Download High-yield production of a human therapeutic protein in

© 2000 Nature America Inc. • http://biotech.nature.com
RESEARCH ARTICLES
High-yield production of a human
therapeutic protein in tobacco chloroplasts
Jeffrey M. Staub1*, Bradley Garcia2, Julie Graves2, Peter T. J. Hajdukiewicz1, Priscilla Hunter1,
Narender Nehra1, Vikram Paradkar2, Michael Schlittler1, James A. Carroll1, Lori Spatola2, Dannette Ward1,
Guangning Ye1, and Douglas A. Russell2
1Monsanto
Company, 700 Chesterfield Village Parkway North, St. Louis, MO 63198. 2Integrated Protein Technologies, Agracetus Campus, Monsanto Company,
8520 University Green, Middleton, WI 53562. *Corresponding author ([email protected]).
© 2000 Nature America Inc. • http://biotech.nature.com
Received 17 November 1999; accepted 12 December 1999
Transgenic plants have become attractive systems for production of human therapeutic proteins
because of the reduced risk of mammalian viral contaminants, the ability to do large scale-up at low cost,
and the low maintenance requirements. Here we report a feasibility study for production of a human therapeutic protein through transplastomic transformation technology, which has the additional advantage
of increased biological containment by apparent elimination of the transmission of transgenes through
pollen. We show that chloroplasts can express a secretory protein, human somatotropin, in a soluble,
biologically active, disulfide-bonded form. High concentrations of recombinant protein accumulation are
observed (>7% total soluble protein), more than 300-fold higher than a similar gene expressed using a
nuclear transgenic approach. The plastid-expressed somatotropin is nearly devoid of complex posttranslational modifications, effectively increasing the amount of usable recombinant protein. We also
describe approaches to obtain a somatotropin with a non-methionine N terminus, similar to the native
human protein. The results indicate that chloroplasts are a highly efficient vehicle for the potential production of pharmaceutical proteins in plants.
Keywords: pharmaceutical, plastid transformation, somatotropin, transplastomic, ubiquitin fusion
The ability to scale up biomass with low maintenance and cost has
made transgenic plants an attractive production vehicle for molecules such as carbohydrates, fatty acids, industrial enzymes,
biodegradable plastics, and pharmaceuticals. For the production of
pharmaceuticals, plants have the additional advantage of minimizing potential viral contaminants such as human immunodeficiency
virus and hepatitis B (refs. 1–4). Recent examples demonstrate the
production of monoclonal antibodies in transgenic plants5,6 and
human antigens in transgenic food crops used for oral immunization7. We report here the feasibility of chloroplast transformation
(transplastomic) technology with several potential advantages over
existing nuclear transgenic systems for production of human therapeutic proteins in a nonfood crop, tobacco. Furthermore, we show
that a normally secreted human protein can be properly expressed in
a biologically active form in transplastomic plants.
For transformation of chloroplasts in higher plants, particle
bombardment is used to introduce transgenes into leaf chloroplasts
where integration of the foreign DNA is directed by homologous
recombination8–11. In a leaf cell, there are as many as 100 chloroplasts with up to 100 plastid DNAs (ptDNA) each, for a total of
∼10,000 genome copies per cell. Stable transformation requires
that all 10,000 ptDNA copies be uniformly converted. This is
achieved through amplification and sorting of transgenic ptDNAs
with elimination of the wild-type copies on selective medium 12.
Because foreign DNA is integrated by homologous recombination,
the location of transgene insertion is predictable and gene expression is uniform among all independently transformed lines.
Furthermore, expression is not subject to gene silencing and is stable over numerous generations. Transplastomic plants also have
the potential for very high amounts of recombinant protein
expression, in part because of the very high ploidy level of the plasNATURE BIOTECHNOLOGY VOL 18 MARCH 2000
http://biotech.nature.com
tid genome. An important point for biotechnology applications is
that field management of transplastomic plants may be simpler
than nuclear transformants, because the plastid genome is only
very rarely transmitted through pollen13,14.
Recombinant therapeutic proteins need to carry all of the same
post-translational modifications and processing events as the native
human protein. In nuclear transgenic systems, this requirement
often necessitates complex protein expression and targeting schemes
to limit expression concentrations and prevent unwanted posttranslational modifications to the heterologous protein. In addition,
expression of nuclear transgenes typically involves random chromosomal insertion, which requires screening large numbers of transgenics to find usable lines. Several bacterial proteins have been successfully expressed in chloroplasts. However, expression of mammalian or secretory proteins has not been tested. Therefore it has
been unknown if chloroplasts could properly express complex
eukaryotic proteins that require controlled disulfide bond formation for biological activity. Although the enzymatic activity of some
nuclear-encoded proteins that are imported into chloroplasts can be
regulated by oxidation/reduction of sulfhydryl groups15, there are no
known plastid-encoded proteins containing disulfide bonds.
