Anthrax Protective Antigen Retargeted with Single#Chain
Variable Fragments Delivers Enzymes to Pancreatic Cancer Cells
The MIT Faculty has made this article openly available.
Please share
how this access benefits you. Your story matters.
Citation
Loftis, Alexander R. et al. "Anthrax Protective Antigen Retargeted
with Single#Chain Variable Fragments Delivers Enzymes to
Pancreatic Cancer Cells." ChemBioChem 21, 19 (June 2020):
2772-2776
As Published
http://dx.doi.org/10.1002/cbic.202000201
Publisher
Wiley
Version
Author's final manuscript
Citable link
https://hdl.handle.net/1721.1/128117
Terms of Use
Creative Commons Attribution-Noncommercial-Share Alike
1
Anthrax protective antigen re-targeted with single-chain variable
fragments delivers enzymes to pancreatic cancer cells
Alexander R. Loftis
+,
[a]Michael S. Santos
+,
[b]Nicholas L. Truex,
[a]Marco Biancucci,
[c]Karla J. F.
Satchell,
[c]and Bradley L. Pentelute*
[a][a] A. R. Loftis+, Dr. N. L. Truex, Prof. Dr. B. L. Pentelute
Department of Chemistry, Massachusetts Institute of Technology 77 Massachusetts Avenue, Cambridge, MA 02139
E-mail: blp@mit.edu [b] Dr. M. S. Santos+
Department of Chemical Engineering Massachusetts Institute of Technology
77 Massachusetts Avenue, Cambridge, MA 02139 [c] Dr. M. Biancucci, Prof. Dr. K. J. F. Satchell
Department of Microbiology-Immunology
Feinberg School of Medicine, Northwestern University 420 E Superior St., Chicago, IL 60611
[+]These authors contributed equally to this work.
Supporting information for this article is given via a link at the end of the document.
Abstract: The non-toxic anthrax protective antigen/lethal factor N-terminal domain (PA/LFN) complex is an effective platform for translocating proteins into the cytosol of cells. Mutant PA (mPA) was recently fused to epidermal growth factor (EGF) to achieve re-targeted delivery of LFN to cells bearing EGF receptors (EGFR), but the requirement of a known cognate ligand limits the applicability of this approach. Here we render practical protective antigen re-targeting to a variety of receptors with mPA-single-chain variable fragment (scFv) fusion constructs. Our design enables targeting of two pancreatic cancer-relevant receptors, EGFR and carcinoembryonic antigen. We demonstrate that fusion to scFvs does not disturb the basic functions of mPA. Moreover, mPA-scFv fusions enable cell-specific delivery of diphtheria toxin catalytic domain and Ras/Rap1-specific endopeptidase to pancreatic cancer cells. Importantly, mPA-scFv fusion-based treatments display potent cell-specific toxicity in vitro, opening fundamentally new routes toward engineered immunotoxins and providing a potential solution to the challenge of targeted protein delivery to the cytosol of cancer cells.
Pancreatic cancer is expected to become the second most lethal cancer by number of cases in the next 10 years.[1] The disease is aggressive, and in 80% of cases cannot be diagnosed in time for surgical resection.[2] Despite advances in diagnosis and chemotherapy, prognosis remains poor, altogether highlighting a need for new therapeutic modalities.
Targeted, efficient intracellular delivery of cytotoxic proteins is a longstanding goal in cancer research. Naturally-derived enzymes, such as the diphtheria toxin A chain (DTA), possess uniquely potent toxicity - one molecule of DTA is sufficient to kill the cell.[3] Others, such as MARTX toxin-derived Ras/Rap1-specific endopeptidase (RRSP), cleave key oncoproteins which are otherwise undruggable (i.e., KRas).[4–9] Though enzyme-based cancer therapy is potentially powerful, intracellular protein delivery remains a challenge due to the plasma membrane, which acts as a barrier to the cytosol where the majority of tumor-sustaining processes take place.[10]
Scheme 1. Delivery of LFN-appended cargoes mediated by mPA-scFv fusion
proteins that bind receptors of interest on the cell surface. A) Schematic representation of monomeric anthrax toxin protective antigen mutant (N682A, D683A; mPA) single-chain variable fragment (scFv) C-terminal fusion (mPA-scFv), linked via a SPGHKTQP linker (composite of PDB 1TZO and 1P4I). The scFvs selected target EGFR and carcinoembryonic antigen (CEA). B) Mechanism of receptor-targeted cell entry of LFN-DTA, which generally applies
to different receptors and LFN-cargo constructs.
