We report on a chemical platform to generate site-specific, homogeneous, antibody-antibody conjugates by targeting and bridging disulfide bonds. A bispecific antibody construct was produced in good yield through simple reduction and bridging of antibody fragment disulfide bonds, using a readily synthesized bis-dibromomaleimide cross-linker. Binding activity of antibodies was maintained, and in vitro binding of target antigens was observed. This technology is demonstrated through linking scFv and Fab antibody fragments, showing its potential for the construction of a diverse range of bispecifics.
We report on a chemical platform to generate site-specific, homogeneous, antibody-antibody conjugates by targeting and bridging disulfide bonds. A bispecific antibody construct was produced in good yield through simple reduction and bridging of antibody fragment disulfide bonds, using a readily synthesized bis-dibromomaleimide cross-linker. Binding activity of antibodies was maintained, and in vitro binding of target antigens was observed. This technology is demonstrated through linking scFv and Fab antibody fragments, showing its potential for the construction of a diverse range of bispecifics.
The limited ability of conventional
monoclonal antibody therapies to induce significant antitumor activity
has led to the development of bispecifics: antibodies that can simultaneously
bind two different antigens. In 2009, catumaxomab became the first
bispecific therapeutic to be clinically approved, combining EpCAM
targeting with T-cell recruitment for the treatment of malignant ascites.[1] Recombinant technologies have produced a diverse
range of bispecific antibodies, generating 45 formats in the past
two decades.[2] Despite this variety of topologies,
the approach is not suited to every protein combination. The fusion
of proteins via their N or C termini can result in a reduction or
loss of bioactivity and variable expression yields can be observed
due to complications in folding and processing.[3−5]An alternative
and potentially more versatile approach to generating
bispecific therapeutics is chemical conjugation. Until now, this has
been a less successful method of producing such conjugates. A fundamental
flaw in the chemical techniques employed in this area has been their
dependence on modifying lysine residues. There is an average of 100
lysine residues per antibody, and their distribution is uniform throughout
the surface topology of the Fab and Fc regions. As such, conjugation
techniques using lysine residues will randomly cross-link to virtually
all areas of the antibody molecule, resulting in a highly heterogeneous
mixture of products with unpredictable properties. One strategy to
overcome this issue is provided by site-directed mutagenesis, which
enables a single nucleophilic cysteine residue to be introduced at
a desired site in an antibody. However, this approach is limited,
as cysteine mutagenesis commonly leads to reduced expression yields
and undesirable properties such as susceptibility to dimerization,
mixed disulfide formation, or disulfide scrambling.[6−8]Recently
the site-specific introduction of chemical linkers has
been reported through unnatural amino acid insertion.[9,10] Using this approach, Schultz et al. described the synthesis of a
homogeneous anti-HER2/anti-CD3 bispecific in good yield.[10] This technology, while elegant, is not readily
transferred; each antibody to be conjugated must undergo prior investigation
to determine appropriate mutation sites, substitution for the unnatural
amino acid is often incomplete, and expression yields are generally
low due to the cellular toxicity of artificial amino acids at the
high concentrations necessary.[11,12] To avoid these difficulties,
an ideal site-directed conjugation technique would use residues natural
to the protein that are revealed for modification only under defined
conditions.Cysteine residues have a low natural abundance in
proteins, and
are often found tied up in disulfide bonds.[13] In the case of antibodies and antibody fragments there are no free
cysteine residues, and site-directed conjugation has been attempted
via interchain disulfide bond reduction and subsequent conjugation
of the free cysteines. However, conjugation of chemical entities to
the generated cysteine residues results in significant physical instability
of conjugates, particularly under times of stress.[14] Furthermore, targeting the cysteine residues responsible
for interchain disulfides using chemical cross-linking reagents results
in poor yields of bispecific due to the formation of homodimers and
intrachain coupling.[15] Therefore, the ideal
solution would be to use reagents that bridge disulfide bonds, maintaining
this key stabilizing feature, and preventing the opportunity for product
heterogeneity.[16−23] Herein we propose a conjugation strategy using simple chemical reagents
that selectively bridge disulfide bonds. Through rapid reduction and
bridging of disulfides, homogeneous bispecific antibodies could be
readily generated with no effect on stability or activity. To demonstrate
the versatility of this chemical conjugation approach to varying antibody
fragment formats, we aimed to generate a homogeneous scFv-Fab conjugate
(Scheme 1).
