Unusual Self-Assembly of the Recombinant Chlamydia trachomatis Major Outer Membrane Protein–Based Fusion Antigen CTH522 Into Protein Nanoparticles

Accepted Manuscript Unusual Self-Assembly of the Recombinant Chlamydia trachomatis Major Outer Membrane Protein (MOMP)-based Fusion Antigen CTH522 into Protein Nanoparticles Fabrice Rose, Kasper Karlsen, Pernille Rønde Jensen, Rasmus Uffe Jakobsen, Grith Krøyer Wood, Kasper Dyrberg Rand, Helene Godiksen, Peter Andersen, Frank Follmann, Camilla Foged PII:

S0022-3549(18)30087-X

DOI:

10.1016/j.xphs.2018.02.005

Reference:

XPHS 1077

To appear in:

Journal of Pharmaceutical Sciences

Received Date: 29 August 2017 Revised Date:

21 January 2018

Accepted Date: 6 February 2018

Please cite this article as: Rose F, Karlsen K, Jensen PR, Jakobsen RU, Wood GK, Rand KD, Godiksen H, Andersen P, Follmann F, Foged C, Unusual Self-Assembly of the Recombinant Chlamydia trachomatis Major Outer Membrane Protein (MOMP)-based Fusion Antigen CTH522 into Protein Nanoparticles, Journal of Pharmaceutical Sciences (2018), doi: 10.1016/j.xphs.2018.02.005. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Unusual Self-Assembly of the Recombinant Chlamydia trachomatis Major Outer Membrane Protein (MOMP)-based Fusion Antigen CTH522 into Protein Nanoparticles

RI PT

Fabrice Rose1, Kasper Karlsen2, Pernille Rønde Jensen1,2, Rasmus Uffe Jakobsen1, Grith Krøyer Wood2, Kasper Dyrberg Rand1, Helene Godiksen2, Peter Andersen3, Frank Follmann3 and

1

SC

Camilla Foged1

Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen

2

M AN U

Universitetsparken 2, DK-2100 Copenhagen Ø, Denmark Department of Vaccine Development, Statens Serum Institut, Artillerivej 5, DK-2300 Copenhagen S, Denmark 3

Department of Infectious Disease Immunology, Statens Serum Institut, Artillerivej 5, DK-2300

TE D

Copenhagen S, Denmark

EP

To whom correspondence should be addressed: Assoc. Prof. C. Foged, Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, DK-

AC C

2100 Copenhagen Ø, Denmark, Telephone: + 45 35336402; FAX: + 45 35336001; E-mail: [email protected]

Keywords: Vaccines; vaccine delivery; protein formulation; protein structure; protein folding/refolding; protein aggregation; preformulation; mass spectrometry; self assembly.

1

ACCEPTED MANUSCRIPT

Abstract Sexually transmitted Chlamydia trachomatis (Ct) infects more than 100 million people annually, and untreated chlamydia infections can cause severe complications. Therefore, there is an urgent

RI PT

need for a chlamydia vaccine. The Ct major outer membrane protein (MOMP) is highly immunogenic but is a challenging vaccine candidate by being an integral membrane protein, and the immunogenicity depends on a correctly folded structure. We investigated the biophysical

SC

properties of the recombinant MOMP-based fusion antigen CTH522, which is tested in early human clinical trials. It consists of a truncated and cysteine-free version of MOMP fused to four

M AN U

variable domains from serovars D-G. In the native state, CTH522 did not exist as a monomer, but showed an unusual self-assembly into nanoparticles with a negative zeta-potential. In contrast to the β-barrel structure of MOMP, native CTH522 contained no well-defined secondary structural elements, and no thermal transitions were measurable. Chemical unfolding resulted

TE D

monomers that upon removal of the denaturant self-assembled into higher order structures, comparable to the structure of the native protein. The conformation of CTH522 in nanoparticles is thus not entirely random and contains structural elements stabilized via denaturant-disruptable

EP

hydrophobic interactions. In conclusion, CTH522 has an unusual quaternary structure of

AC C

supramolecular self-assemblies.

2

ACCEPTED MANUSCRIPT

Chlamydia, caused by infection with Chlamydia trachomatis (Ct), is the most common sexually transmitted bacterial disease in the world.1 It has a high incidence and prevalence globally, especially among young and sexually active individuals, with an estimated 100 million new

RI PT

cases annually.2 Due to the often asymptomatic nature of infection with Ct, the disease remains undiagnosed in many cases for a long period until severe complications appear, including pelvic inflammatory disease, chronic abdominal pain, infertility and ectopic pregnancy.3 Despite the

SC

fact that increased diagnostic efforts and antibiotics treatment have resulted in a higher detection frequency, the spread of Ct has not been reduced, as illustrated by the growing number of

M AN U

incidences in many western countries.2 Chlamydia thus represents a health as well as an economic burden to society. Currently, no effective vaccine exits for Chlamydia, which emphasizes the unmet medical need for an efficacious and safe vaccine-based prevention strategy.

TE D

Ct is an obligate intracellular pathogen, and it is well established that a vaccine against Ct needs to elicit protective B- and T-cell mediated immunity in the genital tract mucosa.4,5 In addition, different serovars of Ct exist, which calls for a cross-protective vaccine. Despite years

EP

of research, it has proven challenging to develop vaccines against Chlamydia. Due to the observation of undesired side effects for traditional vaccines based on whole organisms, the

AC C

subunit vaccine technology has recently become the preferred approach.6 Highly purified, recombinant antigens are used for subunit vaccines, and they are often co-administered with adjuvants to potentiate immunogenicity. In addition, subunit vaccines are combined with immunization strategies that induce high levels of local genital immunity, e.g., mucosal vaccination.7,8

3

ACCEPTED MANUSCRIPT

A number of different surface-exposed antigens have been identified, which are recognized by the immune system during a Ct infection. In particular, the major outer membrane protein (MOMP) is the most well-studied protein, and immunization with MOMP has been shown to

RI PT

protect mice against a subsequent Ct challenge9,10 and to confer partial protection in non-human primates.11 MOMP belongs to a family of proteins found in the outer membranes of Gramnegative bacteria that have a molecular mass of approx. 40 kDa and function as porins.12 Porins

SC

have a general 16-stranded β-barrel structure with an interior membrane-spanning pore. Half of the side chains of the β-barrel are facing the interior of the channel and are very polar, whereas

M AN U

the other half of the side chains are facing the membrane and are very nonpolar. In addition, MOMP has disulfide bonds to other proteins, forming a larger membrane complex. This poses a major challenge in the formulation of MOMP for clinical vaccination purposes, because membrane proteins are notoriously difficult to produce recombinantly at a large scale due to the

TE D

requirement for a correctly folded structure.13 MOMP consists of five constant domains and four surface-exposed variable domains (VDs).14,15 These VDs constitute the molecular basis of the serovar grouping of Ct,16 and they are rich in neutralizing antibody epitopes.15,17 Native MOMP

