Hydrogel Tethering Enhances Interdomain ... - ACS Publications

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

Hydrogel Tethering Enhances Interdomain Stabilization of Single Chain Antibodies Yijia Xiong, Nicole R Ford, Karen A Hecht, Guritno Roesijadi, and Thomas C. Squier Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.7b00512 • Publication Date (Web): 20 Sep 2017 Downloaded from http://pubs.acs.org on September 27, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Bioconjugate Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

Hydrogel Tethering Enhances Interdomain Stabilization of Single Chain Antibodies Yijia Xiong,† Nicole R. Ford,‡,€ Karen A. Hecht,‡,Ʒ Guritno Roesijadi,‡, § and Thomas C. Squier†,*

‡

Marine Biotechnology, Pacific Northwest National Laboratory, Sequim, WA 98383; Department of Microbiology, Oregon State University, Corvallis, OR 97331, †Department of Basic Medical Sciences, Western University of Health Sciences, Lebanon, OR 97355 §

*Corresponding Author: Email: [email protected], Phone: (541) 259-0230

€

Current address: Department of Biology, Chowan University, Murfreesboro, NC 27855 ([email protected]); ƷCurrent address: AstaReal Inc., Bellevue, WA 98004 ([email protected])

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT Here we identify the importance of molecular crowding agents in the functional stabilization of scFv antibodies. Antibodies were tethered through an engineered calmodulin (CaM) binding peptide into a stimulus-responsive hydrogel composed of poly-(ethylene glycol) (PEG) functionalized CaM. Macromolecular crowding is modulated by transient heating, which acts to decrease effective pore sizes. Using a fluorescent ligand bound to the scFv, frequency-domain fluorescence spectroscopy was used to assess the structural coupling between the VH and VL domains and relationships with functional stabilization. There is minimal structural coupling between the VH and VL domains in solution, as is apparent from the substantial rotational mobility for the bound ligand that is suggestive of an independent mobility for the VH and VL domains. In comparison, the hydrogel matrix acts to structurally couple the VH and VL domains, resulting in a reduction in rotational mobility and a retention of ligand binding in the presence of 8.0 M urea. Under these same conditions, ligand binding is disrupted for scFv antibodies in solution. Increases in the stabilization of scFv antibodies in hydrogels is not simply the result of molecular crowding, as decreases in pore size act to destabilize ligand binding. Rather, our results suggest that the functional stabilization of the scFv antibody within the PEG hydrogel matrix includes important factors involving protein solvation that stabilize interdomain interactions between the VH and VL domains necessary for ligand binding.


Page 2 of 24

Page 3 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION Antibody-based immobilization strategies represent a dominant platform for antigen capture, and enable the development of low-cost and field-deployable diagnostic assays aimed at identifying environmental agents and disease biomarkers.1-3 Further, 47 therapeutic monoclonal antibodies against cytokines, toxins, or receptors have been approved by the FDA, which represent more than 20% of the newly developed therapeutics.4-5 In these developmental pipelines, the binding cleft of high-affinity binding single chain antibodies are typically incorporated into the larger and more stable full-length antibody scaffold (e.g., Fab or IgG constructs).6-8 These developmental strategies take advantage of domain-specific functionalities within antibody sequences, allowing variable regions to be identified using the smaller single chain antibodies, including single chain fragment variable (scFv) fusion constructs. The small size of single chain antibodies enables the creation of diverse libraries that facilitate highthroughput selection and affinity maturation to enable the selective recognition of unique target antigens.2, 9 However, observed structural linkages between the variable and constant regions of full-length antibodies create inherent difficulties in transferring binding specificities and affinities between different antibody scaffolds, resulting in substantial increases in developmental time and expense.10-11 Furthermore, batch-to-batch variability, functional stability, and the large size of conventional antibodies limit therapeutic efficacies.12

1:

Figure Fluorescent

ligand

derivatives bind between variable heavy chain domain (VH) and variable light chain domain (VL) in single chain fragment variable (scFv) antibody against TNT. Homology model of scFv fusion protein (panel A) showing approximate position of fluorescent ligand TNB-Alexa488 between VH (N-domain) 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and VL (C-domain) domains, fluorescent ligand derivatives of TNT (i.e., TNB-Alexa488 and TNB-Alexa555; panel B), scFv sequence highlighting 27-amino acid linker sequence connecting VH and VL domains (underlined; panel C), and schematic illustration of scFv antibody conjugated to engineered maltose binding protein (MBP) with an engineered M13 CaM-binding sequence enabling tethering within hydrogel composed of poly-(ethylene glycol) functionalized CaM (panel D).

Single chain fragment variable (scFv) antibody fusion constructs involve engineering a short (~ 27 amino acids) flexible linker between heavy (VH) and light (VL) domains derived from immunoglobulins (Figure 1A,C). The common inability to use the single chain antibody scaffold in field-deployable assays arises because of their lack of conformational stability and poor solubility, which can result in the formation of protein aggregates in solution.13-16 Current approaches that seek to overcome these constraints are limited by the competing requirements to engineer both protein stability and sufficient flexibility within the antigen binding domains to maintain high-affinity ligand binding.17-19 An ability to deploy single chain antibodies would have significant practical advantages, as large amounts of antibodies can be obtained using costefficient and reproducible expression systems.20 In comparison, limiting the broad development of antibodies are drawbacks linked to production technologies, which can result in quality issues involving variable binding selectivity and affinities.21 As a result, there are currently a limited number of high-quality monoclonal antibodies for use in research environments due to the high cost of development (around $50K).21 To overcome current limitations in the application of scFv antibodies, we seek to understand our prior observation that scFv antibodies are functionally stabilized upon incorporation into diatom biosilica.17,22 These biosilica matrices suggest a possible role for molecular crowding in preventing protein unfolding and functional inactivation, which can be investigated using poly(ethylene glycol) diacrylate (PEGDA) hydrogel matrices that covalently incorporate the calcium binding protein calmodulin (CaM) through two engineered cysteines positioned on each of the opposing globular domains.23 In these materials, tethered proteins are engineered to contain a CaM-binding peptide (i.e., M13 from myosin light chain kinase) that, in the presence of calcium, binds to CaM within the hydrogel matrix (Figure 1D). Protein dynamics within these hydrogel matrices can be controlled through transient heating, which acts to decrease effective pore sizes and increase the side-chain rotational dynamics of a test protein (i.e., maltose binding protein; MBP).23 An ability to control effective pore sizes and associated protein dynamics offers an opportunity to explore how crowding agents stabilize a range of proteins, including scFv antibodies. We expect these insights will facilitate the rapid deployment and range of applications for newly developed antibodies. RESULTS High-Affinity Ligand Binding to scFv Antibody. We investigated how hydrogel materials may stabilize scFv antibodies cross-linked to MBP-M13 (MBP*), whose molecular dynamics are known to be sensitive to the pore size within the hydrogel matrix.23 To accomplish this, we cross-linked a high-affinity scFv against trinitrotolulene (TNT; Kd = 0.5 nM)24 to MBP* through addition of a 10-fold molar excess of the amine-reactive cross-linker bis(sulfosuccinimidyl) 4 ACS Paragon Plus Environment

Page 4 of 24

Page 5 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


suberate. The 11.4 Å length of this cross-linker is expected to enhance the independent rotational mobility of the scFv antibody. Upon addition of the bis(sulfosuccinimidyl) suberate cross-linker, there is a disappearance of the individual bands characteristic of scFv and MBP*, and the appearance of cross-linked protein complexes that exhibit a distribution of apparent masses on SDS-PAGE (Figure 2). Intermolecular cross-links between scFv and MBP* result in a large increase in mass, with associated reductions in mobility on SDS-PAGE. In contrast, intramolecular cross-links within either scFv or MBP* result in an increase in their respective mobilities on SDS-PAGE gels, as is expected based on their inability to fully unfold.25