Scheme 1
Strategy for the Production of a Homogeneous
Bispecific through Disulfide
Bridging of Two Antibody Fragments
In prior work we have demonstrated that next generation
maleimides
can be used for the extremely efficient rebridging of disulfide bonds
in Fab and disulfide-stabilized scFv antibody fragments, to produce
fully active, homogeneous protein conjugates in near-quantitative
yields.[20,21] Antibody fragments including Fabs and scFvs
are commonly used in a range of bispecific topologies. Thus, we envisaged
that next generation maleimide based cross-linking reagents could
be used to produce homogeneous bispecific constructs. To this end
homobifunctional linkers were designed, incorporating two dibromomaleimide
moieties linked by a PEG chain, conferring some flexibility to the
molecule (Scheme 2). Using commercially available
dibromomaleimide and diaminePEG, two linkers of distinct length were
readily synthesized. The reaction proceeds under mild conditions in
good yield, requiring only a single purification step.[24]
Scheme 2
Synthetic Route to Linkers
(a) ClCO2Me, NMM,
THF, 97%; (b) For BDBM(PEG)2: NH2CH2CH2(OCH2CH2)2NH2, DCM, 70%. For BDBM(PEG)19: NH2CH2CH2(OCH2CH2)19NH2, DCM, 65%.
Synthetic Route to Linkers
(a) ClCO2Me, NMM,
THF, 97%; (b) For BDBM(PEG)2: NH2CH2CH2(OCH2CH2)2NH2, DCM, 70%. For BDBM(PEG)19: NH2CH2CH2(OCH2CH2)19NH2, DCM, 65%.To examine the feasibility
of this approach to producing a homogeneous
protein–protein conjugate, we decided to first work with the
anti-CEAdisulfide-stabilized scFv fragment. Carcinoembryonic antigen
(CEA) is a cell surface glycoprotein overexpressed in a wide range
of cancers, particularly in colorectal carcinoma.[25] A scFv is the smallest antibody fragment which retains
full binding activity, and is a common component of many bispecific
antibody formats, e.g., BiTEs (Bispecific T-cell Engager) and DARTs
(Dual-Affinity Re-Targetting). Initially, the ability of the linkers
to generate scFv homodimer was investigated (Scheme 3).
Scheme 3
Generation of scFv Homodimer
The scFv was reduced (DTT, 20 equiv with respect to antibody)
and,
following buffer exchange, incubated with 0.5 equiv of linker BDBM(PEG) or BDBM(PEG). The conjugation reaction was monitored
by SDS-PAGE, and a distinct band at ∼50 kDa was observed on
incubation of BDBM(PEG) with
reduced scFv after only 1 h at room temperature (Supporting Information Figure 1). This corresponds to the
molecular weight of scFv dimer. In contrast, the reduced scFv incubated
with linker BDBM(PEG), although
rebridged, reveals only a faint band at ∼50 kDa, suggesting
poor dimer formation.The longer length of linker BDBM(PEG) (19 PEG units, ∼1 kDa), in combination
with its flexibility,
is likely to significantly reduce steric hindrance which creates difficulty
when attempting to link two large proteins. This is observed in the
far greater conversion to homodimer obtained compared to linker BDBM(PEG) (Supporting
Information Figure 1). It was therefore decided that conditions
for homogeneous protein–protein conjugation should be optimized
using linker BDBM(PEG). As
a model system, generation of homogeneous scFv dimer was further pursued
in order to determine ideal conditions for maintaining antibody fragments
in their reduced form while promoting bridging.Following reduction,
the scFv was eluted over a Sephadex column
to remove DTT and to buffer exchange the protein into conditions suitable
for maintaining the reduced antibody (20 mM phosphate buffer, 5 mM
EDTA, pH 7.4). The antibody fragment was then concentrated to no greater
than 1 mg/mL, since higher concentrations promoted disulfide reoxidation.