EP

isolated directly from Ct has been used successfully for vaccination in preclinical animal models.9-11 Although it is highly immunogenic and induces neutralizing antibodies, these

AC C

antibodies are strain-specific, as observed following natural infection, and thus only confer homotypic protection.11 Recombinantly expressed versions of MOMP18 and DNA vaccines encoding MOMP19 have so far been tested with limited success. We recently designed the novel antigen CTH5227 based on a truncated and cysteine-free version of MOMP (amino acids 34-259) fused to four VDs covering neutralizing epitopes from serovars D-G (Supplementary data, Figure S1). Each of these four VDs contains the highly

4

ACCEPTED MANUSCRIPT

conserved species-specific LNPTIAG epitope with flanking serotype-specific regions.15,17 In addition, all reactive cysteine residues in the protein20 have been replaced with serine residues to overcome the purification challenges related to disulfide bond formation. In a recent study, this

RI PT

construct adjuvanted with the cationic adjuvant formulation 1 [CAF01, Statens Serum Institut (SSI), Copenhagen, DK] was shown to stimulate protective immunity in a mouse Ct challenge model.7 CTH522 has entered a phase I first in human trial during 2016 (NCT ID:

SC

NCT02787109). The trial is a double-blind, parallel, randomized and placebo-controlled clinical trial in healthy women aged 18 to 45 years, designed to evaluate the safety of the adjuvanted

M AN U

CTH522 vaccine. The main objective of the current study was to characterize the biophysical properties of the purified CTH522 antigen in solution.

Experimental

TE D

Expression and purification of CTH522

The recombinant fusion protein CTH522 (Supplementary data, Figure S1) was produced in an Escherichia coli expression system (Supplementary data). A Tris buffer was chosen for the

EP

present studies, because the protein in this buffer is used for the clinical study due to optimal stability of CAF01 in Tris buffer. In addition, the choice of buffer was based on a screening

AC C

experiment including 18 different buffers (result not shown).

SEC-HPLC

A Dionex UltiMate 3000 with autosampler (Thermo Fisher Scientific, Waltham, MA, USA) was used for the analysis of CTH522 by SDS size exclusion (SEC)-HPLC. The column was a Yarra 3 µm SEC 3000, 290 Å pore size (Phenomenex, Torrance, CA, USA). The mobile phase was 20

5

ACCEPTED MANUSCRIPT

mM Tris, 2 % (w/v) SDS (pH 7.4). The flow was 0.6 ml/min. The CTH522 standard stored below -70°C was used as calibration standard and injected in volumes of 5, 10, 20, 30 and 35 µl, respectively, to construct a standard curve. The matrix was 25 mM Tris + 10 % (v/v) glycerol.

RI PT

The standard, matrix and samples, respectively, were mixed 1:1 (v/v) with mobile phase to denature the protein. A volume of 20 µl was injected, and the elution time was 30 min. The protein was detected by UV at 280 nm. Data analysis was conducted using the Chromeleon

SC

Chromatography Data System 7.2 software (Thermo Fisher Scientific).

M AN U

Mass spectrometry

Electrospray ionization MS (ESI-MS) coupled to reversed-phase UPLC was used to determine the intact mass of the fusion protein. A sample (90 pmol) of CTH522 in 2.5 mM Tris and 1% glycerol was analysed in duplicate using a customized UPLC-MS system consisting of two

TE D

UPLC pumps (nanoACQUITY ASM and BSM, Waters, Milford, MA, USA) coupled to a Synapt G2Si ESI-QTOF MS through two 6 port-switching valves. The switching valves contained an injection port, a 100 µl sample loop and a reversed-phase trap column (VanGuard

EP

BEH C4 300 Å, Waters). The protein sample was injected via the sample loop and eluted across the trap column and desalted for 4 min at 300 µl/min with Solvent A (MilliQ-water + 0.23%

AC C

formic acid). The protein was subsequently eluded using a gradient from 8 to 95% Solvent B (MeCN + 0.23% formic acid) over 15 min. Acquisition of MS data was done in positive ionization mode using a scan range of 50-2000 m/z and a scan time of 1 s (Waters). The performance and accuracy of the LC-MS analysis was verified using an in-house reference protein of higher mass (71840.7 Da), indicating a mass accuracy of approx. 20 ppm. The mass of the reference protein and CTH522 was calculated from the acquired mass spectra using the

6

ACCEPTED MANUSCRIPT

MaxEnt 1 algorithm of the MassLynx software (Waters) applying the following processing parameters: full width at half maximum 0.550, resolution 1 Da. Peptide mapping experiments were performed by diluting 400 pmol (43.15 µl) CTH522 into 45 µl denaturation buffer [6 M

RI PT

guanidine hydrochloride (GdnHCl), 50 mM NH4CO3, pH 8] and incubation at 60°C for 30 min. Following incubation, 7.7 µl 650 mM CaCl2 was added, after which the solution was diluted with 94.9 µl 50 mM NH4CO3 pH 8. To this mixture, 9.26 µl 5 µM trypsin was added

SC

(substrate:enzyme molar ratio 1:20), to a final volume of 200 µl, and the digestion was performed overnight at 37°C. The sample was loaded onto the LC-MS system described above,

M AN U

however, equipped with a C18 reverse phase trap (VanGuard BEH C418 130 Å). Following trapping for 3 min using a flow of 100 µl/min Solvent A, peptides were eluded onto an analytical column (Waters ACQUITY UPLC BEH C18 Column, 130 Å, 1.7 µm, 1 mm x 100 mm) and separated using a gradient from 8 to 40% Solvent B over 35 min prior to ESI-MS detection. The

TE D

peptides were analyzed using a Synapt G2 ESI-QTOF in positive ionization mode using a scan range of 50-2000 m/z and a scan rate of 1 s (Waters). Peptide identification was done using

EP

PLGS 3.0 (Waters) and data filtration using Dynamx 3.0 (Waters).

Hydrodynamic diameter, zeta-potential and morphology

AC C

The average intensity-based hydrodynamic diameter (dH) and polydispersity index (PDI) of the protein (0.49 mg/ml) in Tris buffer were determined by dynamic light scattering (DLS) using photon correlation spectroscopy. The surface charge was estimated by analysis of the zetapotential (laser-Doppler electrophoresis). The measurements were performed at 25ºC by using a Zetasizer Nano ZS (Malvern Instruments, Worcestershire, UK) equipped with a 633 nm laser and 173º detection optics. For viscosity and refractive index, the values of Tris buffer with 10%

7

ACCEPTED MANUSCRIPT

glycerol were used. Malvern DTS v.7.03 software was used for data acquisition and analysis. A cumulants fitting analysis of the autocorrelation function G(τ) of the scattered intensity was used to calculate the average dH (z-average) using the Stokes-Einstein equation. The particle size

RI PT

distribution was reflected in the PDI, which ranges from 0 for a monodisperse to 1.0 for an entirely heterodisperse dispersion. All measurements were taken in triplicates. For the heat scans, the protein samples were heated from 25°C to 70°C within 5°C steps and 10 min of equilibration

SC

time between each step (size calculations were corrected for viscosity and temperature). Morphological analysis was performed by cryo-transmission EM (cryo-TEM) as previously

M AN U

described 21.

