ISSN 00062979, Biochemistry (Moscow), 2017, Vol. 82, No. 4, pp. 510520. Pleiades Publishing, Ltd., 2017.

Original Russian Text T. V. Komarova, E. V. Sheshukova, E. N. Kosobokova, M. V. Serebryakova, V. S. Kosorukov, V. N. Tashlitsky, Y. L. Dorokhov, 2017, published in
Biokhimiya, 2017, Vol. 82, No. 4, pp. 687699.

Trastuzumab and Pertuzumab Plant Biosimilars:

Modification of Asn297linked Glycan of the mAbs
Produced in a Plant with Fucosyltransferase
and Xylosyltransferase Gene Knockouts
T. V. Komarova1,2#, E. V. Sheshukova1#, E. N. Kosobokova1,3, M. V. Serebryakova2,
V. S. Kosorukov1,3, V. N. Tashlitsky2, and Y. L. Dorokhov1,2*

1
Vavilov Institute of General Genetics, Russian Academy of Sciences,
119991 Moscow; Russia; Email: dorokhov@genebee.msu.su
2
Lomonosov Moscow State University, 119991 Moscow, Russia
3
Blokhin Cancer Research Center, 115478 Moscow, Russia
Received August 1, 2016
Revision received October 14, 2016

AbstractPlant biosimilars of anticancer therapeutic antibodies are of interest not only because of the prospects of their
practical use, but also as an instrument and object for study of plant protein glycosylation. In this work, we first designed a
pertuzumab plant biosimilar (PPB) and investigated the composition of its Asn297linked glycan in comparison with
trastuzumab plant biosimilar (TPB). Both biosimilars were produced in wildtype (WT) Nicotiana benthamiana plant (PPB
WT and TPBWT) and transgenic XTFT N. benthamiana plant with XT and FT genes knockout (PPBXTFT and TPB
XTFT). Western blot analysis with anti1,3fucose and antixylose antibodies, as well as a test with peptideNglycosi
dase F, confirmed the absence of 1,3fucose and xylose in the Asn297linked glycan of PPBXTFT and TPBXTFT.
Peptide analysis followed by the identification of glycomodified peptides using MALDITOF/TOF showed that PPBWT
and TPBWT Asn297linked glycans are mainly of complex type GnGnXF. The core of PPBWT and TPBWT Asn297
linked GnGntype glycan contains 1,3fucose and 1,2xylose, which, along with the absence of terminal galactose and
sialic acid, distinguishes these plant biosimilars from human IgG. Analysis of TPBXTFT total carbohydrate content indi
cates the possibility of changing the composition of the carbohydrate profile not only of the Fc, but also of the Fab portion
of an antibody produced in transgenic XTFT N. benthamiana plants. Nevertheless, study of the antigenbinding capacity
of the biosimilars showed that absence of xylose and fucose residues in the Asn297linked glycans does not affect the abili
ty of the glycomodified antibodies to interact with HER2/neu positive cancer cells.

DOI: 10.1134/S0006297917040137

Keywords: plant, monoclonal antibody, immunoglobulin G, trastuzumab, pertuzumab, glycosylation, XTFT N. benthami
ana with knockout of XT and FT genes, immunotherapy, biosimilars

Like in all eukaryotes, glycosylation is one of the contrast to mammals, plants do not synthesize branched
posttranslational protein modifications in plants. Based and sialylated Nglycans [13]. Recently, our understand
on their structure, Nglycans of plants are divided into ing of mechanisms of protein glycosylation in plants was
highmannose glycans, hybrid, and complex types. In significantly enriched due to the practical necessity of

Abbreviations: ER, endoplasmic reticulum; Fab, fragment antigenbinding; Fc, fragment crystallizable; FT, 1,3fucosyltrans
ferase; GlcNAc, Nacetylglucosamine; HC, antibody heavy chain; HER2 and HER3, human epidermal growth factor receptors 2
and 3; IgG, immunoglobulin G; LC, antibody light chain; MALDITOF/TOF, matrixassisted laser desorption/ionizationtime
offlight/timeofflight tandem mass spectrometry; MAPP, monoclonal antibody produced in plants; PNGase F, peptideNgly
cosidase F; PPB, pertuzumab plant biosimilar; PPBXTFT, pertuzumab plant biosimilar obtained from N. benthamiana XTFT;
TFA, trifluoroacetic acid; TMA, therapeutic monoclonal antibody; TPB, trastuzumab plant biosimilar; TPBXTFT, trastuzum
ab plant biosimilar obtained from N. benthamiana XTFT; XT, 1,2xylosyltransferase; XTFT, transgenic N. benthamiana plants
with knockout of 1,3fucosyltransferase (FT) and 1,2xylosyltransferase (XT) genes.
#
These authors contributed equally to this work.
* To whom correspondence should be addressed.

