Strapped Porphyrins as Model Systems for Atropisomeric Photosensitizer Drugs

The time dependence of atropisomer interconversion has limited the pursuit of single atropisomer drug candidates, even in circumstances where one atropisomer presents favorable biological activity over another. Moderate interconversion energy barriers risk compromising drug stability. As a result, examples of atropisomerically pure drugs in current clinical use are rare. However, in recent years, there has been a shift towards the development of single, stable atropisomer drug candidates with enhanced activity. Consequently, development of methods which effectively restrict rotation in a configuration which favors activity is highly beneficial. The picket fence porphyrin α 4 atropisomer configuration has been previously demonstrated to improve the cell internalization of the pre-clinical drug, redaporfin, applied in photodynamic therapy. In this work, the α 4 configuration was modelled with novel porphyrin photosensitizers through strapped moieties which effectively fixed the atropisomeric configuration. The stable cis - αα configuration demonstrated enhanced cell membrane permeation, effectively predicting the behavior of the α 4 configuration and indicates that strapped porphyrins can serve as stable model systems for the investigation of photoactive drugs.


Introduction
An intrinsic issue for atropisomeric drugs is the time dependence of interconversion with dramatically different racemization half-lives ranging from minutes to years.With the potential to undermine the efficacy and production of stable drug substances, it is perhaps unsurprising that atropisomerism has been dubbed a 'lurking menace' for drug development. [1]Even atropisomers with large interconversion barriers (� 30 kcal.molÀ 1 ), considered class 3 atropisomers, require careful evaluation. [2]As with chiral stereoisomers, pre-clinical assessment of atropisomer drug candidates generally involves consideration of atropisomer pharmacokinetic profiles due to the potential racemization of pure isomers upon metabolism. [1]n cases where an individual atropisomer presents a favorable biological activity, alterations to molecular structure which increases rotational barriers in a manner that preserves metabolic stability are highly desirable.
A prevalent strategy to improve stability involves an increase of steric bulk around the atropisomer axis. [2]Additionally, there are cases where a rotating axis interconversion energy barrier can be increased through creation of a bridge. [3]or photodynamic therapy, a specific tetrapyrrole atropisomer configuration, the α 4 atropisomer, was previously identified to increase cellular internalization and consequently enhanced phototoxicity in vitro and in vivo. [4]The source for this varied efficacy was rationalized in terms of the atropisomer molecular structure, with the distribution of polar substituents oriented to one side of the macrocycle plane, enhancing interaction with the cell-membrane and improving passive permeation.In the present work, restrictive strap moieties were used in novel porphyrin photosensitizers to enforce an α 4 configuration (Figure 1) while effectively locking the axis of rotation for enhanced stability.
Strapped porphyrins feature covalent connections between either β-β- [5] or meso-meso- [6] positions of the porphyrin macrocycle. [7]These systems were originally developed to gain understanding of the structure and activity of biologically active sites.The domed shape of deoxyhemoglobin, known to feature as part of the heme protein, was pursued. [8]Linkage at phenyl substituents of 5,10,15,20-tetraphenylporphyrins gave rise to a variety of strapped porphyrin shapes with names derived from macroscopic objects including 'capped porphyrins' and double strapped 'basket handle' porphyrins. [9]Investigation of biomimetic models was dominated by the pursuit to induce porphyrin distortion. [10]10b] Single strapped meso-linked porphyrins were also prepared to target a cytochrome c oxidase model by Wytko et al. through condensation between a dipyrromethane and dialdehyde. [11]However, a more accurate model was later obtained with a double strapped porphyrin through postfunctionalization of porphyrin atropisomers with linkers. [12]trapped porphyrins have featured in areas beyond biomimetic models as chiral catalysts [13] and scaffolds in molecular engineering [14] as well as potential oxygen storage devices with anthracene straps. [15]hotodynamic therapy (PDT), which effectively combines a photosensitizer drug, light of a specific wavelength and molecular oxygen to mediate a oxidative therapeutic effect, tend to feature the tetrapyrrole motif. [16]However, examples of strapped porphyrins