Site-specific Bioconjugation and Convergent Click Chemistry Enhances Antibody–Chromophore Conjugate Binding Efficiency

Abstract

Photosensitizer (PS)–antibody conjugates (photoimmunoconjugates, PICs) enable cancer-cell targeted photodynamic therapy (PDT). Non-specific chemical bioconjugation is widely used to synthesize PICs but gives rise to several shortcomings. The conjugates are heterogenous, and the process is not easily reproducible. Moreover, modifications at or near the binding sites alter both binding affinity and specificity. To overcome these limitations, we introduce convergent assembly of PICs via a chemo-enzymatic site-specific approach. First, an antibody is conjugated to a clickable handle via site-specific modification of glutamine (Gln) residues catalyzed by transglutaminase (TGase, EC 2.3.2.13). Second, the modified antibody intermediate is conjugated to a compatible chromophore via click chemistry. Utilizing cetuximab, we compared this site-specific conjugation protocol to the non-specific chemical acylation of amines using N-hydroxysuccinimide (NHS) chemistry. Both the heavy and light chains were modified via the chemical route, whereas, only a glutamine 295 in the heavy chain was modified via chemo-enzymatic conjugation. Furthermore, a 2.3-fold increase in the number of bound antibodies per cell was observed for the site-specific compared to non-specific method, suggesting that multiple stochastic sites of modification perturbs the antibody-antigen binding. Altogether, site-specific bioconjugation leads to homogenous, reproducible, and well-defined PICs, conferring higher binding efficiency and probability of clinical success.

Graphical Abstract Text:

Photosensitizer-antibody conjugates (PACs or PICs) enable targeted photodynamic therapy. Widely used non-specific chemical bioconjugation methods for PIC synthesis are limited by their heterogeneity, irreproducibility, and crucially, their adverse effect on antibody structure, binding affinity, and specificity. Alternatively, we employ a site-specific, enzymatic conjugation strategy which yields homogenous, reproducible, and well-characterized PICs with a higher binding efficiency compared to their nonspecific counterpart. The process is broadly applicable to a number of antibodies and chromophores (e.g., photosensitizers or fluorophores), thereby enabling rapid generation of diverse libraries of PICs with tailored functionality that retain the native biological activity of the antibody.

INTRODUCTION

Present clinical applications of photodynamic therapy (PDT) to treat cancer typically occur via systemic administration of unmodified photosensitizers (PSs). The efficacy of this treatment strategy is dependent on the selectivity of both the PS pharmacokinetics and the applied light, as only PS molecules activated by light initiate cytotoxic pathways (1, 2). Although effective in many cases, the maximum tolerated dose is limited by off-target toxicity in neighboring organs due to poor selectivity of PSs when applying diffuse light. Targeted approaches using PS–antibody conjugates (PACs) – commonly referred to as photoimmunoconjugates (PICs) – have steadily been developed to improve treatment selectivity (3–15). Activatable PICs—where the PS fluorescence and photocytotoxicity are quenched until target biomarker binding, cellular internalization, and lysosomal proteolysis—enable microscale tumor selectivity (11) and spare local effector immune cells (15).

Like most antibody–drug conjugates (ADCs), PICs target cell surface proteins frequently overexpressed by tumor cells relative to neighboring tissues and show clinical promise (e.g., clinical trial NCT03769506). The primary components in any PIC are a tumor-targeted antibody and a photodynamic chromophore (or chromophores). In most cases, a chemical linker (or linkers) is included to expedite coupling, although direct conjugation of chromophore to antibody is possible. There are a number of clinical antibodies that target a variety of cell-surface antigens for molecular targeted therapy. Biocompatible and clinically approved PSs are also commercially available, and many next-generation PSs are under preclinical and clinical development (14).