mPA αEGFR COOH
H3N+
H+
LFN DTA (cargo)
Protein synthesis inhibition, cell death 5. Internalization 2. Cleavage 3. Oligomerization 4. Complex formation 6.Translocation 1. Binding
A)
B)
mPA-scFv fusion monomer N682A, D683AX
X X XX X X X X XXmPA αCEA COOH
H3N+
X
X
X X
2
Anthrax is an emerging protein delivery platform.[11–14] Derived from Bacillus anthracis, anthrax lethal toxin has a modular design consisting of an 83 kDa pore-forming protein protective antigen (PA83) which binds to ubiquitously expressed anthrax receptors TEM8 and CMG2.[15,16] After binding, PA83 becomes proteolytically activated, loses a 20 kDa PA20 fragment, and oligomerizes. This PA63 pre-pore binds lethal factor (LF) and internalizes into an endosome.[17–19] At an acidic pH, oligomeric PA63 forms a sodium dodecyl sulfate (SDS)-resistant transmembrane β-barrel pore, then translocates LF from the endosomal lumen to the cytosol.[20–22] The non-toxic N-terminal domain of LF (LFN) alone was found to be sufficient for pore-binding and initiation of translocation, leading to the development of a protein drug delivery platform based on PA/LFN.[11,23] The PA/LFN platform is amenable to cytosolic delivery of a variety of cargoes, which can be easily attached to the C-terminus of LFN.Therapeutic application of anthrax (as is the case for many protein toxins) was initially limited by its non-specificity, as PA natively targets the ubiquitously expressed receptors TEM8 and CMG2.[24,25] However, recent work has established that attachment of targeting ligands to a mutant mPA (N682A D683A) that does not bind TEM8 or CMG2 can re-target PA/LFN-cargo delivery via non-native receptors.[26–28] For example, fusion of mPA to epidermal growth factor can direct LFN-cargo delivery to cells bearing epidermal growth factor receptors (EGFR).[26]
We sought to develop a versatile mPA-targeting ligand strategy against a variety of pancreatic cancer-relevant receptors. We reasoned that single-chain variable fragment antibody fragments (scFvs) were ideal targeting ligands because they provide high affinity binding to a number of antigens and could conceivably be expressed as mPA-scFv fusion proteins.[29] Here we describe our investigations into whether mPA can be re-targeted with scFv fusions to efficiently deliver toxic proteins to pancreatic cancer cells (Scheme 1). For these studies, we selected two specific receptors suitable for re-targeting mPA/LFN. The first, EGFR, is a well-characterized receptor involved in cell growth and proliferation, and is overexpressed on the surface of many types of pancreatic cancer cells.[30] The second, carcinoembryonic antigen (CEA), though widely used as a serum biomarker, exhibits positive expression rates on the surface of pancreatic cancer cells and minimally elsewhere.[31–34] We therefore prepared mPA fusion constructs to an EGFR-targeting scFv (E1v3) (developed using yeast display, see Supplemental Methods, Figure S1) and a previously reported CEA-targeting scFv (sm3e) (mPA-E1v3, mPA-sm3e; Figure S2). [35]
mPA-scFv fusions retain the ability of wild-type PA to become proteolytically activated, oligomerize, and form SDS-resistant pores in response to acid. These basic functions of wild-type PA and mPA-scFv fusions were characterized by SDS-PAGE. PA83 treated with trypsin and exposed to basic pH ran as a ~60 kDa species, corresponding to PA63 (i.e., loss of PA20) (Figure 1 lanes 2-3). However, at acidic pH this trypsinized PA ran as a high molecular weight species, indicating formation of the SDS-resistant oligomeric pore (lanes 4-5). Similarly, trypsin activation of the ~110 kDa mPA-scFv fusions yielded ~90 kDa proteins at basic pH (lanes 7-8, 12-13) and high molecular weight species at acidic pH, indicating formation of SDS-resistant scFv pores (lanes 9-10, 14-15). Of note, trypsin treatment of mPA-scFvs also formed PA-sized products, indicating proteolytic susceptibility of the mPA-scFv fusion linker (lanes 7-10, 12-15).
