Transposon-based technology

enhances the generation of stable and high-producing CHO clones

for industrial production of recombinant hFGF23 protein

Abstract

In the last 15 years, different approaches for gene transfer into mammalian cells have been developed, however it is still challenging to obtain stable, high-producing cell lines for industrial applications. Conventional

L.Strippoli1, I.Albieri2, E.Ghezzi2, D.Ferrante2, P.N.Brusasca2, L.Querin2 and A.Dal Corso2

1Department of Biotechnology and Biosciences, University of Milano-Bicocca2Diasorin Research Center, Gerenzano (VA)

Transposable elements (TE) are DNA segments with the ability to move from one genetic location to

another. TE falls into two major classes according to whether their replication intermediate is RNA

(retrotransposon) or DNA (transposon). DNA transposons move directly as DNA fragment by a conservative

(cut-and-paste) mechanism of transposition. They consist of a single gene encoding the transposase

enzyme, which is flanked by terminal inverted repeats (TIR). The transposase enzyme recognizes the binding

site within TIR and catalyzes the DNA fragment excision from donor locus and subsequently its reinsertion

elsewhere. PiggyBac (PB) transposon is a TE that specifically targets TTAA sites in genome. Recently it has

In the last 15 years, different approaches for gene transfer into mammalian cells have been developed, however it is still challenging to obtain stable, high-producing cell lines for industrial applications. Conventional

methods, based on spontaneous integration of nude DNA, often result in low transgene expression due to plasmid concatemers silencing and/or position effects. To overcome these limitations we used transposons,

as new gene transfer method (Matasci et al., 2011). The PiggyBac (PB) transposon system relies on the ability of transposase to catalyze single transgene integration into actively transcribed regions of genome. Using

both PiggyBac transposon technology and conventional methods, we generated stable CHO cell lines expressing soluble and properly folded recombinant human fibroblast growth factor 23 (hFGF23), a regulator of

phosphate homeostasis and vitamin D metabolism. Our results demonstrate that PB transposition increases the frequency of stable cell lines generation up to 10 fold compared to nude plasmid transfection. In

addition, cell lines establishment is faster and the frequency of high-producing clones is enhanced. Best candidate clones have been adapted to suspension culture and, in batch culture production, they show an

average productivity of 20 mg/L of secreted and soluble hFGF23 protein. In conclusion, the PB transposon system can be considered a quick, powerful alternative to standard method for generation of stable,

high-producing recombinant mammalian cell lines for industrial protein production.

Introduction

Enhancing frequency of high-producing CHO cell lines

To develop hFGF23 high-producing CHO cell lines we cloned different promoters into transposon vector to

compare their strength (prom1, 2, 3, 4). PB transposon vectors expressing hFGF23R179Q sequence under

the control of selected promoters were transfected into CHO-K1 cells using a transposon/transposase molar

ratio of 2,5:1. As a control experiment, cells were transfected with pCDNA3.1_hFGF23R179Q standard

vector (Std).

Results

elsewhere. PiggyBac (PB) transposon is a TE that specifically targets TTAA sites in genome. Recently it has

been harnessed for non-viral transgenes delivery into cultured mammalian cells2.

Fibroblast Growth Factor 23 (FGF23) belongs to the FGF family and, together with FGF19 and FGF21, is part

of the “endocrine FGFs” sub-group. It is secreted by both skeletal osteoblasts and osteoclasts and is involved

in phosphate homeostasis and vitamin D metabolism3. FGF23 is a 251 amino-acid protein with a 24 amino-

acid signal peptide at the N-terminal portion. It has a β-trefoil structure with an intra-molecular disulfide

bond and a unique C-terminal sequence. Enzymes that recognize a 176Arg–X–X–179Arg region

pCDNA3.1_CMV

hFGF23R179Q

PB_prom1

hFGF23R179Q

PBase (ratio 2,5:1)

PB_prom2

hFGF23R179Q

PBase (ratio 2,5:1)

PB_prom3

hFGF23R179Q

PB200 (ratio 2,5:1)

50

0

cells

PB_prom4

hFGF23R179Q

PBase (ratio 2,5:1)The power of this system is that transposase

can mobilize transposons in trans, as long as

their retain the TIR. Therefore, in the molecular

tool, the transposase gene is physically

separated from the TIR and replaced by a gene

of interest. The resultant transposon is carried

by a plasmid vector, while the transposase is

supplied on a helper plasmid. Through the cut

and paste mechanism the gene of interest is

excised from the donor molecule and in single-

copy integrated into the host genome.