510
 TRASTUZUMAB AND PERTUZUMAB PLANT BIOSIMILARS
using plants as biofactories for production of therapeutic
antibodies. In general, antibodies play an important role
in humoral immunity in humans and have been used for a
long time as an efficient tool for treatment of such human
diseases as cancer, viral infections, and inflammatory and
immune diseases [4]. Like most glycoproteins,
immunoglobulins of a healthy human are subject to glyco
sylation in the endoplasmic reticulum (ER) and Golgi
apparatus. Glycosylation of the immunoglobulin deter
mines structural features of Fab (fragment antigenbind
ing) and Fc (fragment crystallizable) portions of the anti
body. In turn, this affects the antibody binding to antigen
and its effector functions [2]. In all immunoglobulin class
es, Nlinked glycans are present both in Fc and Fab por
tions. A glycan linked to Asn297 of the Fc of human IgG1
has a biantenna structure with a core comprising four
GlcNAc (Nacetylglucosamine) residues and three man
nose residues. The heptasaccharide core is supplemented
with fucose branching from a GlcNAc residue, galactose,
and negatively charged Nacetylneuraminic acid [2].
The majority of more than 40 clinically used thera
peutic monoclonal antibodies (TMA) are chimeric
(mouse/human) or humanized murine antibodies that
belong to subclass IgG1 [1, 2, 5, 6]. Therapeutic activity
of TMA may depend on linked oligosaccharide. The gly
cosylation profile of TMA and its carbohydrate composi
tion are largely determined by cell line, method and cul
tivation conditions [7]. It was shown already that it is pos
sible to design glycomodified TMA with increased effi
ciency for treating cancers. Approaches are developed for
glycomodification as a method for boosting the therapeu
tic and biological activity of TMA including: (i) genetic
methods directed toward both optimization of amino acid
sequence of the Fc portion of the TMA and modification
of cell lines producing TMAs; (ii) methods of transient
expression of foreign genes controlling antibody glycosy
lation in producer cells, and (iii) methods of growth con
dition control and selection of the optimal medium com
position for producer cells cultivation. However, high
efficiency of glycomodified TMA is only proven against
malignant blood cells (obinutuzumab (Gazyva) and
mogamulizumab (Poteligeo)), but not against solid and
metastatic tumors [1, 2].
Nevertheless, TMA production in mammalian cell
systems poses potential safety risks to the product, which
are related to contamination of the TMA with cellular
components, mammalian viruses, or DNA having onco
genic activity. On the other hand, plant cells feature pro
tein expression and posttranslational modification mech
anisms similar to those of animal cells (including glycosy
lation). Therefore, plant cells are suitable for production
of proteins of human and animal origin. It is believed that
plant systems can provide safe, efficient, and scalable
production of TMAs [8, 9]. In addition to licensed origi
nal TMA upon expiration of the corresponding patents,
socalled biosimilars, which are usually produced by
BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017
511
other companies, appeared on the market in 2014. The
term biosimilar may also be used to designate mono
clonal antibodies produced in plants (MAPP). At present,
two MAPP types have been studied most successfully
[10]: (i) against infectious diseases, caused for example by
the human immunodeficiency virus (HIV) and Ebola
virus [8], and (ii) anticancer antibodies [1113]. It is
believed that studying the principles of biosimilars syn
thesis in plants will not only reduce the price for such
anticancer antibodies as trastuzumab, rituximab, beva
cizumab, and pertuzumab, but also boost therapeutic
efficiency of TMA preparations because of their directed
glycomodification [13].
MAPPs are of interest not only because of their clin
ical applicability, but also as objects for determining
mechanisms of plant protein glycosylation. Earlier, we
designed a trastuzumab plant biosimilar (TPB) [12].
Trastuzumab being known in clinical practice as
Herceptin, which is produced in Chinese hamster ovary
(CHO) cell culture. Herceptin is designed for treatment
of breast cancer and directed against the oncoprotein
HER2/neu (Human Epidermal growth factor Receptor
2), which is a transmembrane tyrosine protein kinase
ErbB2 with molecular weight of 185 kDa [14]. In this
work, we first describe preparation of another plant
biosimilar, the pertuzumab plant biosimilar (PPB). It rec
ognizes a site in the extracellular moiety of HER2/neu
that is different from the trastuzumab recognition site.
Trastuzumab interacts with subdomain IV (a.a. 480620),
whereas recently introduced into clinical practice per
tuzumab blocks dimerization of HER2 and HER3 via
interaction with subdomain II (a.a. 165310) [15]. As per
tuzumab and trastuzumab block HER2 at different loci,
simultaneous use of these antibodies leads to an additive
effect, which is important for treatment of cancers that
are resistant to trastuzumab [14, 16]. In this work, we
used MAPPs as a model suitable for studying features of
plant protein glycosylation. To do so, we compared the
composition of Asn297linked glycan in HER2/neuspe
cific TPB and PPB produced in wildtype (WT)
Nicotiana benthamiana (TPBWT and PPBWT) and in
transgenic XTFT N. benthamiana plants with a double
knockout of genes of 1,3fucosyltransferase (FT) and
1,2xylosyltransferase (XT) (TPBXTFT and PPB
XTFT) [17].
MATERIALS AND METHODS
Plant cultivation conditions. Wildtype and XTFT
Nicotiana benthamiana plants [17] were grown in a green
house in pots containing a mixture of leaf compost,
humus, peat, and sand with controlled light cycle
day/night 16/8 h at the temperature of 25/18C. Plants
1112 weeks of age having 56 true leaves were used in the
experiments.
512 KOMAROVA et al.
Designing constructs for producing PPB in plant
cells. The nucleotide sequence encoding the light and
heavy chains of PPB, intended for expression in plant
cells of the genus Nicotiana, was obtained based on the
published (DrugBank accession number DB06366)
amino acid sequence of pertuzumab using a back trans
lation tool with simultaneous codon optimization
(http://www.entelechon.de/2008/10/backtranslation
tool/). The amino acid sequence of a signal peptide of the
heavy chain of the P01749 antibody (HVM05_MOUSE)
from Mus musculus was added at the Nterminus of each
sequence. Nucleotide sequences encoding light and
heavy chains of PPB were synthesized by Evrogen
(Russia) and inserted into the pALTA vector (Evrogen).
These sequences contained NcoI and XhoI recognition
sites at the 5 and 3termini, respectively.
To obtain expression vectors for PPB production, a
binary vector, that carries a constitutive 35S promoter of
cauliflower mosaic virus (CaMV) [18] and based on
pCambia 1300 (Cambia, Australia) backbone, was used.
The CaMV 35S terminator was used for transcription ter
mination. The sequence encoding the light or heavy
chain of PPB was placed between the promoter and the
terminator. To obtain expression vectors, fragments
encoding light and heavy PPB chains flanked with NcoI
XhoI sites were inserted into NcoI and SalItreated plas
mid based on pCambia 1300 (Cambia), which contained
35S promoter and 35S terminator. Thus, we obtained
expression vectors: pA16571 carrying the gene of light
chain of PPB, and pA16671 carrying the gene of heavy
chain of PPB. Then, these plasmids were used for trans
forming Agrobacterium tumefaciens strain GV3101.
PPB and TPB production in N. benthamiana, MAPPs
purification, and testing in mice with subcutaneous SKBR
3 human breast cancer xenografts. We described earlier the
procedure for TPB production [12]. To obtain PPBpro
ducing N. benthamiana plants, a suspension of A. tumefa
ciens strains transformed with the plasmids pA16571 and
pA16671, and with a 35SP19 plasmid encoding the anti
silencing protein P19 (GenBank accession number
M21958) of the tomato bushy stunt virus [19], was intro
duced into leaves. PPB was obtained from leaf extract of
N. benthamiana prepared as follows: leaf material was
frozen in liquid nitrogen and ground into powder, and
then 46 volumes of extraction buffer (200 mM sodium
citrate, 0.1% Tween 20, 5 mM EDTA, pH 6.0) were
added. The suspension was incubated on ice with rocking
for 2030 min and centrifuged at 5000g for 15 min. Two
volumes of the extraction buffer were added to the pellet,
and incubation on ice was repeated with rocking for 20
30 min followed by centrifugation at 5000g for 15 min. The
supernatants were filtered through Miracloth and cen
trifuged at 15,000g for 25 min.
At the next step, plant extract was successively filtered
through Millipore filters with pore diameters from 0.8 to