A series of optimization experiments revealed incubation with 0.42
equiv of bis-dibromomaleimide cross-linker BDBM(PEG) for 1 h at room temperature,
or 4 °C overnight, yielded the highest levels of homogeneous
dimer (Supporting Information Figure 2).
In fact, after 1 h approximately 80% of starting material was consumed
(Figure 1[i], lane 3, measured using ImageJ).
After purification of the homodimer by size exclusion chromatography,
an excellent 64% yield of pure, homogeneous scFv–scFv conjugate
was obtained from 1 mg of scFv (Figure 1[iii],
75% yield with respect to the limiting reagent BDBM(PEG)). This yield represents a significant
improvement on previous reports of direct chemical cross-linking of
antibodies using natural amino acids, which have achieved yields in
the 10–40% range.[26−29] Notably the conjugate is homogeneous due to the site-selectivity
of the methodology.
Figure 1
Application of optimized conditions to dimerization of
scFv: [i]
SDS-PAGE analysis (1) scFv; (2) scFv + 20 equiv DTT to afford reduced
scFv (note - scFv with intact disulfide runs faster through gel due
to more compact structure); (3) reduced scFv + 0.42 equiv BDBM(PEG), 1 h rt to afford scFv homodimer (∼50
kDa) and disulfide bridged scFv monomer (∼25 kDa). [ii] Size
exclusion chromatography purification of pure scFv homodimer (peak
a, ∼50 kDa) from monomer starting materials (peak b) and unreacted
linker (peak c) (Superdex 75, GE Healthcare).
Application of optimized conditions to dimerization of
scFv: [i]
SDS-PAGE analysis (1) scFv; (2) scFv + 20 equiv DTT to afford reduced
scFv (note - scFv with intact disulfide runs faster through gel due
to more compact structure); (3) reduced scFv + 0.42 equiv BDBM(PEG), 1 h rt to afford scFv homodimer (∼50
kDa) and disulfide bridged scFv monomer (∼25 kDa). [ii] Size
exclusion chromatography purification of pure scFv homodimer (peak
a, ∼50 kDa) from monomer starting materials (peak b) and unreacted
linker (peak c) (Superdex 75, GE Healthcare).Given the success of our technology in generating homogeneous
scFv
homodimer in high yield, we targeted the generation of a bispecific
antibody conjugate. The monoclonal IgG1 antibody Traztuzumab (Herceptin)
targets the HER2/neu receptor and has successfully been used to treat
HER2+ breast cancerpatients.[30] The Fab
fragment of this clinically relevant antibody can be readily obtained
by enzymatic digest, and its incorporation into bispecific formats
has therapeutic potential.[10] A Fab fragment
contains one variable and one constant region of each of the heavy
and light antibody chains, and so is twice the size of a scFv fragment.
The heavy and light chains are covalently linked via a single interchain
disulfide bond. Thus, we envisaged linking the disulfide-stabilized
scFv to the Fab fragment using BDBM(PEG), generating a conjugate that could simultaneously bind two
different antigens (CEA and HER2) as a proof of concept (Scheme 4).
Scheme 4
Generation of anti-CEA/anti-HER2 scFv-Fab
Bispecific
To efficiently link
these two different antibody fragments, we
decided to first functionalize the scFv with BDBM(PEG). This was cleanly achieved by reducing
the scFv as previously described, purifying by Sephadex column and
subsequently adding 30 equiv of linker to scFv. This reaction was
complete in less than 5 min at room temperature, yielding bromomaleimide
functionalized scFv in quantitative yield (Figure 2[i], lane 3). Excess linker was removed by buffer exchange.