Figure 2: High-Affinity Ligand Binding is Retained Following Cross-linking scFv to Maltose Binding Protein (MBP). Representative fluorescence correlation curves for TNB-Alexa555 (5.0 nM) before (diamonds; black line) and following addition of scFv (1.0 µM; circles, blue line) or scFv-MBP cross-linked protein (1.0 µM; triangles, red line). Inset: SDS-PAGE for scFv (1.5 µg; lane 1), MBP* (2.0 µg; lane 2), a mixture of scFv (1.5 µg) and MBP* (2.0 µg) prior to (lane 3) and following (lane 4) a one-hour incubation with a 10-fold molar excess of the amine reactive cross-linker bis(sulfosuccinimidyl)suberate and quenching for 1 hr with 50 mM Tris (pH 7.5). We also assessed the retention of protein function following cross-linking using fluorescence correlation spectroscopy (FCS) to measure ligand binding. These measurements assess the translational diffusion of the fluorescent analog of trinitrotoluene, i.e., TNB-Alexa555 (Figure 1B) in the absence and presence of added scFv or scFv-MBP*. Because our experiment was conducted using low TNB-Alexa555 concentrations (i.e., 5.0 nM), any decrease in the translational diffusion of TNB-Alexa555 is diagnostic of high-affinity binding. In comparison to TNB-Alexa555 alone, addition of either scFv or the cross-linked scFv-MBP* complex results in a substantial shift in the correlation function toward longer times (Figure 2). These results indicate a retention of high-affinity TNB-A555 binding, irrespective of cross-linking to MBP*. Fitting the autocorrelation functions, we were able to determine the translational diffusion coefficients (Dt). For TNB-Alexa555 in solution, Dt is 2.2 ± 0.4 ×10-6 cm2/s. In comparison Dt decreases to 0.62 ± 0.04 ×10-6 cm2/s upon binding scFv, which is consistent with the expected Dt (i.e., 0.85 ×10-6 cm2/s) based on hydrodynamic calculations using the structure of a 5 ACS Paragon Plus Environment


homologous scFv (i.e., 3GM0) as the basis for a homology model (Figure 1A).26 In comparison, upon incubation of the scFv cross-linked to MBP* with TNB-Alexa555 there is a further 49 ± 12% decrease in Dt (i.e., 0.42 ± 0.02 ×10-6 cm2/s) that is consistent with TNB-Alexa555 binding to the larger scFv-MBP* complex (mass = 84 kDa) relative to the scFv alone (mass = 30 kDa). In all cases, optimal nonlinear least-squares fits were obtained assuming a single translational diffusion coefficient; no improvement in the fit is observed upon addition of a second diffusion coefficient that is characteristic of the unbound TNB-Alexa555. These results demonstrate that essentially all the analyte (ligand) (i.e., 5.0 nM TNB-Alexa555) is bound to the scFv antibody.

Absorbance (280 nm)

Prior to their functional characterization, the complex between scFv and MBP* was isolated using size exclusion chromatography. In these experiments, we first incubated the sample with a 5-fold molar excess of the fluorescent ligand TNB-Alexa488 or TNB-Alexa555. Peak separation is apparent for the scFv-MBP* complex relative to the uncrosslinked MBP* or scFv (Figure 3), where the scFv-MBP* complex co-elutes with either TNB-Alexa488 or TNBAlexa555 prior to the appearance of either MBP* or the scFv. A densitometric analysis of the peak areas associated with the separation demonstrates that: i) 60% of the total scFv antibody is cross-linked to MBP and that ii) 68% of total TNB-Alexa Fluor dye binds to the cross-linked scFv-MBP. The remaining 32% of TNB-Alexa Fluor dye is bound to the uncross-linked scFv (40% of total scFv protein). The modest decrease in stoichiometry of TNB-Alexa Fluor dye binding to the uncross-linked scFv antibodies may be the result of intramolecular cross-linking, which has the potential to interfere with analyte binding. The ability of the cross-linked scFvMBP to bind an equivalent stoichiometry of TNB-Alexa Fluor ligand as the authentic scFv antibody suggests that the cross-linking between the scFv and MBP* does not significantly affect analyte binding to the scFv. scFv-MBP MBP TNB-Alexa555 scFv MBP TNB-Alexa488

Fluorescence

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 24

scFv

Elution Volume Figure 3: Chromatographic Separation of TNB-Alexa555 or TNB-Alexa488 Bound to scFvMBP. Following chemical cross-linking of scFv and MBP* (1 mg/mL), a 5-fold molar excess of TNB-Alexa555 or TNB-Alexa488 (70 µM) was incubated for one hour (25 °C) prior to chromatographic separation. Absorbance (280 nm; blue curve) or fluorescence emission for TNB-Alexa555 (λex = 530 nm; λem = 565 nm) or TNB-Alexa488 (λex = 485 nm; λem = 532 nm) were monitored in 20 mM HEPES (pH 7.5) and 150 mM NaCl.


Page 7 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Molecular Crowding in Hydrogel Matrix Stabilizes scFv Antibody Function. To assess the ability of the hydrogel matrix to functionally stabilize the scFv antibody tethered within the hydrogel, we have measured ligand binding following incubation in the presence of 8.0 M urea. In these experiments TNB-Alexa555 (5 nM) is first incubated with hydrogel samples containing the tethered scFv antibody for 0.5 hr, and following repeated washes the amount of bound TNBAlexa555 was determined based on the remaining fluorescence intensity. Following incubation in the presence of 8.0 M urea for one hour, there is a significant retention of TNB-Alexa555 binding in comparison to control samples (no added urea) (Figure 4D). In comparison, there is a total loss of ligand binding for the scFv in solution upon addition of 8.0 M urea, as evidenced by the essentially identical autocorrelation functions for TNB-Alexa555 (5 nM) irrespective of the addition of an excess of scFv (1 μM) (Figure 4C).

Figure 4: Functionally Stabilized scFv Antibodies in PEG Hydrogels. (Panel A) scFv antibodies (blue/green) with bound ligand (red) cross-linked to MBP* (gray ribbons) bound within the PEG hydrogel matrix (gray beads) through the M13 fusion peptide on MBP* (48 kDa) to CaM (blue) in the presence of calcium (red spheres). CaM forms part of the hydrogel matrix through covalent linkages between introduced cysteines at positions 34 and 110 that are conjugated to poly-(ethylene glycol) diacrylates (PEGDA) prior to initiation of photocrosslinking the PEG matrix. (Panel B) Representation of stimulus-responsive relaxation of hydrogel matrix following transient heating (37 °C for 14 hours), resulting in reduction in average pore sizes.23 (Panel C) Urea (8.0 M) disrupts TNB-Alexa555 binding to scFv in solution [50 mM HEPES, (pH 7.5), 150 mM NaCl, 10 mM CaCl2, 10% (v/v) glycerol, and 0.02% (v/v) Tween 20], demonstrated using fluorescence correlation spectroscopy for TNB-Alexa555 (5 nM) in the absence (solid black curve) and presence (red dashed curve) of scFv (1.0 µM). (Panel D) Fractional binding of TNBAlexa555 to scFv-MBP* in hydrogels following incubation in 8.0 M urea (1 hr at 25 °C) before (Tethered) and following (Tethered + ΔH) transient heating (14 hr, 37 °C) in 20 mM Tris (pH 7.5), 150 mM NaCl, and 2 mM CaCl2. Bound TNB-Alexa555 was determined using ImageJ (https://imagej.nih.gov/ij/) following image collection using a CoolSnap HQ2 camera (Photometrics Inc., Tucson, AZ). To assess the effects of molecular crowding on functional stability, effective pore sizes were decreased through transient heating at 37 ºC,23 which results in a relaxation of the hydrogel 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