0.25 m and applied to a protein G Sepharose column
equilibrated with buffer containing 10 mM Na2HPO4,
137 mM NaCl, 2.7 mM KCl, and 10 mM NaH2PO4,
pH 8.5. The optimal flow rate was 1 ml/min. The extract
was applied 24 times to improve binding. After applica
tion, the column was washed once with phosphatebuffered
saline (PBS) and then with 10 mM Naphosphate buffer,
pH 6.0. Elution was carried out with 10 mM NaH2PO4,
pH 3.0, followed by immediate neutralization with 0.4 M
Na2HPO4. For further purification and removal of viruses,
DNA, and endotoxins, we used filtration through
Sartobind Q nano membrane (Sartorius Stedim Biotech).
Plant extracts were analyzed by electrophoresis in
the absence of mercaptoethanol in 7.5% polyacryl
amide gels followed by Coomassie staining as described
earlier [12]. To estimate antibody yield, we used a kit for
determining immunoglobulin concentration, IgG total
EIABEST (VectorBest, Russia), according to the man
ufacturers instructions.
The procedure for examining suppression of
HER2/neu positive tumor growth in mice with subcuta
neous human breast cancer SKBR3 xenografts using
antiHER2/neu MAPPs was described earlier [12]. In
brief, SKBR3 cell culture was injected subcutaneously
in mice bilaterally (3106 cells per injection). Treatment
was started five days after tumor implantation and carried
out intraperitoneally according to the scheme developed
earlier for similar antibodies: first injection of double dose
(20 mg/kg), 8 following injections (10 mg/kg) every 48 h
(100 mg/kg in total). To estimate antitumor activity of a
preparation, tumors were measured and their growth ratio
compared to the initial tumor size (V0) was estimated
after 1, 7, and 10 days of the treatment completion.
Analysis of carbohydrate composition of antibody
preparations. A 0.45ml sample of 10 mg/ml antibody
solution was mixed with 112.5 l of trifluoroacetic acid
(TFA), mixed, and incubated for 6 h at 100C. After cool
ing to room temperature, the solution was centrifuged for
10 min at 14,000g. A 0.5ml sample of supernatant was
mixed with 1.5 ml of water. Two milliliters of the final
solution were applied to the reversedphase concentra
tion cartridge DIAPACK C16. The first 1.5 ml were dis
carded, the next 0.5 ml was collected, and aliquots of 40,
80, and 120 l were taken. Then, the aliquots and a stan
dard were dried in a SpeedVac vacuum concentrator. The
dried sample was mixed with 20 l of 0.5 M PMP (1
phenyl3methyl5pyrazolone) solution in methanol
and 20 l of 0.3 M KOH and incubated at 70C for 2 h.
The mixture was neutralized with 20 l of 0.3 M
hydrochloric acid, and excessive PMP was removed by
double extraction with 500 l of benzene. The remainder
was evaporated and solubilized in 250 l of acetoni
trilewater mixture (1 : 9). Test mixture was analyzed
using reversedphase HPLC on a Luna C18(2) 4.6
250 mm (5 m) column in the gradient regime with
mobile phase A water, B acetonitrile, D 100 mM
dipotassium phosphate in water (pH 9.12) at 1 ml/min
BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017
 TRASTUZUMAB AND PERTUZUMAB PLANT BIOSIMILARS
flow rate and 25C with UV detection at 260 nm. An
Agilent 1100 gradient chromatograph was utilized com
prising a fourchannel pump, an autosampler, a column
thermostat, and a PDAdetector (photodiode array
detector). Gradient profile: 50% D, 816% B (5 min), 16
16% B (8 min), 1630% B (4 min). Total analysis duration
20.65 min. Chromatogram collection and processing
were done using programs ChemStation and
AutoChrom1200/ACDlabs.
Preparation of glycomodified PPB XTFT and TPB
XTFT in N. benthamiana plants with knockout of 1,2
xylosyltransferase ( XT) and 1,3fucosyltransferase
(FT) genes. Nicotiana benthamiana XTFT plants [16]
were grown from seeds, which were kindly provided by
Dr. V. Klimyuk (Icon Genetics GbmH, Germany) in soil
under controlled illumination conditions with day/night
cycle (16/8 h). Glycomodified antibodies were purified as
described above.
Interaction of PPBWT, TPBWT, PPB XTFT, and
TPB XTFT with SKBR3 and BT474 cancer cells. SK
BR3 (ATCC HTB30) and BT474 Clone 5 (ATCC
CRL3247) cells were incubated in RPMI1640 (PanEco,
Russia) medium containing 10% fetal calf serum
(HyClone, USA). Both cell lines feature elevated
HER2/neu expression. The affinity of PPB and TPB to
HER2/neu antigen on the surface of the SKBR3 or BT
474 cells was assessed using flow cytometry. SKBR3 cells
were used for determination of efficiency of PPB binding
to the antigen. Affinities of PPBWT, TPBWT, PPB
XTFT, and TPBXTFT were compared using BT474
cells. For this purpose, cells of the corresponding lines
were grown in cell culture flasks, washed with Versene
solution (PanEco), detached with 0.05% trypsin solution
containing 0.53 mM EDTA and Hanks salts (PanEco),
and washed twice with PBS, pH 7.4. Then, the necessary
amount of the cell suspension was prepared in PBS of
500,000 cells per sample (50 l). The MAPPs were added
to a sample, and the suspension was incubated at room
temperature for 30 min. The antibodies were tested at con
centrations from 0.01 to 100 g/ml. Each dilution was test
ed in triplicate. Then, the samples were washed once with
PBS. FITCconjugated antihuman IgG immunoglobu
lins (BioRad, USA) were added to samples, and the mix
tures were incubated for 30 min at 4C. After incubation
with the antibodies, the cells were washed twice with 1 ml
of PBS; samples were resuspended in 350 l of 1% forma
lin in PBS. The percentage of cells bound to antibodies was
calculated using a FACSCanto II flow cytometer (Becton
Dickinson, USA) and the FASCDiva software.
Detection of 1,3fucose and 1,2xylose in MAPP
preparations. The MAPP was treated with peptideNgly
cosidase F (PNGase F; NEB, Great Britain) under dena
turing conditions according to the manufacturers
instructions. Heavy chains of the MAPP after PNGase F
treatment were analyzed by electrophoresis in 10% poly
acrylamide gels. The Nlinked glycans were analyzed by
BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017
513
liquid chromatography followed by mass spectrometry of
the peptides. Samples were prepared as follows: purified
preparations of TPB or PPB antibody were reduced with
dithiothreitol and then Salkylated with iodoacetamide.
Then the heavy and light chains were separated by PAGE.
The proteins were transferred onto a PVDF membrane,
and a zone corresponding to the heavy chain of antibody
was cut out for analysis.
Membrane fragments (approximately 5 5 mm) were
washed twice with 10% acetonitrile solution in 0.05 M
NH4HCO3. The protein was digested with trypsin directly
on the membrane: the membrane was immersed in 10 l of
modified trypsin solution (Promega, USA) with concen
tration of 20 g/ml in 80% acetonitrile solution in 0.05 M
NH4HCO3. Trypsinolysis was carried out for 2 h at 37C.
After collecting the supernatant, the peptides were addi
tionally eluted with 20 l of 40% acetonitrile with 0.1%
TFA. The eluates were pooled and concentrated by evap
oration of the solvent. The peptides were solubilized in
0.1% TFA and separated by HPLC. Chromatography was
carried out using a Milichrom A02 apparatus (Econova,
Russia) with a 2 75mm reversedphase ProntoSil120
5C18 AQ column in a linear gradient 080% acetonitrile
in 0.1% TFA. Within the acetonitrile concentration range
540%, 100l fractions were collected. Thus, 20 fractions
of tryptic peptides from TPB or PPB were collected and
analyzed by mass spectrometry. Mass spectra were regis
tered using an UltrafleXtreme MALDI TOF/TOF mass
spectrometer (Bruker Daltonics, Germany) equipped
with a UV laser (Nd) using the positive ion mode with
reflectron. Accuracy of measured monoisotopic masses
was 50 ppm. Spectra were registered in the mass range
6006000 m/z, selecting laser power to reach optimal res
olution. The tandem mode was used to obtain fragmenta
tion mass spectra. Accuracy of fragment ions was better
than 1 Da.
RESULTS
Production of functionally active PPB and TPB in
leaves of wildtype N. benthamiana (PPBWT and TPB
WT) and transgenic N. benthamiana XTFT plants with
knockout of XT and FT genes (PPB XTFT and TPB
XTFT). In the first step of our study, we designed con
structs encoding light (LC) and heavy chains (HC) of
PPB and produced PPB antibody in N. benthamiana
plants (see Materials and Methods). Fullsize PPB
WT antibody were detected at levels comparable to those
of the control TPBWT in three days in N. benthamiana
leaves agroinjected with binary vectors pA16571 and
pA16671 carrying LC and HC of PPB, respectively
(Fig. 1a). As from results of electrophoresis in 7.5% poly
acrylamide gel under nonreducing conditions followed
by Coomassie staining (Fig. 1b), fullsize antibodies
TPBWT (lane 1) and PPBWT (lane 2) are accumulated
514 KOMAROVA et al.

