DFT application for chlorin derivatives photosensitizer drugs modeling

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 68–74

Contents lists available at ScienceDirect

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa

DFT application for chlorin derivatives photosensitizer drugs modeling Neila Machado a,⁎, Carvalho B.G. b, Téllez Soto C.A. c, Martin A.A. c,d, Favero P.P. c,d,⁎⁎ a

Institute of Research and Development, University of Vale do Paraíba, Univap, Shishima Hifumi Ave. 2911, 12244-000 São José dos Campos, São Paulo, Brazil School of Chemical Engineering, University of Campinas, Unicamp, Albert Einstein Ave. 500, 13083-852 Campinas, São Paulo, Brazil c Biomedical Engineering Innovation Center, Biomedical Vibrational Spectroscopy Group, University Brasil, UnBr, Carolina Fonseca st. 235, 08230-030 Itaquera, São Paulo, Brazil d DermoProbes – Research, Innovation and Technological Development, Research and Development Center, Cassiano Ricardo Ave. 601 rooms 73/74, Jardim Aquarius, 12246-870, São José dos Campos, São Paulo, Brazil b

a r t i c l e

i n f o

Article history: Received 16 October 2017 Received in revised form 9 January 2018 Accepted 13 January 2018 Available online xxxx Keywords: Photodynamic therapy Ab-initio VASP Gaussian

a b s t r a c t Photodynamic therapy is an alternative form of cancer treatment that meets the desire for a less aggressive approach to the body. It is based on the interaction between a photosensitizer, activating light, and molecular oxygen. This interaction results in a cascade of reactions that leads to localized cell death. Many studies have been conducted to discover an ideal photosensitizer, which aggregates all the desirable characteristics of a potent cell killer and generates minimal side effects. Using Density Functional Theory (DFT) implemented in the program Vienna Ab-initio Simulation Package, new chlorin derivatives with different functional groups were simulated to evaluate the different absorption wavelengths to permit resonant absorption with the incident laser. Gaussian 09 program was used to determine vibrational wave numbers and Natural Bond Orbitals. The chosen drug with the best characteristics for the photosensitizer was a modified model of the original chlorin, which was called as Thiol chlorin. According to our calculations it is stable and is 19.6% more efficient at optical absorption in 708 nm in comparison to the conventional chlorin e6. Vibrational modes, optical and electronic properties were predicted. In conclusion, this study is an attempt to improve the development of new photosensitizer drugs through computational methods that save time and contribute to decrease the numbers of animals for model application. © 2018 Elsevier B.V. All rights reserved.

1. Introduction Modeling a real system is a highly complex task. Computer tools have been just developed recently [1–14] and for this reason, it is possible to note that there is discrepancies between the amounts of works dedicated to experimental research and those made by computer simulations. Photodynamic Therapy (PDT) investigations are mainly experimental and clinics procedures [15–17]. These analyses demand a long time in laboratory, and also require many supplies and reagents. PDT is an alternative treatment against the cancer, which is a methodology less aggressive. Unlike chemotherapy and radiotherapy, it shows good results without cumulative toxicity, which guarantee the use of this treatment in immunocompromised patients. For this reason, the number of studies about the PDT has increased and is highlighted between new forms of cancer treatment [18,19]. This treatment is

⁎ Corresponding author. ⁎⁎ Correspondence to: Favero, P. P., DermoProbes – Research, Innovation and Technological Development, Research and Development Center, Cassiano Ricardo Ave. 601 rooms 73/74, Jardim Aquarius, 12246-870, São José dos Campos, São Paulo, Brazil. E-mail addresses: [email protected] (N. Machado), [email protected] (P.P. Favero).

https://doi.org/10.1016/j.saa.2018.01.045 1386-1425/© 2018 Elsevier B.V. All rights reserved.

based on the interaction between a drug, which is a photosensitizing substance, an activating light, and molecular oxygen [20,21]. The aim of this study is to shed some light for optimization of photosensitizing drugs via computing modeling in order to guide experiments minimizing the number of photosensitizing to be tested in laboratory. The changes in the photosensitizing molecule conformations helped to perform calculation of stability of the structures in order to find an optimized drug for activation functions. Other properties were also studied such as the vibrational modes and optical features of the illumined molecule. Modeling is a powerful tool for this study because it minimizes the use of chemical synthesis and cellular tests for each proposed drug. Thus, Density Functional Theory (DFT) may give reliable results providing a specific drug with less collateral effects and high efficiency when compared with old types of treatment.