In tandem with this, the Fab fragment was reduced (TCEP, 10 equiv
relative to antibody) and purified from reducing agent into pH 7.4
phosphate buffer containing 5 mM EDTA. The concentrations of bridged
scFv and reduced Fab were adjusted to 37.2 μM (1 mg/mL for the
scFv), and the antibodies mixed in a 2:1 ratio by volume (2 equiv
bridged scFv to 1 equiv reduced Fab). This slight excess of functionalized
scFv was found to be sufficient to promote bridging of the reduced
Fab. The reaction was monitored by SDS-PAGE, and after 1 h at room
temperature a strong band at ∼80 kDa could be observed, corresponding
to the scFv-Fab conjugate. After purification by size exclusion chromatography,
a pleasing 52% yield of homogeneous scFv-Fab bispecific was achieved.
Figure 2
Production
of homogeneous scFv-Fab: [i] SDS-PAGE analysis (1) scFv;
(2) scFv + 20 equiv DTT to afford reduced scFv (note - scFv with intact
disulfide runs faster through gel due to more compact structure);
(3) Reduced scFv + 30 equiv BDBM(PEG) to afford disulfide bridged scFv monomer (note - dibromomaleimide
functionalized scFv runs faster than reduced through gel as compact
structure is recovered through disulfide bridging); (4) Fab; (5) Fab
+10 equiv TCEP to afford component heavy and light Fab chains; (6)
Bridged scFv + Reduced Fab (2:1), 1 h rt to afford scFv-Fab conjugate
(∼80 kDa), reoxidized Fab (∼50 kDa) and scFv starting
material (∼25 kDa) . [ii] Size exclusion chromatography purification
of pure scFv-Fab heterodimer (peak a) from starting materials (peak
b and c) (Superdex 75, GE Healthcare).
Production
of homogeneous scFv-Fab: [i] SDS-PAGE analysis (1) scFv;
(2) scFv + 20 equiv DTT to afford reduced scFv (note - scFv with intact
disulfide runs faster through gel due to more compact structure);
(3) Reduced scFv + 30 equiv BDBM(PEG) to afford disulfide bridged scFv monomer (note - dibromomaleimide
functionalized scFv runs faster than reduced through gel as compact
structure is recovered through disulfide bridging); (4) Fab; (5) Fab
+10 equiv TCEP to afford component heavy and light Fab chains; (6)
Bridged scFv + Reduced Fab (2:1), 1 h rt to afford scFv-Fab conjugate
(∼80 kDa), reoxidized Fab (∼50 kDa) and scFv starting
material (∼25 kDa) . [ii] Size exclusion chromatography purification
of pure scFv-Fab heterodimer (peak a) from starting materials (peak
b and c) (Superdex 75, GE Healthcare).Next, we used enzyme-linked immunosorbent assay (ELISA) to
assess
the immuno-reactivity of the purified conjugates, in comparison to
the unmodified scFv and Fab fragments that bound either CEA or HER2,
respectively (Figure 3[i] and [ii]). Against
CEA, the activity of the scFv homodimer was not only maintained, but
remained high at low dilutions, suggesting dimerization has successfully
increased the avidity of the antibody. Most importantly, the scFv-Fab
conjugate showed comparable antigen binding activity against both
CEA and HER2. Given this success, we wanted to demonstrate that our
heterodimeric conjugate could simultaneously bind its two target antigens.
To achieve this we developed a sandwich ELISA. A 96-well plate was
coated with CEA and the sample to be tested, e.g., scFv-Fab conjugate,
was applied. Subsequent incubation with HER2-Biotin and Extravadin-Peroxidase
would thus only lead to a signal if the sample successfully bound
both CEA and HER2 antigens. Pleasingly, simultaneous binding activity
was confirmed (Figure 3[iii]). Hence we have
successfully produced a homogeneous antibody conjugate with bispecific
ability using our disulfide-bridging technology.
Figure 3
ELISA analysis of conjugates
and unmodified antibody fragments:
[i] ELISA against full length CEA. [ii] ELISA against HER2. [iii]
Sandwich ELISA using full length CEA coated plates and HER2 conjugated
to biotin, followed by Extravadin-Peroxidase.
