinant anti-GD2 Fab fragment

was detected by SDS-PAGE after loading 15 lg of protein

on a 12% gel followed by Coomassie blue staining, and

was used for staining experiments.

Cell Staining

IMR32, HEK293, and HeLa cell lines were stained using

either the anti-GD2 126 mAb or the recombinant biotin-

ylated or biotinylated/histine-tagged anti-GD2 Fab frag-

ments: 3 9 105 cells were stained re-suspending them in

100 lL of 126 mAb supernatant or 100 lL of 20–409

biotinylated r-anti-GD2 supernatant or adding to the cells

0.5–4 lg of biotinylated/polyhistidine-tagged purified

r-anti-GD2. PE-conjugated goat anti-mouse polyclonal

antibody (Southern Biotech, Birmingham, AL) and PE-

conjugated streptavidin (Ancell Corporation, Bayport, MN

USA) were used as secondary reagents. Stained cells were

analyzed by means of a FACSCalibur (BD Biosciences,

Franklin Lakes, NJ) using the CellQuest software.

Results

Design of a Vector System Suitable for the Production

of a Protein Set with Different Tags

The aim of this study is to create an expression vector

system being able to produce sets of tagged proteins, dif-

fering from each other only for the tag, by a convenient and

efficient cloning strategy. The main novelty of this vector

system resides in the presence of an incomplete intron at its

50 end. This 50 incomplete intron is followed by a sequence

that codes for a spacer and, in turn, by a tag (Fig. 1a). The

insertion of a sequence carrying the intron boundary

sequence at its 30 end (Fig. 1b) allows for the reconstitution

of a complete functional intron (Fig. 1c). The splicing of

the complete intron will give rise to an mRNA coding for

the protein of choice joined to the spacer and the tag

(Fig. 1d). Altogether, this strategy theoretically allows for

a timely joining of any given recombinant protein to

diverse tags, provided that the corresponding vectors

Mol Biotechnol

carrying the respective tags are set and available. To

demonstrate experimentally that the above expression

system is indeed efficient and cost-effective, we describe

the production of a recombinant Fab fragment conjugated

by three different tags and demonstrate the functional

features of the products.

Vector Construction

We first produced three vectors each with one of the fol-

lowing tags: one tag that was a biotinylation substrate, one

tag made up of six histidines, or one-third tag made by a

combination of the first two tags. The first vector (pPSI–

BiotC) was built by exploiting the ability of the E. Coli

biotin ligase BirA to attach one biotin molecule to the

lysine of a given sequence [6]. This vector was built by

joining a 50 incomplete intron, one sequence coding for a

peptide with spacer function and one-sequence coding for

the shortest peptide still working as substrate for the BirA

enzyme [5]. The second vector (pPSI–HisTag) was built

exactly as the pPSI–BiotC vector except for the tag

sequence that was designed to code for six histidines. The

third vector (pPSI–BHC) was built as the pPSI–BiotC

vector but adding at the 30 a further shorter spacer and a

polyhistidine tag (see ‘‘Materials and Methods’’ section for

details). The use of these three vectors eventually leads to

the production of biotin-, polyhistidine-, and biotin/poly-

histidine-tagged proteins, respectively.

Construction of a Cell System for Intracellular

Enzymatic Protein Biotinylation

Although polyhistidine-tagged proteins are directly

expressed by the pPSI–HisTag vector, biotin-tagging

requires one further step, namely, the intracellular enzy-

matic biotinylation of the pPSI–BiotC and pPSI–BHC

products. To create a cell system devoted to this kind of

protein biotinylation, we produced a cell line expressing a

variant of the BirA enzyme capable to enter the endo-

plasmic reticulum (see ‘‘Materials and Methods’’ section).

To this purpose, we built a bicistronic construct carrying a

codon-optimized engineered version of the BirA enzyme

(hBirA) [8] followed by the EGFP sequence functioning as

reporter gene (pBiGFP/CD33L–hBirA). The evaluation of

EGFP fluorescence intensity on pBiGFP/CD33L–hBirA

transfected HEK293 cells allowed us to exclude clones

Fig. 1 Schematic drawing of the design of pPSI–BiotC, pPSI–

HisTag and pPSI–BHC vectors. a empty vector: the light gray boxes

indicate the two 15-bp regions flanking the EcoRI site used to clone

the insert using the InFusionTM

Advantage PCR Cloning Kit. b vector

cut with EcoRI and insert to be cloned; the crossed lines indicate the

recombination of the regions of homology between insert and vector.

c complete construct with the reconstituted functional intron.

d immature and mature transcripts. See the ‘‘Results’’ section for

more details

Fig. 2 Drawing of the three variants of the r-anti-GD2 molecules.

a r-anti-GD2 tagged with a peptide substrate for biotin ligase that

enzymatically attaches one biotin to the lysine of the peptide. b r-anti-

GD2 tagged with six histidines. c r-anti-GD2 tagged with a peptide

substrate for biotin ligase and a polyhistidine tag

Mol Biotechnol

with too high, and possibly toxic, hBirA expression. By

proper antibiotic selection, a stable HEK293hBirA?

transfectants was obtained, which was able to enzymati-

cally attach a biotin molecule to the lysine of any BirA

substrate peptide in the secretory pathway, provided that

biotin is added to the cell culture medium.