Figure 1. Characterization of mPA-scFv fusion indicates retention of SDS-resistant pore formation. SDS-PAGE analysis of PA and mPA-scFvs following trypsin treatment and anion exchange chromatography purification (denoted by *). Exposure of PA* and mPA-scFv* constructs to acidic pH led to formation of SDS-resistant oligomer, indicating successful pore formation.
mPA-scFv fusions efficiently translocate a toxin-derived reporter to the cellular cytosol. The enzyme DTA (catalytic A chain of diphtheria toxin) inhibits protein synthesis via ADP-ribosylation of eukaryotic elongation factor 2.[36] As a truncate of full-length diphtheria toxin, DTA cannot reach the cytosol alone. Thus, in combination with PA, the LFN-DTA construct can report on cytosolic delivery via protein synthesis inhibition. In tissue culture, protein synthesis can be accurately measured by measuring the ribosomal incorporation of 3H-leucine spiked into the surrounding cell media. When AsPC-1 pancreatic cancer cells were treated with PA/LFN-DTA, protein synthesis was inhibited at sub-picomolar concentrations of LFN-DTA (Figure 2A). mPA-E1v3/LFN-DTA and mPA-sm3e/LFN-DTA treatments also demonstrated this effect at picomolar concentrations of LFN-DTA. However, mPA-scFv fusions bearing an F427H translocation-disrupting mutation (e.g., mPAH-E1v3) did not.[37] Similarly, treatments containing LFN alone instead of LFN-DTA did not inhibit protein synthesis. Intracellular protein synthesis inhibition, therefore, was dependent on the successful translocation of DTA. Overall, these results suggest that mPA-scFv fusions deliver DTA to the cytosol via the proposed mechanism (Scheme 1). While mPA-scFv fusions demonstrated efficient sub-nanomolar protein synthesis inhibition, we also note that they are less efficient than wild-type PA, which could be attributable to differences in available receptors, receptor dynamics, ligand-receptor interactions, protease activation of mPA-scFv fusions, or a combination thereof.
mPA-scFv fusions rely on their target receptors for activity. To investigate the specificity of the mPA-scFv fusions, we observed their potency on CHO cells, which express anthrax receptors but not EGFR or CEA.[38,39] In combination with LFN-DTA, wild-type PA inhibited protein synthesis, but mPA-E1v3 and mPA-sm3e did not (Figure 2B). This observation highlights the cell surface receptor-specificity attained upon re-targeting mPA with scFv fusions.
mPA-scFv fusions deliver LFN to the cytosol as confirmed by an additional assay (Figure 2C). AsPC-1 cells were treated with LFN alone, PA/LFN, or mPA-scFv/LFN for 24 h. The cytosolic fraction was extracted by digitonin and compared to total lysate by Western blot. An LFN band was present in all PA or mPA-scFv lanes, but not in the LFN-only lane. Immunoblotting against Rab5
pH 8.5 7.5 6.5 5.5 8.5 7.5 6.5 5.5 8.5 7.5 6.5 5.5 PA mPA-sm3e PA* mPA-sm3e* mPA-e1v3 mPA-e1v3* + - - - -- + + + + - - - -- - - + - - - -- - - + + + + - - - - -- - - + - - - -- - - + + + + PA83 PA63 oligomer 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 PA63 mPA63 mPA63-scFv mPA63-scFv oligomer mPA83-scFv
3
Figure 2. mPA-scFv constructs translocate LFN constructs to the cytosol of
pancreatic cancer cells. A) Protein synthesis inhibition in AsPC-1 cells or B) CHO cells treated with PA or mPA-scFvs targeting EGFR or CEA and varying concentrations of LFN or LFN-DTA, then incubated in 3H-leucine media. A
scintillation reagent was added to determine 3H incorporation (n = 3, values
expressed as means ± SD). Counts were normalized to untreated cells. C) AsPC-1 cells were treated with PA construct and LFN. Cytosol was prepared
using extraction with a mild detergent digitonin, and total lysate was prepared using RIPA buffer separately (20 µg total protein loaded). ERK is a cytosolic marker, and Rab5 is an endosomal marker.
Figure 3. EGFR- or CEA-targeted delivery of DTA to cytosol of pancreatic cancer cells, mediated by mPA-scFvs, promotes cell death. A) Dose-response curve indicating relative cell viability of AsPC-1 cells treated with PA or mPA-scFvs targeting EGFR or CEA and varying concentrations of LFN constructs.