1.1. Theoretical Methodology The methodology is based on the DFT implemented in the Vienna Ab-initio Simulation Package (VASP) [22]. The electron–ion interactions between N, O, C, S and H atoms are described by Projector Augmented Wave (PAW) potentials [23,24] and the electron–electron exchange-

N. Machado et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 68–74

69

correlation energy was simulated using the generalized gradient approximation (GGA) [25]. Single-particle orbitals were expressed in a plane-wave basis up to the energy of 400 eV. Atoms were assumed to be in their fully relaxed positions when the forces were smaller than 0.2 eV/Å. The modeling assessment was performed by percentage error calculation, considering as base, the values expressed for the vibrational modes of chlorin e6 in a study by Gladkova et al. (2010) [26]. Thus, the error showing the divergence of results obtained by our simulations was compared with the experimental data given in the literature. As an alternative in this work was also carried out DFT calculations using Gaussian 09 program B3LYP functional with a single basis set 3-21G [27] for determine the vibrational spectra (FTIR) and the Natural Bond Analysis (NBO). All the illustrations were plotted using the Visual Molecular Dynamics (VMD) [28]. 2. Modeling Chlorin In order to make an ideal drug model the Photodithazine drug, which is cited as (chlorin e6), was used as the main structural prototype. The structure has a ring and branches of N- metil-D-glicosamina. The molecule absorbs between 650 and 680 nm [9]. Fig.1 (a) shows the chemical structure of the molecule found in literature [29]. Part (b) and (c) represent the Photodithazine model, in this work it was called by Base Chlorin. The Base Chlorin geometry was built considering the Porphyrin ring as base [29,30]; moreover, carbon chain molecules were used to simulate some radical structures like the branches of the molecule. According to literature [31–33], these branches formed by hydrophobic radicals may improve the penetration of a drug in an animal membrane, which is mainly composed by phospholipids. This geometry conformation may be a possible explanation for the good penetration of this drug in tumor cells. Fig. 2 shows two modified Base Chlorin models with the additional groups. In the right circle of the Fig. 2 shows a model called as Thiol Chlorin; this structure contains a thiol group (-SH) in its right chain. A second model (left circle), Hydrocarbon Chlorin, was suggested with an aliphatic carbon chain functionalized with a thiol group (-S (CH2)6) instead of a simple thiol group. This aliphatic chain is considered as an effective separator [34], thus, it may also be used to generate space between surface of carrier and the drug such as Chlorin. Nanoparticles are often used as drugs carrier [35–41], which are functionalized with different linker such as thiol group and lipoic acid [42,43]. In this context, the present work may help to answer questions about the possible interference of these linkers in the optical properties of the Photodithazine, which is functionalized with these additional groups.

Fig. 2. - Representation of Base Chlorin modified with the additional groups. In the right circle Thiol Chlorin (\ \SH) and in left circle Hydrocarbon Chlorin (\ \S (CH2)6). The large spheres of red, dark blue, yellow and light blue correspond to O, N, S and C atoms, respectively, while the small white spheres represent H atoms. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

To evaluate possible changes in the chlorin geometry due to the biological environment other models were suggested in this work, as can be seen in the Fig. 3. It is known that different biological environment may have distinguished properties such as temperature, pH and also the presence of ions. It is suggested that some kind of drugs may change their conformation when they reach their target cells, for instance, cancer cells target. These geometry changes are important because they may be responsible for the drug accumulation inside the target cell. Thus, this work presents a possible transition of cis-like and trans-like conformation of the drug with presence of double bond in the ring's region close to the its branches (see Fig.1a). 3. Results and Discussions 3.1. Chlorin Structural Analysis The first step of the model validation is shown in the Table 1, which contains results of the molecule energy evaluation. This study was

Fig. 1. (a) Chemical structure of Photodithazine drug [29], (b) drug model called by Base Chlorin (front view) and (c) Base Chlorin (side view). The large spheres of red, dark blue, and light blue correspond to O, N, and C atoms, respectively, while the small white spheres represent H atoms. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