as photosensitizers (PSs) are limited.Regardless of the synthetic route selected, isolation and preparation of stereoisomerically pure strapped porphyrins in reasonable yield remains a challenge. [17]In addition to low synthetic yields, the lack of investigation may be partly due to the primary focus of these systems generally involving macrocycle distortion.With short strap lengths, a non-planar conformational change may occur which quenches fluorescence quantum yields. [18]Macrocycle distortion may also lead to reduced singlet oxygen quantum yield generation, [19] a property detrimental to photodynamic therapy efficacy.Therefore, strap length for integration in PSs must be tuned to ensure planarity.Glycosyl containing single strapped porphyrins have been studied as PSs and compared to hematoporphyrin derivative.The porphyrins were prepared by a BF 3 catalyzed direct condensation of a dipyrromethane and dialdehyde.A mixture of 5,10-and 5,15-single strapped porphyrins were derived with the latter showing enhanced phototoxicity in vitro. [20]Double strapped systems, specifically calix [4]pyrroles, have been dem-onstrated to act as synthetic transmembrane carriers with the cis-αα configuration providing a suitable aromatic cavity for proline transport. [21]The potential for these systems to enhance cellular uptake is high.Despite potentially low yields, interest in strapped porphyrin synthesis continues [22] alongside their role as heme models. [23]n this work, we introduced strapped linkages to porphyrins intended to alter membrane passage.Alkyl and phenyl straps were used to lock rotation of three meso-tetraarylporphyrins and to form double strapped porphyrins.These strapped porphyrins have linkages at the 5,10-and 15,20-meso-positions which direct meso-aryl substituents above or below the macrocycle plane.Depending on strap orientation, they are referred to as cis-αα, cis-αβ and trans-αβ.Such isomers are analogous in configuration to porphyrin atropisomers, α 4, α 2 β 2 and αβαβ, respectively (Figure 1).Polar aryl substituents were incorporated to double strapped porphyrin design to achieve enhanced membrane permeation for the cis-αα isomer.Direct condensation methods were employed to pursue both cis-isomers for in vitro comparison.Structural elucidation was assisted through a combination of NMR spectroscopy and single-crystal X-ray crystallography.Photophysical studies provided insight into the relative planarity between isomers.Membrane passage was determined in two cell lines with the cis-αα isomer exhibiting preferential uptake, illustrating the advantage of the α 4 motif to enhance internalization.

Strapped Porphyrin Synthesis
Double strapped porphyrins were prepared using a direct condensation reaction between a dialdehyde and pyrrole.This approach was favored to ensure cis-isomers were the major product, [9] as we aimed to investigate whether the locked cis-αα configuration, an analogue of redaporfin α 4 , conferred an advantage in membrane passage over other isomers.Linker straps were selected with two requirements.Firstly, the length of the strap needed to be long enough to ensure an absence of macrocycle distortion.Such distortion is inversely proportional to linker length. [8,24]Porphyrin planarity was pursued due to the decrease in singlet oxygen generation which tends to occur with distorted porphyrin macrocycles, [19] inhibiting PS potential. [25]Secondly, it was necessary to ensure that strap length was short enough to prevent flexible s-shaped straps and free rotation. [9]Additionally, the strap chosen was hydrophobic in order to induce amphipathic behavior in the case of cis-αα, with hydrophobic straps on one side of the macrocycle and polar hydrogen bonding groups on the other.The stability of an alkyl chain strap was first targeted with a hexyl chain.This strap was previously used for double strapped porphyrin preparation in the absence of distortion by the Lindsey group with o,o'-linked straps.Increase of an alkyl strap length to eight carbons or more was proposed to lead to undesirable, flexible s-shaped conformations. [9] series of dialdehydes were prepared by alkylation of the appropriate commercially available salicylaldehyde derivative; 5-nitrosalicylaldehye or 5-(methoxycarbonyl)salicylaldehyde, according to the procedure of Scheme 1.The alkylation method was adapted from Charisiadis et al. [26] and featured activation of the salicylaldehyde derivative precursor by stirring for 1 h in the presence of K 2 CO 3 in DMF prior to the addition of the brominated linker.Dialdehydes 1-3 were obtained in adequate yields, suitable for multigram condensation.Similar linker types have been prepared by Wagner et al. [9] and Reddy et al. [27] with comparable yields.