Luminescent quantitation of ATP present in cells indicated relative cell viability. Values were normalized to an untreated control (n = 3, values expressed as means ± SD).
(endosomal marker) revealed a faint band in digitonin-extracted samples and an intense band in total lysates. An intense ERK (cytosolic marker) band was present in all samples. The uniform intensity of ERK across all samples and faintness of Rab5 in digitonin extracts indicates successful fractionation. Therefore, the PA-dependent presence of LFN in these lanes (labeled “Cytosol”) suggests that PA, mPA-E1v3, and mPA-sm3e successfully translocated LFN in all cases.
mPA-scFv fusions combined with LFN-DTA are potent in vitro cancer therapies. AsPC-1 cells were treated with LFN or LFN-DTA in combination with PA, mPA-E1v3 constructs or mPA-sm3e constructs for 72 h, after which the ATP content was measured with a luminescent cell viability assay (CellTiter-Glo) (Figure 3). mPA-E1v3 demonstrated cell proliferation inhibition with sub-picomolar concentrations of LFN-DTA, though with a potency roughly an order of magnitude less than that of wild-type PA/LFN-DTA. mPA-sm3e demonstrated similar potency. Both F427H mutants and LFN control groups lacking DTA had no significant effect on cell viability, indicating that the observed toxicity was due to successful translocation of DTA. These experiments illustrate the potency of the mPA-scFv/LFN-DTA platform against pancreatic cancer cells in vitro and its potential toward development as a novel therapeutic modality.
We aimed to expand our selection of available cytotoxic cargoes based on our understanding of pancreatic cancer. KRAS, NRAS, and HRAS comprise the most frequently mutated family of human oncogenes.[40] Key to pancreatic cancer in particular is an oncogenic mutation in KRAS, which can drive mutations in p53, leading to progression of tumor growth.[41,42] Despite this understanding, traditional approaches to Ras inhibition have struggled to produce a clinically effective therapy.[40]
PA mPA-sm3e mPA-E1v3 LFN 2 ng + + + + + + + + - - + - - - + -- + - - - + - -+ - - - + - - -Cytosol Total LFN ERK Rab5
A)
B)
-16 -14 -12 -10 -8 -6 0.0 0.5 1.0 1.5LF
Nconstruct concentration (log[M])
F
ra
ct
io
n
pr
ot
ei
n
sy
nt
he
si
s
mPA-E1v3 + LFN-DTA mPA-E1v3 + LFN PA + LFN-DTA mPA-sm3e + LFN mPA-sm3e + LFN-DTA mPA(F427H)-sm3e + LFN-DTAPA + LFN-DTA PA(F427H) + LFN-DTA
mPA-sm3e + LFN-DTA LFN-DTA mPA-E1v3 + LFN-DTA
C)
-16 -14 -12 -10 -8 -6 0.0 0.5 1.0 1.5LF
Nconstruct concentration (log[M])
F
ra
ct
io
n
pr
ot
ei
n
sy
nt
he
si
s
CHO cells, human EGFR-, CEA
-AsPC-1 cells, human EGFR+, CEA+
AsPC-1 cells, human EGFR+, CEA+
mPA(F427H)-E1v3 + LFN-DTA
IC
50: 0.7 pM
IC
50: 0.2 pM
IC
50: 0.03 pM
AsPC-1 cells, human EGFR+, CEA+
-18 -16 -14 -12 -10 -8 0
50 100 150
LF
Nconstruct concentration (log[M])
N
or
m
al
iz
ed
lu
m
in
es
ce
nc
e
(%
)
PA + LFN-DTA mPA-sm3e + LFN-DTA PA + LFN mPA-E1v3 + LFN-DTAmPA(F427H)-E1v3 + LFN-DTA mPA-E1v3 + LFN mPA-sm3e + LFN mPA(F427H)-sm3e + LFN-DTA
4
Figure 4. mPA-scFv-mediated delivery of RRSP to pancreatic cancer cells disrupts ERK signalling and causes cell death. A) Proposed mechanism of mPA-scFv-mediated delivery of RRSP to cytosol of EGFR+ cells. B) Western
blot of cytosolic fractions of AsPC-1 cells treated with PA or mPA-scFv targeting EGFR and LFN-RRSP (20 µg total protein loaded). C) Dose-response curve
indicating relative cell viability of AsPC-1 cells treated with PA or mPA-scFv constructs targeting EGFR and LFN constructs. Luminescent quantitation of
ATP present in cells indicated relative cell viability. Values were normalized to an untreated control (n = 3, values expressed as means ± SD).