70

N. Machado et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 68–74

Fig. 3. Representation of Base Chlorin modified (a) chlorin CIS and (b) chlorin TRANS. The large spheres of red, dark blue, and light blue correspond to O, N, and C atoms, respectively, while the small white spheres represent H atoms. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

performed considering two different analyses that are described by EaT and EcT. The first analysis considers the whole molecule, ring + chains + additional groups (Fig. 2), as a unique system, which is described by EaT. In the second analysis, Thiol and Hydrocarbon Chlorin energies were calculated in parts in order to avoid the influence of their additional groups, EbT. For instance, Thiol and Hydrocarbon Chlorin models were optimized dividing the molecule structure in two parts: active region (ring + chains) and additional groups (\\S (CH2)6) or (\\SH), Fig. 2. The aim of this study is to calculate the active region energies of the suggested models without the additional groups influence. The results are shown in Table 1 as EcT, which is given by Eq. (1). ET c ¼ ET a −ET b

ð1Þ

According to the Table 1 Thiol Chlorin can be considered more stable than Hydrocarbon Chlorin structure with difference of −0.7 eV when compared both active region energies (EcT). The energy system (EaT) is used to calculate the wavelengths of Chlorin models, Table 2. Table 2 shows that the most stable system is Chlorin TRANS when compared with the CIS model with difference of −98.4 eV, however, Chlorin CIS absorbs in 670 nm which is the experimental wavelength value. Table 1 Total energy (ETa), Additional Group Energy (ETb), and Active Region Energy (ETc) of Chlorin molecules.

3.2. Vibrational Analysis of Chlorin Theoretical vibrational modes for the Base Chlorin structure were calculated via DFT using the Gaussian 09 program. The calculus was performed using B3LYP functional with a 3-21G-basis. All the theoretical modes were compared with the FTIR experimental data of Base Chlorin [26]. For this study was just considered the theoretical values, which are close to the experimental modes, however, it was possible to find excessive number of internal coordinates. The spectra range analyzed was between 800 and 3400 cm−1. The overestimation of the computed wavenumbers is quite systematic and can be corrected by applying appropriate scaling factors or scaling equations. Applying (0.9613) scaling factor (0.9613), the theoretical wavenumbers are in good agreement with experimental wavenumbers [44]. Table 3 compares theoretical and experimental vibrational modes of Base Chlorin and describes their assignments, which were made through Gaussian calculation. The theoretical spectra can be seen in Fig. 4 with the respective errors ( ) between the theoretical values and the experimental values [26]. The errors presented in Fig.4 shows that the theoretical values converge with the experimental data [26]. The error values found are between 0.07% and 2.2%, which can be considered satisfactory for Gaussian calculations. Base on this vibrational result, it is possible to

Table 3 Theoretical and experimental vibrational modes of Base Chlorin and their respective assignments ν, δ and ρ means stretching, bending and rocking. Theoretical Vibrational Modes (cm−1)a

Experimental Vibrational Modes (cm−1)b

Assignmenta

774 845 929 982 1060 1162

805 840 913 987 1064 1165

a

1221 1299 1401 1488 1550 1637 1712 2873 2926 3004 3258

1217 1306 1377 1500 1542 1601 1719 2867 2924 2966 3303

Twist (COH) δ(CH2) wagg + δ(COH) wagg ν(C\ \C) + ρ(CH3) + δ(CNH) ρ(CH3) + ρ(OH) ν(C\ \C) + ρ(OH) + δ(C-C-C) ν(C_N) + δ(HCH) twist + ρ(CH) + ρ(NH) ρ(NH) + ρ(CH) ν(CN) δ(CH2) wagg δ(CH2) sciss δ(NH2) twist ν(C_O) + δ(NH2) sciss + δ(COH) δ(NH2) sciss ν(CH) methane ν simetric (CH) (CH3) ν assimetric (CH) (CH3) ν(NH)

b

a

a

b

c

Molecule

ET (eV)

ET (eV)

ET (eV)

Base Chlorin Thiol Chlorin Hydrocarbon Chlorin

−830.1 −835.9 −933.7

– −5.58 −104.07

−830.1 −830.3 −829.6

a b c

Theoretical energy values calculated in this work (VASP). Theoretical energy values calculated in this work (VASP). Theoretical energy values calculated in this work (VASP).