The key challenge to efficient and reliable PIC design and synthesis is the linker chemistry to covalently attach the PS to the antibody in a site-specific and homogeneous manner, without compromising the biological activity and specificity of the antibody. Specifically, a method that avoids modifications on the antigen-binding fragment (Fab) in favor of the crystallizable fragment (Fc) are desirable to limit interference with the Fab–receptor binding kinetics. Here, we introduce a simple and robust approach to facilitate site-specific, convergent (stepwise) PIC synthesis using a single-site modification method integrated with click chemistry (Figure 1a) that is compatible with many antibodies and PSs.

Figure 1.

Convergent assembly of antibody-chromophore conjugates in a stepwise fashion via click chemistry. The linker conjugation reaction may be designed from click chemistries like SPAAC, CuAAC, and tetrazine ligation. (a) Site-specific, transglutaminase (TGase)-catalyzed conjugation to glutamine 295 leads to homogeneous PICs, preserving binding specificity and affinity. (b) Non-specific chemical acylation of amines (i.e., lysine and N-termini) on the antibody yields heterogenous mixtures of PICs, with potential for reduction in binding specificity and affinity.

The new PIC synthesis approach introduces a clickable handle (i.e., strained alkyne) to the antibody via site-specific transglutaminase-mediated transamidation of glutamine. The modified antibody intermediate is then linked to the PS with another complementary clickable handle (i.e., azide) via click reaction. This convergent approach allows generation of reproducible conjugates that are easy to characterize and enables generation of libraries of PICs for high-throughput structure—activity relationship studies. In addition, we compared these constructs with the analogous PIC synthesized via non-specific chemical modification of amines followed by click-chemistry conjugation of the chromophore to the stochastically placed handle (Figure 1b). We demonstrate this new approach using anti-epidermal growth factor receptor (EGFR) antibody (cetuximab), and we evaluated the biological activity of these conjugates in three cancer cell lines with varying degrees of EGFR expression. The data indicate that the site-specific technology form well-defined constructs and preserve binding efficiency compared to their non-specific chemical counterpart. Overall, the system is modular, facile to characterize, and enables generation of large and diverse PIC libraries.

Site-specific enzymatic conjugation approaches achieve homogeneous protein modifications (16). Here we utilized microbial transglutaminase (mTGase, E.C. 2.3.2.13) to modify cetuximab at a single site, glutamine (Gln) 295 of the heavy chain (HC) in the Fc region, whereas the light chain (LC) is unmodified (17–19). mTGase catalyzes a transamidation reaction in which an unsubstituted amide in glutaminyl residues is converted into a substituted amide using an amine substrate (20–25). Commercially available and inexpensive, this enzyme has broad specificity towards the amine substrate (26) such as accepting azides (27), strained alkenes and alkynes (28), tetrazines (27), and even bulkier groups such as polyethylene glycol chains (PEG) (29). Therefore, a variety of amine substrates can be attached either in a one-step or stepwise fashion in a reliable and site-specific manner.

The final step in convergent PIC synthesis incorporated click chemistry, which enables biorthogonal, fast, and high yield reactions to prepare conjugate biomolecules (30–32). Remarkably, these features of click chemistry are being tested in humans to synthesize radioimmunoconjugates in vivo—the components are “clicked” together within the body (33). Click chemistry reactions are robust and have been reported to chemically attach various poryphyrin- (34–36) and boron dipyrromethene-class (37) PSs to antibodies. Additionally, diverse chemical moieties of chromophores (e.g., PSs or fluorophores) have been incorporated into antibodies including nanoparticles (38–40), multi-target imaging agents (41), broad wavelength (800 to 1700 nm) dyes (42), and dendrimers or branched polymers to load more PSs (36, 43). Common types of click reactions include copper azide-alkyne cycloadditions (CuAAC), strained promoted azide-alkyne cycloadditions (SPAAC), and tetrazine ligation (i.e., inverse electron-demand Diels–Alder, IEDDA) (44). Because copper may be toxic in vivo and lead to side reactions (e.g., oxidation) (44), SPAAC and IEDDA are preferable alternatives. In order to utilize click reactions, compatible clickable handles need to be conjugated to antibody and as well as the PS. This process has been achieved in proteins using both site specific (i.e., enzymatic) and non-specific (i.e., chemical) methods (45–47). Here, we applied SPAAC click chemistry to assemble the PIC.