CD

Far UV CD measurements were performed in a 0.1 cm quartz cell using a Chirascan

TE D

spectrophotometer (Applied Photophysics, Surrey, UK). The protein concentration was 0.07 and 0.2 mg/ml, respectively, in Tris buffer. All spectra were recorded from 198 to 260 nm at a fixed bandwidth of 0.5 nm with 1 s per point and a step of 1 nm. The measurements were performed at

EP

room temperature (rt) and at specified temperatures in the range of 20-80°C. All spectra were an average of three scans, and the spectra were background-corrected and transformed into molar

AC C

ellipticity (θ) on a per residue basis. The shown spectra have not been smoothed. All measurements were taken in duplicates.

DSC

The thermal stability of the protein was determined by differential scanning calorimetry (DSC), which was performed using a NanoDSC (TA Instruments, Lindon, UT, USA) with a cell volume

8

ACCEPTED MANUSCRIPT

of 299 µl. Thermograms were recorded in the temperature range of 20-90°C at a scan rate of 1°C/min, a cell pressure of 3 atm and an equilibrium time of 600 s between each scan. All samples and references were degassed for 15-20 min before loading. Buffer-buffer scans of Tris

RI PT

buffer were completed prior to loading of the protein solution (0.49 mg/ml) in the sample cell.

Intrinsic fluorescence and SLS

SC

The tertiary and quaternary structures of the protein were studied further by analyzing the intrinsic fluorescence and the static light scattering (SLS) of the protein using Avacta Optim

M AN U

1000 (Avacta Analytical LTd, York, UK). Optim-compatible microcuvette arrays were loaded with 9 µl protein solution (0.49 mg/ml). The samples were excited with 266 nm and 473 nm laser beams, respectively, for 1000 ms with slits of 100 µm. Fluorescence and SLS signals were measured as a function of temperature in the range of 20-90°C at a rate of 0.3°C/min, and a hold

TE D

time of 10 s/well. The spectra were recorded from 240 to 490 nm. Peaks at 266 and 473 nm are specific to the SLS. The changes in the fluorescence were depicted as the barycentric mean

EP

(BCM), which was calculated according to Equation 1:

∑ 

∑ 

AC C


 λ =



Equation 1

where λ is the wavelength, and I is the intensity. The measurements were performed on three different wells and were repeated twice.

Extrinsic fluorescence

9

ACCEPTED MANUSCRIPT

A concentration of 20 µM of 8-anilino-1-naphthalene sulfonate (ANS) in Tris buffer was added to a protein solution at 0.49 mg/ml (final concentrations) and equilibrated for 30 min before the measurements. The 266 nm laser beam of the Optim (for 1000 ms with slits of 100 µm) was used

RI PT

for excitation of ANS, and a spectrum from 290 to 543 nm was recorded in the temperature range of 20-90oC at a rate of 0.3oC/min, and a hold time of 10 s/well. The intensity at 470 nm was plotted as function of temperature. A control with ANS alone was measured, and no

SC

significant influence on the spectrum was found (results not shown). The measurements were

M AN U

performed on three different wells and were repeated twice.

Chemical unfolding

The protein was unfolded with GndHCl at 25oC by sequential dilution of the protein [at a fixed concentration of 0.335 mg/ml (6.2 µM)] with Tris buffer containing increasing concentrations of

TE D

GndHCl (0-6 M, final concentration). A 7 M GndHCl stock solution in Tris buffer was prepared for the dilutions. The GndHCl concentration was determined by refractometry. The refractive index was measured by using an RL3 refractometer (Nr. 28046/01, PZO Warszawa, Warsaw,

EP

PL). The concentration of GndHCl was calculated by fitting to a polynomial function.22 The intrinsic fluorescence of the protein was measured using the Avacta Optim 1000, as described

AC C

above. The BCM value was plotted as a function of the GndHCl concentration to obtain the protein denaturation curve. The pre and post baselines were fitted by linear regression; the intercepts give the baseline values of the parameter at low denaturant concentration (y0) and the maximum value at complete denaturation (ymax), respectively. From the measured parameters and their calculated baseline and maximum values, the fraction of protein in the denatured state (fD)

10

ACCEPTED MANUSCRIPT

as a function of the denaturant concentration was calculated, assuming a simple two-state model for the denaturation process, according to Equation 2:

  

  



Equation 2

RI PT

 = 

Equation 3:   



Equation 3

M AN U

 =

SC

The apparent equilibrium constant KD for the denaturation process was calculated according to

From the KD, the ∆GD was calculated from Equation 4:

TE D

∆ = − ln Equation 4

where R is the gas constant and T is the absolute temperature. The ∆G(H2O) was estimated by

EP

linear extrapolation to zero concentration of GndHCl 23. A least squares analysis was used to fit the data to Equation 5:

AC C

∆ = ∆H$ O − &[GndHCl] Equation 5

where m is a measure of the dependence of the free energy on the GndHCl concentration, and depends on the amount and composition of the polypeptide chain that is freshly exposed to solvent upon unfolding.

Refolding by dialysis 11

ACCEPTED MANUSCRIPT

The protein was first unfolded by diluting the protein at 0.4 mg/ml into 6 M GndHCl, followed by incubation for 1 h. The refolding of the unfolded protein was performed by sequential dialysis for 24 h at 4°C in 10 kDa molecular mass cut off Slide-A-Lyzer dialysis cassettes (Pierce, IL,

RI PT

USA). The protein was dialyzed sequentially against a 300 times larger volume of Tris buffer containing (i) decreasing concentrations of GndHCl (2, 1, 0.5 and 0 M), and (ii) different additives [9 µM of the nonionic surfactant Tween 80, 150 mM NaCl, and 70 µM of the nonionic

SC

surfactant octaethylene glycol monododecyl ether (E8C12), respectively]. Samples were taken at each step and analyzed by DLS, SLS and intrinsic fluorescence, as described above. During the

M AN U

entire refolding process, the protein concentration was measured as described above.

Statistics

Statistical calculations were performed by using GraphPad Prism 6 (GraphPad Software, Inc., La

TE D

Jolla, CA, USA) by a one-way analysis of variance at a 0.05 significance level followed by means comparison by applying the Tukey’s test.