Table 2 Total energy (ETa) of Chlorin CIS and TRANS configuration and their respective wavelengths (λb). Molecule

ETa (eV)

Number of Atoms

λb (nm)

Base Chlorin Chlorin CIS Chlorin TRANS

−830.1 −830.7 −929.1

136 136 136

670 [9] 670 749

Theoretical energy values calculated in this work (VASP). Experimental wavelength of Base Chlorin found in literature [Reshetnickov et al. (2000)] and estimate wavelengths values for Chlorin CIS and TRANS.

b

Theoretical vibrational modes of Base Chlorin and their assignments (Gaussian). Experimental vibrational modes found in literature Gladkova et al. (2010) [8].

N. Machado et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 68–74

71

Fig. 5. DOS plot of Chlorin Base in (black). The values are between −10 eV and 6 eV, with 0 eV at the LUMO (black —). Fig. 4. Theoretical spectra of Base Chlorin - Photodithazine (in red) and errors between theoretical and experimental values, which are presented up to the peaks ( ). The values are between 800 and 3400 cm−1. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

see low divergence values for the proposed structure model, Base Chlorin. Therefore, the optimized Base Chlorin structure proposed was validated as a good model according the vibrational modes. 3.3. Electronic Properties Optical and electronic properties of the structure models were analyzed in the following studies in VASP. They can be seen in Table 4 and Figs. 5 and 6. Absorption sites were identified for each model calculating the change of density that is involved in HOMO-LUMO transition. The found results can describe how changes in functional group of the Base Chlorin may affect the intensity and frequency of its optical adsorption. The main point to be evaluated in this section is the Density of States (DOS) plot, which shows the HOMO-LUMO transition state. Since the HOMO orbital is define. As the last occupied orbital, here our calculations give different occupancies from E1 to E6 with a disperse electron occupancy which varies from 1.78 to 1,7 × 10−4 electrons. In the DOS diagram, the y axis is defined to have arbitrary units and the x axis is defined as the electron

occupancy. Clearly, it is possible to note that there are a maximus in E3 with occupancy of 0.01824 electrons, and correspond a gap of energy of 715 nm with reference to the LUMO orbital with zero electron occupancy. Instead of the molecular orbital with E3 has a lower occupancy; it is near to the experimental value of 670 nm corresponding to the maximum of absorbance in the UV spectra. Table 4 presents the electron occupancy, the band energy and the correlate wavelengths (nm) between the LUMO and the different HOMO states. Fig. 5 shows DOS diagram with different energy states from E1 to E6, as shown in Table 4. As can be seen in Fig. 6, E3 state of Base Chlorin is localized at − 1.732 eV with wavelength of 715 nm. This result was considered as the best value found among all HOMO states (E1 to E6). This evaluation considered as ideal the experimental wavelength of 670 nm, which can be found in the literature [29]. Thus, in the following studies E3 was considered one of the most probable HOMO state for all models, Base, Thiol and Hydrocarbon Chlorin structures. Fig. 6 show the DOS plots for each model and their respective wavelengths (λd) calculated for E3. According to the results obtained in the DOS plots the E3 of the Base Chlorin is −1.732 eV, which is corresponding to a wavelength of 715 nm. Thiol Chlorin has a value of −1.751 eV for its E3 that is equivalent to 708 nm. In addition, Hydrocarbon Chlorin value was of −1.627 eV that is approximately 762 nm of wavelength.

Table 4 Occupied molecular orbitals, band energies and correlates wave lengths (nm).