In contrast to the methodology presented above, traditional chemical conjugation methods for PIC synthesis involve acylation of amines available on lysine residues and N-termini of protein, or alkylation of thiols on cysteine residues. These strategies have been employed to construct ADCs approved by the FDA (48). Nonetheless, chemically synthesized conjugates are not site-specific, heterogenous, and produce positional isomers in a stochastic distribution (i.e., different sites and ratios) (49). Moreover, some of these chemical reactions also suffer from side-reactions that are often underappreciated (45). The multiple modification sites—particularly at or near the binding sites like the complementarity determining region (CDR) in antibodies—may result in structural changes of the antibody and thus alterations in biological functions, including antigen binding capacity and specificity. For instance, Luo and coworkers have found almost all amines in an IgG (76 total, in both light and heavy chains) were acylated using N-hydroxysuccinimide (NHS) ester chemistry (50). Such reactions yield a mixture of conjugates with a widely distributed antibody loading ratio (ALR) ranging from 0 to 8, and within a given ALR, multiple positional isomers exist. From a practical perspective, since the process is kinetically controlled, considerable optimization is required to ensure reproducibility. Furthermore, new protocols must be re-optimized if either antibody or acylation reagent is changed. The protocols developed here overcome all of the challenges associated with chemical conjugation methods.

MATERIALS AND METHODS

The concentrations of the peptides and proteins were determined using UV absorption at 280 nm and extinction coefficients based on amino acid sequences. All aqueous solutions were prepared using Milli-Q water. All cell culture methods were performed with aseptic technique in a biosafety cabinet.

Deglycosylation of antibody using PNGase F.

The deglycosylation reaction contained 3.5 μM cetuximab (Selleckchem, A2000), 1X GlycoBuffer (New England Biolabs, B0701S), and was initiated with 100 units of PNGase F (New England Biolabs, P0708S) at 37°C and for 9.5 hours. To remove excess unreacted reagents, each reaction mixture was desalted using 50 kDa molecular weight cut-off (MWCO) centrifugal filters (Amicon unit, UFC505024) into 100 mM potassium phosphate (pH 7.6) prior to the transamidation reaction.

Transglutaminase-mediated transamidation reaction.

The transamidation reaction contained 100 mM potassium phosphate (pH 7.6), 5 mM dibenzylcyclooctyne-PEG4-amine (compound 1, Fig. 2; Click Chemistry Tools, A103P), 2 μM deglycosylated cetuximab aforementioned, and was initiated with 6.9 μM microbial transglutaminase (mTGase, Ajinomoto, ACTIVA-TI formulation, Uniprot P81453) and incubated at 37°C for 15.5 hours. To remove excess unreacted reagents, each reaction mixture was desalted using 50 kDa MWCO centrifugal filters into 100 mM potassium phosphate (pH 7.6) prior to the click reaction.

Figure 2.

Structures of clickable handles for convergent assembly of PICs via SPAAC click chemistry. The antibody is conjugated through (a) site-specific TGase mediated glutamine transamidation using amine or (b) stochastic amine acylation using NHS ester, each tethered with a strained alkyne for subsequent SPAAC click reaction. (c) Each modified antibody (intermediate) is conjugated to AF647 tethered with a complementary azide clickable handle. (d) The cyclooctyne and azide click together to form a triazole in the final linker structure.

Amine reactive acylation reaction.

The reaction contained 100 mM potassium phosphate (pH 7.5), 155 μM dibenzylcyclooctyne-PEG4-N-hydroxysuccinimide (compound 2, Fig. 2, Broad Pharm, BP-22288), 2.7 μM cetuximab (without deglycosylation), and incubated at 22°C for 3.5 hours. The reaction was then quenched by 2 mM Tris (pH 8). To remove excess unreacted reagents, each reaction mixture was desalted using 30 kDa MWCO centrifugal filters (Amicon unit, UFC503096) into 100 mM potassium phosphate (pH 7.6) prior to the cycloaddition reaction.