RRSP is a recently characterized protease that cleaves and inactivates Ras and Rap1.[4–9] Importantly, cytosolic RRSP was found to process endogenous Ras protein and disrupt downstream ERK signaling, leading to cell rounding and cytotoxicity. Though challenging, targeted cytosolic delivery of this protease would be a promising tool against Ras-driven pancreatic cancers. We therefore envisioned a treatment based on pancreatic cancer cell-targeted Ras cleavage using LFN-RRSP and mPA-E1v3 (Figure 4A).
RRSP is amenable to PA-mediated translocation. As before, 3H-leucine was measured to determine protein synthesis in cells treated with PA and varying concentrations of LFN, LFN-DTA, or LFN-DTA-RRSP (Figure S3). Protein synthesis was inhibited at comparable concentrations of LFN-DTA and LFN-DTA-RRSP. Therefore, C-terminal fusion to RRSP did not appear to prohibit translocation of the DTA reporter.
mPA-E1v3 delivers RRSP to cleave the oncoprotein Ras and disrupt ERK signaling. We prepared an LFN-RRSP fusion to study the effect of RRSP translocation to the cytosol. Using Western blot, we interrogated the effects of mPA-E1v3/LFN-RRSP treatment on Ras signaling in AsPC-1 pancreatic cancer cells (Figure 4B). Immunoblotting the cytosol from PA/LFN-RRSP- or mPA-E1v3/LFN-RRSP-treated cells indicated an RRSP dose-dependent loss of Ras intensity. Further analysis indicated sustained levels of total intracellular ERK but an RRSP dose-dependent loss of downstream phosphorylation (i.e., pERK), indicating disruption of the Ras-associated ERK pathway, reflecting LFN-RRSP activity previously seen using wild-type PA.[8]
Targeted delivery of RRSP using mPA-E1v3 is a potent pancreatic cancer therapy in vitro (Figure 4C). We treated AsPC-1 cells with LFN-RRSP in combination with PA, mPA-EAsPC-1v3, or the translocation-deficient F427H variant for 72 h. Intracellular ATP was quantitated using a luminescent assay (CellTiter-Glo) as an indicator of cell viability. We observed LFN-RRSP dose-dependent cell killing in combination with EGFR-targeting mPA-E1v3, with sub-nanomolar potency. This was roughly ten-fold less potent than PA/LFN-RRSP and roughly ten thousand-fold less potent than PA/LFN-DTA. The mPA-E1v3 F427H translocation-deficient variant in combination with LFN-RRSP demonstrated minimal toxicity, suggesting that any toxicity was due to translocation of RRSP.
The anthrax protective antigen-scFv fusions targeting EGFR (mPA-E1v3) and CEA (mPA-sm3e) described here offer a solution to the challenge of targeted protein drug delivery to the cytosol of pancreatic cancer cells. mPA-scFv fusions enabled cell-specific delivery of two toxic enzymes, DTA and RRSP, demonstrating potent in vitro cancer cell toxicity. Importantly, the susceptibility of the mPA-scFv fusion linker did not prove to be prohibitive, but could be improved in future iterations of this platform. The in vitro success of these strategies point to the broad potential of this re-targeted mPA-scFv/LFN modality for cancer therapy, as this technology could conceivably be directed to a wide array of receptors using previously known scFvs or potentially even full-length antibodies for cell-specific delivery of different therapeutic proteins. Further, we envision that combinations of two highly specific re-targeted mPA-scFv fusions could be combined for precise delivery by leveraging intermolecular complementation.[43] The mPA-scFv/LFN-cargo combinations demonstrated here present a fundamentally new route to engineered immunotoxins and warrant further investigation in tumor models.