Production of Biotin-, Polyhistidine-, and Biotin/

Polyhistidine-Tagged Fab Fragment

As a proof of principle, we validated experimentally our

expression system by producing the Fab fragment of an

anti-GD2 monoclonal antibody, the 126 mAb, tagged with

Fig. 3 Analysis of the staining

efficiency of the biotinylated

r-anti-GD2 molecule. The

GD2? IMR32 neuroblastoma

cell line was stained with

100 lL of 126 mAb supernatant

(a), with 100 lL of supernatant

from HEK293hBirA? cells

stably co-transfected with

pIREShygro/antiGD2VL and

pPSI–BiotC/antiGD2VH,

concentrated 20 times (b) and

40 times (c) (r-anti-GD2), and

with 100 lL of supernatant

from untransfected

HEK293hBirA? cells

concentrated 40 times (d). The

HEK293 and HeLa cells, which

are negative for GD2

expression, were stained with

the 100 lL of 126 mAb

supernatant (e, g) or with

100 lL of supernatant from

HEK293hBirA? cells stably

co-transfected with the

pIREShygro/antiGD2VL and

pPSI–BiotC/antiGD2VH

vectors concentrated 40 times

(f, h) (r-anti-GD2)

Mol Biotechnol

biotin, polyhistidine or the combination of both and

investigated the functional competence of both the protein

and the tags. We first built the constructs for the heavy and

light chain variable regions of the 126 mAb (see ‘‘Materials

and Methods’’ section for details). The VL containing a

leader sequence was attached to the constant region of the

human kappa chain resulting in a mouse/human chimeric

light chain and then cloned into the pIREShygro expression

vector to obtain the pIREShygro/antiGD2VL construct.

The VH, provided with a leader sequence, was attached to

the CH1 domain of the human IgM constant region. This

sequence, which represents a chimeric heavy chain frag-

ment, was cloned into the pPSI–BiotC, the pPSI–HisTag

and the pPSI–BHC vectors, thus leading to the pPSI–

BiotC/antiGD2VH, pPSI–HisTag/antiGD2VH and pPSI–

BHC/antiGD2VH constructs, respectively. The addition of

the inserts was performed using a recombination system,

namely, the InFusionTM

Advantage PCR Cloning Kit

(Clontech). However, any classical approach based on

restriction enzymes can be employed. Finally, the pIRE-

Shygro/antiGD2VL construct (that code was co-transfected

together with either the pPSI–BiotC/antiGD2VH, pPSI–

HisTag/antiGD2VH and pPSI–BHC/antiGD2VH con-

structs (that code for the VH chain) into the HEK293hBirA?

(pPSI–BiotC/antiGD2VH and pPSI–BHC/antiGD2VH) or

into the wild type HEK293 cell line (pPSI–HisTag/

antiGD2VH). Proper antibiotic treatment selected stable

cell lines that released in the supernatant the Fab fragment

of the 126 anti-GD2 mAb (called herein r-anti-GD2),

biotinylated (Fig. 2a), histidine-tagged (Fig. 2b), and bio-

tinylated/polyhistine-tagged (Fig. 3c). The supernatants

collected after 72- or 96-h culturing were used for

r-anti-GD2 purification or immunofluorescence staining

experiments.

Functional Validation

Biotin-tagged r-anti-GD2 was probed by flow cytometric

antigen-recognition experiments. IMR32 cells, known to

express the GD2 antigen on their surface, and HEK293 and

HeLa cells, which do not, were utilized. Cell surface

immunofluorescence of IMR32 cells, stained with con-

centrated (209 or 409) supernatant of HEK293hBirA?

cells expressing biotinylated r-anti-GD2 (Fig. 3b and c),

and subsequently with PE-conjugated streptavidin, was

similar to that of IMR32 cells stained with the 126 mAb

and PE-conjugated anti-mouse IgG (Fig. 3a). Conversely,

negative immunofluorescence was observed for IMR32

cells stained with 409 concentrated supernatant obtained

by culturing, in the presence of biotin, untransfected

HEK293hBirA? cells (Fig. 3d). Negative results were

obtained as well when using the 409 concentrated r-anti-

GD2 supernatant in control cell lines negative for GD2

surface expression (Fig. 3e–h). These data demonstrate that

the biotinylated anti-GD2 Fab fragment produced using our

expression system is highly suitable for providing a sen-

sitive and specific detection tool.