We chose to test whether a human therapeutic protein, human
somatotropin (hST), could be expressed in a biologically active,
disulfide-bonded form in tobacco chloroplasts. The primary use of
hST is in the treatment of hypopituitary dwarfism in children; additional indications are in treatment of Turner syndrome, chronic
renal failure, HIV wasting syndrome, and possibly treatment of the
elderly16. As produced in the pituitary gland, hST enters the secretory system, coincident with removal of its signal peptide and formation of two disulfide bonds. In contrast to most other secretory proteins, somatotropin is not glycosylated in mammals17. Recombinant
333
© 2000 Nature America Inc. • http://biotech.nature.com
RESEARCH ARTICLES
A
A
B
C
B
© 2000 Nature America Inc. • http://biotech.nature.com
D
E
Figure 2. Accumulation of hST protein in transplastomic lines.
Comparative western blot analysis is shown of hST accumulation
from (A) a nuclear hST transgene (Nt-4747) and plastid hST
transgenes (Nt-4838 and Nt-38755) and (B) the Prrn/G10L-driven
plastid transgene in Nt-38794. Plant samples are loaded at 5 µg per
lane of total soluble protein. A dilution series of E. coli-produced hST
(r-hST) is used for the quantitations. Note that incomplete ubiquitin
cleavage in the fusion protein lines results in both unprocessed and
mature forms of hST protein. y, young leaf; m, mature leaf; o, old leaf.
Figure 1. Analysis of transplastomic plants carrying recombinant hST
genes integrated into the plastid genome. (A) hST chimeric genes are
driven by plastid 5′-expression signals, PpsbA or Prrn/G10L. The 3′end mRNA stabilization element, Trps16, is also shown. (B) The
chimeric genes are cloned next to the selectable aadA spectinomycin
resistance gene in the pPRV plastid transformation vectors.
Integration of the transgenes is into the inverted repeat region
between the trnV gene and rps7/3′–rps12 operon. (C) The 2.4 kb wildtype EcoRV/EcoRI DNA fragment (P1) and 0.8 kb hST coding region
(P2) are used as probes for the Southern blot analysis. BamHI
digestion yields a 3.3 kb wild-type ptDNA fragment. Transformants
carry the ∼5.5 kb transplastomic fragment. Note that the size of the
transplastomic fragment differs slightly in each line because of the
different hST transgenes. (D, E) Southern blot analysis of transformed
plants using the P1 probe (D) or the hST coding region P2 probe (E).
somatotropin was one of the first products available from the
biotechnology industry; one current production method is by targeting to the bacterial periplasm18–21. The increasing number of indications for hST warrants an investigation of efficient alternative
modes of therapeutic protein production, and the feasibility study
reported here indicates that chloroplast transformation meets these
requirements.
Results and discussion
Chimeric hST genes for chloroplast expression. In the pituitary
gland, removal of the signal peptide from somatotropin leaves
phenylalanine as the N-terminal amino acid17. As normal translation
in plastids initiates at methionine, we designed a ubiquitin–hST
fusion to yield a phenylalanine N terminus in the final hST product.
Ubiquitin fusion proteins are cleaved by ubiquitin protease immediately downstream of the C-terminal glycine residue of ubiquitin.
This property has allowed production of recombinant proteins containing N-terminal residues other than methionine22. Although
ubiquitin protease is apparently not present in chloroplasts23, we
tested if the ubiquitin–hST fusion would be processed at any point
during synthesis, accumulation, or purification from the plants.
Therefore, for the pMON38755 and pMON38794 constructs, we
fused the yeast ubiquitin in frame to hST, such that cleavage should
generate an N-terminal phenylalanine on hST. The control wrg4838
334
construct carries the full-length cDNA encoding methionine and
then alanine as the first amino acids of hST (Fig. 1).
Expression elements were incorporated to maximize hST accumulation in leaf chloroplasts. In plasmids wrg4838 and
pMON38755, the hST gene is expressed from the plastid psbA promoter and 5′-untranslated region, known to be upregulated in leaves
(PpsbA; ref. 24). Plasmid pMON38794 uses the strong constitutive
chloroplast ribosomal RNA operon promoter (Prrn; ref. 11) with the
ribosome binding site (rbs) region of the bacteriophage T7 gene 10
leader (G10L). This rbs is complementary to the anti-rbs sequence
on the plastid 16SrRNA and was predicted to provide highly efficient
translation signals. The chimeric genes contain the 3′-mRNA stability element from the chloroplast ribosome small subunit 16 gene
(Trps16; ref. 24).