a
LC

HC

b +C c +C

0.25 g/ml
d 1 g/ml
91.9% 99.7%

Fig. 1. PPBWT production in N. benthamiana plants. a) Maps of pA16571 and pA16671 binary vectors carrying genes of light chain (LC)
and heavy chain (HC) of PPB, respectively. RB and LB, right and left borders of TDNA; 35S, Cauliflower mosaic virus (CaMV) 35S pro
moter; T, CaMV 35S terminator of transcription. b) SDSPAGE analysis of total soluble protein of N. benthamiana leaves three days after
agroinjection with binary vectors pA16571 and pA16671 (separation in 7.5% polyacrylamide gel under nonreducing conditions followed by
Coomassie staining). Lanes: 1) leaves injected with binary vectors encoding LC and HC of TPBWT; 2) leaves injected with binary vectors
pA16571 and pA16671; +C) 3 g of trastuzumab (positive control); M) molecular weight markers; *, zone corresponding to fullsize anti
bodies. c) SDSPAGE analysis of protein G Sepharose purified PPBWT isolated from N. benthamiana leaves agroinjected with binary vec
tors pA16571 and pA16671 (separation in 12% polyacrylamide gel under reducing conditions followed by Coomassie staining). Lanes: M)
molecular weight markers; +C) 3 g of trastuzumab (positive control); 1) 2 g of PPBWT from fraction No. 3; 2) 2 g of PPBWT from
fraction No. 4. Zones that correspond to HC (#) and LC (arrowhead) of PPBWT are indicated. d) Analysis of PPBWT affinity for
HER2/neu on the surface of SKBR3 cells according to flow cytometry data. Cells were incubated with antibodies at concentrations from
0.25 to 1 g/ml. Antibody concentrations used are indicated at the top. Percentage of cells bound to antibodies is indicated on the right side
of the histograms. P2, zone of signal detection from antibodybound cells. Count, number of events. Abscissa axis corresponds to FITC fluo
rescence.

BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017

 TRASTUZUMAB AND PERTUZUMAB PLANT BIOSIMILARS 515

Table 1. Dynamics of relative tumor growth in Balb/c nude mice with subcutaneous SKBR3 xenografts after joint
application of TPBWT and PPBWT
Relative tumor growth

Vt1/V0 Vt2/V0 Vt3/V0

Group*, efficiency indicator
days after treatment (after implantation)

1 (22) 7 (28) 10 (31)

Tumor growth control (TGC), saline 20.7 31.9 41.5

TPBWT** 9.3## 18.9## 27.7

TPBWT + PPBWT 4.2### 11.2## 16.9##

Notes: V0 initial tumor volume; Vt tumor volume after treatment; differences are significant (p < 0.05): ##
with the TGC group, ###
between
TPBWT and TPBWT + PPBWT groups.
* 8 mice and 16 tumors in each group.
** Treatment intraperitoneally successively on days 5, 7, 9, 11, 13, 15, 17, 19, and 21 of tumor growth.

in leaves of N. benthamiana. Results of preparative purifi
cation using protein G Sepharose column (Fig. 1c) indi
cate the substantial accumulation and the efficient
assembly of PPBWT in agroinfiltrated N. benthamiana
plants. According to results of ELISA PPBWT yield was
150250 mg per 1 kg of leaf mass.
We studied the ability of PPBWT to bind to the
extracellular moiety of oncoprotein HER2/neu by flow
cytometry using SKBR3 cancer cells. The results shown
in Fig. 1d indicate that PPBWT is a fully functional anti
body that can specifically recognize oncoprotein
HER2/neu on the surface of SKBR3 cancer cells.
To assess in vivo PPBWT activity, we used Balb/c
nude mice with subcutaneous human breast cancer SK
BR3 cells xenografts. Earlier, we showed that TPBWT
had an effect on HER2/neu positive tumor that is similar
to the effect of trastuzumab [12]. Therefore, in this work
we compared efficiency of tumor growth suppression
induced by TPBWT alone or in combination with PPB
WT. In accordance with our expectations, cooperative
application of PPBWT and TPBWT increases suppres
sion of HER2/neupositive tumor growth (Table 1),
which confirms the functional activity of PPBWT.
In the next step, we compared ability of transgenic
XTFT N. benthamiana plants with XT and FT genes
knockout to accumulate PPBXTFT and TPBXTFT.
Figure 2 shows that production of PPBXTFT and
TPBXTFT in transgenic XTFT N. benthamiana
plants is comparable to their production in the wildtype
plants.

a b
TPB PPB TPB PPB C M +C TPB PPB TPB PPB M

Fig. 2. Production of PPBXTFT and TPBXTFT in transgenic XTFT N. benthamiana plants. a) SDSPAGE analysis of total soluble protein
from leaves of wildtype N. benthamiana (WT) and transgenic XTFT plants (separation in 7.5% polyacrylamide gel under nonreducing conditions
followed by Coomassie staining). Lanes: ) intact leaf (negative control); M) molecular weight markers; +) 3 g of trastuzumab (positive con
trol). Zone corresponding to fullsize antibodies is indicated with an arrow. b) Electrophoresis of PPBXTFT and TPBXTFT after purification
on protein G Sepharose column. M) molecular weight markers. Zones corresponding to HC (#) and LC (arrowhead) of PPBWT are indicated.

BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017

516 KOMAROVA et al.

Table 2. Molar content of sugars in preparations of TPBWT and TPBXTFT

Designation Mean# Error, % Relative to TPBWT* Deviation from TPBWT*

TPBWT dmannosePMP 5.002 0.020 1.00

glucosaminePMP 3.046 0.008 1.00
dxylosePMP 0.365 0.042 1.00
LfucosePMP 0.731 0.107 1.00
glucosePMP 0.226 0.037 1.00
dgalactosePMP 0.254 0.153 1.00
darabinosePMP 0.196 0.181 1.00

TPBXTFT dmannosePMP 4.236 0.014 0.847 15.30%

glucosaminePMP 2.462 0.009 0.808 19.20%
dxylosePMP 0.000 0.000 0.000 100.00%
LfucosePMP 0.000 0.000 0.000 100.00%
glucosePMP 0.182 0.104 0.805 19.50%
dgalactosePMP 0.173 0.054 0.681 31.90%
darabinosePMP 0.177 0.080 0.901 9.90%

Note: PMP, 1phenyl3methyl5pyrazolone.

#
Molar content normalized to concentration of 10 mg/ml and MW 150 kDa.
* When comparing content of carbohydrate residues in preparations of TPBWT and TPBXTFT.