EP

Results Production of CTH522

AC C

The primary structure of the CTH522 protein consists of 502 amino acids (Supplementary data, Figure S1), and the theoretical mass of the protein is 53912.6 Da. The clinical CTH522 batch of good manufacturing practice (GMP) quality, produced in an E. coli expression system, was comparable to an in-house standard (control batch), which is a well-characterized batch based on the first pilot-scale production of CTH522, as determined by reducing SDS-PAGE with the main band appearing at approx. 54 kDa (Figure 1A). Employing native, non-reducing PAGE, only a

12

ACCEPTED MANUSCRIPT

smear was observed above the top protein marker (198 kDa), suggesting that the protein exists as higher order self-assemblies (results not shown). The identity was verified by western blotting using anti-CTH522 antibodies (Figure 1B). Additional weak bands were also apparent from the

RI PT

western blotting, with the largest molecular-sized band indicating the presence of CTH522 dimers or aggregated forms, and the weak bands of a size below 54 kDa representing degradation products. SDS-SEC-HPLC analyses showed that the vast majority of the protein eluted as a

SC

monomer (Figure 1C), suggesting that the higher order self-assemblies are, at least partly, held together by SDS-disruptable non-covalent interactions. Based on host cell protein analyses by

M AN U

western blotting, the relative quantity of CTH522 protein was higher than 95% (results not shown).

The mass and the identity of a GMP batch of CTH522 was verified using ESI-MS (Figure 2A).

TE D

The MS analyses indicated a relative abundance of three major forms of CTH522 (53914, 53927 and 53940 Da) of 25%, 100% and 28%, respectively, for this specific GMP batch (Figure 2A). While the 53914 Da species, which correlates with the theoretical mass of unmodified CTH522

EP

(53912.7 Da), is present in the sample, it is not the most abundant species. However, the relative abundances determined by such semi-quantitative analyses varied from batch to batch (results

AC C

not shown). The nature of the modifications found in the two higher mass forms detected in the CTH522 sample was examined further by peptide mapping (results not shown). No modifications were found in the N- and C-terminal parts of the protein, which were intact. This shows that the covalent modifications (+13 Da and +26 Da), which give rise to the two differently sized isoforms of the protein, are located to residues of the protein not covered by the peptide mapping analysis. However, further analyses are needed to determine the exact location

13

ACCEPTED MANUSCRIPT

and identity of these modifications, which are beyond the scope of the present study. In addition, the MS analyses showed the presence of a minor impurity, which was found to have a mass of 16093 Da, likely corresponding to a protein from the E. coli host cells used for recombinant

CTH522 exists as a particle structure in the native state

SC

abundance, relative to the abundance of CTH522.

RI PT

expression of CTH522 (Figure 2B). MS data strongly indicated that this impurity was of low

Measurements of the average dH by DLS revealed that the protein does not exist in a monomeric

M AN U

form in the native state, but rather as supramolecular self-assemblies with an average dH of 55.2 ± 6.5 nm [mean value ± standard deviation (SD), n = 3] and a PDI of 0.37 ± 0.03 (Figure 3A) suggesting a relatively narrow range of species. In general, DLS measurements are very sensitive to the presence of larger submicron particles, which are characterized by high scattering

TE D

intensity. The slope and decay of G(τ) suggest the absence of such submicron aggregates detectable with DLS that could otherwise cause multiple scattering (Figure 3B). The welldefined baseline at high decay times and the y-axis intercept suggest a low signal-to-noise ratio.

EP

The self-assembled structures were visualized by cryo-TEM imaging, and solid ball-like structures were found of an average size corresponding well to the size measured by photon

AC C

correlation spectroscopy (Figure 3C). The shape of the particles appeared spherical with an irregular surface morphology. The zeta-potential was -22.2 ± 5.4 mV (mean value ± SD, n = 3, results not shown).

Conformation of CTH522

14

ACCEPTED MANUSCRIPT

The secondary structure was examined using CD. In contrast to the well-defined β-sheet structure measured for MOMP,13,24 no characteristic structural elements could be identified for CTH522, and the secondary structure was poorly defined (Figure 4A). Concentrating CTH522 to

RI PT

concentrations above approximately 10 mg/ml resulted in protein aggregation (results not shown). This excludes the use of Fourier-transform infrared spectroscopy for further analyses of the secondary structure due to the requirement for high protein concentrations for this type of

SC

analysis. A number of different techniques, including DSC, CD, intrinsic fluorescence, extrinsic fluorescence, SLS and DLS, were used to determine the thermodynamic and structural properties

M AN U

of the protein as a function of temperature. Unexpectedly, no transition could be measured by using any of the methods mentioned above (Figures 4B-H): The DSC thermogram did not show any peaks (Figure 4B), confirming that the protein does not undergo any measurable thermal transitions. No transition could be measured neither using CD in the examined wavelength range

TE D

(Figures 4C-D), suggesting the lack of characteristic secondary structural elements of the protein monomers. For the fluorescence measurements, there was a gradual shift as a function of temperature (Figure 4E): Intrinsic fluorescence was used as a measure of the environment of

EP

aromatic amino acid residues in the protein. The protein has eight tryptophan residues that fluoresce (Supplementary data, Figure S1). In addition, the fluorescence of the 13 tyrosine

AC C

residues and 23 phenylalanine residues of the protein contribute to a minor extent to the intrinsic fluorescence spectrum at the lower emission wavelengths due to the low excitation wavelength (266 nm) of the Optim used for the measurements. The high initial peak wavelength at 20°C [BCM peak position at approx. 347 nm] suggests that the tryptophan residues on average are exposed to the aqueous environment (Figure 4E). A slight red shift was observed as a function of temperature (Figure 4E), eventually resulting in a BCM peak position of approx. 353 nm at

15

ACCEPTED MANUSCRIPT

90°C. This suggests that the tryptophan residues on average become slightly more exposed to the aqueous environment at higher temperatures. The fluorescent probe ANS becomes highly fluorescent in apolar organic solvents, or upon adsorption to solid phases. Thus, interaction of

RI PT

ANS with hydrophobic binding sites of the protein results in an increase in the fluorescence intensity.25 At 20°C, the extrinsic ANS fluorescence intensity was already high (Figure 4F) due to a high binding affinity of ANS for the protein (hydrophobic pockets or aggregates). This

SC

suggests that a significant part of the surface area is apolar and is surface-exposed, even at low temperature. With temperature, a natural and roughly linear decrease in the fluorescence

M AN U

intensity was observed, suggesting a simple thermal quenching of the excited state. From the light scattering intensity as a function of temperature, a decrease could be observed (Figure 4G) that might be explained as a dissociation of the self-assemblies at increasing temperature, which was also observed by DLS measurements (Figure 4H). Thus, the average dH decreased from

TE D

approx. 55 nm at 20°C to approx. 35 nm at 65°C (after 10 min of equilibrium). However, the well-known decrease in pH of the Tris buffer as a function of temperature (approximately 0.3 pH

EP

units for every 10°C) cofounds data interpretation at the higher temperatures (Figures 4B-H).