The chimeric hST genes were cloned into chloroplast transformation pPRV vectors25 linked to the selectable spectinomycin-resistance (aadA) gene. The vectors target insertion of the foreign genes
into the plastid genome between the plastid trnV gene and
rps7/3′–rps12 operon in the duplicated inverted repeat region. The
foreign genes are therefore present in two copies in each recombinant plastid genome.
Identification of transplastomic plants. The chimeric genes were
introduced into tobacco leaf chloroplasts using the biolistic
process11. Transformed shoots were identified by growth on spectinomycin. Uniformly transformed (homoplasmic) plants were
obtained after two rounds of plant regeneration.
To verify that spectinomycin-resistant plants are indeed plastid
transformants, Southern blot analysis was done (Fig. 1). Total cellular DNA from independent transformants was digested with
BamHI, electrophoresed, and transferred to nitrocellulose for probing. The ptDNA flanking region probe (P1) identified a 3.3 kb fragment in the wild-type control plant. In the transplastomic lines,
only transformed genome copies are observed as evidenced by the
∼5.5 kb hybridizing fragment. The transformed lines are homoplasmic, as indicated by no detectable wild-type 3.3 kb genomes
remaining. To prove that the ∼5.5 kb transgenic band contains the
NATURE BIOTECHNOLOGY VOL 18 MARCH 2000
http://biotech.nature.com
© 2000 Nature America Inc. • http://biotech.nature.com
RESEARCH ARTICLES
© 2000 Nature America Inc. • http://biotech.nature.com
A
B
Figure 3. Recombinant chloroplast hST contains the correct disulfide
bonds. (A) Purified chloroplast hST was subjected to RP-HPLC
analysis in the absence (-DTT, black line) or presence (+DTT, gray
line) of the reducing agent, dithiothreitol (DTT). Note the mobility shift
of the reduced sample. (B) The peak, nonreduced chloroplast hST
fraction was collected and ES-MS was used to identify different
protein species and calculate their molecular mass, followed by Nterminal amino acid sequencing. The protein species, P-hST and FhST, represent recombinant chloroplast somatotropin with a proline
or the desired phenylalanine, respectively, at the N terminus. (C) The
purified chloroplast hST was subjected to tryptic digestion followed
by MALDI-MS(oxidized spectrum). Note the correctly paired tryptic
peptides T6 disulfide-bound to T16 (T6S-ST16) and T20 disulfidebound to T21 (T20S-ST21). (D) The tryptic digest was treated with
Tris(2-carboxyethyl)phosphine (TCEP) before mass spectrometry
(reduced spectrum). Note that the disulfide-bound peptides
disappear, whereas signals for the cysteine-containing tryptic
peptides (T6, T16, and T20) become visible. Peptide T21 is not
observed in the spectrum because of the lack of basic residues. T9a
represents a contaminating chymotryptic digestion of peptide T9.
hST gene, the blot was re-probed with the P2 hST coding region
probe. In this analysis, hybridization is detected only in the
transplastomic lines, as expected.
Processed hST accumulates to a high concentration in
transplastomic plants. Accumulation of somatotropin in transplastomic plants was monitored by western blot analysis, using a dilution series of Escherichia coli produced hST for quantitation (Fig. 2).
The transplastomic Nt-4838 line accumulates hST at ∼0.2% total
soluble protein (tsp). The Nt-38755 line accumulated at least fivefold more (>1% tsp) total hST protein species than the Nt-4838 line
that uses the same expression signals. The western analysis of the Nt38794 line (Fig. 2B) indicates a very high expression concentration,
estimated from several samples to be ∼7% tsp in mature leaves of tissue culture-derived plants. This amount of expression is about sevenfold higher than the Nt-38755 lines. Northern blot analysis of
total hST mRNA showed no significant difference among the three
transplastomic lines (data not shown), indicating that differences in
protein accumulation are probably due to enhanced translation
through the G10L or protein stability mediated by a chaperonin-like
effect of ubiquitin22. Leaves of different ages had different hST accumulation patterns, with mature and old leaves having similar
amounts and younger leaves much less hST.
Interestingly, both ubiquitin–hST and processed hST accumulated in the extracts of the Nt-38755 and Nt-38794 lines (Fig. 2).
Processing efficiency ranged from 30 to 80% of total hST protein
species depending on extraction conditions. Ubiquitin processing
was confirmed by loss of signal from the higher molecular weight
hST forms after probing western blots with anti-ubiquitin antibody
(data not shown). These results confirm the utility of the fusion protein approach in chloroplast-expressed proteins. The extra band
observed in the Nt-4838 sample (Fig. 2A) is consistent with an hST
dimer. The additional bands observed in the ubiquitin fusion samples (Fig. 2A, B) could be due to aggregate formation, and/or additional ‘ubiquitination’ of the fusion protein by the plant machinery.