Determination of carbohydrate residues total content
in preparations of antibodies produced in wildtype and
transgenic XTFT N. benthamiana plants. We confirmed
the ability of N. benthamiana plants with 1,2xylotrans
ferase and 1,3fucosyltransferase genes knockout to
alter carbohydrate content of antibodies analyzing the
total amount of carbohydrate residues in preparations of
antibodies. The results presented in Table 2 demonstrate
that TPBXTFT almost completely lacks xylose and
fucose residues. Noteworthy, in this preparation the con
tent of glucosamine, glucose, and galactose residues is
also decreased substantially. This may indicate alteration
of sugar content not only in the Fc portion of the anti
body but also in the Fab.
Analysis of 1,3fucose presence in PPBXTFT and
TPBXTFT Asn297linked glycan. It is well known [20]
that peptideNglycosidase F (PNGase F) cannot cleave
Nglycan when GlcNAc nearest to the Asn residue is
linked to an 1,3fucose residue. We supposed that PPB
XTFT and TPBXTFT lack 1,3fucose and, thus,
Asn297linked glycan should be cleaved from the heavy
chains of antibodies by PNGase F causing reduction of
HC molecular weight, which can be monitored by an
increase in its mobility in a polyacrylamide gel. Figure 3
demonstrates that, in accordance with our expectations,
treatment with PNGase F does not alter the mobility of
HC of TPBWT and PPBWT, whereas it increases the
mobility of TPBXTFT (Fig. 3a) and PPBXTFT
(Fig. 3b) HC. The conclusion about the absence of fucose
and xylose in TPBXTFT and PPBXTFT prepara
tions was further confirmed by their Western blot analysis
using anti1,3fucose and antixylose antibodies (data
not shown).
Massspectrometry of Nglycosylated peptides of
TPB XTFT and PPB XTFT. In the next step, we per
formed peptide analysis of TPBXTFT and PPB
XTFT with subsequent identification of Nglycosylated
peptides using a MALDITOF/TOF mass spectrometer.
After the HPLC separation of peptides obtained upon
tryptic hydrolysis of HC, we analyzed fractions corre
sponding to approximately 15% acetonitrile. Among the
peaks of a light hydrophilic peptides pool signals around
3000 m/z was observed. According to calculations, series
of these peaks were related to peptides of complete and
incomplete trypsinolysis (EEQYNSTYR and TKPREE
QYNSTYR, respectively) with GnGn, GnGnXF, or
GnGnX glycans linked. In case of TPBWT and PPB
WT HC the signals 2970.5, 3102.4, and 3248.4, that cor
respond to GnGn, GnGnX, or GnGnXF3 glycans linked,
were observed (Fig. 4, a and c). In addition, the last two
(3102.4 and 3248.4), which are typical for glycoproteins
containing fucose and/or xylose, were absent in spectra of
the proteins produced in XTFT N. benthamiana plants:
TPBXTFT and PPBXTFT (Fig. 4, b and d). It
should be mentioned that after PNGase treatment of
TPBXTFT HC, only m/z 1190.6 and 1672.9 peaks
were detected, which correspond to nonglycosylated
peptides EEQYNSTYR and TKPREEQYNSTYR
(Fig. 4e). Hence, in case of TPBWT, the 2766.1 and
3248.4 peaks corresponding to EEQYNSTYR and

BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017

 TRASTUZUMAB AND PERTUZUMAB PLANT BIOSIMILARS 517

a C1 C2
b C3 C4

Fig. 3. Analysis of 1,3fucose presence in TPBXTFT and PPBXTFT Asn297linked glycan using PNGase F. a, b) Gel mobility of TPB
XTFT and PPBXTFT, respectively, after PNGase F treatment. Lanes: C1) TPBWT without PNGase F treatment; C2) TPBXTFT
without PNGase F treatment; C3) PPBWT without PNGase F treatment; C4) PPBXTFT without PNGase F treatment; M) molecular
weight markers.

a d
300 TPBWT 3248.4 XTFT
PPB 2970.5
3000

250 GnGnXF3 2500 GnGn

GnGn
GnGn
200 2000 GnGn
2970.5
150 3102.4
1500 2488.1
GnGn
100 GnGn 1000
2488.1 GnGnXF3
50 500
2620.1
Intensity, arbitrary units

2766.1
0 0
2400 2600 2800 3000 3200 2400 2600 2800 3000 3200

b 104 e
2970.5 1.0 1672.9
XTFT
TPB XTFT
TPB
6000 + PNGase
Intensity, arbitrary units

GnGn 0.8
5000

4000 0.6

3000
0.4
2000
GnGn
0.2 1190.6
1000
2488.1
0 0.0
2400 2600 2800 3000 3200 1000 1250 1500 1750 2000 2250 2500 2750 3000 3250

c
PPBWT 3248.4

600
GnGnXF3

400

200 GnGn GnGnX

GnGn GnGnXF3

2766.1 2970.5 3102.4

2488.1
0
2400 2600 2800 3000 3200

Fig. 4. Mass spectrometry of Nglycosylated peptides of TPB and PPB. Nglycosylation profile of TPBWT (a), TPBXTFT (b), PPBWT
(c), and PPBXTFT (d) as well as TPBXTFT treated with PNGase F (e). Underlined mass values correspond to different glycoforms of
TKPREEQYNSTYR peptide. Nglycan abbreviations used correspond to the nomenclature (http://www.proglycan.com).

BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017

518 KOMAROVA et al.

a
0.25 g/ml 1 g/ml 10 g/ml

Fig. 5. Analysis of TPB and PPB affinity for oncoprotein HER2/neu on the surface of BT474 cells by flow cytometry: a) TPBWT; b) TPB
XTFT; c) PPBWT; d) PPBXTFT; e) Herceptin (control). Antibody concentrations used in the experiments are shown at the top.
Percentage of cells bound to antibodies is indicated on the right side of histograms. P2, zone of signal detection from cells bound to antibod
ies. Count, number of events. Abscissa axis corresponds to FITC fluorescence.

BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017

 TRASTUZUMAB AND PERTUZUMAB PLANT BIOSIMILARS
TKPREEQYNSTYR with GnGnXF3 glycan linked were
still present (data not shown). In addition, fragmentation
spectra for m/z 3248.4 and 2970.5 peaks were obtained.
They confirmed (i) correspondence of these peaks to gly
cosylated forms of TKPREEQYNSTYR peptide and (ii) a
standard form of glycoantenna, GnGnXF3 and GnGn.
Thus, Fig. 4 demonstrates that Asn297linked gly
cans of TPBWT and PPBWT represent mainly complex
glycans GnGn and GnGnXF3, whereas glycans of TPB
XTFT and PPBXTFT GnGn contain neither Asn297linked glycans (Fig. 4c). Asn297linked glycans
xylose nor fucose.
Interaction of PPBWT, TPBWT, PPB XTFT, and
TPBXTFT with BT474 cancer cells. To assess the effi
ciency of interaction of biosimilars with antigens on the
surface of cancer cells, we utilized BT474 clone 5 cells,
which are resistant to Herceptin (trastuzumab):
trastuzumab recognizes oncoprotein HER2/neu on the
surface of these cells, binds to it, but does not inhibit
growth of tumor because there is no expression of p27, an
inhibitor of a cyclindependent kinase, in these cells [21].
Using flow cytometry, we showed that there is no signifi
cant difference between ability of TPBWT, TPBXTFT,
and Herceptin to bind to cancer cells (Fig. 5, a, b, and e).
Preparations of PPBWT and PPBXTFT also bind to
HER2/neupositive cells with the same efficiency (Fig. 5,
c and d). Therefore, we conclude that production of gly
comodified biosimilars, PPBXTFT and TPBXTFT,
in XTFT plants with XT and FT genes knockout does
not affect their ability to interact with the oncoprotein
HER2/neu on the surface of BT474 cancer cells.
DISCUSSION
Fundamental studies of posttranslational modifica
tion of proteins in plants were significantly accelerated
after discovery of their ability to produce active biosimi
lars of human therapeutic antibodies [1, 2]. At present, of
18 TMA intended for malignant tumors treating [2], only
rituximab [9], trastuzumab [12, 22, 23], and cetuximab
[24] plant biosimilars have been obtained and their carbo
hydrate profiles analyzed. Furthermore, an idiotypic vac
cine for treatment of nonHodgkins lymphoma was
designed based on fullsize antibodies [25].
Our study represents the first stage of the research
aiming the elucidation of plantproduced protein glyco
sylation features using human antibodies as an example.
We have shown that PPB, a biosimilar of pertuzumab that
interacts with one of the HER2/neu extracellular
domains different from the trastuzumab recognition site
[26], could be produced in plants. Analysis of total solu
ble protein from agroinjected leaves and subsequent
purification of antibodies on a protein G Sepharose col
umn demonstrated high level of PPBWT production
comparable with that for TPB (Fig. 1, b and c). PPB is a
functional antibody able to specifically recognize
BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017
519
HER2/neu oncoprotein on the surface of SKBR3 can
cer cells (Fig. 1d) and to boost anticancer activity of TPB
(Table 1). The reason for decreased (compared to TPB)
ability of PPB to bind to BT474 cells (Fig. 5, c and d) is
not clear yet and requires further investigation.
Glycoanalysis of PPBWT using PNGase F revealed
1,3fucose in Asn297linked glycan (Fig. 3). Mass spec
trometry of PPBWT shows that it contains predomi
nantly complex GnGn, GnGnX, and GnGnXF3
of PPBWT are of GnGntype, which agrees with previ
ously published results for trastuzumab [27], cetuximab
[24], and rituximab [9] produced in a plant. Our results
(Table 2) also suggest that preparation of TPBXTFT
antibody produced in XTFT plants with knockout of XT
and FT genes not only lacks xylose and fucose residues,
but also contains reduced amounts of glucosamine, glu
cose, and galactose residues. This may indicate altered
sugar content not only in Fc, but also in the Fab portion.
However, we did not find significant differences between
TPBWT and TPBXTFT affinity toward antigen. Also,
in accordance with previously reported data, Asn297
linked glycan of biosimilars produced in transgenic
XTFT N. benthamiana plants lacks 1,2xylose and
1,3fucose residues (Fig. 4, b and d), which does not
affect the ability of PPBXTFT and TPBXTFT to
interact with BT474 cancer cells containing oncoprotein
HER2/neu on the surface (Fig. 5).
Acknowledgements
This work was supported by the Russian Science
Foundation (project No. 161400002).
REFERENCES
1. Sheshukova, E. V., Komarova T. V., and Dorokhov, Y. L.
(2016) Plant factories for the production of monoclonal
antibodies, Biochemistry (Moscow), 81, 11181135.
2. Dorokhov, Y. L., Sheshukova, E. V., Kosobokova, E. N.,
Shindyapina, A. V., Kosorukov, V. S., and Komarova, T. V.
(2016) Functional role of carbohydrate residues in human
immunoglobulin G and therapeutic monoclonal antibod
ies, Biochemistry (Moscow), 81, 835857.
3. Loos, A., and Steinkellner, H. (2014) Plant glycobiotech
nology on the way to synthetic biology, Front. Plant Sci., 5,
523.
4. Gronski, P., Seiler, F. R., and Schwick, H. G. (1991)
Discovery of antitoxins and development of antibody
