Chinese Journal of Chemical Physics  2020, Vol. 33 Issue (1): 69-74

#### The article information

Zhuang Chen, Yang-yi Liu, Xiao-xiao He, Jin-quan Chen

Ultrafast Excited State Dynamics of Biliverdin Dimethyl Ester Coordinate with Zinc Ions

Chinese Journal of Chemical Physics, 2020, 33(1): 69-74

http://dx.doi.org/10.1063/1674-0068/cjcp1911193

### Article history

Accepted on: November 27, 2019
Ultrafast Excited State Dynamics of Biliverdin Dimethyl Ester Coordinate with Zinc Ions
Zhuang Chena , Yang-yi Liua , Xiao-xiao Hea , Jin-quan Chena,b
Dated: Received on November 1, 2019; Accepted on November 27, 2019
a. State Key Laboratory of Precision Spectroscopy, East China Normal University, Shanghai 200241, China;
b. Collaborative Innovation Center of Extreme Optics, Shanxi University, Taiyuan 030006, China
Abstract: As one of the biological endogenous pigments, biliverdin (BV) and its dimethyl ester (BVE) have extremely weak fluorescence in solution with quantum yield less than 0.01%. However, the situation reverses with the addition of zinc ions. The strength for fluorescence of BVE-Zn$^{2+}$ complex is greatly enhanced and fluorescence quantum yield can increase to $\sim$5%. Herein, we studied ultrafast excited state dynamics of BVE-Zn$^{2+}$ complex in ethanol, $n$-propanol, and DMSO solutions in order to reveal the mechanism of fluorescence quantum yield enhancement. The results show that BVE can form a stable coordination complex with zinc with 1:1 stoichiometry in solution. BVE is structurally and energetically more stable in the complex. Using picosecond time-resolve fluorescence and femtosecond transient absorption spectroscopy, we show that smaller non-radiative rate constant of BVE-Zn$^{2+}$ complex in DMSO is the key to increasing its fluorescence quantum yield and the excited state decay mechanism is also revealed. These results provide valuable information about the fluorescence property change after BVE binding to metal ions and may provide a guidance for the study of phytochromes or other fluorescence proteins in which BV/BVE acts as chromophores.
Key words: Biliverdin    Zinc    Fluorescence    Quantum yield    Femtosecond transient absorption    Excited state dynamics
Ⅰ. INTRODUCTION

The phytochromes are cellular protein which is sensitive to the light and regulate massive physiological reaction processes in plants, animals, microorganisms [1]. Although, the chromophores of the phytochrome vary with the host organism, they all have linear tetrapyrrole structure [2]. In general, the chromophores are mainly phytochromobilin (P$\Phi$B) and phycocyanobilin (PCB) in plants [3], but they are bilirubin (BR) and biliverdin (BV) in animals [4].

BV is produced by degradation of heme-iron protoporphyrin IX complex, and it can be further reduced to bilirubin by biliverdin reductase [5, 6]. BV has two deprotonated carboxyl groups and tends to acquire a distorted helical conformation that is stabilized mainly by intramolecular hydrogen bonding between unprotonated and protonated N atoms in pyrroles (FIG. 1) [7, 8]. BV exhibits absorption and emission band maxima at 664 and 670 nm [9], respectively, and demonstrates a fluorescence quantum yield (QY) of less than 0.01% [10].

 FIG. 1 (A) Structure of BV/BVE and (B) proposed structure of 1:1 BVE-Zn$^{2+}$ complex

Earlier studies on BV using steady spectroscopy, theoretical calculation and NMR were focused on its physical and chemical properties such as conformation, polarity, fluorescence QY [4, 11]. Recent studies on BV were mainly about its association with proteins or polymeric materials [12-14]. However, the excited state dynamics of BV/BVE remains unknown, especially when it complexes with other materials. It was reported that P$\Phi$B/PCB/bilirubin show similar excited state dynamics after binding to apoproteins or metal ions [15, 16]. We speculate that BV may have similar properties since it is a member of linear tetrapyrrole family. Zinc is one of trace elements in human body that has great significance to our health [17-19]. Therefore, research on BV/BVE and zinc ions complex is not only important for physiology and medical treatment, but also could provide information to understand the excited state dynamics after BV/BVE binds to proteins.

In this work, BVE-Zn$^{2+}$ complex is studied in ethanol, $n$-propanol, and DMSO solutions by using steady state, picosecond time-resolved fluorescence (time-correlated single photon counting, TCSPC), and femtosecond transient absorption (TA) spectroscopy. The bound stoichiometry of BVE-Zn$^{2+}$ complex is determined by titrating different concentrations of zinc ions with BVE. With the help of time-resolve spectroscopy, we show that the smaller non-radiative rate constant is the key to BVE fluorescence QY enhancement, and a complete picture of excited state decay mechanism is accomplished. With the understanding of the excited state decay mechanism of BVE-Zn$^{2+}$ complex, we can obtain useful information to further study fluorescence proteins using BV/BVE as chromophores in the future.