ELISA analysis of conjugates
and unmodified antibody fragments:
[i] ELISA against full length CEA. [ii] ELISA against HER2. [iii]
Sandwich ELISA using full length CEA coated plates and HER2 conjugated
to biotin, followed by Extravadin-Peroxidase.In a final investigation, we assessed the binding of our
antibody
conjugates to a CEA-positive cell line, A375CEA, and a HER2-positive
cell line, BT-474 (controls shown in Supporting
Information Figure 3). The monomer controls demonstrate that
anti-CEAscFv binds only to A375CEA cells and anti-HER2Fab only to
BT474 cells (Figure 4, monomer). Application
of our scFv homodimer conjugate reveals that CEA binding activity
and selectivity is maintained after conjugation, with a shift in fluorescence
being observed only on the A375CEA cell line (Figure 4, homodimer). Following this success, we tested our bispecific
construct (Figure 4, heterodimer). In this
case, scFv and Fab were detected in both the CEA-positive cell line
A375CEA and the HER2-positive BT474. This demonstrates that the antibody
fragments have maintained their distinct selectivity and binding activity
in vitro, while being successfully chemically linked.
Figure 4
Flow cytometry based
binding assay of unmodified scFv and Fab (monomer),
scFv dimer (homodimer), and scFv-Fab conjugate (heterodimer) to a
CEA-positive cell line (A375CEA) and HER2-positive cell line (BT474).
Flow cytometry based
binding assay of unmodified scFv and Fab (monomer),
scFv dimer (homodimer), and scFv-Fab conjugate (heterodimer) to a
CEA-positive cell line (A375CEA) and HER2-positive cell line (BT474).
Conclusion
To date, the generation
of bispecifics for the clinic by chemical
cross-linking has been unsuccessful, due to low yields and product
heterogeneity. Here we have presented the rapid production of two
homogeneous conjugates in high yield: an anti-CEAscFv homodimer and
an anti-CEA/anti-HER2scFv-Fab heterodimer. Our chemical conjugation
approach uses bis-dibromomaleimidePEG linker readily
synthesized over two steps from the commercially available dibromomaleimide.
Through targeting and bridging the disulfide bond of antibody fragments,
conjugates were produced which exhibit retention of activity by ELISA
and cell binding assays. This platform has the potential to enable
the facile generation of bispecifics from a range of antibody fragment
formats, and could be readily translated to other protein conjugates
of choice, exploiting the versatility of the chemical conjugation
approach.
Experimental Procedures
-PEG2-bis-3,4-dibromomaleimide
(BDBM(PEG)2)[24]
2,2′-(Ethylenedioxy)bis(ethylamine)
(0.2 mL, 1.38 mmol) was added to a stirred solution of N-(methoxycarbonyl)-3,4-dibromomaleimide[24] (1.08 g, 3.45 mmol) in DCM (8 mL). After 20 min EtOAc (80 mL) was
added and the organic layer extracted with saturated NH4Cl solution (2 × 40 mL) and H2O (3 × 40 mL).
The organic layer was dried with MgSO4 and concentrated
in vacuo. Purification by column chromatography (gradient elution
from 1:9 to 3:7 EtOAc/Petrol) yielded BDBM(PEG) as a pale yellow solid (600 mg, 70%): mp 116–117
°C; 1H NMR (500 MHz, CDCl3) δ 3.82
(4H, t, J = 5.6 Hz), 3.65 (4H, t, J = 5.6 Hz), 3.58 (4H, s); 13C NMR (125 MHz, CDCl3) δ 164.2 (C), 129.8 (C), 70.4 (CH2), 68.0 (CH2), 39.3 (CH2); IR (solid) 3489, 2913, 1784, 1720,
1597; LRMS (ES+) 651 ([81,81,81,81M+Na], 20), 649 ([81,81,81,79M+Na], 70), 647 ([81,81,79,79M+Na], 100),
645 ([81,79,79,79M+Na], 70), 643 ([79,79,79,79M+Na], 20); HRMS (ES+) calculated for C14H12N2O6Na79Br4 642.7327,
observed 642.7355.