Band number

Band energies

Occupaon

Transion

Wave length

probabilies

(nm)

192

−3.5360

1.77618

E6

402

193

−2.9045

0.50481

E5

506

194

−2.3639

0.04424

E4

648

195

−2.1841

0.01824

E3

715

E2

196

−1.8363

0.00323

197

−1.5280

0.00069

198

−1.2448

0.00017

199

−0.9692

0.00004

200

−0.8829

0.00003

201

−0.6900

0.00001

202

−0.4513

0.00000

895 1158

E1

LUMO

1564

72

N. Machado et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 68–74

Fig. 6. DOS plot of Base, Thiol, and Hydrocarbon Chlorin structures, in black, red and green, respectively. The values are between −10 eV and 6 eV, with 0 eV at the LUMO (black —). All E3 and LUMO state are represented with dashed lines (…) and their respective wavelengths (λd). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.4. Charge of Density (HOMO-LUMO). The purpose of the charge density studies is to determine if the found E3 state is the best HOMO between the other theoretical states E1 and E2, which are the nearest states of the LUMO. Thus, in order to complement the DOS study the HOMO - LUMO states of the Base Chlorin were illustrated in Fig. 7. As can be seen in Fig. 7 the charge of density localized on E3 with 0.01824 of electron occupancy is well distributed on porphyrin ring and is very similar to LUMO. This evaluation is an indication of wave function overlap. Thus, when compared with the other states HOMO (E1 and E2), the E3 state shows to be the most probable state that will relocate to LUMO state. LUMO state of Base Chlorin lost charge in its ring on the right side. E3 was elected as the best occupied state for the first molecule and it will be also used as base for the following models, thus, the comparison of the results will be possible. Thiol Chlorin and Hydrocarbon Chlorin models are shown in Fig. 8. As mentioned previously only the E3-HOMO state will be compared to the LUMO. According to Fig. 8 the density of charge (E3-HOMO) of Thiol Chlorin structure is localized on its porphyrin ring and the LUMO state is placed between the ring and left chain. Thus, it is possible to note that there was a transition between the states with a wavelength of 715 nm,

Fig. 6. This overlapping region of charges is an important feature that is responsible for the site transition. The HOMO state of the Hydrocarbon Chlorin shows that there was a displacement of charge density from the porphyrin ring (LUMO) to its left chain (E3-HOMO). When compared with the Base Chlorin results is possible to see that the aliphatic chain may inactive the porphyrin ring which is considered the center active of molecule. Table 5 shows previous theoretical and experimental wavelengths, and their respective optical transition areas. In order to find the transition area some assumptions were made in this study. First of all, theoretical wavelength of 715 nm was considered equivalent to the experimental of 670 nm, which can be found in literature. Moreover, the optical transition area was estimated for each structure model calculating the area under the DOS curve, Fig. 6, between E3 and LUMO transition state. Table 5 shows that Hydrocarbon Chlorin absorbs 36% in 762 nm, more than Base Chlorin, against the absorption of 19.6% in the Thiol model, which has a gain in the wavelength of 708 nm. The second wavelength found is close to the value obtained for the Base Chlorin calculus via DFT. 3.5. Natural Bond Orbital Analysis (NBO) Donor-acceptor interactions through the Second Order Perturbation Theory Analysis of the Fox Matrix in NBO basis [45–47] at the B3LYP/3-

Fig. 7. Base Chlorin occupation level in E1, E2, E3, and LUMO, respectively. The large spheres of red, dark blue, and light blue correspond to O, N, and C atoms, respectively, while the small white spheres represent H atoms. Grey regions represent the electronic charge density (HOMO-LUMO). Isosurface: 0.0015. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

N. Machado et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 68–74

73

Fig. 8. Thiol Chlorin and Hydrocarbon Chlorin occupation level in E3, and LUMO. The large spheres of red, dark blue, yellow and light blue correspond to O, N, S and C atoms, respectively, while the small white spheres represent H atoms. Grey regions represent the electronic charge density HOMO–LUMO. Isosurface: 0.0015. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Table 5 Theoretical and experimental wavelengths λa and λb in (nm) and its respective optical transition area in (%). Molecule

λa (nm)

λb (nm)

Transition Area (%)

Base Chlorin Thiol Chlorin Hydrocarbon Chlorin

715 708 762

670 633 714

100 119.6 136

a

Theoretical wavelengths values from E3 to LUMO (VASP). b Experimental wavelength of Base Chlorin found in literature [Reshetnickov et al. (2000)] and estimates wavelengths values for Thiol and Hydrocarbon.