For comparison of expression systems, nuclear transgenic plants
were generated that express hST from two different sets of expression signals (Table 1 and Fig. 2A). The wrg4747 and wrg4776 conNATURE BIOTECHNOLOGY VOL 18 MARCH 2000
http://biotech.nature.com
C
D
structs express hST using the strong figwort mosaic virus promoter
or the cauliflower mosaic virus 35S promoter, respectively. The
wrg4747 construct uses a chloroplast transit peptide to post-translationally target hST to chloroplasts (FMV:CTP-hST), whereas the
wrg4776 construct targets the hST through the endoplasmic reticulum to the secretory pathway (35S:ER-hST). Transgenic lines for
both constructs were obtained through particle bombardment.
Expression of hST was quantitated by enzyme-linked immunosorbent assay (ELISA) to be less than 0.025% tsp. This amount of
expression is at least 300-fold lower than the pMON38794 lines,
proving a potentially significant advantage of the chloroplast expression system for the production of hST.
Analysis of processed hST protein. A purification scheme for
chloroplast-expressed hST was defined using the high-expressing
Nt-38794 line. Briefly, soluble protein was extracted from greenhouse-grown leaves, the pH of the protein extract was lowered, and
acid-insoluble material was removed by centrifugation. The supernatant was passed through an anion exchange column, and hST
protein was eluted in high salt followed by pH adjustment.
Purification steps were monitored by ELISA and western blot
analysis for processed hST. The purified hST protein fraction was
further characterized by reverse-phase HPLC (RP-HPLC), electroTable 1. Expression concentrations of chimeric hST genes in
transgenic plants.
Plasmid
Gene location Protein location Expression concentration (∼% tsp)
wrg4747
wrg4776
wrg4838
pMON38755
pMON38794
Nucleus
Nucleus
Chloroplast
Chloroplast
Chloroplast
Chloroplast
ER
Chloroplast
Chloroplast
Chloroplast
ND–0.025a
0.004, 0.008a
0.2a
1.0b
7.0b
aProtein
expression in several nuclear transgenic lines for wrg4747, or two
independent lines for wrg4838 and wrg4776, were quantitated by ELISA assay.
Western blot analysis gave similar values. ND, None detected.
bProtein expression from ubiquitin fusion lines was quantitated by western blot
analysis and includes both unprocessed and processed hST species.
335
© 2000 Nature America Inc. • http://biotech.nature.com
RESEARCH ARTICLES
© 2000 Nature America Inc. • http://biotech.nature.com
A
Figure 4. Plastid-expressed hST is biologically active. A positive
growth response of the Nb2 cell line is indicative of the presence of
somatotropin in the medium. An extract from wild-type control plants
(+, blue) shows no activity in the assay. Growth response is observed
in the remaining samples: (open diamond, green) recombinant E. coli
produced hST; (open triangle, brown) hST derived from pMON38794
plants; (square, pink) null plants spiked with E. coli hST; (filled
diamond, teal) hST derived from wrg4838.
spray ionization mass spectrometry (ES-MS), and N-terminal
amino acid sequencing.
The RP-HPLC profile of the purified hST fraction from the Nt38794 plastid sample (Fig. 3A) had the same mobility as native,
refolded E. coli-produced 22-kDa hST (data not shown). The peak
plastid protein fraction was collected and then subjected to ES-MS
analysis (Fig. 3B) that calculated the precise molecular mass of protein species in the fraction. This analysis identified two abundant
protein species, for which N-terminal amino acid sequence was subsequently determined.
The most abundant peak (21,977 Da; P-hST) represents hST
with an N-terminal proline residue. The P-hST protein species could
arise from ubiquitin processing that removes one additional amino
acid beyond the desired phenylalanine N terminus, or a secondary
protease activity that removes the terminal phenylalanine. However,
the lower abundance 22,124 Da (F-hST) species does represent processing of ubiquitin–hST to the desired phenylalanine N terminus.
The purified hST sample was treated with the reducing agent
dithiothreitol (DTT), which is predicted to change the retention time
in the RP-HPLC profile if disulfide bonds are present. The RP-HPLC
profile (Fig. 3A, gray line) shows this mobility shift after DTT reducing treatment, indicating that disulfide bonds do exist in the chloroplast hST sample. To confirm the presence of chloroplast hST with
disulfide linkages identical to native human somatotropin26, the sample was subjected to tryptic digestion followed by peptide mass mapping using matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS). Correct pairing of the disulfides of hST
would yield tryptic peptides T6 bound to T16 and T20 bound to T21.