matrix (Figure 4B).23 Reducing the pore size acts to reduce the amount of bound TNBAlexa555 (Figure 4D). These observations suggest that the stabilization of the scFv within the hydrogel matrix does not principally arise through molecular crowding; rather other mechanisms contribute to the ability of the PEG crowding agent to stabilize high-affinity ligand binding in the presence of chaotropic agents (i.e., 8 M urea). It is of considerable interest to understand the basis for the functional stabilization of scFv antibodies through their tethering within these hydrogel matrices. In this respect hydrophilic polymers such as PEG are commonly thought to modify protein function through an excluded volume effect, whereby the polymeric matrix is inert and reduces the available volume to favor a more compact protein conformation.27 However, PEG is not an inert matrix, and is able to form transient, weak interactions with embedded proteins, which may oppose protein conformational flexibility to stabilize active conformations.28-31 These soft interactions have the effect of increasing diffusional constraints that are akin to the molecular interactions observed within the cellular milieu, acting to increase the effective viscosity. In addition, PEG can modify protein solvation, which has the potential to affect protein dynamics and ligand binding affinity.32-34 Hydrogel Matrix Causes Large Increase in TNB-Alexa488 Lifetime and Restricts Protein Rotational Mobility. To gain further insight into the importance of how PEG acts to maintain scFv function under conditions commonly associated with protein inactivation, we have measured the excited state lifetime and rotational dynamics of TNB-Alexa488 bound to the scFv antibody in solution and within the hydrogel matrix. TNB-Alexa488 was used for these experiments due to its known environmental sensitivity and the longer excited state lifetime in comparison to TNB-Alexa555. This permits measurements of possible changes in protein solvation and overall protein rotational dynamics for the tethered scFv-MBP* within the hydrogel matrix. In this respect, the excited state lifetime of the Alexa488 dye assesses possible changes in water structure, as the induced dipole and charge distribution of the excited state are sensitive to solvation.35 Complementary measurements of rotational dynamics of the independent rotational dynamics of the Alexa488 dye and the overall rotational mobility of the scFv protein offers a means to sort out the effects of conformational restriction relative to alterations in water solvation on protein function.

Figure 5: PEG Hydrogels Increase in Fluorescence Lifetime of TNB-Alexa488 Bound to scFvMBP*. Frequency-domain fluorescence lifetime data showing phase shift (□,○,Δ,∇) or 8 ACS Paragon Plus Environment

Page 8 of 24

Page 9 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


modulation (■,●,▲,▼)(panel A) and calculated mean lifetimes (panel B) for TNB-Alexa488 bound to scFv-MBP* in solution (□,■) or in PEG hydrogels with (Tethered) (Δ, ▲,∇,▼) or without (No Tether)(○, ●) CaM polymerized within the hydrogel matrix. Tethered scFv-MBP* prior to (Δ, ▲) (Tethered) or following (∇,▼) (Tethered + ΔH) transient heating. We used frequency-domain fluorescence spectroscopy to measure the excited state lifetime of TNB-Alexa488 bound to scFv-MBP* in solution and tethered within the hydrogel matrix. Using sinusoidally modulated light to excite the Alexa488 chromophore, we measure the phase delay and loss of modulation as a function of the modulation frequency (Figure 5). In comparison to TNB-Alexa488 bound to scFv-MBP* in solution, there are large alterations in the frequency-response for TNB-Alexa488 bound to scFv-MBP* that is tethered within the hydrogel matrix that are indicative of a large increase in the excited state lifetime. A nonlinear least squares fit to the data for TNB-Alexa488 bound to scFv-MBP* in solution is dominated by a single component with a mean fluorescence lifetime of 0.5 ns (Table S1 in Supporting Information). In comparison, the mean fluorescence lifetime of TNB-Alexa488 bound to the scFv antibody tethered within the hydrogel requires two lifetime components with approximately equal amplitudes. The short component is similar to that observed for TNB-Alexa488 bound to the scFv-MBP* in solution. A larger component centered near 4 ns is indicative of an altered environment, and is consistent with changes in water solvation. Tethering did not affect the excited state lifetime, as essentially identical frequency-responses (and mean lifetimes following nonlinear least-squares fitting of the data) are observed irrespective of the presence of CaM within the hydrogel. However, transient heating does result in a small increase in the excited state lifetime. These latter results suggest that transient heating alters the hydrogel structure to modify the environment around the TNB-Alexa488 binding site on the scFv antibody.

Figure 6: Hydrogels Restrict the Rotational Dynamics of scFv-MBP*. Frequency-domain fluorescence anisotropy measurements of differential phase (Panel A) and modulated anisotropy (Panel B) for TNB-Alexa488 bound to scFv-MBP* in solution (■) or in PEG hydrogels with (Tethered) (▲, ∇) or without (No Tether)(●) CaM polymerized within the hydrogel matrix. Rotational dynamics of tethered scFv-MBP* was analyzed prior to (▲) (Tethered) or following (∇) (Tethered + ΔH) transient heating (14 hr, 37 °C). Buffer is 20 mM Tris (pH 7.5), 150 mM NaCl, and 2 mM CaCl2. Frequency-domain fluorescence anisotropy measurements of the rotational dynamics of TNB-Alexa488 permitted an assessment of how the hydrogel matrix influences the structural dynamics of the scFv. The differential phase and modulated anisotropy are measured as a 9 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 24

function of the modulation frequency of the exciting light for scFv-MBP* in solution and within the hydrogel matrix. Large changes in the differential phase and modulated anisotropy are apparent upon tethering the scFv antibody within the hydrogel matrix (Figure 6). Importantly, the measured rotational dynamics are relatively insensitive to the presence of the tether, as very similar frequency responses are apparent for the scFv antibody irrespective of the presence of the tethering to CaM within the PEG matrix. These results indicate that the tether does not restrict the dynamics of the scFv antibody. Transient heating of the hydrogel matrix alters the anisotropy decay curves, which is consistent with prior results that suggest that transient heating results in a relaxation of the PEG matrix that: i) increases protein side chain mobility and ii) decreases the effective pore size within the hydrogel matrix (Figure 4B).23 Table 1: Rotational Correlation Times for TNB-Alexa488 Bound to scFv-MBP* Sample Condition φ1 (ns) φ2 (ns)

N.A.

1.51 (1.45 – 1.55)

No Tether (PEG)

0.43 (0.40 – 0.46)

9.4 (8.6 – 10.6)

Tether (PEG + CaM*)

0.74 (0.69 – 0.79)

9.5 (8.8 – 10.4)

Solution

Tether + ΔH (PEG + CaM*)

0.53 13.3 (0.51 – 0.55) (12.3 – 14.5) Brackets are range of reliable rotational correlation times obtained from error surfaces (see Figure 7). A single rotational correlation time (φrot = 1.5 ns) is sufficient to describe the anisotropy decay for TNB-Alexa488 bound to scFv-MBP* in solution (Table 1). There is a small (two-fold) 2 improvement in the goodness of fit (i.e., χ R ) upon inclusion of two rotational correlation times, where a shorter component is consistent with an independent rate of probe motion (i.e., φ1 = 0.3 ± 0.1 ns)(Table S2 in Supporting Information). However, irrespective of the model used to fit the data, the rate of rotational motion for the slower rate constant (i.e., φ2 < 2 ns) is substantially faster than is expected for the overall rotational motion of the scFv antibody in solution (1/6Dr = 17 nsec)26. These results suggest that the motion of the two domains in the scFv are not tightly coupled and undergo independent rotational motion. Upon incorporation of the scFv-MBP* into the hydrogel matrix, two rotational correlation times are required to describe the anisotropy decay, which correspond to the independent rotational motion of the TNB-Alexa488 probe (φ1) and the overall rotational dynamics of the scFv antibody (φ2). The requirement for two rotational rates is apparent from 2 the greater than 60-fold improvement in the χR (Table S2). In comparison to the scFv in solution (φ2 < 2 ns), the longer component for the scFv-MBP* tethered within the hydrogel is 10 ACS Paragon Plus Environment

Page 11 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


substantially larger (i.e., φ2 > 9 ns). This is consistent with a structural coupling between the two domains of the scFv, such that the VH and VL domains now move together. In comparison, the VH and VL domains of the scFv antibody move independently of one another in solution.