-PEG20-bis-3,4-dibromomaleimide
(BDBM(PEG)19)
O,O′-Bis(2-aminoethyl)octadecaethylene
glycol (50 mg, 0.06 mmol) was added to a stirred solution of N-(methoxycarbonyl)-3,4-dibromomaleimide[24] (44 mg, 0.14 mmol) in DCM (2 mL). The reaction mixture
was left at room temperature overnight, and then the solvent removed
in vacuo. Purification by column chromatography (gradient elution
from DCM to 20:1 DCM/MeOH) yielded BDBM(PEG)19 as a pale
yellow oil (53 mg, 65%). 1H NMR (600 MHz, CDCl3) δ 3.81 (4H, t, J = 6.0), 3.67–3.59
(76H, m); 13C NMR (150 MHz, CDCl3) δ 164.0
(C), 129.6 (C), 70.7 (CH2), 70.2 (CH2), 67.7
(CH2), 39.0 (CH2); IR (oil, cm–1) 2865, 1722; LRMS (EI) 1394 ([81,81,81,81M+NH4], 20), 1392 ([81,81,81,79M+NH4], 70), 1390
([81,81,79,79M+NH4], 100), 1388 ([81,79,79,79M+NH4], 70), 1386 ([79,79,79,79M+NH4], 20), 1346 (40), 1300 (10); HRMS (EI) calculated for C48H8081,81,79,79Br4N2O3NH4 1390.2192, observed 1390.2181.
Disulfide Bridging
of scFv with Linker BDBM(PEG)2 to Generate scFv–scFv
Homodimer
Anti-CEA ds-scFv
in PBS (1 mg/mL, 37.2 μM) was reduced with DTT (20 equiv relative
to scFv, 37.2 mM stock in PBS) for 1 h at room temperature. DTT was
then removed using a desalting column (PD-10, GE Healthcare) and the
reduced scFv buffer exchanged into conjugation buffer (20 mM phosphate
buffer, 5 mM EDTA, pH 7.4). The scFv was concentrated to approximately
1 mg/mL and linker BDBM(PEG) added (0.42 equiv relative to scFv, 3.72 mM stock in DMF). After
1 h at room temperature or overnight at 4 °C, the reaction was
purified by size exclusion (500 mL Superdex 75 column, GE Healthcare).
Disulfide Bridging of scFv with Linker BDBM(PEG)19 to
Generate scFv–scFv Homodimer
Anti-CEA ds-scFv
in PBS (1 mg/mL, 37.2 μM) was reduced with DTT (20 equiv relative
to scFv, 37.2 mM stock in PBS) for 1 h at room temperature. DTT was
then removed using a desalting column (PD-10, GE Healthcare) and the
reduced scFv buffer exchanged into conjugation buffer (20 mM phosphate
buffer, 5 mM EDTA, pH 7.4). The scFv was concentrated to approximately
1 mg/mL and linker BDBM(PEG) added (0.42 equiv relative to scFv, 3.72 mM stock in water). After
1 h at room temperature or overnight at 4 °C, the reaction was
purified by size exclusion (500 mL Superdex 75 column, GE Healthcare).
Disulfide Bridging of scFv and Fab to Generate scFv-Fab Heterodimer
Anti-CEA ds-scFv in PBS (1 mg/mL, 37.2 μM) was reduced with
DTT (20 equiv relative to scFv, 37.2 mM in PBS) for 1 h at room temperature.
DTT was then removed using a desalting column (PD-10, GE Healthcare)
and the reduced scFv buffer exchanged into conjugation buffer (20
mM phosphate buffer, 5 mM EDTA, pH 7.4). The scFv was concentrated
to approximately 1 mg/mL and linker BDBM(PEG) added (30 equiv relative to scFv, 37.2 mM in water).