Figure 3C shows the MALDI mass spectrum of the tryptic digest with
the correctly paired disulfide-bound peptides labeled. Interestingly,
no improperly paired cysteines are observed in this spectrum. After
treatment of the tryptic digest with the reducing agent tris(2-carboxyethyl)phosphine (TCEP), the disulfide-bound peptides disappear from the spectrum (Fig. 3D). Instead, the cysteine-containing
tryptic peptides T6, T16, and T20 are observed in their reduced form.
Tryptic peptide T21 is not detected by the mass spectrometer because
it does not contain any basic residues. The additional tryptic peptides
in both samples are consistent with native hST, indicating that the
proper recombinant protein was produced.
Formation of the two correct disulfide bonds and the complete
lack of mispaired cysteines in the chloroplast-derived hST was striking because there are no known, plastid-expressed proteins that have
336
B
Figure 5. Chloroplast transgenes are maternally inherited. (A)
Seeds derived from the reciprocal crosses of transplastomic line
Nt-4838 and wild-type (WT) tobacco were surface-sterilized and
plated onto MS salts medium containing 500 µg ml-1 spectinomycin
dihydrochloride. The progeny are labeled with the female recipient
( ) crossed to the pollen donor (
). Resistant seedlings are
uniformly green whereas sensitive seedlings are bleached white.
Photographs were taken ∼10 days after seedling germination. (B)
Western blot analysis is shown of hST accumulation in pooled
seedlings from the reciprocal crosses. Plant samples are loaded at
5 µg per lane of total soluble protein and labeled as in (A); 10 ng of
E. coli-produced hST (r-hST) is included as a control.
disulfide bonds. Further, eukaryotic secretory proteins are normally
routed through the endoplasmic reticulum where disulfide bond
formation occurs. Our results suggests that chloroplasts have the
machinery needed to fold complex eukaryotic secretory proteins in
the soluble chloroplast stroma compartment, presumably using the
chloroplast thioredoxin system15 or a recently discovered chloroplast
protein disulfide isomerase27. This is also distinct from E. coli, where
recombinant proteins tend to accumulate within inclusion bodies,
and then require solubilization and refolding.
Plastid-expressed hST is bioactive. Proper disulfide pairing indicates that the chloroplast hST protein should be biologically active.
To test this hypothesis in vitro, a rat lymphoma cell line, Nb2, which
proliferates in the presence of somatotropin, was used. Proliferation
of this cell line is proportional to the amount of somatotropin in the
culture medium, until saturation is reached. The ion exchange column eluate from transplastomic Nt-4838 and Nt-38794 plants or
identically treated wild-type plants was added to the Nb2 cell culture
medium. As control, E. coli produced, refolded hST was used (Fig.
4). The wild-type plant extract showed no activity in this assay, indicating that there is no endogenous plant compound capable of stimulating growth of the Nb2 cell line. In contrast, the Nt-4838 and Nt38794 extracts both stimulated proliferation of the cell line to an
NATURE BIOTECHNOLOGY VOL 18 MARCH 2000
http://biotech.nature.com
© 2000 Nature America Inc. • http://biotech.nature.com
© 2000 Nature America Inc. • http://biotech.nature.com
RESEARCH ARTICLES
equal extent as the positive controls: either wild-type plant extract
that had been spiked with purified E. coli hST or the pure hST alone.
The Nb2 cell results show that the chloroplast-derived somatotropin is biologically active. Previous studies of refolded recombinant somatotropin produced in E. coli showed equivalent pharmacokinetics of the protein with either an N-terminal methionine or
phenylalanine28. In this study, ubiquitin cleavage of the fusion protein
in Nt-38794 lines generated predominantly P-hST, suggesting that
this species is also bioactive. The hST from Nt-4838 extracts was also
characterized. Amino acid analysis (data not shown) indicated >95%
protein species with alanine at the N terminus. This result suggests
that a methionine aminopeptidase activity generated the alaninehST, which is also bioactive. A similar aminopeptidase activity exists
in E. coli29. This finding in plastids may be exploited in the future as
an alternative means to generate a nonmethionine N terminus.
Biological containment of plastid transgenes. Plastids are not
transmitted through pollen in most crop species, therefore, the
introduction of genes through chloroplast genetic engineering in
these plants is a potential route to gene containment. To demonstrate biological containment of the chloroplast transgenes, we
examined the seed progeny of reciprocal crosses between the
transplastomic lines and wild-type tobacco. Two assays were used
for this analysis: phenotypic resistance to the selected antibiotic
resistance marker and accumulation of recombinant hST by western
blot analysis. An example using the Nt-4838 line is shown in Figure
5. When crosses are performed using Nt-4838 as the female parent
and either wild-type tobacco pollen (Nt-4838 × WT ) or Nt-4838
self pollen (Nt-4838 × Nt-4838 , Fig. 5A), all of the progeny
seedlings tested showed a uniform resistant green phenotype. In
contrast, when Nt-4838 is used as the pollen donor in a cross to
wild-type tobacco (WT X Nt-4838 ), all seedlings are bleached
white similar to the self-progeny from wild-type tobacco controls
(WT X WT ), indicating antibiotic sensitivity.