Guanidine-induced unfolding into monomeric CTH522

AC C

Guanidine-induced unfolding was measured to gain more information about this rather unusual conformational stability of the protein. The unfolding process was observed as a change in the intrinsic fluorescence (red shift) as a function of the GndHCl concentration (Figure 5A). A clear shift to a higher wavelength of the maximum peak was observed, and a shoulder appeared at a lower wavelength (at approx. 300 nm) due to emission from tyrosine and phenylalanine residues. This suggests that that the aromatic amino acid residues are exposed to the aqueous environment

16

ACCEPTED MANUSCRIPT

due to complete protein unfolding. Assuming a two-state unfolding process, an unfolding curve was constructed by plotting the fraction of unfolded protein (fD) as a function of the GndHCl concentration (Figure 5B). The ∆GD was calculated for each of the tested concentrations of

RI PT

GndHCl. From the extrapolation of the linear fit to zero concentration of GndHCl, the ∆G(H2O) of CTH522 was estimated to be 17.3 kJ mol-1 (Figure 5C, intersection with the y-axis), representing the conformational stability of the protein. The m value was 10.3 kJ mol-1 M-1

SC

(Figure 5C, the slope of the line fit). The midpoint of the denaturation curve (∆G = 0) was found at a GndHCl concentration of 1.7 M (Figure 5C, intersection with the x-axis). The intrinsic

M AN U

fluorescence spectrum for the protein denatured with GndHCl showed a clear difference compared to the fluorescence spectrum recorded after heating (Figure 5D). In contrast to the spectrum of the GndHCl-induced denatured protein (Figure 5D, black curve), the spectrum of the temperature-induced denatured protein decreased in intensity (Figure 5D, red curve), but overall,

TE D

the two spectra had a similar appearance.

Unfolded CTH522 refolds into nanoparticulate structures

EP

Refolding studies were conducted in order to improve the understanding of the conformation of the self-assembled structures. A dilution-based refolding approach was applied, and the

AC C

denatured protein was dialyzed in different dilution steps using Tris buffer supplemented with GndHCl at concentrations around the midpoint value of the unfolding curve (2, 1, 0.5 and 0 M GndHCl). The photon correlation spectroscopy technique was used to monitor the size at the different dialysis steps. As expected, the scattering intensity was low due to dissociation and unfolding of CTH522 at 2 M GndHCl (Table 1). At 1 M, the average dH stabilized at approx. 40 nm, and below this concentration, the size stabilized at approx. 50 nm. Upon refolding, the

17

ACCEPTED MANUSCRIPT

protein existed in a multimeric form. The average dH after complete refolding (58.1 ± 3.9 nm) was slightly higher than the initial average dH (55.2 ± 6.5 nm). Different excipients were added to test if refolding could be improved and the monomeric conformation of the protein could be

RI PT

stabilized. The different conditions were (i) refolding in the presence of two different surfactants commonly used for protein formulation and membrane protein solubilization [Tween 80 and E8C12], and (ii) refolding at high ionic strength by addition of salt (150 mM NaCl). When these

SC

conditions were tested, none of them could improve the final state of the protein (Table 1). Thus in the presence of salt and the E8C12 surfactant, the size was significantly higher (103.7 and

M AN U

276.0 nm, respectively). A similar behavior was observed for the intrinsic fluorescence and the SLS measurements (Figure 6 and 7). At high concentration (2 and 6 M GndHCl), the initial BCM had a high value (above approx. 355 nm), confirming the presence of the unfolded state of the protein at these two concentrations, compared to the lower concentrations (from 0 to 1 M),

TE D

where the BCM values where more close to approx. 345 nm. At 0.5 and 1 M, the SLS measurements showed an increase of aggregation with temperature for all tested conditions.

EP

Discussion

In this study, the recombinant fusion protein antigen CTH522 was expressed in E.coli and

AC C

purified. The manufacturing process complies with current GMP. Unexpectedly, the protein was shown to self-assemble into homogeneous higher-order spherical and compact particulate structures with a mean dH of approx. 55 nm. For comparison, the monomer is expected to have a dH of approx. 5-7 nm, calculated using empirical relationships and assuming a globular conformation of the protein.26-28 It is expected that these nanoparticle structures are formed during the expression and purification of the protein. They are probably stabilized via

18

ACCEPTED MANUSCRIPT

hydrophobic interactions, and hydrophobic interactions are likely the main driving force for particle formation for the reasons discussed further below. These data suggest that the protein exists entirely as a relatively narrow distribution of higher-order structures.

RI PT

The particles had a high negative zeta-potential (-22.2 mV). The calculated theoretical pI value is 4.97, confirming that the protein has a net negative surface charge at pH 9.0. The net negative zeta-potential may be attributed to surface-exposure of the relatively high content of

SC

acidic residues (48 aspartic acid and glutamic acid residues, 9.6%, Supplementary data, Figure S1). This high negative zeta-potential also indicates that the protein nanoparticles may be

M AN U

colloidally stable, with repulsive forces existing between the nanoparticles and preventing their aggregation.29,30 The high negative zeta-potential implies that the protein nanoparticles may adsorb strongly to cationic adjuvants, e.g., CAF01, facilitated mainly by attractive electrostatic interactions.31

TE D

No thermal transitions were observed for CTH522, suggesting that the protein has less tertiary structure than typical globular proteins. Using CD, we did not identify any characteristic secondary structural elements in the protein. This result might be a consequence of the

EP

particulate structure, as CD is a global low-resolution method usually applied to measure the secondary structure of proteins in solution, either as monomers or as homogenous multimeric

AC C

states. Therefore, we cannot exclude that the measured lack of secondary structural elements might be a consequence of the particulate nature of the protein. However, the CD data may also suggest that the protein does not exist in a porin conformation. This is in contrast to the welldefined β-strand structure measured for MOMP, which shares many biochemical properties with the classical porin proteins.13,24 This may be a result of the replacement of the cysteine residues in MOMP with serine residues in CTH522 that hinders the formation of covalent bonds in the

19

ACCEPTED MANUSCRIPT

protein, which can stabilize the conformation of the protein. Studies have shown that the disulfide bonds are essential for the folding and function of MOMP.20,32 The β-barrel structure of MOMP thus becomes disrupted upon removal of the cysteine residues.

RI PT

The SOPMA33 and NetSurfP34 algorithms were applied to model the secondary structural elements from the primary structure of CTH522. Mainly random coil (approx. 38 and 55% using SOPMA and NetSurfP, respectively) was estimated for CTH522, but also β-strand (approx.