Strain promoted cycloaddition click reaction (SPAAC).

The reaction contained 100 mM potassium phosphate (pH 7.5), 121 μM Alexa Fluor® 647 azide (AF647, compound 3, Fig. 2, Click Chemistry Tools, 1299), 1.5 μM cetuximab either modified by mTGase or acylated as aforementioned, and incubated at 37°C for 3.5 hours. To remove excess unreacted reagents, each reaction mixture was desalted using 30 kDa MWCO centrifugal filters into 100 mM potassium phosphate (pH 7.6) prior to analysis.

Determination of chromophore-antibody loading ratio.

The loading ratio was calculated based on the ratio of the absorption of the chromophore and antibody in similarity to prior literature (7, 11). Briefly, the absorbance spectra of PIC solutions were collected from 220 to 750 nm (NanoDrop spectrophotometer, ND-1000). The concentration of the chromophore AF647 was calculated using the absorbance of the conjugate at 650 nm and extinction coefficient of 270,000 M⁻¹cm⁻¹ provided by the vendor. The protein concentration was estimated by subtracting the chromophore AF647 contribution at 280 nm, and the calculated extinction coefficient based on its amino acid sequence (217,440 M⁻¹cm⁻¹). For the non-specific chemical acylation method, the chromophore and antibody concentration were 20.7 ± 0.3 μM and 4.90 ± 0.03 μM, respectively (a ratio of 4.22 ± 0.07). For the site-specific transglutaminase-mediated method, the chromophore and antibody concentration were 2.20 ± 0.10 μM and 2.5 ± 0.00 μM, respectively (a ratio of 0.88 ± 0.04). A summary of each construct used in this study is provided in Table 1.

Table 1.

Antibody chromophore conjugate characteristics. ALR: Antibody Loading Ratio.

	Site-specific	Non-specific
Target (antibody)	EGFR (cetuximab)
Chromophore (Ex/Em (nm))	Alexa Fluor® 647 (648/671)
Conjugation Method	Glutamine	Amine
Step 1	Transglutaminase	NHS ester
Number of modification sites	2	76
Step 2	Click Chemistry
ALR (Chromophore:Antibody)	0.88 ± 0.04	4.22 ± 0.07

Characterization of the conjugates by SDS-PAGE.

Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was performed using a Bio-rad Mini-PROTEAN 3 system. First, the reaction mixture was incubated with SDS-Sample Buffer at 80°C for 10 minutes. For reducing and non-reducing gels, 4× reducing SDS sample buffer (Boston Bioproducts, BP-110R) and 2× nonreducing SDS sample buffer (Bio-rad, 1610737) were used, respectively. Second, the samples were loaded into 12% tris-tricine precast protein gels (Biorad, 4561044). Precision Plus Protein™ Dual Xtra Prestained Protein Standards (Bio-rad, 1610377) were used for mass calibration. Electrophoresis was then performed at 200 V for 20 min. The gel was stained by Coomassie R250 and then destained using 10% acetic acid and 40% methanol. The gels were imaged using an iBright FL1000 Imaging system (Thermo Fisher Scientific).

Characterization of conjugates by isoelectric focusing (IEF).

Isoelectric focusing was performed using a Bio-rad Criterion system. First, the reaction mixture was mixed in 1:1 dilution with IEF sample buffer (Bio-rad, 1610763). Second, the samples were loaded into pH 3–10 Criterion IEF precast gel and placed into the cell. Then, 1× Anode buffer (Biorad, 1610761) and 1× Cathode buffer (Biorad, 1610762) were placed in the upper and lower chamber of the criterion cell, respectively. Third, IEF was run initially at 100 V for 1 hour to initiate desalting of the sample, followed by higher voltage at 250 V for 1 hour to mobilize the antibody, and lastly 500V for 30 minutes to complete electrofocusing. The gel was stained by Coomassie R250 and Crocein Scarlet, and then destained using 10% acetic acid and 40% methanol. The gels were imaged using an iBright FL1000 Imaging system.

Cancer cell models.

Ovarian cancer cell line NIH:Ovcar-3 (Ovcar3, HTB-161) and breast cancer cell line T-47D (HTB-133) were purchased from the American Type Culture Collection and cultured in RPMI 1640 (Gibco, 11875093) completed with 10% heat-inactivated fetal bovine serum (FBS, Hyclone™ GE Healthcare Life Sciences, SH30071.03HI) and 1% penicillin/streptomycin antibiotic (Pen/Strep, Fisher BioReagents, BP295950). A human primary high-grade serous ovarian cancer line (Powder, Cellaria Biosciences) was cultured in Renaissance Essential Tumor Medium (RETM) and RETM Supplement (Cellaria Biosciences, CM-0001), plus 6.3% FBS and 1% mL Pen/Strep. Cells were grown in a humidified atmosphere at 5% CO₂ and 37°C using T75 Flasks (Thermo Scientific, 1256685) according to their recommended culturing protocols. During passaging, cells were lifted with 0.25% trypsin EDTA cocktail (Corning, 25053CI), washed in phosphate buffered saline (PBS, Gibco, 70011069), and resuspended at 10,000 cells/mL. Cells were plated at 3,000 cells per well in a 96-well plate (Perkin Elmer, LLC 6055302) and allowed to grow for 3 days. Three wells were plated for each group.

Antibody binding assessment via flow cytometry.

Cells were plated in triplicate at 3,000 cells per well in a 96-well plate (Perkin Elmer, LLC 6055302) and allowed to grow for 3 days. Cells were lifted with trypsin, washed, and suspended in 50 μL PBS with 0.25 μg of Cet-AF647 per replicate and incubated in non-internalization conditions at 4°C for 30 minutes in the dark. The samples were washed once and resuspended in 300 μL of PBS for analysis via flow cytometry (ThermoFisher, Attune NxT). A 635 nm laser was used to measure the auto-fluorescence of the unstained control group; the PMT voltage in that channel was adjusted so that the mean AF signal was small (~10–50 arbitrary fluorescence units). The rest of the samples were analyzed with identical instrument settings at 500 μL/min flow rate. For each cell line, a gate was placed around the population of cell-sized objects based on forward- vs. side-scatter plot analysis. This population was then displayed as a histogram of fluorescence intensity from the fluorescence measurement channel. The median fluorescence intensity from each replicate was recorded as the measurement of the average brightness of the cells stained with each Cet-AF647 conjugate.

Antibody internalization assessment via confocal microscopy.

Cells were plated at 10,000 cells per well in a 96-well plate and allowed to grow for 24 hours. Wells were stained in-well with 1 μg of Cet-AF647 in 50 μL PBS at 37°C for 1 hour in the dark. After 45 minutes, 50 μL of Hoechst 33342 (Life Technologies, H3570) at 10 μg/mL was added to each well to bring the final staining concentration to 5 μg/mL. The plate was immediately returned to the incubator for the final 15 minutes. The wells were washed twice in PBS before imaging via laser scanning confocal microscope (Zeiss, LSM 800). A 640 nm excitation laser was used to excite AF647, and a 405 nm laser was used for the Hoechst nuclear stain. Images of the AF647 and Hoechst fluorescence were collected using identical laser power and PMT gain using a 10× magnification, 0.3 numerical aperture air objective.

RESULTS AND DIscussion

Both site-specific mTGase-mediated and non-specific amine-NHS ester processes successfully conjugated cetuximab with the fluorophore (Fig. 3 and Supporting Information Figures S1–S2). In terms of site-specificity, for the enzymatic method, only the HC was modified, which is consistent with modification at a single site, glutamine 295, as reported by several laboratories (18, 19). Whereas, for the chemical approach, both the HC and LC of the antibody were modified, also consistent with the 76 reactive lysine residues and N-termini amines (50). Interestingly, we observed a faint fluorescent band at 37 kDa indicating mTGase was self-modified (Fig. 3a). Such self-modification has been reported in mammalian and other bacterial TGases (17).

Figure 3.

Analysis of cetuximab–AF647 conjugates synthesized via either transglutaminase-mediated conjugation (Gln-mTGase) or amine-reactive NHS ester (Amine-NHS) chemistries, with (a) reducing SDS-PAGE and (b) IEF. Both fluorescence intensity and migration patterns by IEF indicated the degree of modification. Complete gels with controls are provided in Figures S1–S3. HC, Heavy Chain; LC, Light Chain; pI, isoelectric point.

Under the described reaction conditions, the chromophore–antibody loading ratio for the site-specific enzymatic and non-specific chemical methods were 0.9 and 4.2, respectively, as determined by ultraviolet-visible spectroscopy (Table 1). Further optimization of the enzymatic transamidation reaction is possible by increasing the reaction time, temperature, or enzyme concentration to achieve a maximum ALR of 2 as reported (17–19). The degree of modification is also consistent with the IEF results. Due to the four strong negatively charged sulfonates on AF647, the isoelectric point of the modified antibody via the site-specific method (pI 6.5–7.5) was lower than the unmodified antibody (pI 7.2–8, Fig. S3). Moreover, multiple degrees of antibody modification for the non-specific chemical route (pI 5–6, Fig. 3b) were observed compared to the site-specific approach. It is worth noting that the intrinsic charge heterogeneity of the unmodified antibody contributed to the multiple bands observed for both enzymatic and chemical conjugates. As is well documented in the literature, these antibody charge variants are due to various post-translational modifications, such as deamidation, oxidation, glycosylation, or other reactive metabolites (51–59).

We sought to analyze the binding activity of each construct in three cancer cell lines with varying EGFR expressions: Ovcar3 (high EGFR), Powder (moderate EGFR), and T-47D (low EGFR). Under non-internalization conditions, cells stained with the amine-conjugated constructs were brighter than those stained with glutamine conjugates (Figures 4a and S4). This measurement is the product of both the number of antibodies bound per cell and the number of dyes per antibody. Normalizing by the ALR of each conjugate reveals a relative measurement of each conjugate’s binding capacity to EGFR (Figure 4b), which may then be compared across conjugates and cell lines. Under saturated staining conditions (more antibodies than receptors), this represents a relative measurement of the EGFR expression. The fold difference in relative affinity among cell lines as measured by each conjugate was consistent (p = 0.1116). For example, the 2.15-fold higher binding by the glutamine-conjugated constructs over the lysine-conjugated constructs to Ovcar3 cells was not statistically different from the 2.24-fold increase measured using Powder cells (p = 0.7700). Therefore, we combined the results from all 3 cell lines to assess the overall difference in binding efficiency. We found that the absolute expression of all cell lines measured by the glutamine-conjugated constructs was 2.29 ± 0.21-fold greater (p < 0.0001) than the measurement made by the amine-conjugated constructs (Figure 4c). Clearly, the natural binding properties of EGFR are degraded by non-specific conjugation strategies.

Figure 4.

Binding activity of cetuximab–AF647 conjugates. (a) 3 cell lines with high (Ovcar3), medium (Powder), and low (T-47D) EGFR expression were stained with cetuximab–AF647 conjugates prepared via non-specific chemical (Amine-NHS) or site-specific enzymatic (glutamine-microbial transglutaminase, Gln-mTGase) conjugation. The median fluorescence intensity per cell was recorded by flow cytometric analysis (n = 3 biological replicates, Figure S5). (b) Data from a normalized by ALR (Table 1) reveals the difference in relative EGFR expression of each cell line as measured by each conjugate is the same (p = 0.1116, one-way ANOVA). However, the absolute EFGR expression measured by amine-NHS conjugates is less than that measured by Gln-mTGase conjugates indicating that the site-specific Gln-mTGase chemistry better preserves the binding affinity of the antibody than the non-specific amine-NHS chemistry. (c) The fold-difference in EGFR expression measured by the two conjugates across 3 cell lines was averaged and normalized to the amine-NHS group. The result is a 2.29 ± 0.21-fold higher EGFR affinity (****p < 0.0001, unpaired, two-tailed student’s t-test) of the site-specific conjugates over the amine-NHS group.