The accumulation of hST in the seed progeny follows the same
pattern as antibiotic resistance, as expected. Seedlings from each
cross were pooled, protein extracted, and western blot analysis performed. This analysis (Fig. 5B) shows that hST protein accumulates
only in the seed progeny when the Nt-4838 line is used as the female
parent. Lack of hST accumulation or antibiotic resistance when the
transplastomic line is used as pollen donor indicates strict maternal
inheritance of the plastid transgenes and that no functional transgenes reside in the nuclear genome of the transplastomic plants.
We describe here a novel expression system used to obtain biologically active human therapeutic protein in plant plastids.
Chloroplasts have been used previously to obtain extremely high
concentrations of recombinant bacterial proteins30. Recently, a
plastid-encoded synthetic biopolymer gene modeled after a portion of mammalian elastin proteins expressed abundant RNA but a
very low recombinant protein amount in plastids compared to a
similar nuclear-expressed transgene31. The 7% hST protein accumulation concentration observed here proves for the first time the
feasibility of plastids for expression of human proteins.
Transplastomic lines expressing constitutively high concentrations
of hST showed phenotypically normal growth throughout development, and were fertile. In contrast, the insertion properties of
nuclear transgenic systems requires screening large numbers of
plants to identify optimal overexpressing lines with normal growth
phenotypes, with a chance of subsequent co-suppression 32. Even
so, recombinant trout somatotropin had previously been
expressed in transgenic tobacco and Arabidopsis plants, although
at only ∼0.1 % tsp concentrations33. Similarly, we could not produce significant concentrations of hST from two different nuclear
gene constructs in tobacco. Additional experiments are underway
to test the maximum concentrations of properly processed hST
possible using the chloroplast expression system.
NATURE BIOTECHNOLOGY VOL 18 MARCH 2000
http://biotech.nature.com
Experimental protocols
Construction of chloroplast transformation vectors. The human somatotropin and ubiquitin fusion genes were synthesized using E. coli preferred
codons and the yeast ubiquitin sequence, respectively (GenBank accession
AF205361, AF205360). The promoter/leader was fused to hST at an NcoI site
containing the translational start codon. The 3′-end was fused downstream
of the translational stop codon using a blunted HindIII site.
The plastid Prrn/G10L expression signal carries the Prrn promoter for the
plastid-encoded polymerase (positions 1–101; ref. 9) upstream of the rbs
region from the bacteriophage T7 G10L. The G10L rbs region34 was from
position +22 of the mRNA to the start codon; the C at position +23 was converted to G to destroy a XbaI site and the AT at positions +62, 63 were converted to CC to create an NcoI site at the translation start. The plastid PpsbA
and Trps16 expression signals24 were obtained as an EcoRI/NcoI and
XbaI/HindIII DNA fragments, respectively. The chimeric hST genes were
cloned into the KpnI and HindIII polylinker sites of pPRV112B (wrg4838
and pMON38755) or pPRV111B (pMON38794) (ref. 25).
Generation and analysis of transplastomic plants. Chloroplast transformation, selection of transplastomic lines and Southern blot analysis was carried out as described11,24. Several transplastomic events from bombardment
of each construct were identified by Southern blot analysis. Western blot
analysis was performed as described24, except that proteins were electrophoresed on 4–20% SDS–polyacrylamide gels (Sigma, St. Louis, MO).
Anti-hST (Biodesign, Kennebunk, ME) and anti-ubiquitin antisera
(Boehringer Mannheim; Indianapolis, IN) were used at 1:1000 and 1:500
dilutions, respectively. Protein was quantitated by the Pierce Coomassie Plus
kit (bovine IgG standard) or BioRad Coomassie kit (BSA standard). All
transplastomic lines were initially characterized for protein expression concentrations and were found to be nearly identical as expected for plastidencoded transgenes. Two lines of each were chosen for further analysis.
Purification and characterization of hST proteins. Chromatography was
performed using an anion exchange column equilibrated with 50 mM Tris,
10 mM NaCl, pH 8.6, and eluted with 500 mM NaCl, 5 mM Tris, pH 5.0, with
subsequent pH adjustment of eluted protein peaks to a range of 8.7–9.3 using
NaOH. Protein was then run on a C18 RP-HPLC column equilibrated with
1% trifluoroacetic acid, 50% acetonitrile, and eluted using a 2% linear acetonitrile gradient, followed by a linear 10% acetonitrile gradient.