To probe the functional competence of the polyhistidine

tag as well, r-anti-GD2 collected from the supernatant

of wild-type HEK293 co-transfected with pIREShygro/

antiGD2VL and pPSI–HisTag/antiGD2VH was purified on

nickel-charged resin columns. Upon running the eluate on a

polyacrylamide gel and staining with Coomassie blue,

bands at the expected molecular weight were observed

(Fig. 4). Finally, to validate the functionality of the r-anti-

GD2 with both biotin and histidine tags, the supernatant of

HEK293hBirA? cell line co-transfected with pIREShygro/

antiGD2VL and pPSI–BHC/antiGD2VH constructs was

collected and purified on nickel-charged resin columns.

The supernatant that passed through the column was saved

for further tests. The eluate was concentrated 109 and used

to stain the IMR32 cell line. Flow cytometric measure-

ments of these samples further stained with PE-conjugated

streptavidin as a secondary reagent (Fig. 5c–e) demon-

strated a dose-sensitive binding to the GD2 antigen. Fur-

thermore, the flow through supernatant was concentrated

109 and used to stain the IMR32 cell line (Fig. 5b). No

binding to the cell surface was observed, demonstrating an

efficient binding of the polyhistidine tag to the resin.

Altogether, these data demonstrate the ability of expression

vectors designed in this study to produce fully functional

proteins attached to fully functional tags.

Discussion

In this study, we describe a vector design that allows the

production of differently tagged variants of a protein of

Fig. 4 Biochemical analysis of the histidine-tagged and biotiny-

lated/histidine-tagged r-anti-GD2 molecules. SDS-PAGE gel elec-

trophoresis of purified histidine-tagged r-anti-GD2 in both reducing

and non-reducing conditions (a). Comparison of the reduced

forms of histidine-tagged and biotinylated/histidine-tagged r-anti-

GD2 (b)

Mol Biotechnol

interest in a time- and cost-effective way: the amplified

DNA fragment coding for the selected protein is directly

cloned, without any further modification, into different

vectors, previously built and ready-to-use, each carrying a

particular tag that will be attached to the given protein at its

C terminus. The vector design is based on the reconstitu-

tion, upon the insert cloning, of a functional intron whose

splicing leads to the attachment of the tag to the C terminus

of the protein deprived of the stop codon. This expression

system has been validated by producing a biotinylated, a

polyhistidine-tagged and a biotinylated/polyhistidine-tag-

ged variant of the Fab fragment of the 126 mAb that binds

the antigen GD2. The GD2 molecule is a ganglioside

expressed on the surface of tumor cells of neuroectodermal

origin such as neuroblastoma, an embryonal tumor origi-

nating from sympathetic neural crest cells that represents

6–8% of all childhood cancers [10]. GD2 expression pat-

tern makes it a good target for cancer immunotherapy [11].

The biotinylated and biotinylated/histidine-tagged r-anti-

GD2 variants were produced using intracellular enzymatic

biotinylation, an approach that allows the secretion of the

biotinylated protein directly in the culture supernatant. In

staining experiments, both the biotinylated r-anti-GD2 and

the biotinylated/histidine-tagged r-anti-GD2 showed an

Fig. 5 Analysis of the staining

efficiency of the biotinylated/

polyhistidine-tagged r-anti-GD2

molecule. The GD2? IMR32

neuroblastoma cell line was

stained with the 126 mAb

supernatant (a), with 50 lL

column flow through

supernatant, concentrated 10

times (b), with 0.5 lg (c), 1 lg

(d), 2 lg (e), and 4 lg (f) of the

purified biotinylated/histidine-

tagged r-anti-GD2 Fab fragment

Mol Biotechnol

excellent binding capability of the Fab to the GD2 antigen

and a very good biotin–streptavidin binding. The polyhis-

tidine sequence has been shown to be fully functional as

well, since polyhistidine-tagged r-anti-GD2 was easily

purified. Altogether, our experiments demonstrate the

validity of our expression vector system for both the pro-

tein to be tagged and the tag itself. The feasibility of our

system was validated for tagging proteins at their C ter-

minus. Nevertheless, the same design concept can be used

to build vectors that tag the protein at the N terminus

(Fig. 6). In addition, our expression vector system allows

the design of vectors that join more than one tag to the

protein, sequentially at the same terminus, separated by a

small spacer.