Figure 7. Interdomain Stabilization of scFv by Hydrogel Matrix. Normalized chi-squared error surfaces for rotational dynamics for TNB-Alexa488 segmental mobility (φ1) and overall protein rotational dynamics of scFv-MBP* (φ2) for scFv-MBP* in solution (blue solid line) or in PEG hydrogels with (Tethered Hydrogel) (black lines) or without (No Tether)(red solid line) CaM polymerized within the hydrogel matrix. For tethered hydrogels, scFv dynamics were measured prior to (solid black line) (Tethered) or following (dashed black line) (Tethered + ΔH) transient heating (14 hr, 37 °C). Buffer is 20 mM Tris (pH 7.5), 150 mM NaCl, and 2 mM CaCl2. λex = 488 nm; emitted light was collected subsequent to an HQ535/50 bandpass filter (Chroma Technologies). Fitting parameters below the dashed horizontal line correspond to acceptable values within one standard deviation of the optimal fit (χ2N = 1.00). To obtain a quantitative understanding of how the hydrogel matrix affects the structural coupling within the scFv antibody, we undertook a rigorous analysis of the errors, which accounts for any correlations between the fitting parameters (i.e., measured amplitudes and rates of motion). In an analysis of the overall rate of rotational motion the correlation time (i.e., φ2) is systematically varied, and all other parameters are optimized in fitting the data (Figure 7B). The increase in the rotational correlation time associated with overall protein rotational mobility observed upon transient heating is statistically significant, as judged by the non-overlapping error surfaces. In comparison, prior to transient heating, the error surfaces are virtually superimposable irrespective of whether the scFv-MBP* is tethered within the hydrogel matrix. These measurements demonstrate that while tethering does not significantly affect the overall protein dynamics, that there are reductions in rates of overall motion following transient heating at 37 ºC that are consistent with reductions in the pore diameters that arises from a relaxation of the hydrogel structure (see Figure 4B). Under these latter conditions there is an increase in the segmental rotational dynamics of the TNB-Alexa488 ligand that approaches that observed in the absence of tethering (Figure 7A), suggesting that transient heating acts to disrupt protein associations with the PEG hydrogel matrix. These latter results are consistent with prior NMR measurements that have demonstrated that the side chain rotational dynamics of tethered proteins within the hydrogel matrix increase following transient heating.23



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

DISCUSSION Our results indicate that underlying principles of antibody stabilization by the hydrogel matrix include direct interactions between the PEG matrix and water, acting to reduce rates of water reorganization. These changes in water structure result in a greater than 4-fold increase in mean fluorescence lifetime (i.e., τ¯ ) (Figure 5; Table S1). Increases in τ¯ arise because of the sensitivity of the induced dipole of the excited state to solvation, and reflect changes in solvent structure. Consistent with our results, solvent reorganization has previously been suggested to underlie the stabilization of the bioactive protein conformation that arises upon addition of crowding agents.29, 34, 36-37 Mechanisms of stabilization include enhancements in hydrogen bonds and other electrostatic interactions that stabilize the native state. Further, favorable attractive interactions between the PEG matrix and the scFv antibody act to stabilize the bioactive protein structure and counteract reductions in conformational entropy arising from excluded volume effects associated with molecular crowding.28 Under these conditions, weak (quinary) interactions between the scFv-MBP* and the PEG matrix act to restrict domain mobility and stabilize the structural coupling between the VL- and VH-domains of the scFv antibody. These results are broadly consistent with prior observations that increases in noncovalent interactions within polyacrylamide hydrogel matrices enhance protein stability.38 Quinary interactions within the hydrogel matrix are sensitive to transient heating, which acts to modulate conformational dynamics, whereby the fast rotational mobility of the TNBAlexa488 chromophore increases and the overall mobility of the scFv antibody is reduced due to a reduction in the effective pore diameter of the matrix (Figures 4B and 7). These latter results imply that protein stabilization does not arise through macromolecular crowding per se (excluded volume effects), but rather arises due to protein solvation effects that act to favor transient interactions between the scFv antibodies and the hydrogel matrix. This interpretation is consistent with recent molecular dynamics simulations that suggest an important role for lowaffinity biomolecular interactions in stabilizing protein structure within PEG mixtures, which act to reduce protein conformational flexibility to favor bioactive conformations.28, 30 Further, PEG maintains a random coil structure when covalently bound to proteins, creating a soft interactive surface that functions to self-organize proteins and increase thermal stability.31 The challenge has been to retain the dynamics needed for high-affinity substrate binding, which is commonly diminished when proteins are derivatized by PEG.33 We have demonstrated that tethering scFv antibodies within a hydrogel matrix solves this problem, as sufficient protein dynamics are retained to maintain high-affinity substrate binding under conditions that stabilize the scFv antibody against common protein denaturants (i.e., 8.0 M urea). The stabilization of scFv antibody structure and associated retention of ligand binding in the presence of 8.0 M urea has a number of important practical applications for diagnostics, enabling analyte sensing under conditions that lyse cells and inactivate bacterial and viral pathogens.39-40 Advantages include an ability to rapidly select and deploy high-affinity scFv antibodies that overcome current immobilization strategies that commonly functionalize surfaces, which have insufficient surface area for efficient ligand capture and commonly result in a loss of function.3, 41-42 By tethering antibodies within a porous hydrogel matrix, our design increases the available surface area for ligand binding while maintaining binding affinity, further increasing the avidity of ligand binding through the increased three-dimensional density.43 As 12 ACS Paragon Plus Environment

Page 12 of 24

Page 13 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


scFv antibodies are stable under conditions that denature target proteins, these materials should enable rapid selection of antibodies against linear peptides. Deployment of multiple antibodies that all recognize linear epitopes on target proteins will increase the reliability of diagnostic assays that require an ability to distinguish between closely related strains.41, 44 As proteins can be immobilized in either sheets or micron hydrogel beads (see Figure S1 in Supporting Information), these hydrogel materials have the potential to overcome current limitations involving the rapid clearance of scFv antibodies in the circulation to enable their use in therapeutics. Further, an ability to directly deploy scFv antibodies avoids down-stream engineering that commonly involves moving regions identified in scFv antibodies into fulllength antibody scaffolds, which commonly result in decreases in binding specificity.9 As 8.0 M urea acts to both lyse cells and inactivate pathogens,39 these conditions have the added advantage of reducing the need for biosafety containment. In summary, we have demonstrated that incorporation of scFv antibodies into selfassembled hydrogel biomaterials enhances the structural coupling between the heavy (VH) and light (VL) domains (Figure 8). High-affinity ligand binding is retained in the presence of chaotropic agents (i.e., 8.0 M urea), which unfold protein antigens to enable scFv recognition of linear epitopes. As these hydrogel materials exhibit superior mass transfer, permitting proteins and other ligands unrestricted access,23 our results suggest a range of future applications that include the creation of bifunctional sensors (with fluorescence donor and acceptor probes on different antibodies) capable of undergoing fluorescence resonance energy transfer (FRET) upon binding to protein antigens for diagnostic applications.

Figure 8: Hydrogels Stabilize Interdomain Coupling in Single Chain Variable Fragment (scFv) Antibodies. PEG within hydrogel matrix acts as a crowding agent to stabilize the structural coupling between VL and VH-domains, resulting in the retention of ligand binding under environmental conditions normally associated with protein denaturation (e.g., the presence of chaotropic agents). Illustration depicts homology model of scFv antibody, with VL domain (top; C-domain) and VH domain (bottom; N-domain) connected through a 27 amino acid linker. Ligand is TNB-Alexa488 (oxygen, red; nitrogen, blue; sulfur, yellow; carbon, gray). Protein backbone structure highlighting helical (pink) and sheet (orange) secondary structures.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