Electrospray ionization mass spectrometry was performed using a
Micromass LCT time-of-flight mass spectrometer with samples prepared in
50% methanol, 2% acetic acid. N-terminal sequencing was performed using
standard Edman degradation. Nb2 cell proliferation assays were performed
as described35.
Tryptic digests were performed on RP-HPLC purified material solubilized in 0.2 M NH4HCO3, pH 8.0, by addition of a 1/20 (vol/vol) of 1 mg
ml-1 trypsin (Sigma T8642). After incubation for 14 h at 22°C, the sample
was dried in a Savant Speed-Vac (Savant, Hicksville, NY) and then resuspended in water. Reduction of disulfide bonds was performed by the addition of 0.5 µl of 20 mM TCEP at room temperature for 30 min to a 5 µl
aliquot of the resuspended tryptic digest. MALDI-MS used a Voyager DESTR time-of-flight mass spectrometer (PerSeptive Biosystems,
Framingham, MA). The samples (0.5 µl) were placed on a stainless-steel
sample plate with 1 µl of matrix solution and 0.25 µl of 3% trifluoroacetic
acid. The matrix solution was 12 mg ml-1 α-cyano-4-hydroxycinnamic
acid (99%, Aldrich Chemical Co., Milwaukee, WI) dissolved in 3:1:1:0.1
acetonitrile:methanol:water:trifluoroacetic acid. Spectra were collected in
the reflected mode. All masses observed corresponded to the monoisotopic species and were calculated using public PAWS software (www.proteometrics.com).
Analysis of maternal inheritance. Analysis of maternal inheritance by
phenotypic resistance to the selective antibiotic was performed as described9.
Aerial portions of eight-day-old green or bleached seedlings were collected
and used for extraction of total soluble protein. Note that bleached seedlings
of this age are not dead and can be recovered on medium lacking antibiotics.
Acknowledgments
We thank Robin Weinberg for a ubiquitin–hST fusion plasmid, Titik Dian for
assistance with protein purification, Dawn Dufield for mass spectrometry
analysis, Skooter Jennings for amino acid analysis, and Pal Maliga, Rutgers
University, for the pPRV plasmids. We also thank Michael Montague for a critical reading of the manuscript and Ganesh Kishore, Michael Montague, Julio
Baez, Doug Taylor, and Stephen Padgette for support throughout this work.
337
© 2000 Nature America Inc. • http://biotech.nature.com
© 2000 Nature America Inc. • http://biotech.nature.com
RESEARCH ARTICLES
1. Goddijn, O.J.M. & Pen, J. Plants as bioreactors. TIBTECH 13, 379–387 (1995).
2. Mason, H.S. & Arntzen, C.J. Transgenic plants as vaccine production systems.
TIBTECH 13, 388–392 (1995).
3. Cramer, C.L. et al. Bioproduction of human enzymes in transgenic tobacco. Ann.
NY Acad. Sci. 792, 63–72 (1996).
4. Ma, J.K.-C. & Hein, M.B. Antibody production and engineering in plants. Ann. NY
Acad. Sci. 792, 73–81 (1996).
5. Zeitlin, L. et al. A humanized monoclonal antibody produced in transgenic plants
for immunoprotection of the vagina against genital herpes. Nat. Biotechnol.
16,1361–1364 (1998).
6. McCormick, A.A. et al. Rapid production of specific vaccines for lymphoma by
expression of the tumor-derived single-chain Fv epitopes in tobacco plants.
Proc. Natl. Acad. Sci. USA 96, 703–708 (1999).
7. Tacket, C.O. et al. Immunogenicity in humans of a recombinant bacterial antigen
delivered in a transgenic potato. Nat. Med. 4, 607–609 (1998).
8. Svab, Z., Hajdukiewicz, P.T.J. & Maliga, P. Stable transformation of plastids in
higher plants. Proc. Natl. Acad. Sci. USA 87, 8526–8530 (1990).
9. Svab, Z. & Maliga, P. High-frequency plastid transformation in tobacco by selection for a chimeric aadA gene. Proc. Natl. Acad. Sci. USA 90, 913–917 (1993).
10. Sikdar, S.R., Serino, G., Chaudhuri, S. & Maliga, P. Plastid transformation in
Arabidopsis thaliana. Plant Cell Reports 18, 20–24 (1998).
11. Sidorov, V.A. et al. Stable chloroplast transformation in potato, use of green fluorescent protein as a plastid marker. Plant J. 19, 209–216 (1999).