A common obstacle for the preparation of double strapped porphyrin cis-isomers is low yields.This was somewhat met by the one-flask synthesis developed by the group of Lindsey. [28]owever, the alkyl linker dialdehyde 1 demonstrated limited solubility in solvents acceptable for this strategy including CH 2 Cl 2 or CHCl 3 .As a result, condensation was first attempted using adapted Adler-Longo conditions [29] of refluxing dialdehyde and pyrrole in propionic acid, originally optimized for direct condensation of pyrrole and dialdehyde by Momenteau et al. [30] Conditions are shown in Scheme 2. After 3 h of heating at 150 °C, purification by chromatography to remove pyrrole impurities proceeded.A mixture of three porphyrins was indicated by TLC.The combined porphyrin yield was inadequate (< 0.5 %) to allow complete characterization following separation.Mass spectrometry analysis (MALDI) indicated that each fraction corresponded to the mass of the targeted strapped porphyrin 4 (Scheme 2).In an attempt to improve the yield, Lindsey conditions [9] were used with high dilution.Marginally greater solubility of dialdehyde 1 was observed in CH 2 Cl 2 over CHCl 3 and was, thus, selected for condensation with BF 3 • OEt 2 as a catalyst.After 24 h stirring at rt under argon, DDQ was added as an oxidation step and neutralized with TEA.Two fractions were separated by chromatography.
Characterization confirmed that two strapped porphyrin 4 isomers were obtained using the Lindsey conditions.A low isolated yield (< 1%) in comparison to similar o,o'-linked strapped porphyrins (15 %) was attained. [9]This was likely caused by the low solubility of the dialdehyde 1 in CH 2 Cl 2 which led to incomplete condensation of starting material.This was apparent from a dialdehyde peak visible at 10.3 ppm in 1 H NMR following porphyrin separation, contaminating the cis-αα-4 porphyrin.The low solubility of the dialdehyde 1 across a variety of organic and aqueous solvents prevented chromatographic purification of the porphyrin.Metalation with zinc using Zn(OAc) 2 in CH 2 Cl 2 sufficiently altered the porphyrin polarity relative to the dialdehyde 1 to allow recrystallization from ethyl acetate.The pure Zn-porphyrin 4 subsequently underwent demetallation with TFA to yield the cis-αα-4.Dialdehyde solubility limited condensation and ultimately led to extensive purification steps which likely contributed to a reduced yield.
Elucidation of the strapped porphyrins 4 configuration was possible due to 1 H NMR and single-crystal X-ray crystallography.Mass spectrometry (MALDI) analysis and absorption spectra were identical for both strapped porphyrin fractions following separation.The order of elution of similar double strapped porphyrins on silica gel has previously been reported as transαβ followed by cis-αβ and cis-αα. [17,30]For the two porphyrin fractions prepared using Lindsey conditions, retention factors aligned with the latter fraction of the Adler-Longo condensation on TLC.This suggested that the two porphyrins were the cisisomers, which conforms with the almost exclusive formation observed by the Lindsey group during method development for the condensation. [9]While the 1 H NMR spectra of the alkyl strap porphyrins (Figure S4 and S5) were similar, they displayed distinctive resonances for β-pyrrolic protons.The first fraction displayed eight magnetically equivalent β protons as a singlet.This is generally a feature more consistent with the trans-αβ symmetry. [27,30]However, the methylene protons of the hexyl strap did not exhibit the characteristic upfield shift associated with the diamagnetic anisotropy typical of trans configurations. [31]The second fraction β-pyrrolic protons split into two singlet resonances in agreement with anticipated symmetry of the cis-αα configuration. [30]However, 1 H NMR could not ambiguously distinguish the two isomers' configuration alone.