21G level were carried out to rationalize the factors that contribute to the total conformational energy. Table 6 describes the donor and acceptor orbitals and the energy in Kcal·mol−1. The most representative interactions are shown in the NBO/B3LYP calculations for Base Chlorin molecule, as shown in Fig. 9. Table 6 and Fig. 9 show that the most energetic electronic transfer comprises the atoms forming the central ring due to the higher electronic density around itself. This result corroborates with the literature which describes the branches as an improver of penetration drugs in the animal membrane [11–13] and the ring as an active center of the molecule where reactions occurs due to the interaction with the light, this phenomenon can be seen in the Photodynamic Therapy [1,2,5,10,17]. This

absorbed light triggers reactions that make the molecular oxygen toxic reacting with the vital cellular components leading to localized cellular death [2,5]. Conclusion The best Chlorin derivative model was chosen analyzing different properties such as energy total of the system and the gain of absorption intensity. The gains were calculated based on optical and electronic molecule features. The first Base Chlorin derivative model, Thiol Chlorin, showed to be more stable than Hydrocarbon model, when compared their energy total values. When compared their gain of intensities, the second model had a better result with an absorption of 36% while Thiol Chlorin obtained a value of 19.6%. However, theoretical wavelengths

Table 6 - Second Order Perturbation Theory of Fock matrix in NBO basis for the Base Chlorin alpha electrons where LP, BD, and DB* mean lone pair, 2-center bond, and 2-center antibond, respectively. Donor NBO (i)

Acceptor NBO (j)

Kcal/mol

Molecular Structure

LP O 80 BD* C 86 - N 131 LP C 98 LP C 98 LP O 74 BD* C 96 - C 97 BD* C 93 - N 133 BD* C 89 - C 92 BD* C 87 - C 88 LP O 81 BD H 46 - O 81 BD* C 95 - N 134 BD* C 86 - N 131 LP C 98 LP N 132 LP O 68 BD C 96 - C 97

BD* H 46 - O 81 BD* C 87 - C 88 BD* O 67 - C 109 BD* C 95 - N 134 BD* O 75 - C 128 BD* C 85 - C 99 BD* C 89 - C 92 BD* C 90 - C 91 BD* C 90 - C 91 BD* H 46 - O 81 BD* O 82 - C 118 BD* C 100 - C 101 BD* C 85 - C 99 BD* C 96 - C 97 BD* C 89 - C 92 BD* O 67 - C 109 LP C 98

332.33 161.41 147.26 145.32 123.5 120.66 118.99 97.71 91.25 86.19 85.38 75.91 68.79 67.17 52.42 51.81 50.32

Right Chain (H-bond) Ring Branch Ring Ring Left Chain (COO−) Ring Ring Ring Ring Right Chain Right Chain (H-bond) Ring Ring Ring Ring Branch Ring (COO−) Ring

Fig. 9. Partial structure of Chlorin showing the atom numbering (DFT: B3LYP/3-121G). The large spheres of red, dark blue, and light blue correspond to O, N, and C atoms, respectively, while the small white spheres represent H atoms. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

74

N. Machado et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 68–74

were also compared with experimental ones in order to determine the best structure. Thiol Chlorin had a wavelength of 708 nm and Hydrocarbon model generates a value of 763 nm. In this case the first one is closest of Base Chlorin wavelength that is 715 nm. The present work suggests different modified chlorin e6 to be used in photodynamic therapy. The structures based on the chlorin e6 were determined with appropriate characterization of optical and vibrational properties aim to confirm the experimental infrared spectrum values. The NBO results agree with the ring electronic structure where there are several atoms with double bond and nitrogen atoms with free electron pairs. These results can save experimental time as well as laboratory goods. Thus, it may help to focus efforts in systems, which are more likely to be successful in an experimental field.