SC

27/25%) and α-helix (approx. 26/20%) were predicted (results not shown). The same algorithms were used to predict the secondary structure of MOMP [GenBank Accession Number

M AN U

3329133],35 and similar results were found (results not shown). The secondary structures of CTH522 and MOMP were also estimated using the PRED-TMBB (Hidden Markov Model) algorithm for a β-barrel structure.36,37 Using the PRED-TMBB algorithm, 13 β-strands were predicted for MOMP (results not shown), compared to 16 predicted β-strands reported in the

TE D

literature 32. Using the same program, 14 β-strands were estimated for CTH522, and a large part of the protein (amino acids 331-502) was predicted to constitute an extracellular part, corresponding to the surface-exposed VDs. When these estimates are compared to the

EP

experimental data, it is evident that the β-barrel structure of MOMP is not conserved for CTH522, despite rather similar probabilities of the proteins for forming a β-strand conformation.

AC C

This might be due to the lack of cysteine residues in CTH522, which are essential for intra- and intermolecular disulfide bond formation, and/or the absence of the natural hydrophobic membrane environment that can stabilize a β-barrel structure.24,32 No thermal transitions, but rather a gradual shift in the tertiary structure as a function of temperature, were detected by measurement of the intrinsic and extrinsic fluorescence measurements. The shift in BCM peak position as a function of temperature suggests that the

20

ACCEPTED MANUSCRIPT

aromatic amino acid residues on average become slightly more exposed to the aqueous environment at higher temperatures. This suggests that reorganization event(s) may take place as a function of temperature, eventually resulting in dissociation of the self-assemblies in a way that

RI PT

increases the exposure of aliphatic side chains to the aqueous environment. This was supported by DLS experiments showing a decrease in the average dH as a function of temperature. These gradual changes of the physical properties collectively suggest a polypeptide structure with

SC

simple repetitive conformations and not a well-defined porin structure. The ANS fluorescence intensity was already high at rt due to strong binding affinity of ANS for the protein

M AN U

(hydrophobic pockets or aggregates). This suggests that a significant part of the surface area is apolar and is exposed, even at low temperature. This may be explained by the high number of aliphatic and aromatic amino acid residues present in the primary sequence of the protein (Supplementary data, Figure S1). These residues are localized primarily in the transmembrane

TE D

parts of MOMP.32 However, for the extrinsic ANS fluorescence, a linear decrease in the fluorescence intensity as a function of temperature was observed, suggesting a simple thermal quenching of the excited state. We cannot exclude that temperature-induced pH changes of the

EP

applied Tris-buffer may influence the results.

However, denaturation could be measured upon addition of GndHCl, suggesting that the

AC C

polypeptide structure has a folded conformation. Although the mechanism(s) behind this is not completely understood, it is thought that one of the ways GndHCl acts is by forming hydrogen bonds with the protein backbone and the polar side chains, eventually increasing the hydrophilicity of the protein and reducing the hydrophobic contribution towards maintaining the folded state.38 A reversible unfolding transition was measured as a function of the GndHCl concentration using intrinsic fluorescence, which reflects the disruption of the tertiary structure

21

ACCEPTED MANUSCRIPT

of the protein. The abruptness of the unfolding transition is indicative of a highly cooperative transition, and suggests a two-state transition.23 At 1.7 M GndHCl, 50% unfolding was observed. Assuming a two-state transition, the ∆G(H20) was calculated. The value of 17.3 kJ mol-1

RI PT

suggests that the K for folding is approx. 103, suggesting a probability of unfolding of 10-3. This value for the probability of unfolding is relatively high, as compared to the values of globular proteins (usually in the range of 10-4 to 10-7). This indicates that the conformational stability of

SC

CTH522 is relatively low and that the folded structure of CTH522 is stabilized via relatively weak interactions. The slight tendency of a shoulder in the beginning of the denaturation curve

M AN U

(the shift of the fluorescence) before the unfolding transition might represent the dissociation of the particle structure into lower-order structures.

The protein was refolded upon sequential removal of GndHCl by dialysis. The protein refolded by self-assembly into nanoparticle structures with physical properties very

TE D

similar to the properties of the native state protein. A number of different excipients are generally used to stabilize therapeutic proteins in solution. We tested two surfactants commonly used for the stabilization of therapeutic proteins and membrane proteins. These two surfactants did not

EP

influence the self-assembly behavior of the protein, and the monomer could not be stabilized sufficiently to avoid self-assembly into nanoparticle structures. Addition of NaCl did not

AC C

influence the process either, suggesting that hydrophobic interactions, and not electrostatic interactions, are the main driving forces for particle formation. Based on the minimal information the present data provides regarding the biophysical conformation of the protein, the conclusion of these studies is that CTH522 has an unusual quaternary structure of supramolecular self-assemblies. The self-assembly of protein antigens into higher-order structures may be favorable from production as well as immunogenicity points

22

ACCEPTED MANUSCRIPT

of view, and there is increased focus on self-assembling peptide and protein structures for vaccination purposes.39,40 However, further follow-up studies are needed to fully understand the biophysical conformation of the protein and to clarify the relationship between structure and

RI PT

immunogenicity of CTH522.

Supplementary material

SC

Supplementary data associated with this article can be found, in the online version, at doi: .

M AN U

Conflict of interest

FF and PA are co-inventors of a patent relating to chlamydia vaccines (US 20140275478 A1). All rights have been assigned to SSI, a Danish non-profit governmental institute, of which KK, HG,

Acknowledgements

TE D

GKW, PA and FF are also employees. All other authors report no potential conflicts.

This work was supported by Innovation Fund Denmark [previously The Danish National

EP

Advanced Technology Foundation (grant number 069-2011-1)]. We acknowledge the Danish Agency for Science, Technology and Innovation for funding the Zetasizer Nano ZS and The

AC C

Danish National Advanced Technology Foundation and the Danish Ministry of Science, Technology and Innovation for funding the nano-DSC. The Drug Research Academy, University of Copenhagen, is kindly acknowledged for funding the NanoDrop. The funding sources had no involvement in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; nor in the decision to submit the paper for publication. We thank the Core Facility for Integrated Microscopy, Faculty of Health and Medical Sciences, University of

23

ACCEPTED MANUSCRIPT

Copenhagen, for the cryo-TEM studies. We are grateful to C. G. Jensen and R. Søe, Department of Vaccine Development, SSI, for designing and constructing the vectors, as well as expressing

RI PT

CTH522.

Reference List

SC

1. Rey-Ladino J, Ross AG, Cripps AW 2014. Immunity, immunopathology, and human vaccine development against sexually transmitted Chlamydia trachomatis. Hum Vaccines Immunother 10:2664-2673.

M AN U

2. WHO 2012. Global incidence and prevalence of selected curable sexually transmitted infections - 2008. Geneva, Switzerland: World Health Organization. 3. Elwell C, Mirrashidi K, Engel J 2016. Chlamydia cell biology and pathogenesis. Nat Rev Microbiol 14:385-400. 4. Brunham RC, Rey-Ladino J 2005. Immunology of Chlamydia infection: implications for a Chlamydia trachomatis vaccine. Nat Rev Immunol 5:149-161.