EXPERIMENTAL PROCEDURES Alexa-TNB Synthesis: TNB-Alexa488 or TNB-Alexa555 were synthesized by conjugating TNB-sulfonic acid (Sigma-Aldrich) to either Alexa Fluor 488 or Alexa Fluor 555 cadaverine (Life Technologies),45 essentially as previously described.45 Fluorescent ligands were purified using Supelclean LC-18 SPE columns45 or by C-18 reversed phase high performance liquid chromatography (RP-HPLC) with water/methanol gradients. Purity of the AF555-TNB was verified by analytical-scale C-18 RP-HPLC using AF555-TNB as a standard, as previously described.17 Protein Expression and Purification: The gene encoding CaM (16.7 kDa) from chicken with mutations T34C and T110C was overexpressed in E. coli (strain BL21) grown in LB broth with 100 µg/mL ampicillin (Sigma-Aldrich), essentially as previously described.46 Following cell lysis, CaM was purified using calcium-dependent hydrophobic binding on a phenylsepharose column, as described.47 Concentrations of CaM were determined using the absorption coefficient (ε277 nm = 3029 M-1 cm-1).48 Following purification, CaM was buffer exchanged into HEPES buffer (50 mM HEPES pH 7.5, 150 mM NaCl, and 10 mM CaCl2) and concentrated 10 mg/mL or 0.6 mM) using an Amicon Ultra 10K MWCO (Millipore Corp.; Bedford, MA) prior to functionalization. Alternatively, CaM was buffer exchanged into milliQ water, flash frozen in liquid nitrogen, and lyophilized for storage at -80 °C. As previously described,23 the gene encoding maltose binding protein (MBP) (42.5 kDa) derived from E. coli was engineered to contain a 23 amino acid linker sequence (2.5 kDa) (NSSSNNNNNNNNNNLGDDDDKVP) prior to the CaM-binding sequence (M13) from skeletal myosin light chain kinase (3.0 kDa) (KRRWKKNFIAVSAANRFKKISSSGAL) at the C-terminus using a pMal vector from New England Biolabs (Ipswich, MA). MBP* (48 kDa) was overexpressed in E. coli (strain BL21) grown in LB broth with 100 µg/mL ampicillin (Sigma-Aldrich) following addition of 1 mM isopropylthio-β-galactoside (IPTG); following 4-5 hours of growth, cells were harvested and frozen (-80 °C). Cells were resuspended in 20 mM HEPES (pH 7.5), 200 mM NaCl, and 1 mM EDTA (Binding buffer) prior to lysis; MBP* was purified using prepacked MBPTrap HP affinity columns (GE Healthcare Life Sciences; Oregon City, OR) where lysate was loaded and washed with 20 mM HEPES (pH 7.5), 200 mM NaCl, 1mM EDTA prior to elution in 10 mM maltose in Binding buffer. MBP* concentration was determined using ε280 nm = 70,400 M-1 cm-1. Purified MBP* was flash frozen in liquid nitrogen and stored at -80°C. An E. coli clone with a plasmid encoding the scFv antibody against TNT, as well as the purified scFv antibody, was provided by Dr. Ellen R. Goldman at the Naval Research Laboratory. Expression and purification was done essentially as previously described.24 Briefly, E. coli, containing the scFv plasmid was inoculated into 50 mL of sterile Terrific Broth (TB) (Millipore Sigma; St. Louis, MO) with 2% (w/v) glucose and was grown at 30 °C overnight. Media was transferred to 500 mL of fresh TB and grown for 3 hours at 25 °C prior to induction with IPTG (1 mM). Following 3 hours of expression, E. coli was pelleted by centrifugation at 6000 × g for 20 minutes. Cells were resuspended in 14 mL of ice-cold 100 mM Tris-HCl (pH 7.5) and 0.75 M sucrose. Hen egg lysozyme (0.1 mg/mL) was added, followed by drop-wise addition of 28 mL of 1 mM EDTA (pH 7.5). Following gentle mixing on ice for 15 min, MgCl2 was added (24 mM), cells were swirled for another 15 min, and the E. coli was pelleted (20,000 14 ACS Paragon Plus Environment

Page 14 of 24

Page 15 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


× g for 30 minutes). Concentrated IMAC buffer was added to lysate to a final concentration of 20 mM Na2HPO4, 0.4 M NaCl, and 20 mM imidazole (pH 7.5) prior to the addition of 0.5 mL of pre-washed Ni-NTA sepharose (GE Healthcare Life Sciences; Oregon City, OR) that was previously equilibrated in IMAC buffer. The IMAC resin was then packed into an empty column (Poly-Prep disposable 2 mL column, Bio-Rad laboratories, Hercules, CA), washed twice with IMAC buffer, and eluted with 5 mL of elution buffer (IMAC + 500 mM imidazole). Protein was concentrated to 200 μL using an Amicon 10 kDa ultrafiltration device (Millipore, Billerica, MA) and further purified by gel filtration on a Superdex 75 or 200 HR 10/300 column (GE Healthcare Life Sciences; Oregon City, OR) operating in PBS [20 mM Na2HPO4, (pH 7.5) and 150 mM NaCl]. Functionalization of Calmodulin: Poly(ethylene glycol) diacrylate (PEGDA) (10 kDa) was conjugated to cysteines in the engineered CaM (Cys34 and Cys110) by a Michael-type addition, essentially as previously described.23, 49-52 Briefly, following addition of tris(2carboxyethyl)phosphine (TCEP) ( 2 mM) in HEPES buffer for 1 hour at 23 °C, the cysteines in CaM (0.6 mM) were conjugated to PEGDA (6.5 mM) following overnight incubation (> 20 hours). Excess unreacted PEGDA was removed from the PEG-modified CaM using a phenylsepharose column, as described for purification of CaM (see above). Concentrated PEGmodified CaM was buffer exchanged into milliQ water using a Zeba Desalt Spin Column (Thermo Scientific; Rockford, IL). Raffinose (36 mM) (MP Biomedicals) was added to PEGmodified CaM (i.e., CaM*) (10 mg/mL or 0.6 mM), and following equilibration for 1 hour at 23 °C the sample was flash frozen in liquid nitrogen, and subsequently lyophilized for storage at -80 °C. Cross-linking and Chromatographic Separation: A mixture of scFv (0.4 mg/mL; 30.1 kDa mass implies 13 μM) and MBP* (0.5 mg/mL; 54 kDa mass of complex implies 13 uM) in (10 mM phosphate (pH 7.4,) 2.7 mM KCl, 150 mM NaCl) (PBS) was incubated for one hour with 150 μM of the amine reactive cross-linker bis(sulfosuccinimidyl)suberate (BS3) (ThermoFisher Scientific, Waltham, MA), which has an 11.4 Å spacer arm. The cross-linking reaction was stopped upon addition of 50 mM Tris (pH 7.5). Following the chemical crosslinking of scFv and MBP*, a 5-fold molar excess of TNB-Alexa555 or TNB-Alexa488 (70 µM) was incubated with the sample for one hour (25 °C). Excess TNB-Alexa555 or TNB-Alexa488 was removed by size exclusion chromatography using a Superdex 75 10/300GL prepackaged column and an AKTA FPLC (GE Healthcare Life Sciences; Oregon City, OR); eluent absorbance (280 nm) and fluorescence emission was monitored for either TNB-Alexa555 (λex = 530 nm; λem = 565 nm) or TNB-Alexa488 (λex = 485 nm; λem = 532 nm). Buffer was 20 mM HEPES (pH 7.5) and 150 mM NaCl. Hydrogel Formation: Formation of hydrogel sheets involved mixing scFv-MBP* with PEGDA in the presence of (when indicated) 10% (mol/mol) CaM* (HEPES buffer, pH 7.5) prior to addition of the photoinitiator 2-hydroxy-1-[4-(hydroxyethoxy)phenyl]-2-methyl-1-propanone (4.5 mM) (i.e., IGRACURE 2959; Ciba Specialty Chemicals; BASF Schweiz AG; Basel, Switzerland), essentially as described previously.23 Polymerization involved a 20 minute incubation in the presence of ultraviolet light (λex = 365 nm). For fluorescence experiments, gels were formed in 200 µL aliquots within sealed cuvettes (pathlength = 0.5 cm).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 24