Figure 5. Clone formation assay. A) Crystal violet staining of clones obtained after 1 week neomycin selection of

CHO-K1 transfected cells (see text). Cells were plated at the density of 500 cells/dish. B) Count of stained clones

obtained plating 500 cells/dish for each transfection condition (see text). Results were represented as fold

increase over clones originated from Sdt vector. (n=5, ***p < 0.001).

Figure 6. Classification of hFGF23 producing

clones. Quantitative screening of cell lines

generated with the different transposon plasmids,

constructed for the evaluation of promoters

strength (see text). Clones were classified based

on hFGF23 expression level. 100 clones for each

experimental condition were grown in a 96-well

plate for 2 days. Cell supernatants were analyzed

with an automated prototype immunoassay for

A B

Gene of interestTIR TIR

TIR TIR

Transposase

Transposon plasmid

TransposaseHelper plasmid

Excision

GOI

Reintegration

Host genome

A

B

C

Figure 1. The PiggyBac transposon system. A) DNA transposon structures; B-C)

Transposon system used for transgene delivery: components and mechanism of action.

Std prom prom prom prom0

5

10

15

20

25

30

*** ***

*** ***

Fo

ld in

crea

se

1 2 3 4

50

100StdProm1Prom2Prom3

Prom4

Clo

nes

nu

mb

er (

%)

hFGF23 Batch production

The best producing clones were obtained using transposons with promoter 3 and 4. Two clones for each

construct and for standard plasmid were adapted to serum-free suspension culture. Both adhesion and

suspension cultures were tested for hFGF23 productivity by commercial ELISA assay. The clones in the latter

condition showed an increased yield (Table 1). In Western blot analysis hFGF23 protein showed a molecular

weight of 32 kDa, probably due to glycosylation pattern.

bond and a unique C-terminal sequence. Enzymes that recognize a 176Arg–X–X–179Arg region

proteolytically cleave FGF23 protein between 179Arg and 180Ser. However, the active form of FGF23

corresponds to the full-length protein4. The occurrence of mutations at either 176Arg or 179Arg, destroys

the Arg–X–X-Arg motif and renders the protein resistant to proteolytic cleavage4.

NH2 COOH

1 24

176R-X-X-R179

251

Cleavege site

FGF homology region

Signal peptide

25118017925

Inactive

N-terminal peptideInactive

C-terminal peptide

M 1 2 3 a b c

37 kDa

50 kDa

25 kDa

20 kDa

15 kDa

wt R179Q mutant

ResultsVector used Clone ID

Adherent

culture

yield

(µg/ml)

Suspension

serum-free

culture yield

(µg/ml)

pCDNA3.1_

hFGF23R179Q

a 6,2 13

with an automated prototype immunoassay for

detection of hFGF23 (Diasorin). Values were

normalized based on clones cells number

(Dojindo proliferation assay) and days of culture.

50 kDa

M A S

A S

A B

A B

Figure 2. hFGF23 structure. A) Polypeptide structure; B) Western blot of hFGF23 in supernatant from CHO cell lines expressing hFGF23 wildtype (1,2 and 3)

or R179Q mutated form (a, b and c).

0-80 80-500 500-1000 >10000

ng/106cells/day

Clo

nes

nu

mb

er (

%)

Generation efficiency of stable CHO clones

CHO-K1 cells were transfected with conventional method or with PiggyBac transposon system. To allow a

comparison between this two methods, the hFGF23 coding sequence, mutated at R179 position, was cloned

into pCDNA3.1, a standard expression plasmid (Std), or into PB transposon vector. Different

transposon/transposase molar ratio (2.5:1, 5:1) were tested for the evaluation of optimal transfection

conditions. As a control experiment, cells were also transfected with PB_hFGF23R179Q alone (0).