Acknowledgements
Financial support for this work was provided by an NSF CAREER grant (CHE-1351807) to B.L.P., the MIT/NIGMS Biotechnology Training Program (T32GM008334-28) to A.R.L., a PanCan/FNLCR KRAS post-doctoral fellowship (to M.B.), NIH H+
LFN RRSP
Ras cleavage, disruption of ERK pathway, cell death 5. Internalization 2. Cleavage 3. Oligomerization 4. Complex formation 6. Translocation 1. EGFR binding Ras pERK ERK [LFN-RRSP] (log[M]) PA mPA-E1v3 -7 -8 -9 -10 -11 -12 -7 -8 -9 -10 -11 -12 - - - + + + + + + + - - - + - -- - - + + + + + + - - +
-A)
B)
C)
IC50: 40 pM IC50: 0.03 pM IC50: 300 pM-18
-16
-14
-12
-10
-8
-6
0
50
100
150
LF
Nconstruct concentration (log[M])
N
or
m
al
iz
ed
lu
m
in
es
ce
nc
e
(%
)
PA + LF N-DTA PA + LFN-RRSP mPA(F427H)-E1v3 + LFN-RRSP mPA-E1v3 + LFN-RRSPAsPC-1 cells, human EGFR+, CEA+
AsPC-1 cells, human EGFR+, CEA+
X
X
XXXX
XXXX XXXX
Cell viability assay
5
grant R01AI092825 and the Northwestern Medicine Catalyst Fund (to K.J.F.S.).Keywords: anthrax toxin • scFv • pancreatic cancer • intracellular delivery • protein therapeutics
[1] L. Buscail, B. Bournet, P. Cordelier, Nat. Rev. Gastroenterol. Hepatol. 2020, DOI 10.1038/s41575-019-0245-4.
[2] A. McGuigan, P. Kelly, R. C. Turkington, C. Jones, H. G. Coleman, R. S. McCain, World J. Gastroenterol. 2018, 24, 4846–4861.
[3] M. Yamaizumi, E. Mekada, T. Uchida, Y. Okada, Cell 1978, 15, 245–250.
[4] I. Antic, M. Biancucci, K. J. F. Satchell, Proteins Struct. Funct. Bioinforma. 2014, 82, 2643–2656.
[5] M. Biancucci, K. J. F. Satchell, Oncotarget 2015, 6, 18742– 18743.
[6] I. Antic, M. Biancucci, Y. Zhu, D. R. Gius, K. J. F. Satchell, Nat. Commun. 2015, 6, 7396.
[7] M. Biancucci, A. E. Rabideau, Z. Lu, A. R. Loftis, B. L. Pentelute, K. J. F. Satchell, Biochemistry 2017, 56, 2747– 2757.
[8] M. Biancucci, G. Minasov, A. Banerjee, A. Herrera, P. J. Woida, M. B. Kieffer, L. Bindu, M. Abreu-Blanco, W. F. Anderson, V. Gaponenko, et al., Sci. Signal. 2018, 11, eaat8335.
[9] S. Y. Jang, J. Hwang, B. S. Kim, E.-Y. Lee, B.-H. Oh, M. H. Kim, J. Biol. Chem. 2018, 293, 18110–18122.
[10] L. Johannes, D. Decaudin, Gene Ther. 2005, 12, 1360– 1368.
[11] A. E. Rabideau, B. L. Pentelute, ACS Chem. Biol. 2016, 11, 1490–1501.
[12] X. Liao, A. E. Rabideau, B. L. Pentelute, ChemBioChem 2014, 15, 2458–2466.
[13] J. A. Young, R. J. Collier, Annu. Rev. Biochem. 2007, 76, 243–265.
[14] A. E. Rabideau, X. Liao, G. Akçay, B. L. Pentelute, Sci. Rep. 2015, 5, 1–11.
[15] K. A. Bradley, J. Mogridge, M. Mourez, R. J. Collier, J. A. T. Young, Nature 2001, 414, 225–229.
[16] H. M. Scobie, G. J. A. Rainey, K. A. Bradley, J. A. T. Young, Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 5170– 5174.
[17] K. E. Beauregard, R. J. Collier, J. A. Swanson, Cell. Microbiol. 2000, 2, 251–258.
[18] L. Abrami, S. Liu, P. Cosson, S. H. Leppla, F. G. van der Goot, J. Cell Biol. 2003, 160, 321–328.
[19] J. Mogridge, K. Cunningham, D. B. Lacy, M. Mourez, R. J. Collier, Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 7045– 7048.