In limited cases, micron sized hydrogel beads were formed. Hydrogel bead formation was similar to the conditions associated with sheet formation, with the exception that CaM was functionalized using a small molecular weight PEG (575 Da). Conditions for bead formation are inspired from those described by Murphy and co-workers,53 and required CaM to be bufferexchanged pure milliQ water using a Zeba spin desalting column with the 7,000 Da molecular weight cutoff (#89891; ThermoFisher Scientific; Waltham, MA). When forming hydrogel beads, lyophilized CaM derivatized with PEG was dissolved at 22.5% to 25% (w/v) in Tris buffer [20 mM tris(hydroxymethyl)aminomethane (pH 7.5), 150 mM NaCl, and 10 mM CaCl2], incubated at room temperature for one-half hour, and illuminated with 365nm UV for 20 min. Beads were resuspended in Tris buffer and stored at 4 ºC. Absorption and Fluorescence Emission Spectra: Unless otherwise indicated, all absorption spectra were taken on an Evolution 300 spectrophotometer (ThermoFisher Scientific; Waltham, MA). Fluorescence spectra were measured on a FluoroMax-2 fluorometer (Horiba Jobin Yvon; Albany, NY). All measurements were made in PBS, unless otherwise mentioned. Integrated fluorescence emission spectra were used for all quantitative analysis. Fluorescence Correlation Spectroscopy (FCS): FCS was used to assess the binding of Alexa555-TNB with the scFv antibody, which involved assessing the rates of translational diffusion (Dt) for TNB-Alexa555 (0.5 nM) in the absence and presence of the scFv or scFvMBP*. FCS measurements were collected in phosphate buffered saline (PBS)(10 mM phosphate (pH 7.4), 2.7 mM KCl, 150 mM NaCl) with 0.05% Tween 20 (v/v) and 20% (v/v) glycerol, essentially as previously described.54 Briefly, excitation involved a 532 nm diode pumped solid state laser (Laserglow Technology, Toronto, Canada) to excite TNB-Alexa555, where the laser power was 45 μW measured at the entry port of the microscope. Light was focused using a 60× objective lens (Fluo Apo 60X VC, Nikon, Melville, NY) onto a spot 50 µm above the surface of a glass bottom fluorescence microplate (Greiner bio-one, Monroe, North Carolina). Fluorescence emission was collected using the same objective, separated by a z532/633rpc dichroic mirror and a HQ560/50 emission filter (ChromaTechnology, Bellows Falls, VT) and coupled into a 50 µm diameter optical fiber which also served as the pinhole. The fluorescence was sent to a pair of SPCM-AQR-14 avalanche photodiodes (Perkin-Elmer Optoelectronics, Vaudreuil, Canada) after a cube beam splitter (Thorlabs, Newton, NJ). The output of the avalanche photodiodes was fed into a Flex02-01D multi-tau correlator (Correlator.com, Bridgewater, NJ), and the fluorescence correlation curves were calculated in real time. Calibration used the stand fluorophore rhodamine 6G (R6G), whose translational diffusion coefficient is 4.1 x 10-6 cm2/s.55 Typically, accumulation times for each sample were 5 minutes. Correlation curves were collected for TNB-Alexa555 (5.0 nM) in the absence and presence of excess scFv or scFv-MBP* (i.e., 1 µM), and fit as previously described.54 The relationship between the translational diffusion time (τDt) and translational diffusion coefficient (Dt) is: 𝜏𝐷𝑡 =

𝜔2

4𝐷𝑡

.

(1)

Frequency-Domain Fluorescence Measurements: The fluorescence lifetime and rotational dynamics of TNB-Alexa488 were measured using an ISS K2 frequency domain fluorometer (Champaign, IL), as described previously.17, 56-57 Samples incubated with TNBAlexa488 were excited using a 488 nm laser diode with emitted light detected subsequent to an 16 ACS Paragon Plus Environment

Page 17 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


HQ535/50 band-pass filter (Chroma Technology Corporation, Bellow Falls, VT). All measurements were taken at 25 °C. Fluorescein was used as a lifetime standard (τref = 4.0 ns) (http://www.iss.com/resources/reference/data_tables/ StandardsLEDsLaserDiodes.html). Analysis of Fluorescence Lifetime Intensity Decays: The frequency domain fluorescence lifetime data were analyzed by fitting the time-dependent decay, I(t), of fluorescence to a sum of exponentials using nonlinear least-squares, as previously described:58-59 𝑛

𝐼(𝑡) = � 𝛼𝑖 𝑒 −𝑡/𝜏𝑖 𝑖=1

(2)

where αi values represent the pre-exponential factors, τi values represent the decay times, and n is the number of exponential components required to describe the decay. The intensity decay law is obtained from the frequency response of amplitude-modulated light and is characterized by the frequency-dependent values of the phase and the extent of demodulation. The values are compared with the calculated values from an assumed decay law until a minimum of the reduced squared deviation (χR2) is obtained. After the measurement of the intensity decay, the mean lifetime was calculated: 𝜏� = ∑𝑛𝑖=1 𝛼𝑖 𝜏𝑖

(3)

Errors in the differential phase and modulated anisotropy were assumed to be 0.2° and 0.004, 2 respectively. Weighted residuals ( χ R ) were calculated as the difference between measured and fit data divided by the error of individual measurements (0.2° or 0.004 for phase shift and modulation data respectively). Analysis of Fluorescence Anisotropy Decays: Time-resolved anisotropies were measured from the differential phase and modulated anisotropy, as previously described.59 Anisotropy decays were fit to a sum of exponentials: 𝑟(𝑡) = 𝑟𝑜 𝑥 ∑𝑛𝑖=1 𝑔𝑖 𝑒 −𝑡/𝜑𝑖 ,

(4)

where ro is the limiting anisotropy in the absence of rotational diffusion, φi are the rotationial correlation times, and gi × ro are the amplitudes of the total anisotropy loss associated with each correlation time. Parameter values were determined by minimizing the reduced chi-squared χ2R , and error surfaces were constructed for each rotational correlation time, which provides a conservative estimate of the recovered values.60 Briefly, the values of χ2R determined using a range of fixed rotational correlation time φi, (with all other parameters variable) were divided by the minimum value of χ2R (min) obtained for the optimal fit. Typically, the normalized chisquared (χ2N ) is calculated as: χ2N = χ2R �χ2R (min), which facilitates comparison of different fitting parameters. Error ranges are calculated using the F-statistic for one standard deviation from the mean.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Homology Model of scFv Antibody: A homology model of the structure of the scFv antibody against TNT was based on the structure of a homologous single chain antibody made against methamphetamine (i.e., 3GM0.pdb).61 The scFv against methamphetamine was created by joining the light and heavy chain variable domains of the parent mAb6H4 (accession number DQ38153 and DQ381542) with a 15 amino acid linker (Gly4Ser)3, which reduced the mass from about 150 kDa to 27.4 kDa (GenBank Accession FJ821514). A sequence comparison demonstrates that the scFv antibodies have 71% sequence identity, with major differences involving the interdomain linker connecting the VH and VL domains, which contains 15 amino acid for 3GM0.pdb or 27 amino acids for the scFv against TNT (Figure S2). The homology model was created using the SWISS-MODEL server, where the sequence of scFv antibody against TNT was uploaded to the ExPASy web server and searched with Blast and HHBlits against the SWISS-MODEL template library (SMTL) (http://swissmodel.expasy.org/).62-64 The quality of found templates was predicted from the target-template alignment. Templates with the highest sequence identity (> 70) and sequence similarity (>0.5) are suitable for model building, where highest global model quality estimate (GMQE) was 0.73 for the homology model. Model visualization, rendering and the superposition of the TNB-Alexa488 onto the protein homology model were done using the molecular visualization software PyMOL (https://www.pymol.org/). Theoretical Correlation Times: The theoretical diffusion coefficients for the homology model of the scFv antibody was calculated using the program Hydropro10, where the solvent density was assumed to be 1.0 g/mL, the viscosity was 1.3 cP, and the temperature was 20 ºC.26, 65 Calculated translational (Dt) and rotational (Dr) diffusion coefficients are, respectively, 8.5 × 10-7 cm2/s and 9.5 × 106 sec-1, respectively. The calculated rotational correlation time (φr) for the overall rotational dynamics of the scFv antibody is 17 ns, assuming φr = 1/6Dr.66 In comparison, Dr = 1.074 × 107 sec-1 for 3GM0.pdb (φr = 16 ns). AKNOWLEDGMENTS We thank Dr. Ellen R. Goldman for providing both clones for the scFv and purified protein, and assistance in preparing the fluorescent TNT surrogate compounds. This work was supported by the Defense Threat Reduction Agency.

SUPPORTING INFORMATION The supporting information is available free of charge on the ACS Publications website at DOI: 10.102/acs.bioconjchem.7b00334.