After 10 min excess linker was removed by buffer exchange (repeat
3 times, Amicon Ultra-4 Centrifugal Filter Units, 10 kDa cutoff) into
conjugation buffer, and the bridged scFv concentrated to approximately
1 mg/mL. Meanwhile, HerceptinFab in Borate buffer (1 mg/mL, 25 mM
sodium borate, 25 mM NaCl, 1 mM EDTA, pH 8.0) was reduced with TCEP
(10 equiv relative to Fab, 37.2 mM in borate buffer) for 1 h at room
temperature. TCEP was then removed using a desalting column (PD-10,
GE Healthcare) and the reduced Fab buffer exchanged into conjugation
buffer (20 mM phosphate buffer, 5 mM EDTA, pH 7.4). The reduced Fab
concentration was adjusted to 37.2 μM, and bridged scFv (37.2
μM) mixed with the reduced Fab in a 2:1 ratio by volume, respectively.
After 1 h at room temperature or overnight at 4 °C, the reaction
was purified by size exclusion (500 mL Superdex 75 column, GE Healthcare).
Authors: Chan Hyuk Kim; Jun Y Axup; Anna Dubrovska; Stephanie A Kazane; Benjamin A Hutchins; Erik D Wold; Vaughn V Smider; Peter G Schultz Journal: J Am Chem Soc Date: 2012-06-06 Impact factor: 15.419
Authors: Cristina Marculescu; Hanno Kossen; Rachel E Morgan; Patrick Mayer; Sally A Fletcher; Berend Tolner; Kerry A Chester; Lyn H Jones; James R Baker Journal: Chem Commun (Camb) Date: 2014-07-11 Impact factor: 6.222
Authors: Moorthy S S Palanki; Abhijit Bhat; Ben Bolanos; Florence Brunel; Joselyn Del Rosario; Danielle Dettling; Mark Horn; Rodney Lappe; Ryan Preston; Annette Sievers; Nebojsa Stankovic; Gary Woodnut; Gang Chen Journal: Bioorg Med Chem Lett Date: 2012-12-05 Impact factor: 2.823
Authors: Felix F Schumacher; Muriel Nobles; Chris P Ryan; Mark E B Smith; Andrew Tinker; Stephen Caddick; James R Baker Journal: Bioconjug Chem Date: 2011-01-27 Impact factor: 4.774
Authors: Mark E B Smith; Felix F Schumacher; Chris P Ryan; Lauren M Tedaldi; Danai Papaioannou; Gabriel Waksman; Stephen Caddick; James R Baker Journal: J Am Chem Soc Date: 2010-02-17 Impact factor: 15.419
Authors: Felix F Schumacher; Vishal A Sanchania; Berend Tolner; Zoë V F Wright; Chris P Ryan; Mark E B Smith; John M Ward; Stephen Caddick; Christopher W M Kay; Gabriel Aeppli; Kerry A Chester; James R Baker Journal: Sci Rep Date: 2013 Impact factor: 4.379
Authors: Martin C Boonstra; Berend Tolner; Boudewijn E Schaafsma; Leonora S F Boogerd; Hendrica A J M Prevoo; Guarav Bhavsar; Peter J K Kuppen; Cornelis F M Sier; Bert A Bonsing; John V Frangioni; Cornelis J H van de Velde; Kerry A Chester; Alexander L Vahrmeijer Journal: Int J Cancer Date: 2015-06-22 Impact factor: 7.396
Authors: Mark B Richardson; Kristin N Gabriel; Joseph A Garcia; Shareen N Ashby; Rebekah P Dyer; Joshua K Kim; Calvin J Lau; John Hong; Ryan J Le Tourneau; Sanjana Sen; David L Narel; Benjamin B Katz; Joseph W Ziller; Sudipta Majumdar; Philip G Collins; Gregory A Weiss Journal: Bioconjug Chem Date: 2020-04-30 Impact factor: 4.774
Authors: Ross J Taylor; Mauricio Aguilar Rangel; Michael B Geeson; Pietro Sormanni; Michele Vendruscolo; Gonçalo J L Bernardes Journal: J Am Chem Soc Date: 2022-07-14 Impact factor: 16.383
Authors: Calise Bahou; Daniel A Richards; Antoine Maruani; Elizabeth A Love; Faiza Javaid; Stephen Caddick; James R Baker; Vijay Chudasama Journal: Org Biomol Chem Date: 2018-02-21 Impact factor: 3.876