Alkyl-strapped porphyrin configurations were both confirmed as cis-isomers by single-crystal X-ray crystallography.Crystals were grown from the chromatographic fractions and structures could be conclusively assigned as cis-αβ for the first fraction to elute and cis-αα for the second (Figure 2 S1).The order of elution was in agreement with literature sources. [17,27,30]The cis-αβ-4 isomer porphyrin core adopted a primarily planar conformation (Figure 2A).The asymmetric unit was identified as two unique half molecules.The cis-αα-4 isomer (Figure 2B  subunits were identical for the cis-αβ-4 isomer as 2.31 Å.The cis-αα-4 isomer showed small deviation between the two bonding distances, 2.33 and 2.31 Å, respectively, consistent with the observed minor deformation.Crystal packing is presented in Figure S13.Interconversion of the straps at rt was not observed in solution by 1 H NMR over approximately one week and the isomer configuration assigned to each fraction were the sole components of each. To improve the double strapped porphyrin yield, dialdehydes with greater solubility under condensation conditions were pursued with phenyl-linked dialdehydes 2 and 3 (Scheme 1).A more rigid strap linker was anticipated to lead to improved strapped porphyrin yields. [9]Both dialdehydes were acquired with similar success to dialdehyde 1 but enhanced chloroform solubility was observed.This facilitated the Lindsey condensation method with dialdehyde 2. Selective porphyrin isomer formation was indicated by TLC after condensation and an initial column chromatography to remove pyrrole impurities. 1 H NMR analysis indicated a single double strapped porphyrin with a cis configuration due to two characteristics (Figure S6); βpyrrolic protons were observed as two singlet signals.Additionally, the splitting observed for methylene proton peaks at 5.67 and 5.00 ppm, correlated to an equivalent 13 C resonance at 71.79 ppm by HSQC (Heteronuclear Single Quantum Coherence) analysis.Thus, these methylene carbons were indicated to be potentially orientated in the same environment relative to the macrocycle plane as a cis-αα isomer.
Selective formation of a single double strapped porphyrin is unlikely.The Lindsey group [9] was confronted by the same issue during the preparation of a series of m,m'-linked strapped porphyrins.They discovered through 1 H NMR analysis that the porphyrins were rapidly interconverting.This was unlikely to be the case with strapped-porphyrin 5, as the Lindsey group had only observed interconversion with longer alkyl strap units. [9]herefore, it may be possible that the other isomer formed but to such a lower extent that it was not possible to isolate during purification.The preferred formation of cis-αα over cis-αβ has been previously established, initially by Momenteau et al. [30] for the Adler-Longo condensation and later confirmed for the Lindsey condensation method by the Lindsey group. [9]he use of the rigid linker in dialdehyde 2 proved beneficial for the Lindsey condensation with an isolated yield of 8.1 %.In order to prepare the other isomer, the Adler-Longo condensation method was employed.Three strapped porphyrins were indicated by TLC (Table 1).Repetitive chromatography purification was required to isolate the three fractions.A substantially lower combined yield, 1.25 %, was observed for this procedure.The elution order, previously established for porphyrins with similar straps [27] and comparison of 1 H NMR, confirmed the dominant Lindsey condensation product as cis-αα.The 1 H NMR of trans-αβ-5 (Figure S8) indicated the characteristic features of symmetry associated with trans double strapped porphyrins, such as eight magnetically equivalent β-pyrrolic protons and shielding of the strap's methylene protons due to its position stretching the porphyrin plane. [31]Both cis-isomers were formed in a higher proportion with respect to the trans-αβ, a common outcome of these reaction conditions. [30]he stability of the strapped-porphyrin cis-αα-5 isomer was probed.The cis-αα isomer was selected due to the lower thermal stability observed by Urbani et al. [17] An initial attempt was made to induce strap rotation through heating a sample of the strapped porphyrin for 24 h in toluene at 111 °C.However, no conversion was observed.Subsequent attempts in DMF (153 °C) and DMSO (189 °C) lead to no observable conversion by TLC.The strapped porphyrin was deemed stable against conversion at elevated temperatures.
The number of double strapped porphyrins formed varied depending on the condensation protocol employed (Table 1).Lindsey condensation conditions resulted in formation of only cis configurations.The Adler-Longo reaction yielded the cis configuration with preference over trans isomer formation. [30]his trend, which conforms with similar strapped systems in the literature, has prompted alternative preparative strategies specific to the trans isomer including post-functionalization of the porphyrin with the strap [30] or condensation between a bisdipyrromethene and dialdehyde. [15]In both the Lindsey and Adler-Longo condensations, the cis-αα isomer dominated, consistent in the preparation of double strapped porphyrins with both respective strategies in this work. [9,30] consistent issue encountered with the isolation of each strapped porphyrin was low solubility across a variety of solvents, which hindered characterization during preparation.TFA was employed to solubilize all porphyrins for 1 H NMR analysis via dication salt formation.A new double strapped porphyrin 6 was targeted in an attempt to improve solubility.The strap length of dialdehyde 2 was increased to attempt to alter the number of porphyrins accessible from the Lindsey condensation.However, Adler-Longo conditions were necessary to form multiple porphyrins: cis-αα-6 and trans-αβ-6.The strapped porphyrin cis-αβ-6 was not observed to form.Carboxylic acid functionality was introduced by hydrolysis using KOH as a base of methyl-ester groups following condensation.The solubility of the novel porphyrins was significantly higher than that of strapped porphyrins 5. Structural characterization of strapped porphyrins 6 was confirmed by single-crystal X-ray crystallography (Figure 3, Figure S13C,D and Table S1).Saddle distortion [32] of the porphyrin macrocycle was apparent with the COOMe-trans-αβ-6 isomer.