The two highest performing clones in serum-free suspension culture were compared in 7 days batch

production (Sarstedt-miniPerm and Flask) (Table2). The culture supernatants were harvested and hFGF23

protein was purified by ion exchange as previously described4. The purified protein was tested in

comparison with a commercial mammalian hFGF23 on Liaison CLIA analyzer using an automated prototype

immunoassay (Diasorin) (Fig. 8).

hFGF23R179Qb 3,1 6

PB_prom3

hFGF23R179Q +

PBase (2,5:1)

a 3,4 8

b 2,1 3

PB_prom4

hFGF23R179Q +

PBase (2,5:1)

a 6,9 11

b 4,3 7

Clone Miniperm

yield (µg/ml)

Flask

yield (µg/ml)

pCDNA3.1_

hFGF23R179Q2,3 ± 1,2 22,1 ± 4,3

PB_prom4

Table 1.

37 kDa

25 kDa

20 kDa

15 kDa

Figure 7. Adhesion vs suspension culture. A) Western blot of hFGF23

in supernatant from a clone in adherent (A) or serum-free suspension

(S) culture. B) Bright field microscopy of the same clone in adherent

(A) and serum-free suspension (S) culture. Magnification 20x, scale bar

50 um.

Clones number

Std 0 2,5 50

2

4

6

8

10

12

Fol

d in

crea

se

*** ***

ApCDNA3.1_

hFGF23R179Q

PB_

hFGF23R179Q

PB_

hFGF23R179Q

PBase (ratio 2,5:1)

PB_

hFGF23R179Q

PBase (ratio 5:1)

25

00

cell

s

50

0

cell

s

Figure 8. Comparison of commercial and

purified hFGF23. Different protein dilution

were analyzed on Liaison CLIA analyzer

Commercial vs purified hFGF23

200000

400000

600000Commercial hFGF23Purified hFGF23

Lia

iso

n S

ign

al (

RL

U)

References: 1) Matasci M, et al. (2011) Biotechnol Bioeng., 108(9):2141-50 2) Ding S, et al. (2005) Cell, 122(3):473-83 3) Fukumoto S, Yamashita T. (2007) Bone, 40(5):1190-5 4) Shimada T, et al. (2002) Endocrinology 143(8):3179-82

Conclusions

PiggyBac transposable elements provide a powerful and efficient method for the establishment of stable

CHO cell lines for recombinant proteins production.

→ PB transposon increases up to 10-fold the generation of stable transfected clones compared to standard

gene delivery strategy, also shortening the establishment process

→Changing PB promoter strength enhances the frequency of high-producer clones

→Exploiting this system we are able to generate hFGF23 producing CHO cell lines with an average yield of

20mg/L in batch culture

→The obtained purified protein is correctly detected by both commercial ELISA assay and CLIA automated

Diasorin immunoassay with comparable results to commercial mammalian hFGF23.

Figure 3. Clone formation assay. Crystal violet staining of clones obtained after 1 week

neomycin selection of CHO-K1 transfected cells. Transfection with standard vector was

compared with several molar ratio of transposon to helper plasmid (0:1, 2,5:1 or 5:1). Total

amount of DNA in each transfection was constant. Cells were plated at different density

(500, 2500, 5000 and 50000 cells/plate).

Figure 4. Generation efficiency and survival of clones. A) Count of

stained clones obtained plating 500 cells/dish for each transfection

condition (see text). Results was expressed as fold increase over clones

originated from TE alone. (n=5, ***p < 0.001). B) Percentage of survival

clones after 2 weeks in selective medium. 100 clones were picked for

each transfection condition.

PB_prom4

hFGF23R179Q

+ PBase (2,5:1)

4,3 ± 1,5 24,0 ± 5,1

Table 2. Clones Survival

Std 0 2,5 50

50

100

Clo

nes

su

rviv

al r

ate

(%)

B

50

00

0

cell

s

50

00

cell

s

were analyzed on Liaison CLIA analyzer

(Diasorin). Results are represented as RLU

(relative light units) on inverse of dilution

factor.0.0 0.2 0.4 0.6 0.8 1.0

0

1/FD

Lia

iso

n S

ign

al (

RL

U)