12. Maliga, P. Towards plastid transformation in flowering plants. TIBTECH 11,
101–106 (1993).
13. Scott, S.E. & Wilkinson, M.J. Low probability of chloroplast movement from
oilseed rape (Brassica napus) into wild Brassica rapa. Nat. Biotechnol. 17,
390–392 (1999).
14. Daniell, H., Datta, R., Varma, S., Gray, S. & Lee, S.-B. Containment of herbicide
resistance through genetic engineering of the chloroplast genome. Nat.
Biotechnol. 16, 345–348 (1998).
15. Ruelland, E. & Miginiac-Maslow, M. Regulation of chloroplast enzyme activities
by thioredoxins: activation or relief from inhibition? Trends Plant Sci. 4, 136–141
(1999).
16. Tritos, N.A. & Mantzoros, C.S. Recombinant growth hormone: old and novel
uses. Am. J. Med. 105, 44–57 (1998).
17. Chawla, R.K., Parks, J.S. & Rudman, D. Structural variants of human growth hormone: biochemical, genetic and clinical aspects. Ann. Rev. Med. 34, 519–547
(1983).
18. Cronin, M.J. Pioneering recombinant growth hormone manufacturing: pounds
produced per mile of height. J. Pediatr. 131, S5–S7 (1997).
19. Goeddel, D.V. et al. Direct expression in Escherichia coli of a DNA sequence cod-
338
ing for human growth hormone. Nature 281, 544–548 (1979).
20. Gray, G.L., McKeown, K.A., Jones, A.J.S., Seeburg, P.H. & Heyneker, H.L.
Pseudomonas aeruginosa secretes and correctly processes human growth hormone. Bio/Technology 2, 161–165 (1984).
21. Becker, G.W. & Hsiung, H.M. Expression, secretion and folding of human growth
hormone in Escherichia coli. FEBS Lett. 204, 145–150 (1986).
22. Baker, R.T. Protein expression using ubiquitin fusion and cleavage. Curr. Opin.
Biotechnol. 7, 541–546 (1996).
23. Vierstra, R.D. Proteolysis in plants: mechanisms and functions. Plant Mol. Biol.
32, 275–302 (1996).
24. Staub, J.M. & Maliga, P. Translation of psbA mRNA is regulated by light via the 5′untranslated region in tobacco plastids. Plant J. 6, 547–553 (1994).
25. Zoubenko, O.V., Allison, L.A., Svab, Z. & Maliga, P. Efficient targeting of foreign
genes into the tobacco plastid genome. Nucleic Acids Res. 22, 3819–3824
(1994).
26. Lei, J., Chen, D.A. & Regnier, F.E. Rapid verification of disulfide linkages in
recombinant human growth hormone by tandem column tryptic mapping. J.
Chromatogr. 808, 121–131 (1998).
27. Kim, J. & Mayfield, S.P. Protein disulfide isomerase as a regulator of chloroplast
translational activation. Science 278, 1954–1957 (1997).
28. Moore, J.A. et al. Equivalent potency and pharmokinetics of recombinant human
growth hormones with or without an N-terminal methionine. Endocrinology 122,
2920–2926 (1988).
29. Meinnel, T., Mechulam, Y. & Blanquet, S. Methionine as translation start signal: a
review of the enzymes of the pathway in Escherichia coli. Biochimie 75,
1061–1075 (1993).
30. McBride, K.E. et al. Amplification of a chimeric Bacillus gene in chloroplasts
leads to an extraordinary level of an insecticidal protein in tobacco.
Bio/Technology 13, 362–365 (1995).
31. Guda, C., Lee, S.-B. & Daniell, H. Stable expression of a biodegradable proteinbased polymer in tobacco chloroplasts. Plant Cell Reports 19, 257–262 (2000).
32. Smyth, D.R. Gene silencing: cosuppression at a distance. Curr. Biol. 7, R793–795
(1997).
33. Bosch, D., Smal, J. & Krebbers, E. A trout growth hormone is expressed, correctly folded and partially glycosylated in the leaves but not the seeds of transgenic
plants. Transgenic Res. 3, 304–310 (1994).
34. Studier, F.W., Rosenberg, A.H., Dunn, J.J. & Dubendorff, J.W. Use of T7 RNA
Polymerase to direct expression of cloned genes. Methods Enzymol. 185, 60–89
(1990).
35. Dattani, M.T. et al. G120R, a human growth hormone antagonist, shows zincdependent agonist and antagonist activity on Nb2 cells. J. Biol. Chem. 270,
9222–9226 (1995).
NATURE BIOTECHNOLOGY VOL 18 MARCH 2000
http://biotech.nature.com