Strapped porphyrin photophysical and photochemical characterization
Evaluation of the photophysical properties of the double strapped porphyrins provided valuable insight into the relative planarity of strapped porphyrin isomer macrocycles.Results are presented in Table 2 and Figure S14.Strapped porphyrins 4 with moderate length, hexyl straps were indicated by single-crystal X-ray crystallography to share a lack of significant distortion.Overlapping absorption spectra (Figure S14) confirmed that there were no major changes in the porphyrin macrocycle conformation between the isomers which would alter photophysical properties.If macrocycle distortion was present, a bathochromic shift would be anticipated. [8,19]The cis-isomers of strapped porphyrin 5 also exhibited overlapping absorption spectra with the Soret band (λ Soret ) at 428 nm; however, the trans isomer demonstrated a bathochromic shift of 10 nm to 434 nm.This would indicate that when the phenyl straps of porphyrin 5 are linked between positions 5,15 and 10,20, some distortion is introduced to the porphyrin macrocycle.A similar shift was observed for the trans-αβ-6 isomer, indicating distortion, which was confirmed by single-crystal Xray crystallography (Figure 3B).The propensity of these rigid linkers to induce slight distortion has previously been observed by Reddy et al. [27] An accompanying indication of distortion was not only a bathochromic shift but significant reduction in molar absorptivity.With the introduction of a rigid linker in strapped porphyrins 5 and 6, lower molar absorption coefficients were observed relative to strapped porphyrins 4; however, this was particularly the case for trans-αβ-5 and trans-αβ-6 isomers.Reddy et al. [27] rationalized the changes by deviations in the intensity of Soret and Q bands, specifically with a trans-αβ isomer, due to changes in degeneracy of the π-molecular orbitals of the porphyrin ring, resulting in red shifts and a significant decrease in intensity.Fluorescence quenching tends to accompany these signs of distortion for absorption spectra. [18]Indeed, the absolute values of fluorescence quantum yield (Table 2, Φ F ) of the double strapped porphyrins cis-αβ-5 and trans-αβ-6, recorded using an integrating sphere, were both lower relative to the respective cis-αα isomers.
The consequences of distortion for the trans-αβ-6 were apparent with a slightly lower singlet oxygen quantum yield (Table 2, Φ Δ ) value.A decrease in singlet oxygen quantum yields has been established as relevant to the planarity of the porphyrin macrocycle. [19]Determination of the singlet oxygen quantum yield was performed by employing two methods: steady-state and time-resolved techniques to detect phosphorescence emission from singlet oxygen.The main difference between these techniques is the use of continuous (CW) or pulsed light to induce excited states of the molecules.The values of Φ Δ of trans-αβ-6 and cis-αα-6 presented are the average of three independent measurements using either technique.For strapped porphyrins 4, the steady-state method gave values slightly higher than 1 and the pulsed method values ca.1.3 (not shown in Table 2).The increase of the Φ Δ value obtained using pulsed laser (light peak power of the order of MW compared to mW of CW light) indicates that reaction paths other than energy transfer and/or non-linear effects might be present, [33] possibly elicited by the high density of orbitals due to the straps.For this reason, only values obtained with the steady-state technique were reported in Table 2, with minor differences observed between the Φ Δ for porphyrins 4.
Nevertheless, with both values observed as close to 1, a lack of significant deviation from planarity was indicated for the pair of isomers.Previous investigation of redaporfin atropisomers has indicated that distribution of polar substituents relative to the porphyrin macrocycle influences cellular internalization, [4] thus, insight into porphyrin shape is critical.