preparations for clinical uses from 1890 to 1990, Mol.
Immunol., 28, 13211332.
5. Kunert, R., and Casanova, E. (2013) Recent advances in
recombinant protein production: BACbased expression
vectors, the bigger the better, Bioengineered, 4, 258261.
6. Walsh, G. (2014) Biopharmaceutical benchmarks 2014,
Nat. Biotechnol., 32, 9921000.
520 KOMAROVA et al.
7. Niwa, R., and Satoh, M. (2015) The current status and
prospects of antibody engineering for therapeutic use: focus
on glycoengineering technology, J. Pharm. Sci., 104, 930
941.
8. Hiatt, A., Whaley, K. J., and Zeitlin, L. (2014) Plant
derived monoclonal antibodies for prevention and treat
ment of infectious disease, Microbiol. Spectr., 2, AID0004
2012.
9. Holtz, B. R., Berquist, B. R., Bennett, L. D., Kommineni,
V. J. M., Munigunti, R. K., White, E. L., Wilkerson, D. C.,
Wong, K.Y. I., Ly, L. H., and Marcel, S. (2015)
Commercialscale biotherapeutics manufacturing facility
for plantmade pharmaceuticals, Plant Biotechnol. J., 13,
11801190.
10. Gleba, Y. Y., Tuse, D., and Giritch, A. (2014) Plant viral
vectors for delivery by Agrobacterium, Curr. Top. Microbiol.
Immunol., 375, 155192.
11. Komarova, T. V., Baschieri, S., Donini, M., Marusic, C.,
Benvenuto, E., and Dorokhov, Y. L. (2010) Transient
expression systems for plantderived biopharmaceuticals,
Expert Rev. Vaccines, 9, 859876.
12. Komarova, T. V., Kosorukov, V. S., Frolova, O. Y., Petrunia,
I. V., Skrypnik, K. A., Gleba, Y. Y., and Dorokhov, Y. L.
(2011) Plantmade trastuzumab (Herceptin) inhibits
HER2/Neu+ cell proliferation and retards tumor growth,
PLoS One, 6, e17541.
13. Stoger, E., Fischer, R., Moloney, M., and Ma, J. K.C.
(2014) Plant molecular pharming for the treatment of
chronic and infectious diseases, Annu. Rev. Plant Biol., 65,
743768.
14. Rimawi, M. F., Schiff, R., and Osborne, C. K. (2015)
Targeting HER2 for the treatment of breast cancer, Annu.
Rev. Med., 66, 111128.
15. Barthelemy, P., Leblanc, J., Wendling, F., Wissler, M.P.,
and Bergerat, J.P. (2014) Pertuzumab and solid tumors:
perspectives, Bull. Cancer (Paris), 101, 11141121.
16. Chung, C., and Lam, M. S. H. (2013) Pertuzumab for the
treatment of human epidermal growth factor receptor type
2positive metastatic breast cancer, Am. J. Health Syst.
Pharm., 70, 15791587.
17. Strasser, R., Stadlmann, J., Schahs, M., Stiegler, G.,
Quendler, H., Mach, L., Glossl, J., Weterings, K., Pabst,
M., and Steinkellner, H. (2008) Generation of glycoengi
neered Nicotiana benthamiana for the production of mono
clonal antibodies with a homogeneous humanlike Ngly
can structure, Plant Biotechnol. J., 6, 392402.
18. Arce, A. L., Cabello, J. V., and Chan, R. L. (2008) Patents
on plant transcription factors, Recent Pat. Biotechnol., 2,
209217.
19. Dorokhov, Y. L., Frolova, O. Y., Skurat, E. V., Ivanov, P. A.,
Gasanova, T. V., Sheveleva, A. A., Ravin, N. V., Makinen,
K. M., Klimyuk, V. I., Skryabin, K. G., Gleba, Y. Y., and
Atabekov, J. G. (2006) A novel function for a ubiquitous
plant enzyme pectin methylesterase: the enhancer of RNA
silencing, FEBS Lett., 580, 38723878.
20. Tretter, V., Altmann, F., and Marz, L. (1991) PeptideN4
(Nacetylbetaglucosaminyl)asparagine amidase F cannot
release glycans with fucose attached alpha 13 to the
asparaginelinked Nacetylglucosamine residue, Eur. J.
Biochem. FEBS, 199, 647652.
21. Kute, T., Lack, C. M., Willingham, M., Bishwokama, B.,
Williams, H., Barrett, K., Mitchell, T., and Vaughn, J. P.
(2004) Development of Herceptin resistance in breast can
cer cells, Cytom. Part J. Int. Soc. Anal. Cytol., 57, 8693.
22. Garabagi, F., McLean, M. D., and Hall, J. C. (2012)
Transient and stable expression of antibodies in Nicotiana
species, Methods Mol. Biol. Clifton NJ, 907, 389408.
23. Grohs, B. M., Niu, Y., Veldhuis, L. J., Trabelsi, S.,
Garabagi, F., Hassell, J. A., McLean, M. D., and Hall, J.
C. (2010) Plantproduced trastuzumab inhibits the growth
of HER2 positive cancer cells, J. Agric. Food Chem., 58,
1005610063.
24. Castilho, A., Gruber, C., Thader, A., Oostenbrink, C.,
Pechlaner, M., Steinkellner, H., and Altmann, F. (2015)
Processing of complex Nglycans in IgG Fcregion is
affected by core fucosylation, mAbs, 7, 863870.
25. Tuse, D., Ku, N., Bendandi, M., Becerra, C., Collins, R.,
Langford, N., Sancho, S. I., LopezDiaz de Cerio, A.,
Pastor, F., Kandzia, R., Thieme, F., Jarczowski, F., Krause,
D., Ma, J. K.C., Pandya, S., Klimyuk, V., Gleba, Y., and
ButlerRansohoff, J. E. (2015) Clinical safety and
immunogenicity of tumortargeted, plantmade IdKLH
conjugate vaccines for follicular lymphoma, BioMed Res.
Int., 2015, 648143.
26. Franklin, M. C., Carey, K. D., Vajdos, F. F., Leahy, D. J.,
De Vos, A. M., and Sliwkowski, M. X. (2004) Insights into
ErbB signaling from the structure of the ErbB2pertuzum
ab complex, Cancer Cell, 5, 317328.
27. Garabagi, F., Gilbert, E., Loos, A., McLean, M. D., and Hall,
J. C. (2012) Utility of the P19 suppressor of genesilencing
protein for production of therapeutic antibodies in Nicotiana
expression hosts, Plant Biotechnol. J., 10, 11181128.
BIOCHEMISTRY (Moscow) Vol. 82 No. 4 2017