We feel that our labor-saving technique for readily

tagging diverse molecules to antibodies represents an

appealing task in current times where mAbs are becoming

widely utilized both in vitro and in vivo. As an example,

we envisage the application of our expression system for

the production of recombinant antibodies with different

antigenic specificities by means of expression vectors each

tagged with domains for multimerization (e.g., NC1 tri-

merization region of collagen XVIII, p53 multimerization

domain) and/or effector proteins potentially useful for

tumor immunotherapy (e.g., IL2, TNF-alpha). Interest-

ingly, our expression vector system can be used to produce

Fab fragments as well, known to have a better diffusion

than whole mAbs and faster clearance. In addition, chi-

merization of the Fab by means of a human constant

region, as we did, and more in general the production of

humanized Fabs, warrants a reduced in vivo antigenicity of

the compound.

Acknowledgments This study was supported by grants from

Compagnia di San Paolo 4824 SD/CV, 2007.2880, AIRC (IG-10698),

Associazione ‘‘Davide Ciavattini’’ Onlus and Fondazione Maria

Piaggio Casarsa. The authors would like to thank Fondazione

Internazionale in Medicina Sperimentale (FIRMS) that provided

financial and administrative assistance.

References

1. Barat, B., & Wu, A. M. (2007). Metabolic biotinylation of

recombinant antibody by biotin ligase retained in the endoplas-

mic reticulum. Biomolecular Engineering, 24, 283–291.

2. Predonzani, A., Arnoldi, F., Lopez-Requena, A., & Burrone, O.

R. (2008). In vivo site-specific biotinylation of proteins within the

secretory pathway using a single vector system. BMC Biotech-nology, 8, 41.

3. Cuesta, A. M., Sanchez-Martin, D., Sanz, L., Bonet, J., Compte,

M., Kremer, L., et al. (2009). In vivo tumor targeting and imaging

with engineered trivalent antibody fragments containing colla-

gen-derived sequences. Public Library of Science One, 4, e5381.

4. Deyev, S. M., & Lebedenko, E. N. (2008). Multivalency: The

hallmark of antibodies used for optimization of tumor targeting

by design. BioEssays, 30, 904–918.

5. Beckett, D., Kovaleva, E., & Schatz, P. J. (1999). A minimal

peptide substrate in biotin holoenzyme synthetase-catalyzed

biotinylation. Protein Science, 8, 921–929.

6. Chapman-Smith, A., & Cronan, J. E. (1999). The enzymatic

biotinylation of proteins: A post-translational modification of

exceptional specificity. Trends in Biochemical Sciences, 24,

359–363.

7. Occhino, M., Ghiotto, F., Soro, S., Mortarino, M., Bosi, S.,

Maffei, M., et al. (2008). Dissecting the structural determinants of

the interaction between the human cytomegalovirus UL18 protein

and the CD85j immune receptor. Journal of Immunology, 180,

957–968.

8. Mechold, U., Gilbert, C., & Ogryzko, V. (2005). Codon opti-

mization of the BirA enzyme gene leads to higher expression

and an improved efficiency of biotinylation of target proteins

in mammalian cells. Journal of Biotechnology, 116, 245–

249.

Fig. 6 Schematic drawing of the notional design of the N-terminal

tagging variant of the vector system described. a empty vector: the

light gray boxes indicate the two 15 bp regions flanking the EcoRI

site used to clone the insert with the InFusionTM

Advantage PCR

Cloning Kit. b vector cut with EcoRI and insert to be cloned; the

crossed lines indicate the recombination of the regions of homology

between insert and vector. c complete construct with the reconstituted

functional intron. d immature and mature transcripts. Of note, the

primers used to amplify the insert to be cloned in the N tagging vector

are different from that used to amplify the insert to be cloned in the C

tagging vector

Mol Biotechnol

9. Occhino, M., Raffaghello, L., Burrone, O., Gambini, C., Pistoia,

V., Corrias, M. V., et al. (2004). Generation and characterization

of dimeric small immunoproteins specific for neuroblastoma

associated antigen GD2. International Journal of MolecularMedicine, 14, 383–388.

10. Maris, J. M., Hogarty, M. D., Bagatell, R., & Cohn, S. L. (2007).

Neuroblastoma. Lancet, 369, 2106–2120.

11. Gray, J. C., & Kohler, J. A. (2009). Immunotherapy for neuro-

blastoma: turning promise into reality. Pediatric Blood andCancer, 53, 931–940.

Mol Biotechnol