TE D

5. Brunham RC, Rappuoli R 2013. Chlamydia trachomatis control requires a vaccine. Vaccine 31:1892-1897. 6. Mabey DC, Hu V, Bailey RL, Burton MJ, Holland MJ 2014. Towards a safe and effective chlamydial vaccine: lessons from the eye. Vaccine 32:1572-1578.

AC C

EP

7. Olsen AW, Follmann F, Erneholm K, Rosenkrands I, Andersen P 2015. Protection Against Chlamydia trachomatis Infection and Upper Genital Tract Pathological Changes by Vaccine-Promoted Neutralizing Antibodies Directed to the VD4 of the Major Outer Membrane Protein. J Infect Dis 212:978-989. 8. Lorenzen E, Follmann F, Boje S, Erneholm K, Olsen AW, Agerholm JS, Jungersen G, Andersen P 2015. Intramuscular Priming and Intranasal Boosting Induce Strong Genital Immunity Through Secretory IgA in Minipigs Infected with Chlamydia trachomatis. Front Immunol 6:628. 9. Pal S, Theodor I, Peterson EM, de la Maza LM 2001. Immunization with the Chlamydia trachomatis mouse pneumonitis major outer membrane protein can elicit a protective immune response against a genital challenge. Infect Immun 69:6240-6247. 10. Sun G, Pal S, Weiland J, Peterson EM, de la Maza LM 2009. Protection against an intranasal challenge by vaccines formulated with native and recombinant preparations of the Chlamydia trachomatis major outer membrane protein. Vaccine 27:5020-5025. 24

ACCEPTED MANUSCRIPT

11. Kari L, Whitmire WM, Crane DD, Reveneau N, Carlson JH, Goheen MM, Peterson EM, Pal S, de la Maza LM, Caldwell HD 2009. Chlamydia trachomatis native major outer membrane protein induces partial protection in nonhuman primates: implication for a trachoma transmission-blocking vaccine. J Immunol 182:8063-8070.

RI PT

12. Nikaido H 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 67:593-656. 13. Cai S, He F, Samra HS, de la Maza LM, Bottazzi ME, Joshi SB, Middaugh CR 2009. Biophysical and stabilization studies of the Chlamydia trachomatis mouse pneumonitis major outer membrane protein. Mol Pharm 6:1553-1561.

SC

14. Stephens RS, Sanchez-Pescador R, Wagar EA, Inouye C, Urdea MS 1987. Diversity of Chlamydia trachomatis major outer membrane protein genes. J Bacteriol 169:3879-3885.

M AN U

15. Baehr W, Zhang YX, Joseph T, Su H, Nano FE, Everett KD, Caldwell HD 1988. Mapping antigenic domains expressed by Chlamydia trachomatis major outer membrane protein genes. Proc Natl Acad Sci U S A 85:4000-4004. 16. Fitch WM, Peterson EM, de la Maza LM 1993. Phylogenetic analysis of the outermembrane-protein genes of Chlamydiae, and its implication for vaccine development. Mol Biol Evol 10:892-913.

TE D

17. Stephens RS, Wagar EA, Schoolnik GK 1988. High-resolution mapping of serovarspecific and common antigenic determinants of the major outer membrane protein of Chlamydia trachomatis. J Exp Med 167:817-831. 18. Carmichael JR, Pal S, Tifrea D, de la Maza LM 2011. Induction of protection against vaginal shedding and infertility by a recombinant Chlamydia vaccine. Vaccine 29:52765283.

AC C

EP

19. Pal S, Barnhart KM, Wei Q, Abai AM, Peterson EM, de la Maza LM 1999. Vaccination of mice with DNA plasmids coding for the Chlamydia trachomatis major outer membrane protein elicits an immune response but fails to protect against a genital challenge. Vaccine 17:459-465. 20. Yen TY, Pal S, de la Maza LM 2005. Characterization of the disulfide bonds and free cysteine residues of the Chlamydia trachomatis mouse pneumonitis major outer membrane protein. Biochemistry 44:6250-6256. 21. Rose F, Wern JE, Ingvarsson PT, van de Weert M, Andersen P, Follmann F, Foged C 2015. Engineering of a novel adjuvant based on lipid-polymer hybrid nanoparticles: A quality-by-design approach. J Control Release 210:48-57. 22. Pace CN 1986. Determination and analysis of urea and guanidine hydrochloride denaturation curves. Methods Enzymol 131:266-280.

25

ACCEPTED MANUSCRIPT

23. Shirley BA 1995. Urea and guanidine hydrochloride denaturation curves. Methods Mol Biol 40:177-190.

RI PT

24. Sun G, Pal S, Sarcon AK, Kim S, Sugawara E, Nikaido H, Cocco MJ, Peterson EM, de la Maza LM 2007. Structural and functional analyses of the major outer membrane protein of Chlamydia trachomatis. J Bacteriol 189:6222-6235. 25. Hawe A, Sutter M, Jiskoot W 2008. Extrinsic fluorescent dyes as tools for protein characterization. Pharm Res 25:1487-1499.

SC

26. Wilkins DK, Grimshaw SB, Receveur V, Dobson CM, Jones JA, Smith LJ 1999. Hydrodynamic radii of native and denatured proteins measured by pulse field gradient NMR techniques. Biochemistry 38:16424-16431. 27. Dill KA, Ghosh K, Schmit JD 2011. Physical limits of cells and proteomes. Proc Natl Acad Sci U S A 108:17876-17882.

M AN U

28. Erickson HP 2009. Size and shape of protein molecules at the nanometer level determined by sedimentation, gel filtration, and electron microscopy. Biol Proced Online 11:32-51. 29. Park Y, Huang R, Corti DS, Franses EI 2010. Colloidal dispersion stability of unilamellar DPPC vesicles in aqueous electrolyte solutions and comparisons to predictions of the DLVO theory. J Colloid Interface Sci 342:300-310.

TE D

30. Missana T, Adell A 2000. On the Applicability of DLVO Theory to the Prediction of Clay Colloids Stability. J Colloid Interface Sci 230:150-156. 31. Hamborg M, Jorgensen L, Bojsen AR, Christensen D, Foged C 2013. Protein antigen adsorption to the DDA/TDB liposomal adjuvant: effect on protein structure, stability, and liposome physicochemical characteristics. Pharm Res 30:140-155.

AC C

EP

32. Findlay HE, McClafferty H, Ashley RH 2005. Surface expression, single-channel analysis and membrane topology of recombinant Chlamydia trachomatis Major Outer Membrane Protein. BMC Microbiol 5:5. 33. Geourjon C, Deleage G 1995. SOPMA: significant improvements in protein secondary structure prediction by consensus prediction from multiple alignments. Comput Appl Biosci 11:681-684. 34. Petersen B, Petersen TN, Andersen P, Nielsen M, Lundegaard C 2009. A generic method for assignment of reliability scores applied to solvent accessibility predictions. BMC Struct Biol 9:51. 35. Stephens RS, Kalman S, Lammel C, Fan J, Marathe R, Aravind L, Mitchell W, Olinger L, Tatusov RL, Zhao Q, Koonin EV, Davis RW 1998. Genome sequence of an obligate intracellular pathogen of humans: Chlamydia trachomatis. Science 282:754-759.