Cellular Internalization
Strapped porphyrin cell-internalization varied in two cancer cell lines depending on configuration.Picket fence porphyrin atropisomer internalization has been observed to vary according to α 4 > α 3 β > α 2 β 2 > αβαβ. [4]The strapped porphyrins 6 were selected for evaluation as they shared higher DMSO solubility.Measurement of strapped porphyrin fluorescence from the supernatant obtained after cell lysis confirmed the enhanced uptake (> 90 fold increase) of strapped-porphyrin internalization cis-αα-6 relative to trans-αβ-6 when both short (4 h) and long (24 h) incubation times were used (Figure 4A,B).At the porphyrin concentration evaluated (4 μM), there was nearly a complete absence of trans-αβ-6 cell internalization observed in both cell lines.
In terms of distribution of polar substituents relative to the porphyrin macrocycle, the cis-αα, cis-αβ and trans-αβ isomers of double strapped porphyrins are analogues to porphyrin atropisomers α 4 , α 2 β 2 , and αβαβ, respectively.The increased cell-internalization of cis-αα relative to trans-αβ follows the pattern of uptake previously observed with redaporfin atropisomers; enhanced internalization for the α 4 atropisomer relative to the other atropisomers of the drug mixture. [4]As with the α 4 atropisomer, the cis-αα polar substituents are oriented to one side of the porphyrin macrocycle hydrophobic core.This induces amphipathicity which is expected to confer an advantage for passive membrane permeation.The presence of straps effectively locks this favorable configuration to ensure stability against interconversion.Incorporation of the α 4 molecular configuration proved successful as a transferable motif for the design of macromolecules with enhanced cell-internalization.This is particularly critical for strapped-systems where molecular weight (> 1k Da) tends to exceed that of unstrapped counterparts (< 1 kDa).

Conclusion
Restrictive strap moieties effectively stabilized a favorable atropisomer configuration to enhance cell-internalization.A series of double strapped porphyrins were prepared using direct condensation reactions which favor the formation of cisisomers.Characterization of isomer configuration was assisted with NMR spectroscopy and X-ray crystallography methods.Photophysical evaluation indicated relative planarity of strapped porphyrin isomers and the appropriate method to evaluate cellular uptake.Internalization was assessed across two cancer cell lines and revealed enhanced membrane passage of the cis-αα isomer.This isomer was formed in greatest yield following the Lindsey condensation method.Lack of interconversion at elevated temperatures confirmed cis-αα stability.The structure of cis-αα is analogous to the α 4 atropisomer of redaporfin in terms of distribution of polar substituents.Straps concomitantly prevented atropisomer interconversion and enforced a structure which would potentiate cellular uptake.This development of strategies, which can impose high interconversion barriers while exploiting the benefit of a single atropisomer configuration, endorses the potential of atropisomer drugs candidates.

Experimental Section
General synthetic and analytical methods: All chemicals were commercially sourced and used without further purification.Analytical TLC was performed using silica gel 60 (fluorescence indicator F254, pre-coated sheets, 0.2 mm thick, 20 cm × 20 cm; Merck) plates and visualized by UV irradiation (λ = 254 nm).Column chromatography was carried out using Fluka Silica Gel 60 (230-400 mesh; Merck).NMR spectra were recorded on a Bruker AV 600, Bruker Advance III 400 MH or a Bruker DPX400 400 MHz or an Agilent 400 spectrometer.Accurate mass measurements (HRMS) were carried out using a Bruker microTOFÀ Q™ ESI-TOF mass spectrometer.Mass spectrometry was performed with a Q-Tof Premier Waters MALDI quadrupole time-of-flight (Q-TOF) mass spectrometer equipped with Z-spray electrospray ionization (ESI) and matrix-assisted laser desorption ionization (MALDI) sources in positive mode with trans-2-[3-(4-tert-butylphenyl)-2-methyl-2propenylidene]malononitrile as the matrix.ESI mass spectra were acquired in positive modes as required, using a Micromass time-offlight mass spectrometer (TOF) interfaced to a Waters 2960 HPLC or a Bruker microTOFÀ Q III spectrometer interfaced to a Dionex UltiMate 3000 LC.Atmospheric pressure chemical ionization (APCI) experiments were performed on a Bruker microTOF-Q III spectrometer interfaced to a Dionex UltiMate 3000 LC.Melting points were measured using an automated melting point meter, SMP50 (Stuart), and are uncorrected.