Acknowledgments The authors acknowledge financial support from the Brazilian agencies CAPES, FINEP and FAPESP (Project FINEP 01.10.0661-00, FAPESP 2011/13250-0, FAPESP 2013/17247-9, FAPESP 2014/05975-2, CAPES 88887068264/2014-00), of the Institute of Research and Development, University of Vale do Paraíba. References [1] Helgi I. Ingólfsson, et al., Computational ‘microscopy’ of cellular membranes, J. Cell Sci. 129 (2) (2016) 257–268. [2] Helgi I. Ingólfsson, et al., The power of coarse graining in biomolecular simulations, Wiley Interdiscip. Rev.: Comput. Mol. Sci. 4 (3) (2014) 225–248. [3] Sander Pronk, et al., GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit, Bioinformatics (2013) btt055. [4] Wojciech Kopeć, Jelena Telenius, Himanshu Khandelia, Molecular dynamics simulations of the interactions of medicinal plant extracts and drugs with lipid bilayer membranes, FEBS J. 280 (12) (2013) 2785–2805. [5] Martin Karplus, J. Andrew Mccammon, Molecular dynamics simulations of biomolecules, Nat. Struct. Mol. Biol. 9 (9) (2002) 646–652. [6] Siewert J. Marrink, et al., The MARTINI force field: coarse grained model for biomolecular simulations, J. Phys. Chem. B 111 (27) (2007) 7812–7824. [7] Siewert J. Marrink, D. Peter Tieleman, Perspective on the Martini model, Chem. Soc. Rev. 42 (16) (2013) 6801–6822. [8] James C. Phillips, et al., Scalable molecular dynamics with NAMD, J. Comput. Chem. 26 (16) (2005) 1781–1802. [9] N.C.F. Machado, et al., Assessment of penetration of ascorbyl tetraisopalmitate into biological membranes by molecular dynamics, Comput. Biol. Med. 75 (2016) 151–159. [10] Andrew G. Mercader, Eduardo A. Castro, in: A.K. Haghi (Ed.), Nanoscience and Computational Chemistry: Research Progress, CRC Press, 2013. [11] B. Gregatti, et al., Phenylalanine ab initio models for the simulation of skin natural moisturizing factor, Spectrochim. Acta A Mol. Biomol. Spectrosc. 106 (2013) 73–79. [12] B.G. Carvalho, et al., Analysis of DNA nanosensors interactions via density functional theory, Sens. Lett. 13 (2015) 318–323. [13] Jan Langmaier, et al., Extreme basicity of biguanide drugs in aqueous solutions: ion transfer voltammetry and DFT calculations, J. Phys. Chem. A 120 (37) (2016) 7344–7350. [14] Maurizio Muniz-Miranda, Francesco Muniz-Miranda, Alfonso Pedone, Raman and DFT study of methimazole chemisorbed on gold colloidal nanoparticles, Phys. Chem. Chem. Phys. 18 (8) (2016) 5974–5980. [15] Luciana C. Silva, et al., Molecular analysis of apoptosis pathway after photodynamic therapy in breast cancer: animal model study, Photodiagn. Photodyn. Ther. 14 (2016) 152–158.