26

ACCEPTED MANUSCRIPT

36. Bagos PG, Liakopoulos TD, Spyropoulos IC, Hamodrakas SJ 2004. PRED-TMBB: a web server for predicting the topology of beta-barrel outer membrane proteins. Nucleic Acids Res 32:W400-W404.

RI PT

37. Bagos PG, Liakopoulos TD, Hamodrakas SJ 2005. Evaluation of methods for predicting the topology of beta-barrel outer membrane proteins and a consensus prediction method. BMC Bioinformatics 6:7. 38. Tsumoto K, Ejima D, Kumagai I, Arakawa T 2003. Practical considerations in refolding proteins from inclusion bodies. Protein Expr Purif 28:1-8.

SC

39. Eskandari S, Guerin T, Toth I, Stephenson RJ 2016. Recent advances in self-assembled peptides: Implications for targeted drug delivery and vaccine engineering. Adv Drug Delivery Rev.

AC C

EP

TE D

M AN U

40. Matsuurua K 2014. Rational design of self-assembled proteins and peptides for nano- and micro-sized architectures. RSC Adv 4:2942-2953.

27

ACCEPTED MANUSCRIPT

Figure legends FIGURE 1. Representative SDS-PAGE run under reducing conditions at a sample volume of 20 µl. Molecular weight (Mw) of the standard (control) batch (left), compared to the current good

RI PT

manufacturing practice (GMP) batch (right), both run in triplicates (A). Verification of the identity of CTH522 by representative WB analysis (B). SDS-SEC-HPLC of three different GMP

(approx. 54 kDa). Arrows 3-7: degradation products.

SC

batches of CTH522. (C). Arrow 1: CTH522 dimers or aggregated forms. Arrow 2: main peak

M AN U

FIGURE 2. Mass analysis of the GMP batch of CTH522 by LC-ESI-MS. Representative mass spectrum recorded for CTH522 with the corresponding charge-state deconvoluted mass spectrum indicated in the insert (A). MS analyses indicated that CTH522 existed as at least three major species (53914 Da, 53927 Da and 53940 Da) of relative abundances of 25%, 100% and 28%,

TE D

respectively. Representative mass analysis by LC-ESI-MS of the main protein impurity in CTH522 (B). The resulting charge state distribution is shown above. The charge-state deconvoluted mass spectrum is shown in the insert. The impurity was found to have a mass of

AC C

of CTH522.

EP

16093 Da, corresponding to a protein from the E. coli host cells used for recombinant expression

FIGURE 3. Characterization of CTH522 by DLS (A-B) and cryo-TEM (C). Representative intensity-based size distribution (A) and the correlation function (B) of the size distribution. Representative cryo-TEM image of CTH522 (C). Scale bar = 200 nm.

28

ACCEPTED MANUSCRIPT

FIGURE 4. Temperature effect on CTH522 in Tris buffer (pH 9.0). Representative far UV CD spectra of CTH522 at 0.07 mg/ml (red line) and 0.2 mg/ml (black line) (A). Representative DSC thermogram of CTH522 (B). The curve has been normalized to the molar content.

RI PT

Representative far UV CD spectra of CTH522 measured at 20°C (black line), 30°C (red line), 40°C (orange line), 50°C (green line), 60°C (blue line), 70°C (purple line) and 80°C (pink line) (C). Secondary structure as a function of temperature for CTH522 (0.2 mg/ml) measured by CD

SC

at 208 nm (red dots), 215 nm (blue dots), and 222 nm (black dots) (D). The tertiary structure as function of temperature measured by intrinsic fluorescence (E) and extrinsic fluorescence (ANS

M AN U

fluorescence intensity) (F). The quaternary structure was measured by static light scattering (G) and DLS (H) of CTH522 as a function of temperature. Data represent mean values ± SD of triplicate measurements.

TE D

FIGURE 5. Unfolding of CTH522 with GndHCl, measured by intrinsic fluorescence. Intrinsic fluorescence spectrums of CTH522 at different GndHCl concentrations (A). The arrow marks the direction of increasing GndHCl concentration. Fraction of unfolded protein as a function of

EP

the GndHCl concentration (B). Fitting of the ∆G values as a function of the GndHCl concentration (C). Comparison of the intrinsic fluorescence spectra at 25°C (black curve) and

AC C

90°C (red curve) (D). Data represent mean values ± SD of technical replicates (n = 3).

FIGURE 6. Refolding of CTH522 measured by intrinsic fluorescence. Refolding of CTH522 by dialysis against Tris buffer (A) supplemented with (i) 9 µM Tween 80 (B), (ii) 150 mM NaCl (C), and (iii) 70 µM E8C12 (D).

29

ACCEPTED MANUSCRIPT

FIGURE 7. Refolding of CTH522 measured by static light scattering at 266 nm. Refolding of CTH522 by dialysis against Tris buffer (A) supplemented with (i) 9 µM Tween 80 (B), (ii) 150

AC C

EP

TE D

M AN U

SC

RI PT

mM NaCl (C), and (iii) 70 µM E8C12 (D).

30

ACCEPTED MANUSCRIPT TABLE 1: Denaturation and dialysis-mediated refolding of CTH522 measured by using DLS. Intensity-weighted hydrodynamic particle diameter (dH) and polydispersity index (PDI) of CTH522 at different concentrations of guanidine hydrochloride (GndHCl). Data represent mean values ± standard deviation (n = 3, technical replicates). PDI [GndHCl] (M) Additive dH (nm) None

55.2 ± 6.5

0.37 ± 0.03

1.0

None

42.0 ± 0.9

0.22 ± 0.00

0.5

None

54.0 ± 3.9

0.0 (after denaturation)

None

58.1 ± 3.9

1.0

Tween 80a

50.8 ± 7.9

0.5

Tween 80

64.5 ± 1.6

0.42 ± 0.05

0.0 (after denaturation)

Tween 80

54.6 ± 8.2

0.33 ± 0.05

1.0

NaClb

0.5

NaCl

0.0 (after denaturation)

NaCl

1.0

E8C12c

0.5

E8C12

a

E8C12

AC C

EP

9 µM 150 mM c 70 µM b

0.22 ± 0.04

0.31 ± 0.02

0.36 ± 0.03

SC

M AN U

TE D

0.0 (after denaturation)

RI PT

0.0 (before denaturation)

46.8 ± 2.2

0.36 ± 0.02

80.7 ± 10.5

0.36 ± 0.11

103.7 ± 32.6

0.40 ± 0.09

45.0 ± 0.6

0.20 ± 0.02

52.4 ± 1.0

0.18 ± 0.01

276.0 ± 65.9

0.53 ± 0.05

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT