Porphyrin-modified phosphoramidites as nucleoside phosphoramidite analogs have been achieved by attaching porphyrins to the sugar rings at 1′ carbon as a total replacement of the nucleobase unit (Morales-Rojas and Kool, 2002) or by tethering the porphyrin to the nucleoside phosphoramidite via amino functionalization of 3′ sugar carbon (Balaz et al., 2005). In regard to the former strategy, in 2002, Kool et al. (Morales-Rojas and Kool, 2002) synthesized a porphyrin nucleoside by reacting 3′,5′-bis-O-toluoyl-protected deoxyribose-C1-carboxaldehyde with benzaldehyde and dipyrromethane under Lindsey-like condensation conditions (Lindsey et al., 1987) (i.e., condensation of aldehydes and pyrroles in the presence of strong acid catalysts and subsequent oxidation) followed by the deprotection of the toluoyl groups. Selective 5′-O-tritylation and 3′-O-phosphitylation afforded the requisite 2-cyanoethyl phosphoramidite monomer (Figure 1A). This porphyrin phosphoramidite was introduced onto the growing ON sequence at either the terminal positions or the interior backbone. Duplex structures formed by a pair of self-complementary ON strands modified with porphyrin macrocycles at the termini demonstrated higher thermal stability (∆T _m = 14.7°C) as compared to the unmodified congener. End stacking, via aromatic interactions, of porphyrins with the terminal nucleobases was posited to be the reason for the enhanced stabilization of the double helix. In contrast, internally tethered porphyrins (attached to the phosphate backbone) of the DNA sequence intercalate within the duplex and lead to a slight destabilization of the double helix in comparison to the perfectly complementary analog. Interestingly, when a sequence with a nucleobase mismatch was used, the internal porphyrin modification led to the enhancement of duplex stability (versus the unmodified control).

In 2005, Balaz et al. synthesized 3′-tetraarylporphyrin-thymidine-5′-phosphoramidites and used them as building blocks in the synthesis of ON–porphyrin conjugates (Balaz et al., 2005). It was reportedly the first example of porphyrins incorporated onto the 3′ position of ONs. Meso-trispyridylphenylporphyrin carboxylic acid and meso-tetraphenylporphyrin carboxylic acid were coupled with 3′-amino-5′-dimethoxytritylthymidine using 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and 4-Dimethylaminopyridine (DMAP) as coupling reagents (Figure 1B). The DMT group was removed with a 4% solution of p-toluenesulfonic acid followed by the reaction with 2-cyanoethyl-N,N-diisopropylchloro phosphoramidite to offer the desired 3′-porphyrin-thymidine 5′-phosphoramidites. Two separate self-complementary sequences, 5′-ACGCGCGT-3′ and 3′-T-5′-GCGCGCA-3′ were chosen for the study where the terminal thymidine residues bear the porphyrin moiety. The former sequence was synthesized via automated DNA synthesis in the reverse 5′ to 3′ direction. The latter sequence was synthesized in the typical 3′–5′ direction up to the penultimate 5′-guanosine residue, after which, the terminal porphyrin-thymidine phosphoramidite was introduced via a 5′–5′ coupling (Figure 1C).

Palladium catalyzed cross coupling is an attractive method to attach porphyrins off the nucleobase units since both partners are aromatic. Indeed, over the decades, there have been numerous reports of utilizing Palladium catalyzed coupling to generate porphyrinoid- and pyrrole-nucleobase conjugates illustrating the broad scope of these reactions to create new C-C bonds between nucleobases and porphyrin-related moieties (Sessler et al., 2006; Torres et al., 2007; Pierce et al., 2008). In particular, Sonogashira coupling between alkyne-modified porphyrins and iodo-modified uridine nucleosides are frequently employed by researchers to generate porphyrin-based phosphoramidites.

The Stulz research group, in 2007, carried out Sonogashira coupling between 5-iodo deoxyuridine and alkyne functionalized tetraphenylporphyrin to synthesize a phosphoramidite porphyrin-nucleotide (Figure 2A) (Fendt et al., 2007). Subsequently, up to 11 tandem porphyrin units were covalently tethered to a DNA strand via standard solid-phase ON synthesis. After the conjugation and placement in aqueous media, hydrophobic stacking led to the porphyrin units interacting with each other resulting in the broadening of the absorption bands and quenching of the emission. Moreover, when complementary sequences were hybridized, the porphyrin moieties aligned themselves in the major groove of the double helix. Further, the chirality imparted by the helicity of the DNA domain resulted in a circular dichroism signal in the porphyrin Soret band region.

In 2010, Albinsson et al. used repetitive phenylethynylene linkers to tether a porphyrin moiety at a certain distance from a 39-mer DNA core (Börjesson et al., 2010). First, the researchers synthesized a deoxyuridine porphyrin nucleoside phosphoramidite via two Sonogashira coupling reactions. The first Sonogashira coupling to yield triisopropylsilyl ether (TIPS) protected phenylethynylene nucleoside involved tris(dibenzylideneacetone)dipalladium (0) as the Pd catalyst, CuI as a copper (I) cocatalyst, triphenylarsine (AsPh₃) as a dative ligand, and triethylamine (NEt₃) as a base as well as the reaction solvent. In the second copper-free Sonogashira coupling reaction, to generate porphyrin phenylethynylene nucleoside, the copper (I) cocatalyst was excluded to prevent Cu-metalation of the porphyrin core. In addition, the copper-free coupling was optimized by using 1,4-diazabicyclo [2.2.2]octane (DABCO) as the base and dimethylformamide (DMF)/tetrahydrofuran (THF) mixture as the reaction solvent (Figure 2B). After phosphitylation, the resultant phosphoramidite was then used as a building block to tether porphyrins to the DNA sequence followed by the zinc metalation of the porphyrin core. When the Zn porphyrin-DNA conjugate was mixed with liposomes dispersed in water, the hydrophobic effect kept the porphyrin embedded in the lipid bilayer, and the DNA domain stayed in the aqueous medium. Further hybridization with a complementary strand generated a duplex architecture on the lipid surface. More importantly, the use of doubly modified strands resulted in the generation of two porphyrin anchors that enables the DNA duplex to be oriented tangential to the membrane surface (Figure 2C). Importantly, binding of the porphyrins into the lipid bilayer protected the porphyrin from the bulk aqueous solution, precluded the hydrophobic stacking among porphyrin units, and made the system redox-active facilitating electron transfer between the excited-state porphyrin and a benzoquinone electron acceptor, placed within the bilayer.

Again in 2013, Albisson et al. used a similar DNA-Porphyrin-Lipid architecture to employ free-base porphyrin and YO-PRO-1 (YO) as acceptor and donor dyes, respectively, in a DNA-based FRET architecture (Woller et al., 2013). The YO dyes were intercalated in the DNA duplex with a spacing of ∼10 Å and upon excitation at 483 nm (where only the YO dyes absorb), the YO dyes transferred energy directionally to the porphyrin embedded in the lipid bilayer via FRET (Figure 2D). Prior to energy transfer to the porphyrin unit, the YO dyes transfer energy among themselves via homoFRET. In this artificial light-harvesting assembly, when the porphyrin emission was monitored at 700 nm, the antenna effect facilitated by the YO dyes was calculated to be 12 (with a 20:1 ratio of YO to porphyrin). This multi-component self-assembly clearly demonstrates how the programmability of DNA assembly, along with the use of other supramolecular interactions, can generate a well-defined light-harvesting antenna.

The Häner group in 2014 developed porphyrin phosphoramidites that would completely replace a nucleoside component at the specific site of attachment (Vybornyi et al., 2014). Specifically, a 5,15-bisphenyl-substituted porphyrin was modified such that one arm of the porphyrin possessed a phosphoramidite unit, and the other arm bore a DMT-protected hydroxyl group. Using the solid-phase synthesis method, ON sequences (ON1 through ON6) bearing 0, 1, or 2 porphyrin units were synthesized. ON1, ON3, and ON5 represent the same sequence except for the sites where nucleobases are substituted by the porphyrins. These ONs (ON1-ON3) are complementary to another set of ONs with the same sequence (ON2, ON4, and ON6). Hybridization of the complementary sequences generated duplexes with up to 4 porphyrins next to each other (Figure 3). Compared to the unmodified duplexes the substitution of nucleobases with the porphyrin units resulted in the reduction of the melting point (up to 12°C per modification). Further, the resulting H-aggregation of the porphyrins led to fluorescence quenching.

The reaction of an activated carboxylic acid with an amine generates a stable amide bond and is referred to as amide (or peptide) coupling. Carboxylic acids can be activated in situ by reacting with carbodiimides (e.g., dicyclohexylcarbodiimide, DCC) and aminium/phosphonium derivatives of benzotriazole (e.g., HATU, BOP) or can be first transformed to storable activated esters (e.g., N-hydroxysuccinimido esters). An early example of a Porphyrin-DNA conjugate was reported by Meunier et al., in 1993 who used amide bond-formation (Casas et al., 1993). In this paper, the researchers conjugated a cationic manganese (III) porphyrin (Mn-TMPyP) that is also appended with a carboxylic acid functional group onto ONs at the 5′ position. For this, they activated the carboxylic acid functional group with 1-1′-carbonyldiimidazole followed by the reaction with 1-hydroxybenzotriazole (HOBt). Unmodified ON sequences do not have primary aliphatic amines that react efficiently with the activated porphyrin ester. Hence, a 5′-amino-modified ON sequence with a C6 linker was used to react with the activated porphyrin to afford the porphyrin-DNA conjugate (Figure 4A). In terms of function, the cationic Mn-TMPyP acts as an endonuclease to hydroxylate C-H bonds of DNA sugars at low concentrations. Hence, when hybridized to the AUG initiation codon of the HIV-1 TAT gene, the porphyrin-DNA conjugates were able to cleave the target sequence showing promise as a potential antiviral agent.

In 2017, Das et al. (Kumari et al., 2017) and Campidelli et al. (Chatelain et al., 2017) independently reported the self-assembly of multi-branched Porphyrin-DNA hybrid molecules into elaborate nanostructures. In both manuscripts meso-tetra(4-carboxyphenyl)porphyrin was functionalized, with 4 single-stranded ON sequences generating a porphyrin molecule linked covalently with four ON arms. To achieve tetra-ON-porphyrin hybrids, Campidelli et al. first modified meso-tetra(4-carboxyphenyl)porphyrin to a tetra azido derivative via amide coupling with a 2-azidoethanamine linker, followed by copper catalyzed alkyne-azide cycloaddition (CuAAC) reaction with alkyne modified ONs, whereas Das et al. first modified meso-tetra (4-carboxyphenyl) porphyrin to a tetra amino derivative, followed by a reaction with imidazole activated 5′-phosphate group on the ONs. Incubation with another similar tetra substituted porphyrin-DNA hybrid with complementary sequences facilitated the formation of extensive nanostructure networks (Figure 4B). In the specific work of Das et al. the porphyrin-DNA hybrid nanostructures were shown to exhibit specific antibacterial activity against Gram-positive bacteria, Staphylococcus aureus, when irradiated with white light due to generation of reactive oxygen species (ROS). The conversion of non-fluorescent dihydrorhodamine to fluorescent rhodamine was monitored to measure the production of ROS. The viability of growing bacterial cultures was measured (in colony forming units, CFU) with free porphyrin and porphyrin nanonetwork in light and dark conditions. The percentage CFU count was above 90% for all culture samples in dark. However, the percentage CFU count dropped to 10 ± 4%, and 2 ± 1.5% in the cultures mixed with free porphyrin and porphyrin nanonetwork, respectively when exposed to white light (100 W bulb) for 30 min. From these studies the authors concluded that the self-assembled porphyrin nanonetwork generates more ROS as compared to the free porphyrin system, and therefore is more effective in killing the bacteria. Here, the uniform dispersion of the porphyrin moieties separated by negatively charged DNA duplexes was believed to preclude the self-quenching of the porphyrin units thereby enhancing the efficiency of ROS generation.

In 2019, Pathak et al. used a guanine quadruplex (G-quadruplex) scaffold to generate nanostructures capable of arraying multiple porphyrins in close spacing (Pathak et al., 2019). In particular, a short G-rich sequence (G₄T₂G₄) was modified with a single porphyrin unit at an interior thymidine residue via coupling of an ON-appended N-hydroxy succinimidyl (NHS) ester and an amino porphyrin (Figure 4C). Next, three mono-azido per-O-methylated β-cyclodextrin (PMβCD) hosts were covalently attached to the remaining porphyrin arms via CuAAC. The PMβCD hosts hydrophobically encapsulate the porphyrin core, thereby masking the chromophore from the aqueous environment which was shown to be essential to preclude dye aggregation and fluorescence quenching. In addition, the intramolecular encapsulation also kept the porphyrin from disrupting the secondary structure of the DNA self-assembly. The self-assembly of the ON sequence resulted in directional G-quadruplex nanowires decorated with hundreds of bright porphyrin dyes with 1–2 nm spacing. These masked porphyrin assemblies were upto 180-fold brighter than the analogs lacking the PMβCD encapsulation. It is of note that Hamilton and colleagues in 2007 employed similar amine-NHS ester coupling to tether amino-functionalized porphyrin with NHS-ester modified G-rich ON sequences. Their work involved an investigation of the effect of porphyrin stacking interactions in the modulation of G-quadruplex assembly and stability (Jayawickramarajah et al., 2007).

Another highly selective reaction that can be conducted post ON-synthesis involves the reaction between hydrazides and carbonyl compounds to generate hydrazones. The hydrazide to hydrazone transformation can typically be considered as a bioorthogonal reaction as carbonyl compounds and hydrazides react selectively in a biological milieu without interfering with other biological metabolites. However, caution has to be applied to ensure the naturally occurring aldehydes/ketones and other highly reactive electrophiles are in low concentration to avoid undesired side reactions.

In 1996, Hélène et al. investigated the site-specific DNA damage of single-stranded, double-stranded, and triplex nucleic acids by tethering chlorins to corresponding DNA sequences followed by irradiation (Boutorine et al., 1996). Chlorins are porphyrin derivatives wherein one of the β,β′-pyrrolic double bonds are saturated, resulting, inter alia, in modulated optical and redox properties (Taniguchi and Lindsey, 2017). In this system, the DNA sequence provided target specificity with the chlorin acting as a photosensitizer. The researchers synthesized a chlorin-type photosensitizer, CHEVP (chlorin derived from heptaethylvinylporphyrin) by photochemical conversion of the corresponding vinyl porphyrin precursors. Next, the 5′-phosphate of the ON was activated by reacting with 2,2′-dipyridyldisulfide, triphenylphosphine, and 4-dimethylaminopyridine to yield 4-dimethylaminopyridinium derivative of the ON (Figure 5A). The activated ON was reacted with an adipic acid dihydrazide linker to obtain a hydrazide functional group at the terminal end. Finally, the aldehyde on the CHEVP moiety was reacted with the hydrazide functionalized ON to offer the required ON-chlorin conjugate. The ON-CHEVP (sequence: 5′-TTC​TTC​TCC​TTT​CT-3′) was incubated with a single strand, duplex, and hairpin loop structures to form a duplex and two triplex structures. When these hybrid assemblies were irradiated using 668 nm light under aerobic conditions, site-specific DNA cleavage at guanine positions was observed due to singlet oxygen generation by the photosensitized chlorin. As piperidine-labile photo-adducts were formed due to oxidative damage to guanine sites, the DNA cleavage was more pronounced when the samples were treated with 1 M piperidine.

Copper catalyzed click reaction between an alkyne and azide functionalities has revolutionized the way how chemical species are connected. This reaction is facile and can be performed under ambient aqueous conditions. In this reaction, terminal azide and alkyne groups react with each other and the major product is typically the 1,4-disubstituted 1,2,3-triazoles. Since alkyne and azide functional groups are uncommon in natural systems and the reaction proceeds fast to generate a stable triazole ring, this reaction is bioorthogonal. However, one challenge posed by CuAAC, especially for in vivo studies, is the toxicity of the Cu catalyst. To circumvent this issue, numerous copper-free versions of alkyne-azide click chemistry (e.g., strain promoted cycloadditions) have been developed for more biocompatible applications. Another popular reaction for tethering two components is the Diels-Alder cycloaddition, which is a pericyclic reaction between an alkene and a conjugated diene to yield a six-membered ring.

In 2014, Wellner and Wagenknecht synthesized porphyrin-ON conjugates via post-synthetic modification using both CuAAC and Diels-Alder cycloaddition reactions (Wellner and Wagenknecht, 2014). The team developed a uridine phosphoramidite functionalized at the 2’ ribose OH with either an alkyne group for post-synthesis CuAAC or a furan moiety for post-synthesis Diels-Alder cycloaddition, and thus achieved ON sequences with internal uridine modifications. They also functionalized meso-tetra-(4-N-methylpyridyl)porphyrin (TMPP) with azide and maleimide units for CuAAC and Diels-Alder cycloaddition, respectively. As seen in Figure 5B, CuAAC generates a simple triazole linker whereas, the Diels-Alder cycloaddition results in a linker with a tricyclic portion. These two linkers had a significant effect on the interaction between the ON backbone and TMPP. When the porphyrin-ON conjugates were hybridized with their complementary sequences to form DNA duplexes (Figure 5C), the bathochromic shift in the Soret and Q bands of the porphyrin units in UV-vis spectra along with the presence of a prominent negative peak at the porphyrin Soret region in CD spectra were considered to be the signature of intercalation of the porphyrin units in the DNA duplex. In contrast to the CuAAC product, in the tricyclic Diels-Alder product, the intercalation of TMPP in the DNA duplex was hindered (due to steric restrictions), which was evidenced by the absence of the bathochromic shift of the Soret peak and a significantly diminished CD peak, as compared to CuAAC product.

In 2014, Campidelli et al. functionalized meso-tetra(4-carboxyphenyl)porphyrin with 2-azidoethanamine trifluoroacetic acid salt to obtain the tetra azido derivative of the porphyrin dye (Clavé et al., 2014). The group then synthesized an ON (5′ GGA-GCT-GCA-GTT-CAU-propargyl 3′) to undergo CuAAC with the porphyrin dye to finally achieve a multi-branched porphyrin-DNA scaffold (Figure 6A). The resulting porphyrin-DNA conjugate was sequentially hybridized with a complementary 14-mer DNA strand containing a thiol group at the 3′ end and then incubated with gold nanoparticles (AuNPs). TEM analysis showed the hybrid nanostructures mostly possessing three gold nanoparticles instead of four most likely due to steric hindrances.

In 2018, Wang et al. used a porphyrin-DNA scaffold to develop energy transfer assemblies by tethering different numbers of ON sequences on to Zn-tetra-(phenylethynyl)porphyrin (ZnTPEP) and employing complementary sequences to the ON arms to dial in AlexaFluor 546 (A546) dyes (Figure 6B, top) (Anderson et al., 2018). An azido-modified DNA strand was reacted with the tetra-alkyne modified Zn-porphyrin via CuAAC resulting in five different porphyrin–DNA architectures (1-Arm, cis-2-Arm, trans-2-Arm, 3-Arm, and 4-Arm). Hybridization of the DNA arms with A546 modified complementary sequences led to the formation of the corresponding five different FRET assemblies (Figure 6B, bottom). The donor (A546) and acceptor (ZnTPEP) were shown to display effective FRET, demonstrating the utility of porphyrin-DNA hybrid constructs in light-harvesting applications. The experimental energy transfer efficiencies in the assemblies, 1-Arm, cis-2-Arm, trans-2-Arm, 3-Arm, and 4-Arm were found to be 75, 72, 85, 87, and 87%, respectively. While the energy transfer might depend on the actual orientation and microenvironment of the fluorophores in the assemblies, it was found that, in general, the energy transfer efficiency increased with an increase in the number of donor dyes in the FRET architecture.

The conjugation of two components using the mild reaction between an azide and a phosphine is named the Staudinger ligation. This reaction is one of the most important reactions in the bioorthogonal chemistry toolkit since it is rapid and highly chemoselective. While this reaction is compatible with biological conditions, the susceptibility of phosphines for oxidation in the presence of air and their poor water solubility have somewhat limited its scope. The seminal step of the reaction is the formation of a phosphorus ylide and is driven by the expulsion of N₂. The ylide may subsequently be trapped by a neighboring acyl group to afford a stable amide bond and an attached phosphine oxide. Among the reported Staudinger variations, the traceless method (where, in addition to the formation of an amide bond there is loss of the phosphorus-containing unit) introduced by Raines et al. is particularly amenable for the conjugation of peptides with other entities (Figure 7A) (Nilsson et al., 2001; Bednarek et al., 2020).

In 2010, Umezawa et al. used the traceless Staudinger ligation method for the conjugation of four peptide units onto a porphyrin core. In this paper, the researchers synthesized 5,10,15,20-tetrakis (3-azidophenyl) porphyrin by condensation of 3-azidobenzaldehyde with equimolar pyrrole (Umezawa et al., 2010). Also, the phosphine-containing amino acid partner was prepared according to the literature using a suitable phosphinothiol reagent (Soellner et al., 2002). Last, the ligation of the tetra azido porphyrin to the peptides was carried out in an anhydrous solvent to prevent early-stage hydrolysis. An example of one such reaction is shown in Figure 7B. Here, two equivalents of the amino acid derivative (for each azide group on the porphyrin) were introduced to give the tetra-substituted amino acid-porphyrin conjugate (Umezawa et al., 2010).

The key features of CuAAC include selectivity (azides are essentially absent in natural biological systems) and applicability (i.e., this reaction works in a broad range of pH and temperature and is significantly faster than the Staudinger ligation). Hence, CuAAC has become an ideal option when considering the ligation of porphyrin photosensitizers to peptides. Indeed, researchers have used CuAAC facilitated porphyrin-peptidic chimeras for applications like PDT (Ke et al., 2012; Sokolova et al., 2013; Acherar et al., 2015; Bryden et al., 2015; Ladomenou et al., 2016; Gazzali et al., 2017), PACT (Feese et al., 2019), and tumor imaging or targeting (Bakleh et al., 2009).

Ikawa et al., in 2009 synthesized a derivative of N-fused porphyrin (NFP) and conjugated this porphyrin with a nona-arginine (R9) peptide tail by “click” chemistry for the first time. As a result, N-fused tetraphenylporphyrin-R9 conjugate was water-soluble and possessed a two-step protonation behavior. The researchers selected NFP as the photosensitizer owing to its absorption of near-infrared (NIR) light over 1,000 nm. This NIR window has been identified as a wavelength suitable in the biomedical field for diagnostic and therapeutic applications due to less interference by tissues, blood, and bio (macro)molecules. Further, R9 is a member of a class of peptides with cell-penetrating ability. The synthesis of N-fused tetraphenylporphyrin-R9 conjugate was carried out by reacting mono-alkyne functionalized N-fused tetraphenylporphyrin with azido modified R9 attached on a Rink Amide resin. Cleavage from the resin was achieved by using trifluoroacetic acid (TFA) (Figure 7C) (Ikawa et al., 2009).

Another example of using CuAAC to generate porphyrin-peptide conjugates was reported by the Eggleston group, in 2016. These researchers used TPP as a template to prepare a mono-alkyne functionalized TPP for light triggered drug delivery and PDT applications. Before conducting the conjugation, zinc metal was utilized to prepare a zinc-porphyrin complex to avoid sequestration of the copper catalyst by the free-base macrocycle. The peptide conjugation partner was synthesized via Fmoc based peptide synthesis on Rink Amide resin, using benzotriazol-1-yloxytripyrrolidinophosphoniumhexafluorophos phate (PyBOP) activation. The original peptide was chosen based on the cell-penetrating peptide (CPP) sequence derived from the transcriptional activator (Tat) protein from the human immunodeficiency virus 1 (HIV-1). This decapeptide, known as HIV-1 Tat (48–57), has the sequence GRKKRRQRRR. CPPs are a class of “carrier” systems consisting of 8–30 amino acid residues that can translocate across biological membranes and transport different molecular cargos. In this work, the clickable peptide was synthesized via direct diazo transfer on the resin-bound Tat sequence using imidazole-1-sulfonyl azide hydrochloride, followed by cleavage from the resin (Figure 7D) (Goddard-Borger and Stick, 2007). For the CuAAC reaction, the azide functionalized peptide was reacted with an excess amount of mono-alkyne functionalized TPP in the presence of CuSO₄ and sodium ascorbate in a DMSO/H₂O/tBuOH mixture, and the requisite N-linked conjugate was successfully obtained after removal of the zinc metal (Dondi et al., 2016).

As an application of this work, in addition to classical PDT, the authors also focused on a novel mechanism of drug activation, namely, PCI. PCI is a photo-induced technique that enables drugs to escape the endosomal and lysosomal pathways and reach the targeted intracellular sites to maximize drug efficacy. In particular, many drugs are uptaken by endocytic mechanisms and are trapped in endo/lysosomes rendering them less effective. In PCI, a photosensitizer can localize into the endo/lysosomal membrane and rupture the membrane via singlet oxygen production, as a result of a sub-lethal light dosage. This rupturing event does not directly kill the cells but rather enables entrapped co-drugs to escape and reach their desired target. A critical requirement for such lysosomotropic agents is that they should be amphiphilic. As shown in Figure 7E, the CPP-TPP conjugate has an amphiphilic character as it includes a hydrophobic photosensitizer core (that can embed within the lipid bilayer of endosomal membranes) and a hydrophilic peptide arm that can associate with the exterior of the lipid layer. To confirm uptake and endolysosomal localization of their CPP-TPP conjugate, the authors used confocal microscopy on MC28 rat fibrosarcoma cells and the results showed the predicted sub-cellular localization (Dondi et al., 2016).

While CuAAC reactions are efficient at labeling biomolecules, the conjugation of porphyrins to peptides using this method has some limitations. First, these reactions need to be preceded by a metalation step (typically via the pre-formation of the Zn complex). Also, copper catalysts may be toxic to bacterial and mammalian cells, limiting applications of this reaction where cell viability is essential. On the other hand, the conjugation of alkynes with azides can also be catalyzed utilizing ring strain; specifically, strained cyclooctyne partners react with terminal azides without using copper (I) catalysis. However, it should be noted that the latter reactions have slower reaction rate constants compared to CuAAC. (Agard et al., 2004; Lang and Chin, 2014).

The Eggleston group has also investigated the SPAAC strategy for preparing CPP-TPP conjugates. Specifically, the researchers used dibenzocyclooctyne (DBCO) to synthesize mono-DBCO functionalized TPP. The complementary conjugation component, N-terminus azide functionalized CPP, was prepared using a similar strategy developed by this group and discussed above (Figure 7D). In the last step, the reaction of two equivalents of mono-DBCO functionalized TPP with N-terminal azido Tat peptide proceeded in DMSO to give the desired CPP-TPP conjugate (Figure 7F) (Dondi et al., 2016). In addition to demonstrating the expected sub-cellular localization, this CPP-TPP conjugate also showed the highest phototoxicity when compared to other conjugation strategies (e.g., CuAAC, thiol-maleimide, etc.), due to the extended triazole-based linker between hydrophobic porphyrin and the polycationic hydrophilic peptide, which reportedly is highly beneficial. In particular, phototoxicity was examined in monolayer cell culture of MC28 human breast cancer cells and the porphyrin dose required to induce 50% toxicity (LD₅₀) after 5 min illumination of blue light was ∼40 nM (compared to μM range when conducted in the dark). These exciting results make this conjugate a potential candidate for PDT applications. Last, to investigate the PCI effect, a protein toxin, saporin (which is a 30 kDa ribosome inactivating protein) was used as the co-drug. These experiments showed that the cell viability was significantly reduced (∼3-fold reduction) when the saporin was also introduced as compared to when the CPP-TPP conjugate alone was illuminated. Importantly, the viability reduction using saporin alone was small (∼10%) (Dondi et al., 2016).

In further work, the same group conjugated TPP derivatives to the C-terminal region of bombesin (BBN [7–14]) using different conjugation strategies including SPAAC, because this polypeptide can target the gastrin-releasing peptide receptor (GRPR), a protein that is overexpressed in several tumors, especially in human prostate cancer (PC-3) cells. First, mono-amino functionalized TPP was reacted with DBCO acid to prepare mono-DBCO functionalized TPP. The polypeptide, on the other hand, was assembled by Fmoc solid phase peptide synthesis on Rink Amide resin. Acylation of the N-terminal amino acid was performed on resin-bound BBN [7–14] using 11 azido undecanoic acid to afford the required azido functionalized polypeptide. Finally, the two components were conjugated via SPAAC to prepare TPP-BBN [7–14]. The cell-uptake studies (on PC-3 and HeLa cells) of this conjugate were attempted by confocal microscopy. However, the results showed no uptake in either cell line. Importantly, agglomeration in the cell media was observed (even after several washes with Hank’s Balanced Salt Solution). Hence, a likely cause for the poor uptake many be due to low solubility of the compound (inter alia, as a result of the long aliphatic linker). (Ciaffaglione et al., 2021).

One powerful tool for carbon-carbon double bond formation that is widely used in synthetic chemistry is the metal-catalyzed olefin metathesis. Overall, this reaction involves the scissoring of double bonds of olefins and the distribution of the fragments to produce new carbon-carbon double bonds. The reaction is catalyzed by ruthenium olefin metathesis catalysts such as the well-known Grubbs’ catalysts (Delaude and Noels, 2005; Kirshenbaum and Arora, 2008). Owing to the good activity and commercial availability of Grubb’s catalysts, olefin metathesis has been explored for the labeling of proteins and peptides. In terms of conjugating peptides to porphyrins, Sol and coworkers demonstrated, in 2008, that a glucosylated porphyrin dimer could be conjugated to a pseudo-pentapeptide, containing an arginyl-glycyl-aspartyl (RGD) sequence. In this work, the dimeric triglucosylated porphyrin derivative was synthesized according to previous reports, and the peptide unit was prepared using Fmoc-based solid-phase peptide synthesis. Aiming to insert two porphyrin units on the RGD sequence, the dimeric porphyrin was conjugated to the allylGlycyl-RGD-allylGlycine linear peptide on Rink Amide MBHA resin using the cross-metathesis reaction (Figure 8). Various reaction conditions were tested, including modulating the ratio of reactants, reaction time, temperature, and the use of Grubbs’ I and II catalysts. The best results were obtained when 4 equivalents of the porphyrin unit versus the peptide unit was utilized in the presence of Grubbs’ II catalyst. Upon reaction completion, the product was cleaved and deprotected from the resin (Sol et al., 2008).

The impetus for developing the dimeric porphyrin attached to the RGD peptide was for the potential application in PDT. Here, the authors used the dimeric porphyrin system to increase the amount of photosensitizer that can be activated by light. The two allyl glycine arms were included to provide conformational flexibility to attach porphyrins through olefin metathesis Moreover, the RGD sequence was included since it is established to bind to α_υβ₃ integrin (an extracellular matrix protein that is overexpressed in various cancers). Further, the design included the glucosyl units to engender amphipathic character that could enhance passive uptake of the photosensitizer. The sugar appendages may also modulate active uptake of the photosensitizer due to higher glucose uptake by cancer cells. (Chaleix et al., 2004; Chen et al., 2004; Sol et al., 2008).

The efficiency of ¹O₂ production was first evaluated by a trapping reaction using ergosterol acetate. In this technique, the ¹O₂ production efficacy of the dimeric porphyrin compound was compared to well-known ¹O₂ producers, such as hematoporphyrin (HP). The results showed that this conjugate has the same efficacy as HP which means that it could be a promising candidate for PDT applications (Böcking et al., 2000). In addition, the anticancer activity of this dimeric porphyrin conjugate against K562 cells was also evaluated and the photocytotoxicity results showed that dead cell counts in the presence of the dimer were lower than those with Photofrin (a clinical PDT photosensitizer) (Sol et al., 2008).

Maleimide chemistry is an attractive choice for bioconjugation due to the relative ease of accessing functionalized maleimides and their excellent reactivity with thiols. The thiol side-chain containing Cys residues are generally in low abundance within most peptides and proteins, hence making them an ideal choice for fairly selective bioconjugations. In addition, thiol functional groups are significantly superior nucleophiles than amino groups and, in particular, react faster with the double bond of a maleimide under physiological conditions. Taken together, these factors engender thiol-maleimide ligation a salient tool for peptide bioconjugation (Figure 9A) (Wang et al., 2012; Ravasco et al., 2019; Chu et al., 2021).

In 2012, Liu et al. utilized the thiol-maleimide ligation strategy for the attachment of a photosensitizer protoporphyrin (PpIX) to a lipopolysaccharide (LPS) binding antimicrobial peptide Y113WF (YVLWKRKRKFCFI-amide). The goal of this work was to investigate the bioconjugate for PACT activity and intracellular fluorescent imaging (Liu et al., 2012). PACT is a method that utilizes non-toxic (in the dark) photosensitizers to kill microbes by nonspecific oxidative damage by ROS only upon irradiation with light. In fact, energy transfer from the photosensitizer to oxygen can ultimately lead to the production of various reactive oxygen species (H₂O₂, OH^●, O₂ ^●−, ¹O₂) that are lethal to microbial pathogens (Bullous et al., 2011; Pereira et al., 2014; Meng et al., 2015; Orosz et al., 2017). Briefly, bis-maleimide functionalized PpIX was reacted with peptide Y113WF in DMSO in presence of DIPEA to achieve dimeric PpIX-Y113WF conjugate. Next, the effectiveness of fluorescent imaging and PACT on Gram-negative bacterial strains based on this conjugation was evaluated (Figures 9B,C). Due to the high binding affinity of the Y113WF peptide sequence to the LPS component, the photosensitizer was able to produce localized ROS. Specifically, these studies demonstrated that the dimeric PpIX-peptide conjugate was ca. 99% lethal to bacteria at 0.5 μM concentration upon light irradiation (30 J/cm²). Furthermore, the authors showed that the efficacy of bacterial photoinactivation was concentration-dependent (and could reach 99.9%). As importantly, the dimeric PpIX-Y113WF conjugate also demonstrated the ability to image, in real-time, bacterial strains via fluorescence emission. Further photo-inactivation experiments and fluorescence imaging studies illustrated that the dimeric conjugate can selectively target bacterial strains over mammalian cells—an important requisite for PACT (Liu et al., 2012).

Another example of thiol-maleimide ligation was shown by the Eggleston group who conjugated CPPs to TPP. Here, the maleimide functionalized TPP was first synthesized via a peptide coupling reaction between mono-amino TPP and 3-maleimidopropionic acid. The complementary thiol-containing peptide was obtained by reaction of Fmoc-L-Cys with resin-bound peptide using HATU activation. In the final step, the N-terminal functionalized cysteine residue on the peptide was reacted with two-fold excess of the maleimide functionalized TPP to offer the CPP conjugate in excellent yield (Figure 9D) (Dondi et al., 2016). After showing, via fluorescence microscopy, that the maleimide tethered CPP-TPP conjugate could indeed localize in lysosomes of MC28 rat fibroblasts, the phototoxicity of the conjugate was tested. In particular, illumination of the conjugate in monolayer culture of MC28 cells showed an LD₅₀ in the range of 100 nM which was higher than for CPP-TPP conjugates formed by SPAAC (∼40 nM) (Strain-Promoted Alkyne-Azide Cycloaddition section) but was much lower than the μM range obtain without illumination. Also, PCI studies showed that the cell viability was significantly reduced (∼three-fold reduction, similar to SPAAC ligation) (Dondi et al., 2016).

A complementary reaction that can be used to link porphyrinoids to cysteine residues of peptides is the thiol-haloacetamide ligation. In particular, bromo- and iodo-acetamides have been used as electrophilic sites (Chaloin et al., 2001; Arai et al., 2003) with the thiol serving as a nucleophile. Since the thiol on cysteine is more nucleophilic than hydroxyl and amino side-chains on other amino acids, this reaction is selective for cysteine ligation (Giuntini et al., 2011). The principal advantage of thiol-haloacetamide reaction over the thiol-maleimide reaction is the greater stability of the thioether linkage product in the former reaction. The thioether linkage produced from the thiol-maleimide reaction is considered less-stable since it can slowly undergo a retro-Michael reaction to reform the maleimide group which can then react further with other biomolecules. In order to overcome this problem, the thiosuccinimide ring product can be hydrolyzed after conjugation. On the other hand, the thiol-haloacetamide reaction has a slower reaction rate and also requires a higher pH to proceed. This makes the thiol-haloacetamide reaction less chemoselective, because at pH > 7.5 free primary amino groups can compete with thiols to react with the haloacetamide (Schelté et al., 2000; Szijj et al., 2018).

In 2003, Arai et al. tethered a porphyrin to a pentapeptide to mimic natural multiple porphyrin clusters. The pentapeptide sequence (CKVKV) includes alternating hydrophilic and hydrophobic residues and forms a random coil structure in aqueous trifluoroethanol. However, upon conjugation with the porphyrin at the N-terminal cysteine, the pentapeptide assembles into a β-sheet structure due to porphyrin-based pi-stacking. This tendency to form a β-sheet structure is interesting due to its resemblance to natural porphyrin self-assemblies (Chaloin et al., 2001; Arai et al., 2003). The preparation of the bromoacetamide linked porphyrin partner involved first the synthesis of mono-nitro porphyrin using a modified Lindsey method followed by reduction (via SnCl₂) to afford aniline containing porphyrin. Condensation of this porphyrin with bromoacetic acid along with EDC•HCl provided the required porphyrin conjugation agent. The bromo acetamido-linked porphyrin was next reacted with the Cys-containing pentapeptide to afford the porphyrin-peptide conjugate, see Figure 9E (Arai et al., 2003).

The reaction of amine groups with activated carboxylic acids can often be slow and lead to the generation of by-products such as N-acyl ureas (Clarke and Boyle, 1999). One method to circumvent these byproducts is to react amines with isothiocyanate derivatives, resulting in thiourea linkages (Smith et al., 2011). Indeed, isothiocyanate (NCS) functionalized porphyrins can be reacted with amines under mild conditions with no by-products formed (Staneloudi et al., 2007). As an example, in 2005, Vazquez et al. conjugated an NCS linked porphyrin to poly-L-lysine to investigate the effect of the number and nature of amino acid conjugates on cellular uptake. Here, the mono NCS-tethered porphyrin was first obtained by exposing mono-amino porphyrin to 1,1′-thiocarbonyldi-(2H)-pyridone. Next, appropriately protected peptides were conjugated to the NCS porphyrin in a mixture of DMF/triethylamine and the requisite product was isolated after deprotection with TFA (Figure 10A). Finally, the dark toxicity (toxicity in absence of light irradiation) and cellular uptake of the porphyrin-peptide conjugate were investigated in human Hep2 cells. The results showed low dark toxicity of this conjugate at a concentration of 10 μM (IC₅₀ > 250 μM) (Sibrian-Vazquez et al., 2005).

The reaction between amines on peptides and activated carboxylic acids on porphyrin derivatives have been extensively investigated (Tomé et al., 2004; Gabriel et al., 2007; Giuntini et al., 2011; Shi et al., 2011; Dosselli et al., 2013; Orosz et al., 2013; Li et al., 2015; Jiang et al., 2016; Chen et al., 2018; Liu et al., 2018; Ciaffaglione et al., 2021). For example, in 2011, Mező et al. conjugated one and two tetrapeptides to tri-cationic meso-tri (4-N-methylpyridyl)-mono- (4-carboxyphenyl) porphyrin (TMPCP) and bi-cationic meso-5,10-bis(4-N-methylpyridyl)-15,20-di- (4-carboxyphenyl)porphyrin (BMPCP) respectively. Their goal was to design porphyrin-peptide conjugates with enhanced cellular uptake and delivery and to facilitate DNA binding. The conjugation partners, TMPCP, BMPCP, and tetrapeptide Ac-Lys (Ala-D-Ala-Ala)-NH₂ were prepared using standard literature procedures (Habdas and Boduszek, 2005). The amide bond formation was investigated using two different strategies. In the first method, TMPCP and the tetrapeptide were reacted in the presence of DIEA/BOP in DMF (Figure 10B) which resulted in a good yield. While reaction between BMPCP and two tetrapeptides produced many side products. In the second method, EDC was used as the activation reagent in lieu of BOP. Here, the reaction between BMPCP and two tetrapeptides resulted in higher reaction yield (Figure 10C), while no significant improvement was observed for the reaction between TMPCP and a single tetrapeptide. Moreover, DNA binding of porphyrin-peptide conjugates was evaluated by spectroscopic methods due to changes in photophysical properties of bound porphyrins upon interaction with anionic DNA. Here, results from a combination of absorption spectra, fluorescence decay, fluorescence energy transfer, and CD spectroscopy demonstrated that both di- and tri-cationic porphyrins bind to DNA by two different mechanisms including intercalation and external binding. The authors concluded that the tricationic structure and peptide appendages without negative charge are preferable for binding to DNA (Mező et al., 2011).

An active area of research is where porphyrin-biomacromolecule conjugates have applications in artificial light harvesting. Towards this goal, the Solladie group has developed model systems using peptides attached to multiple porphyrins prepared via amide coupling (Solladié et al., 2000, 2001; Aubert et al., 2002; Aziat et al., 2008). Their synthetic strategy has focused on in-situ carbodiimide-based activation of a carboxylic acid functionality on porphyrins to build a monomer, and subsequent oligomerization via further amide couplings. For example, a porphyrin bearing one carboxylic acid arm was reacted with L-lysine (protected with Boc and allyl groups on its N- and C-terminus, respectively) in the presence of N,N′-dicyclohexylcarbodiimide (DCC) and HOBT in DCM. The resulting lysine-porphyrin (P(H ₂ P) ₁ ) was metallated with zinc acetate to afford P(ZnP) ₁ . Next, an iterative synthetic strategy was employed to prepare dimers, tetramers, and octamers (Figure 11A). First, one of the coupling partners was prepared by deprotection of the amine group on P(ZnP) ₁ (with 1 M TMSCl). Next, the other partner was obtained by deprotection of the carboxylic acid end on P(ZnP) ₁ using tetrakistriphenylphosphine palladium (0). Finally, the amine and carboxylic acid-deprotected versions of P(ZnP) ₁ were reacted together using DCC and HOBT to afford the porphyrin dimer P(ZnP) ₂ . The syntheses of the tetramer P(ZnP) ₄ and octamer P(ZnP) ₈ were carried using similar procedures (Solladié et al., 2000).

Key requirements for artificial light-harvesting systems are efficient light funneling and charge separation processes. Since α-polypeptides can form defined secondary structures (especially α-helices), the multiple porphyrin units could serve as the light capture moieties and the helical peptide conformation could align the porphyrins to promote exciton migration (Solladié et al., 2020). Regarding the charge separation step, a supramolecular strategy was employed wherein the metallated porphyrins could also serve as hosts for fulleropyrrolidines (bearing pyridine, PyC₆₀ or imidazole, ImC₆₀, sidearms) that are well-known for their electron-accepting capacity (Nair et al., 2014). When the group studied the photophysical capabilities of these complexes (Figure 11B), it was observed that the lifetime of the charge-separated state was enhanced when higher generation oligomers were used, with the P(ZnP) ₈ -ImC ₆₀ ensemble achieving a rather long lifetime of 0.84 ms (Solladié et al., 2020).

In addition to in situ activation of carboxylic acids, it is often useful to prepare activated, yet stable, carboxylic acid esters that can be stored and used as needed. For example, in 2017, Zou et al. prepared N-hydroxysuccinimidyl esters of glutaric acid tethered porphyrins (TPP-G-NHS) that were then reacted with the L-phenylalanine-L-phenylalanine (FF) dipeptide, under basic conditions to afford peptide-porphyrin conjugate (TPP-G-FF) (Figure 12A). The goal for preparing such conjugates was to self-assemble peptide−porphyrin nanodots (PPP-NDs) for photothermal applications, wherein light energy absorption leads to heat energy release in a controlled manner, resulting in cancer cell death. Notably, this method is different from PDT where the absorbed light energy leads to the generation of ROS (Zou et al., 2017). In this work, the researchers used the FF peptide since it is biodegradable, biocompatible, has functional versatility, and low immunogenicity (Yan et al., 2010). Moreover, FF can co-assemble with different functional molecules through non-covalent interactions to form stable nanomaterials with controllable morphology (Zou et al., 2014, 2016). Nanodot formation was driven by the tendency of hydrophobic porphyrins to form J-aggregates via strong π-stacking interactions (Figure 12B). These interactions resulted in the quenching of porphyrin fluorescence thus switching the system towards deactivation via heat loss. The PPP-NDs exhibited high stability (even after dilution and irradiation) and afforded a relatively high light-to-heat conversion efficiency (54%). Further, in vivo studies on mice models demonstrated that the PPP-NDs accumulate in tumors and are efficient at inhibiting tumor growth.

While most applications of porphyrin-peptide conjugates rely on the photophysical properties of the porphyrins, the Hamilton group elegantly demonstrated that these macrocycles could also be used as scaffolds to project peptides in a defined manner to develop targeted protein surface binders. Indeed, this group has shown that such porphyrin-peptide systems can selectively recognize various protein surfaces (inter alia via charge complementarity), including cytochrome c (cyt c), myoglobin, and ferrodoxin (Jain and Hamilton, 2000; Zhou et al., 2006). The porphyrin-peptide protein surface binders are based on tetra-meso-carboxylphenylporphyrin (TCPP) as the hydrophobic core. TCPP was conjugated with two different peptides from a pool of five peptides. Specifically, the coupling step involved the reaction of 1 equivalent of TCPP with an excess of the two different peptides in the presence of PyBOP, which yielded six different products for each reaction: X₄, Y₄, X₃Y, XY_3, and two geometrical isomers of X₂Y₂. These individual products were isolated (and deprotected) for each set of coupling reactions leading to a total of 35 unique porphyrin-peptide conjugates decorated with charges varying from +8 to −8 and having various hydrophobic appendages (Figure 12C) (Baldini et al., 2004). With this sizable pool of porphyrin-peptide derivatives, a rapid screening strategy was developed (based on the porphyrin fluorescence intensity change) to identify high-binding-affinity porphyrin receptors for the targeted proteins (Baldini et al., 2004; Zhou et al., 2006). Figure 12D shows an illustration representing the sensing platform where strong binders are fluorescence quenched. From these experiments, the authors found that charge complementarity was critical with e.g., ferrodoxin interacting with positively charged porphyrin conjugates such as porphyrin-Lys₂-Lys 1,5 pentadiamine₂ whilst cytochrome c shows higher binding to the negatively charged receptors (e.g., porphyrin-Asp₄).