[20] E. L. Benson, P. D. Huynh, A. Finkelstein, R. J. Collier, Biochemistry 1998, 37, 3941–3948.
[21] A. M. Friedlander, J. Biol. Chem. 1986, 261, 7123–7126. [22] C. Petosa, R. J. Collier, K. R. Klimpel, S. H. Leppla, R. C.
Liddington, Nature 1997, 385, 833–838.
[23] S. Zhang, A. Finkelstein, R. J. Collier, Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 16756–16761.
[24] A. E. Frankel, E. P. Tagge*, M. C. Willingham†, Semin. Cancer Biol. 1995, 6, 307–317.
[25] S. Liu, S. Netzel-Arnett, H. Birkedal-Hansen, S. H. Leppla, Cancer Res. 2000, 60, 6061 LP – 6067.
[26] A. Mechaly, A. J. McCluskey, R. J. Collier, MBio 2012, 3, e00088-12.
[27] A. J. McCluskey, A. J. Olive, M. N. Starnbach, R. J. Collier, Mol. Oncol. 2013, 7, 440–451.
[28] N.-I. Zahaf, A. E. Lang, L. Kaiser, C. D. Fichter, S. Lassmann, A. McCluskey, A. Augspach, K. Aktories, G. Schmidt, Sci. Rep. 2017, 7, 41252.
[29] Z. A. Ahmad, S. K. Yeap, A. M. Ali, W. Y. Ho, N. B. M. Alitheen, M. Hamid, Clin. Dev. Immunol. 2012, 2012, 980250.
[30] M. Oliveira-Cunha, W. G. Newman, A. K. Siriwardena, Cancers 2011, 3, 1513–1526.
[31] S. Hammarström, Semin. Cancer Biol. 1999, 9, 67–81. [32] N. Beauchemin, A. Arabzadeh, Cancer Metastasis Rev.
2013, 32, 643–671.
[33] Q. Meng, S. Shi, C. Liang, D. Liang, W. Xu, S. Ji, B. Zhang, Q. Ni, J. Xu, X. Yu, Onco. Targets Ther. 2017, 10, 4591– 4598.
[34] J. Zhou, L. Hu, Z. Yu, J. Zheng, D. Yang, M. Bouvet, R. M. Hoffman, J. Surg. Res. 2011, 171, 631–636.
[35] C. P. Graff, K. Chester, R. Begent, K. D. Wittrup, Protein Eng. Des. Sel. 2004, 17, 293–304.
[36] T. Honjo, Y. Nishizuka, O. Hayaishi, I. Kato, J. Biol. Chem. 1968, 243, 3553–3555.
[37] J. Sun, A. E. Lang, K. Aktories, R. J. Collier, Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 4346–4351.
[38] T. S. Bringman, P. B. Lindquist, R. Derynck, Cell 1987, 48, 429–440.
[39] E. Eftekhar, F. Naghibalhossaini, Mol. Biol. Rep. 2014, 41, 459–466.
[40] B. Papke, C. J. Der, Science 2017, 355, 1158–1163. [41] M. A. Collins, F. Bednar, Y. Zhang, J.-C. Brisset, S. Galbán,
C. J. Galbán, S. Rakshit, K. S. Flannagan, N. V. Adsay, M. Pasca di Magliano, J. Clin. Invest. 2012, 122, 639–653. [42] J. P. Morton, P. Timpson, S. A. Karim, R. A. Ridgway, D.
Athineos, B. Doyle, N. B. Jamieson, K. A. Oien, A. M. Lowy, V. G. Brunton, et al., Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 246–251.
[43] S. Liu, V. Redeye, J. G. Kuremsky, M. Kuhnen, A. Molinolo, T. H. Bugge, S. H. Leppla, Nat. Biotechnol. 2005, 23, 725– 730.
6
Entry for the Table of Contents
Cell-specific cytosolic delivery of proteins is highly challenging. Here we demonstrate that mutants of anthrax-derived protective antigen (mPA) can be expressed as fusions to scFvs for targeted delivery of potent protein therapeutics to the cellular cytosol. Targeting of EGFR or carcinoembryonic antigen with this strategy led to selective delivery to and death of pancreatic cancer cells.
X H+ Lethal factor N-terminal domain X X Toxic protein Pancreatic cancer cell death EGFR
or CEA
PA-scFv fusion (anti-EGFR or anti-CEA)
X X X X XX X X XX