Strapped porphyrins 4:
1,6-Bis(2-formyl-4-nitrophenoxy)hexane (1.09 g, 2.62 mmol) was dissolved in CH 2 Cl 2 (750 mL) and pyrrole (351.26mg, 5.24 mmol) was added to the solution.The mixture was purged with Ar for 15 minutes prior to the addition of BF 3 • OEt 2 (520.15mg, 3.66 mmol).The reaction mixture was shielded from light and stirred overnight at rt. DDQ (2.082 g, 9.17 mmol) was added to the solution and stirred for an additional 2 h.A color change of orange/red to black/purple was observed.The mixture was quenched by the addition of TEA (318.14 mg, 3.14 mmol).The solution volume was reduced to half by removal of solvent under reduced pressure.The resulting crude residue was filtered through a short silica column (CH  (20 mL) and stirred at rt. Zn(OAc) 2 • 2H 2 O (126.2 mg, 0.575 mmol) dissolved in MeOH (3 mL), was added to the solution and stirred overnight at rt. Removal of solvent occurred under reduced pressure.The resulting residue was dissolved in CH 2 Cl 2 and washed sequentially with H 2 O (× 3).The organic layer was dried over MgSO 4 and the solvent removed under reduced pressure to yield a red solid.Recrystallization from EtOAc (75 °C) was set-up to remove 1,6-bis(2-formyl-4nitro-phenoxy)hexane 1 impurity.Cooling of the solution to rt resulted in dialdehyde 1 precipitation which was separated by filtration.EtOAc of the filtrate was removed under reduced pressure.Column chromatography with CH 2 Cl 2 :EtOAc (99 :  (20 mL) and TFA (1.5 mL) at rt for 1 h.The solvent was removed under reduced pressure and the resulting residue dissolved in CH 2 Cl 2 , washed sequentially with sat.aq.NaHCO 3 twice and deionized H 2 O.The organic layer was dried over MgSO 4 and the solvent removed under reduced pressure.Column chromatography (CH 2 Cl 2 ) purification was used to yield a red solid, cis-αα-4 (14 mg, 0.137 mmol, 0.52 %).M.p. > 300 °C; R f = 0.29 (CH 2 Cl 2 :EtOAc, 98 : 2, v/v); 1  1,3-Bis(2-formyl-4-nitrophenoxy)xylene (904.1 mg, 2.07 mmol) was dissolved in propionic acid (40 mL) at 100 °C.An addition of pyrrole (277.75 mg, 4.14 mmol) was made to the solution.The mixture was shielded from light and refluxed at 150 °C for 4 h.A color change of yellow to black/purple was observed.CH 3 CH 2 CO 2 H was evaporated by distillation under high vacuum.The resulting crude residue was filtered through a short silica column (CH 2 Cl 2 ).The complete purification was performed through a repetitive silica gel column chromatography.An initial column with CH 2 Cl 2 : EtOAc (98 : 2, v/v) separated the trans-αβ-5 as the first fraction and a mixture of the cis strapped porphyrins which was followed by a second column (CHCl 3 :EtOAc 96 : 4, v/v) to separate the mixture.

Refinement details cis-αβ-4:
For the half molecules in the asymmetric unit, pyrrole hydrogens were located and refined with restraints (DFIX).One of these has disorder in one nitro group position, modelled in two positions, 63 : 37 % occupancy, with restraints (SADI, SIMU).There was also disorder in the hexyl chain linker (O56-C60) and this was modelled in two positions, 52 : 48 % occupancy, with restraints (SADI, ISOR).
COOMe-trans-αβ-6: Very weak data from a poorly diffracting sample leading to a high R(int).Data limited to 0.87 Angstroms.One phenyl ester was modelled in two locations, 53 : 47 % occupied with geometric (SIMU) and displacement (SIMU, ISOR) restraints.PLATON SQUEEZE [40] used to subtract the diffuse solvent component resulting in the removal of 190 electrons/683 cubic Angstrom void per asymmetric unit.This is approx.7 EtOH molecules.