[16] Aleksandra Kawczyk-Krupka, et al., Photodynamic therapy in colorectal cancer treatment—the state of the art in preclinical research, Photodiagn. Photodyn. Ther. 13 (2016) 158–174. [17] Anyanee Kamkaew, et al., Scintillating nanoparticles as energy mediators for enhanced photodynamic therapy, ACS Nano 10 (4) (2016) 3918–3935. [18] Beatriz Rodríguez-Amigo, et al., Chapter 2. Photosensitisers for photodynamic therapy: state of the art and perspectives, Photodynamic Medicine: From Bench to Clinic 2016, pp. 23–62. [19] S.B. Brown, E.A. Brown, I. Walker, The present and future role of photodynamic therapy in cancer treatment, Lancet Oncol. 5 (8) (2004) 497–508. [20] T.J. Dougherty, et al., Photodynamic therapy, J. Natl. Cancer Inst. 90 (12) (1998) 889. [21] D. Kessel, Photodynamic therapy: from the beginning, Photodiagn. Photodyn. Ther. 1 (1) (2004) 3–7. [22] G. Kresse, J. Furthmuller, Comput. Mater. Sci. 6 (15) (1996) 15–50. [23] P.E. Blöchl, Physics 50 (1994) 17953. [24] G. Kresse, J. Joubert, Physics 59 (1999) 1758. [25] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Reb. Lett. 78 (1997) 1396. [26] Gladkova OL, M.V. Parkhats, A.N. Gorbachova, S. Terekhov, FTIR spectra and normalmode analysis of chlorin e(6) and its degradation-induced impurities, Spectrochim. Acta A Mol. Biomol. Spectrosc. 76 (3–4) (2010 Aug) 388–394. [27] M.J. Frisch, et al., Gaussian 09, Revision E.01, Gaussian, Inc., Wallingford CT, 2009. [28] W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, J. Mol. Graph. 14 (1996) 33–38. [29] A.V. Reshetnickov, G.V. Ponomarev, O.Y. Abakumova, T.A. Tsvetkova, A.V. Karmenyan, A.G. Rebeko, R.P. Baum, Novel drug form of chlorin E6, Proc. SPIE 3909 (2000) 124–130. [30] E.S. Nyman, P.H. Hynninen, Research advances in the use of tetrapyrrolic photosensitizers for photodynamic therapy, J. Photochem. Photobiol. B Biol. 73 (2004) 1–28. [31] S.J. Singer, Garth L. Nicolso, The fluid mosaic model of the structure of cell membranes, Science 175 (4023) (1972) 720–731. [32] Garth L. Nicolso, The fluid—mosaic model of membrane structure: still relevant to understanding the structure, function and dynamics of biological membranes after more than 40 years, Biochim. Biophys. Acta 1838 (2014) 1451–1466. [33] Gerrit van Meer, Cellular lipidomics, EMBO J. 24 (2005) 3159–3165. [34] M. Preuss, et al., J. Comput. Chem. 25 (2004) 112–122. [35] Dev Kumar Chatterjee, Shan Fong Li, Yong Zhang, Nanoparticles in photodynamic therapy: an emerging paradigm, Adv. Drug Deliv. Rev. 60 (2008) 1627–1637. [36] Yvette Niamien Konan, Robert Gurny, Eric Alleman, State of the art in the delivery of photosensitizers for photodynamic therapy, J. Photochem. Photobiol. B Biol. 66 (2002) 89–106. [37] R. Rotomskis, G. Streckytė, Karabanovas, Nanodariniai diagnostikai ir gydymui: nanomedicinos link, Medicina 42 (2006) 542–558. [38] A.C. Anselmo, S. Mitragotri, An overview of clinical and commercial impact of drug delivery systems, J. Control. Release 190 (2014) 15–28. [39] V. Sokolova, M. Epple, Bioceramic nanoparticles for tissue engineering and drug delivery, Tissue Engineering Using Ceramics and Polymers, Second edition 2014, pp. 633–647. [40] Hui Zheng, Sai Li, Yuji Pu, Yusi Lai, Bin He, Zhongwei Gu, Nanoparticles generated by PEG-Chrysin conjugates for efficient anticancer drug delivery, Eur. J. Pharm. Biopharm. 87 (3) (2014) 454–460. [41] Anil Kumar, Xu Zhang, Xing-Jie Liang, Gold nanoparticles: Emerging paradigm for targeted drug delivery system, Biotechnol. Adv. 31 (5) (2013) 593–606. [42] M.E. Wieder, et al., Intracellular photodynamic therapy with photosensitizernanoparticle conjugates: cancer therapy using a ‘Trojan horse’, Photochem. Photobiol. Sci. 5 (2006) 727–734. [43] L.J. Lucas, K.C. Hewitt, Nanobiophotonics for Molecular Imaging of Cancer: Au- and Ag-based Epidermal Growth Factor Receptor (EGFR) Specific Nanoprobes, Proc. SPIE, 8234, Plasmonics in Biology and Medicine, IX, 2012 82340C. [44] Anthony P. Scott, Leo Radom, Harmonic vibrational frequencies: an evaluation of Hartree-Fock, Møller-Plesset, quadratic configuration interaction, density functional theory, and semi empirical scale factors, J. Phys. Chem. 100 (41) (1996) 16502–16513. [45] J.P. Foster, F. Weinhold, Natural hybrid orbitals, J. Am. Chem. Soc. 102 (24) (1980) 7211–7218. [46] Alan E. Reed, Frank Weinhold, Natural bond orbital analysis of near-Hartree–Fock water dimer, J. Chem. Phys. 78 (6) (1983) 4066–4073. [47] Alan E. Reed, Robert B. Weinstock, Frank Weinhold, Natural population analysis, J. Chem. Phys. 83 (2) (1985) 735–746.