ABBREVIATIONS CaM, calmodulin; CaM*, engineered CaM with PEG conjugated at Cys34 and Cys110; Dt, translational diffusion coefficient; Dr, rotational diffusion coefficient; FPR, fluorescence photobleaching recovery; FRAO, fluorescence recovery after photobleaching; M13, CaMbinding peptide derived from skeletal muscle myosin light chain kinase 18 ACS Paragon Plus Environment

Page 18 of 24

Page 19 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(KRRWKKNFIAVSAANRFKKISSSGAL); MBP, maltose-binding protein; MBP*, fusion protein between MBP and M13; PEG, polyethylene glycol; PBS, phosphate buffered saline or 10 mM NaH2PO4 (pH 7.4), 137 mM NaCl, and 2.7 mM KCl; PEG-CaM, PEG conjugated calmodulin; PEGDA, polyethylene glycol diacrylate; scFv, single chain fragment variable antibody; TNB, trinitrobenzene; TNB-Alexa488 or TNB-Alexa555, fluorescent analogs of trinitrobenzene linked to either the Alexa488 or Alexa555 chromophore through cadaverine linkages; TNT, trinitrotoluene. REFERENCES:

1. Tiller, K. E., and Tessier, P. M. (2015) Advances in Antibody Design. Annu. Rev. Biomed. Eng. 17, 191-216. 2. Kuhn, P., Fuhner, V., Unkauf, T., Moreira, G. M., Frenzel, A., Miethe, S., and Hust, M. (2016) Recombinant antibodies for diagnostics and therapy against pathogens and toxins generated by phage display. Proteomics Clin. Appl. 10, 922-948. 3. Barbosa, A. I., and Reis, N. M. (2017) A critical insight into the development pipeline of microfluidic immunoassay devices for the sensitive quantitation of protein biomarkers at the point of care. Analyst 142, 858-882. 4. Torre, B. G., and Albericio, F. (2017) The pharmaceutical industry in 2016. An analysis of FDA drug approvals from a perspective of the molecule type. Molecules 22, 368. 5. Saeed, A. F., Wang, R., Ling, S., and Wang, S. (2017) Antibody Engineering for Pursuing a Healthier Future. Front Microbiol. 8, 495. 6. Nelson, A. L. (2010) Antibody fragments: hope and hype. mAbs 2, 77-83. 7. Braren, I., Greunke, K., Umland, O., Deckers, S., Bredehorst, R., and Spillner, E. (2007) Comparative expression of different antibody formats in mammalian cells and Pichia pastoris. Biotechnol. Appl. Biochem. 47 (Pt 4), 205-214. 8. Braren, I., Blank, S., Seismann, H., Deckers, S., Ollert, M., Grunwald, T., and Spillner, E. (2007) Generation of human monoclonal allergen-specific IgE and IgG antibodies from synthetic antibody libraries. Clin. Chem. 53, 837-844. 9. Sapparapu, G., Planque, S. A., Nishiyama, Y., Foung, S. K., and Paul, S. (2009) Antigenspecific proteolysis by hybrid antibodies containing promiscuous proteolytic light chains paired with an antigen-binding heavy chain. J. Biol. Chem. 284, 24622-24633. 10. Armacost, K., and Acevedo, O. (2014) Exploring the Aldol Reaction using Catalytic Antibodies and "On Water" Organocatalysts from QM/MM Calculations. J. Am. Chem. Soc. 136, 147-156. 11. Smirnov, I., Carletti, E., Kurkova, I., Nachon, F., Nicolet, Y., Mitkevich, V. A., Debat, H., Avalle, B., Belogurov, A. A., Jr., Kuznetsov, N., et. al. (2011) Reactibodies generated by kinetic selection couple chemical reactivity with favorable protein dynamics. Proc. Natl. Acad. Sci. U.S.A. 108, 15954-15959. 12. Coppock, M. B., Warner, C. R., Dorsey, B., Orlicki, J. A., Sarkes, D. A., Lai, B. T., Pitram, S. M., Rohde, R. D., Malette, J., Wilson, J. A., et. al. (2017) Protein catalyzed capture agents with tailored performance for in vitro and in vivo applications. Biopolymers 108 (2). doi: 10.1002/bip.22934. 13. Kim, D. Y., Hussack, G., Kandalaft, H., and Tanha, J. (2014) Mutational approaches to improve the biophysical properties of human single-domain antibodies. Biochim. Biophys. Acta 1844, 1983-2001. 19 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

14. Hugo, N., Lafont, V., Beukes, M., and Altschuh, D. (2002) Functional aspects of co-variant surface charges in an antibody fragment. Protein Sci. 11, 2697-2705. 15. Nichols, P., Li, L., Kumar, S., Buck, P. M., Singh, S. K., Goswami, S., Balthazor, B., Conley, T. R., Sek, D., and Allen, M. J., (2015) Rational design of viscosity reducing mutants of a monoclonal antibody: hydrophobic versus electrostatic inter-molecular interactions. mAbs 7, 212-230. 16. Bradbury, A. (2003) scFvs and beyond. Drug Discov. Today 8, 737-739. 17. Xiong, Y., Ford, N. R., Hecht, K. A., Roesijadi, G., and Squier, T. C. (2016) Dynamic stabilization of expressed proteins in engineered diatom biosilica matrices. Bioconjug. Chem. 27, 1205-1209. 18. Walters, B. T., Jensen, P. F., Larraillet, V., Lin, K., Patapoff, T., Schlothauer, T., Rand, K. D., and Zhang, J. (2016) Conformational destabilization of immunoglobulin G increases the low pH binding affinity with the neonatal Fc receptor. J. Biol. Chem. 291, 1817-1825. 19. Sinha, N., and Smith-Gill, S. J. (2005) Molecular dynamics simulation of a high-affinity antibody-protein complex: the binding site is a mosaic of locally flexible and preorganized rigid regions. Cell. Biochem. Biophys. 43, 253-273. 20. Monnier, P. P., Vigouroux, R. J., and Tassew, N. G. (2013) In vivo appllications of single chain Fv (variable domain) (scFv) fragments. Antibodies 2, 193-208. 21. Weller, M. G. (2016) Quality issues of research antibodies. Anal. Chem. Insights 11, 21-27. 22. Ford, N. R., Hecht, K. A., Hu, D., Orr, G., Xiong, Y., Squier, T. C., Rorrer, G., and Roesijadi, G. (2016) Antigen binding and site-directed labeling of biosilica-immobilized fusion proteins expressed in diatoms. ACS Synthetic Biology 5, 193-199. 23. Beech, B. M., Xiong, Y., Boschek, C. B., Baird, C. L., Bigelow, D. J., McAteer, K., and Squier, T. C. (2014) Controlled activation of protein rotational dynamics using smart hydrogel tethering. J. Am. Chem. Soc. 136, 13134-13137. 24. Liu, J. L., Zabetakis, D., Acevedo-Velez, G., Goldman, E. R., and Anderson, G. P. (2013) Comparison of an antibody and its recombinant derivative for the detection of the small molecule explosive 2,4,6-trinitrotoluene. Anal. Chim. Acta 759, 100-104. 25. Griffith, I. P. (1972) The effect of cross-links on the mobility of proteins in dodecyl sulphate-polyacrylamide gels. Biochem. J. 126, 553-560. 26. Ortega, A., Amoros, D., and Garcia de la Torre, J. (2011) Prediction of hydrodynamic and other solution properties of rigid proteins from atomic- and residue-level models. Biophys. J. 101, 892-898. 27. Breydo, L., Sales, A. E., Frege, T., Howell, M. C., Zaslavsky, B. Y., and Uversky, V. N. (2015) Effects of polymer hydrophobicity on protein structure and aggregation kinetics in crowded milieu. Biochemistry 54, 2957-2966. 28. Wirth, A. J., and Gruebele, M. (2013) Quinary protein structure and the consequences of crowding in living cells: leaving the test-tube behind. Bioessays 35, 984-993. 29. Mukherjee, S. K., Gautam, S., Biswas, S., Kundu, J., and Chowdhury, P. K. (2015) Do macromolecular crowding agents exert only an excluded volume effect? A protein solvation study. J. Phys. Chem. B 119, 14145-14156. 30. Candotti, M., and Orozco, M. (2016) The differential response of proteins to macromolecular crowding. PLoS Comput. Biol. 12, e1005040. 31. Cattani, G., Vogeley, L., and Crowley, P. B. (2015) Structure of a PEGylated protein reveals a highly porous double-helical assembly. Nat. Chem. 7, 823-828.