Photophysical and photochemical characterization: Absorption spectra of strapped porphyrins were recorded on an Agilent Cary5000 UV-VIS-NIR spectrometer.The absorption coefficients (ɛ) were obtained from absorption measurements of three solutions of different concentrations, diluted from two stock solutions with independently measured masses.The slope of the plot absorption vs. concentration was used to calculate ɛ.Further photophysical characterization of strapped porphyrins were carried out with fluorescence emission and excitation spectra recorded on the Horiba-JY and Fluoromax4 spectrofluorometer.Strapped-porphyrin absolute values of Φ F , obtained for each strapped porphyrin, were measured using the absolute method with a Hamamatsu Quantaurus QY absolute photoluminescence quantum yield spectrometer model C11347 (integration sphere).Solutions with absorbance < 0.1 were used.The quantum yield of singlet oxygen (Φ Δ ) was recorded using a steady-state technique exciting at 424 nm and collecting the phosphorescence of singlet oxygen between 1240 and 1340 nm on a Horiba-Jovin-Yvon Spex Fluorog 3-2.2 spectrophotometer coupled to the near-infrared Hamamatsu R5509-42 photomultiplier, cooled to 193 K in a liquid nitrogen chamber.A time-resolve technique was also used exciting the molecules with a Nd:YAG pulsed laser of 8 ns at 355 nm and the decay detected at 1270 nm using the same detector reported above.The singlet oxygen reference was phenalenone in DMSO (Φ Δ = 1, steady-state), determined in this work, and in MEOH (Φ Δ = 0.98, time-resolved). [41]A long-pass filter (Newport, 10LWF-1000-B, AE23) was used to avoid fluorescence from porphyrins.The quantum yields, using both techniques, were calculated using the Equation ( 1): (1) Where R and S are reference and sample, respectively.In the steady-state method the value of Iis the area of the phosphorescence emission and in the time-resolved case, I is the slope obtained from the phosphorescence signal dependence with laser energy.
Cellular internalization in vitro: 4T1 and CT26 cells were seeded (40,000 cells/well) in 24-well plates and let to adapt for 24 h.For strapped porphyrins 6, cells were incubated with single isomers, diluted from DMSO stock solutions, at concentration of 4 μM for 4 and 24 h.Cells were lysed (DMSO: Triton X-100, 4 % (v/v)) and fluorescence detection of the supernatant was performed with a microplate reader (Biotek Synergy HT) using 420/50 nm excitation and 645/40 nm emission filters.A calibration curve of each isomer was prepared in lysis buffer in order to quantify the absolute amount of internalized isomer.Protein quantification was carried out using the Pierce™ BCA Protein Assay Kit according to the manufacturer's instructions (Thermo Scientific).
Supporting Information (see footnote on the first page of this article): Spectroscopic data of all compounds and X-ray crystallographic data.
, Figure S13A,B and Table Scheme 2. (A) Condensations of dialdehyde and pyrrole to prepare double strapped porphyrins via the (i) Adler-Longo and (ii) Lindsey condensation methods.(B) Target strapped porphyrins.

Figure 2 .
Figure 2. View of the molecular structure in the crystal of (A) cis-αβ-4 and (B) cis-αα-4 isomers with plan (atomic displacements at the 50 % probability level) and side views (schematic pipe view for clarity).(A) Symmetry generated complete molecule of cis-αβ-4 from one of the unique half molecules of the asymmetric unit.(B) cis-αα-4 molecular structure with hexane solvent omitted.Major occupied moiety of each isomer represented only and hydrogen atoms are omitted for clarity.

Figure 3 .
Figure 3.View of the molecular structure in the crystal of (A) COOMe-cis-αα-6 with CH 2 Cl 2 solvent omitted and (B) COOMe-trans-αβ-6 with symmetry generated complete molecules from the unique half molecules of the asymmetric unit.Major disordered moiety of each isomer represented.Atomic displacements at the 50 % probability level and H atoms are omitted for clarity.

Figure 4 .
Figure 4. Cellular internalization of strapped porphyrins.(A,B) Quantification of strapped porphyrins 6 in the supernatant of cancer cells obtained after cell lysis.Cancer cells were incubated with strapped porphyrins 6 for indicated timepoints followed by cell lysis of (A) 4T1 and (B) CT26 cells.Bars indicate the mean � SEM of three independent experiments with statistical significance evaluated using two-way ANOVA in comparison to the cis-αα porphyrin, * p < 0.05 ** p < 0.01 and *** p < 0.001.

Table 1 .
Synthesis of double strapped porphyrins.

Table 2 .
Photophysical and photochemical properties of strapped porphyrins.
[a] λ Soret and ɛ recorded in THF.All other measurements recorded in DMSO.[b] Single representative experiment.