The choice of ALR normalization is valid under two important assumptions. First, the surface protein expression should not vary between groups stained with each conjugate. This is achieved simply by culturing and handling cells under identical conditions, i.e., in the same flask or plate. Second, quenching effects arising from antibody conjugation should be negligible. In this case, AF647 conjugation onto the antibody corresponds with increasing fluorescence brightness between ALR ≈ 1–4 (60), so no correction for quenching effects is necessary for a first-order estimate. In general, chromophore quenching as a function of ALR is not predictable and varies by molecule. For example, conjugates of cetuximab and BPD exhibit quenching at ALR ≥ 2 (7). If this is the case, a correction based on the reduced quantum yield of the quenched construct should be included in the analysis scheme.

The biological activity of the antibody was also assessed via confocal fluorescence microscopy. The three aforementioned cell lines were stained under physiological conditions with each antibody conjugate and a nuclear stain (Figure 5). AF647 fluorescence was normalized to the dye-loading ratios of each conjugate. Signal was clearly dependent on the cellular EGFR expression. Fluorescence signal was apparent in the cytoplasm of Ovcar3 cells, indicating cellular uptake of each conjugate. This was also observed for the Powder line, albeit with minimal detectable signal. No signal was observed in the low EGFR-expressing T-47D cell line as expected. Unnormalized images are provided in Figure S5.

Figure 5.

Laser-scanning confocal microscopy of (a) Ovcar3 (EGFR high), (b) Powder (EGFR medium), and (c) T-47D (EGFR (low) cancer cell lines. Cells were incubated with non-specific amine-NHS conjugated (left) or site-specific Glutamine (Gln)-mTGase conjugated (middle) cet–AF647 constructs, or no antibody (right). All cells were stained with Hoechst 33342 nuclear stain. AF647 fluorescence is normalized to the ALR of each conjugate and represents the amount of antibody uptake per cell. AF647 fluorescence intensity for each cell line is indicated via 16-bit color bar. Scale bar, 50 μm.

The data suggest that chemical acylation of amines results in conjugation of multiple copies of the chromophore that may alter the structure of the antibody and thus binding affinity and specificity; moreover, in extreme cases, complete abolishment of the binding activity. Such changes may also lead to nonspecific binding to other cell surface targets and thus off-target toxicity. This further emphasizes the need for site-specific technology to construct PICs.

In summary, there are several attributes for the convergent assembly of PICs in a chemo-enzymatic and site-specific manner that are worth noting. First, the method is widely applicable to other antibodies, particularly IgG1. This synthetic route allows control of the degree of modification and enables reproducible PICs that are easy to characterize. Second, the system is modular. Convergent assembly allows a variety of chromophores to be incorporated including photosensitizers or dyes. Similarly, a variety of IgGs can be evaluated, which enable rapid generation of diverse libraries of PICs for target validation with the ultimate goal of forming clinical candidates. Finally, to increase the payload further, multiple photosensitizers can be appended at a single site; for example using dendrimer (36) or branched polymer (43) technology. The facile, uniform, and site-specific PIC synthesis approach reported here may stimulate new and creative, next-generation multi-PS and bright fluorophore PICs for multi-organelle and hybrid fluorescence-guided surgery–PDT applications.

Supplementary Material

figure s1-s5

Figure S1. Antibody conjugate characterization: non-reducing SDS-PAGE.

Figure S2.Antibody conjugate characterization: reducing SDS-PAGE.

Figure S3.Antibody conjugate characterization: Isoelectric Focusing.

Figure S4.Antibody conjugate characterization: Flow Cytometry.

Figure S5.Antibody conjugate characterization: Confocal Microscopy.