Page 20 of 24

Page 21 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


32. Clop, E. M., Chattah, A. K., and Perillo, M. A. (2014) Water and membrane dynamics in suspensions of lipid vesicles functionalized with poly(ethylene glycol)s. J. Phys. Chem. B 118, 6150-6158. 33. Nolan, V., Sanchez, J. M., and Perillo, M. A. (2015) PEG-induced molecular crowding leads to a relaxed conformation, higher thermal stability and lower catalytic efficiency of Escherichia coli beta-galactosidase. Colloids Surf. B. Biointerfaces 136, 1202-1206. 34. Matsue, S., Fujii, T., and Miyawaki, O. (2001) Effects of water activity and aqueous solvent ordering on thermal stability of lysozyme, alpha-chymotrypsinogen A, and alcohol dehydrogenase. Int. J. Biol. Macromol. 28, 343-349. 35. Wang, L., Gaigalas, A. K., and Reipa, V. (2005) Optical properties of Alexa 488 and Cy5 immobilized on a glass surface. BioTechniques 38, 127-132. 36. Iordachescu, M., and Imai, R. (2008) Trehalose biosynthesis in response to abiotic stresses. J Integr Plant Biol 50, 1223-1229. 37. Iordachescu, M., and Imai, R. (2011) Trehalose and abiotic stress in biological systems. In Abiotic Stress in Plants - Mechanisms and Adaptations, Shanker, A., and Venkateswarlu, B., Eds. INTECH: 2011. 38. Kisley, L., Serrano, K. A., Guin, D., Kong, X., Gruebele, M., and Leckband, D. E. (2017) Direct imaging of protein stability and folding kinetics in hydrogels. ACS Appl. Mater. Interfaces 9, 21606-21617. 39. Roberts, P. L., and Lloyd, D. (2007) Virus inactivation by protein denaturants used in affinity chromatography. Biologicals 35, 343-347. 40. Glatter, T., Ahrne, E., and Schmidt, A. (2015) Comparison of different sample preparation protocols reveals lysis buffer-specific extraction biases in gram-negative bacteria and human cells. J. Proteome Res. 14, 4472-4485. 41. Vidic, J., Manzano, M., Chang, C. M., and Jaffrezic-Renault, N. (2017) Advanced biosensors for detection of pathogens related to livestock and poultry. Vet. Res. 48, 11. doi: 10.1186/s13567-017-0418-5. 42. Chan, J. F., Sridhar, S., Yip, C. C., Lau, S. K., and Woo, P. C. (2017) The role of laboratory diagnostics in emerging viral infections: the example of the Middle East respiratory syndrome epidemic. J. Microbiol. 55, 172-182. 43. Van Audenhove, I., and Gettemans, J. (2016) Nanobodies as versatile tools to understand, diagnose, visualize and treat cancer. EBioMedicine 8, 40-48. 44. Butler, D. (2014) Ebola experts seek to expand testing. Nature 516, 154-155. 45. Goldman, E. R., Egge, A. L., Medintz, I. L., Lassman, M. E., and Anderson, G. P. (2005) Application of a homogenous assay for the detection of 2,4,6-trinitrotoluene to environmental water samples. The Scientific World Journal 5, 446-451. 46. Allen, M. W., Urbauer, R. J., Zaidi, A., Williams, T. D., Urbauer, J. L., and Johnson, C. K. (2004) Fluorescence labeling, purification, and immobilization of a double cysteine mutant calmodulin fusion protein for single-molecule experiments. Anal. Biochem. 325, 273-284. 47. Yao, Y., Schoneich, C., and Squier, T. C. (1994) Resolution of structural changes associated with calcium activation of calmodulin using frequency domain fluorescence spectroscopy. Biochemistry 33, 7797-7810. 48. Strasburg, G. M., Hogan, M., Birmachu, W., Thomas, D. D., and Louis, C. F. (1988) Sitespecific derivatives of wheat germ calmodulin. Interactions with troponin and sarcoplasmic reticulum. J. Biol. Chem. 263, 542-548.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

49. Boschek, C. B., Squier, T. C., and Bigelow, D. J. (2007) Disruption of interdomain interactions via partial calcium occupancy of calmodulin. Biochemistry 46, 4580-4588. 50. Boschek, C. B., Sun, H., Bigelow, D. J., and Squier, T. C. (2008) Different conformational switches underlie the calmodulin-dependent modulation of calcium pumps and channels. Biochemistry 47, 1640-1651. 51. King, W. J., Pytel, N. J., Ng, K., and Murphy, W. L. (2010) Triggered drug release from dynamic microspheres via a protein conformational change. Macromol. Biosci. 10, 580-584. 52. Murphy, W. L., Dillmore, W. S., Modica, J., and Mrksich, M. (2007) Dynamic hydrogels: translating a protein conformational change into macroscopic motion. Angewandte Chemie 46, 3066-3069. 53. King, W. J., Toepke, M. W., and Murphy, W. L. (2011) Facile formation of dynamic hydrogel microspheres for triggered growth factor delivery. Acta Biomater. 7, 975-985. 54. Xiong, Y., Chen, B., Smallwood, H. S., Urbauer, R. J., Markille, L. M., Galeva, N., Williams, T. D., and Squier, T. C. (2006) High-affinity and cooperative binding of oxidized calmodulin by methionine sulfoxide reductase. Biochemistry 45, 14642-14654. 55. Muller, C. B., Loman, A., Pacheco, V., Koberling, F., Willbold, D., Richtering, W., and Enderlein, J. (2008) Precise measurement of diffusion by multi-color dual-focus fluorescence correlation spectroscopy. Europhysics Letters 83, 46001. 56. Gratton, E., and Limkeman, M. (1983) A continuously variable frequency cross-correlation phase fluorometer with picosecond resolution. Biophys. J. 44, 315-324. 57. Lakowicz, J. R., and Maliwal, B. P. (1985) Construction and performance of a variablefrequency phase-modulation fluorometer. Biophys. Chem. 21, 61-678. 58. Bevington, P. R. (1969) Data Reduction and Error Analysis for the Physical Sciences. McGraw-Hill: New York. 59. Hunter, G. W., and Squier, T. C. (1998) Phospholipid acyl chain rotational dynamics are independent of headgroup structure in unilamellar vesicles containing binary mixtures of dioleoyl-phosphatidylcholine and dioleoyl-phosphatidylethanolamine. Biochim. Biophys. Acta 1415, 63-76. 60. Beechem, J. M., and Haas, E. (1989) Simultaneous determination of intramolecular distance distributions and conformational dynamics by global analysis of energy transfer measurements. Biophys. J. 55, 1225-1236. 61. Celikel, R., Peterson, E. C., Owens, S. M., and Varughese, K. I. (2009) Crystal structures of a therapeutic single chain antibody in complex with two drugs of abuse-Methamphetamine and 3,4-methylenedioxymethamphetamine. Protein Sci. 18, 2336-2345. 62. Arnold, K., Bordoli, L., Kopp, J., and Schwede, T. (2006) The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling. Bioinformatics 22, 195-201. 63. Biasini, M., Bienert, S., Waterhouse, A., Arnold, K., Studer, G., Schmidt, T., Kiefer, F., Gallo Cassarino, T., Bertoni, M., Bordoli, L., et. al. (2014) SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information. Nucleic Acids Res. 42, W252W258. 64. Kiefer, F., Arnold, K., Kunzli, M., Bordoli, L., and Schwede, T. (2009) The SWISSMODEL Repository and associated resources. Nucleic Acids Res. 37, D387-D392. 65. Dorsey, N. E. (1940) Properties of Ordinary Water Substance. Reinhold: New York. 66. Lakowicz, J. R. (1999) Principles of Fluorescence Spectroscopy. Kluwer Academic/Plenum Publishers.


Page 22 of 24

Page 23 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


TABLE OF CONTENTS GRAPHIC (TOC)

TOC Figure: Hydrogels Stabilize Interdomain Coupling in Single Chain Variable Fragment (scFv) Antibodies. Molecular crowding stabilizes structural coupling between VL and VHdomains to enhance ligand binding under environmental conditions normally associated with protein denaturation (e.g., the presence of chaotropic agents).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment

Page 24 of 24

Hydrogel Tethering Enhances Interdomain ... - ACS Publications

Recommend Documents