Ⅱ. EXPERIMENTS A. Sample preparation

Biliverdin dimethyl ester (Aladdin) was dissolved in different solvent for steady state, TCSPC and TA experiments. No vortex or ultrasound is used in order to avoid thermo-isomerization at room temperature. All solvents (ethanol, $n$-propanol, and DMSO) used in the experiment were purchased from Sigma-Aldrich. BVE-Zn$^{2+}$ complex solutions with different ratios were prepared by first preparing BVE and Zn$^{2+}$ solutions and then mixing those two solutions. The ratios are $c$(BVE):$c$(Zn$^{2+}$) = 1:0, 1:0.5, 1:1, 1:2, 1:4, 1:5, 1:10.

B. Spectroscopy measurements

Steady-state absorption and fluorescence spectra were recorded by a UV-Vis spectrophotometer (TU1901, Beijing Purkinje General Instrument Co. Ltd.) and a FluoroMax-4 spectrofluorometer (Horiba, Jobin Yvon), respectively.

Fluorescence lifetimes were measured on a home-built TCSPC system. The excitation light was from a picosecond continuum fiber laser (SC400-pp-4, Fianium, UK) with a repetition rate of 20 MHz, and the fluorescence was recorded by a TCSPC module (PicoHarp 300, PicoQuant) and a MCP-PMT (R3809U-50, Hamamatsu). A monochromator (7ISW151, Sofn Instruments) was used to select the emission wavelength.

The transient absorption (TA) spectra were measured by a transient absorption spectrometer (Helios-EOS fire, Ultrafast System) coupled with of a mode-locked Ti-sapphire laser system (Astrella, Coherent) with a 1 kHz Ti:sapphire amplifier (Astrella, Coherent Inc.). A fraction of the fundamental beam was used to produce pump beams via an optical parametric amplifier (OPerA Solo, Coherent Inc.). The 640 nm pump pulse intensity was limited to less than 1.5 GW/cm$^2$. The samples were hold in a 2 mm fused silica cuvette. The instrument response function (IRF) was determined to be $\sim$120 fs by measuring solvent responses under the same experimental conditions. All measurements were performed at room temperature.

We use the reference method to measure the QY and the equation is as follows:

 $\begin{eqnarray} Q = Q_ {\rm{R}} \cdot\frac{{ {\frac{I}{{ \rm{OD}}}}}}{{ {\frac{{I_ {\rm{R}} }}{{ {\rm{OD}}_ {\rm{R}} }}}}} \cdot {\frac{{n^2 }}{{n_ {\rm{R}}{^2} }}} \end{eqnarray}$ (1)

where $Q_ \rm{R}$ is the QY of reference sample. In this work, we use Nile Blue A as reference sample and its QY is 27% at 630 nm [20]. $I$ is the fluorescence integrated intensity of the complex from 640 nm to 850 nm, and $\rm{OD}$ is the absorbance of the complex at 630 nm. $I_ \rm{R}$ is the fluorescence integrated intensity of Nile Blue A from 640 nm to 850 nm, and $\rm{OD}_ \rm{R}$ is the absorbance of Nile Blue A at 630 nm. $n$ is the refractive index of the complex solvent and $n_ \rm{R}$ is the solvent refractive index of Nile Blue A.

Except the above formula, QY can also be noted like this:

 $\begin{eqnarray} Q = \frac{{k_{ \rm{RAD}} }}{{k_{ \rm{EXC}} }} \end{eqnarray}$ (1)

where $k_{ \rm{RAD}}$ is radiative rate, $k_{ \rm{EXC}}$ is excited state decay rate, and they have the following relationship:

 $\begin{eqnarray} k_{ \rm{EXC}} = k_{ \rm{RAD}} + k_{ \rm{NR}} \end{eqnarray}$ (3)

where $k_{ \rm{NR}}$ is non-radiative rate, and also $k_{ \rm{EXC}}$ is the reciprocal of $\tau_{ \rm{ \rm{EXC}}}$:

 $\begin{eqnarray} k_{ \rm{EXC}} = \frac{1}{{\tau _{ \rm{EXC}} }} \end{eqnarray}$ (4)

Combining Eqs. (2), (3) and (4), we can get:

 $\begin{eqnarray} k_{ \rm{RAD}} & = & Q \cdot k_{ \rm{EXC}} \end{eqnarray}$ (5)
 $\begin{eqnarray} k_{ \rm{NR}} & = & (1 - Q) \cdot k_{ \rm{EXC}} \end{eqnarray}$ (6)
Ⅲ. RESULTS AND DISCUSSION A. Steady-state spectra

In order to determine the bound stoichiometry of BVE-Zn$^{2+}$ complex, we performed concentration titration in ethanol and DMSO. FIG. 2 (A) and (B) show the absorption spectra and emission spectra of the complex in ethanol, respectively. As shown in FIG. 2(A), BVE has two characteristic absorption peaks centered at 376 nm and 670 nm in the absence of Zn$^{2+}$. The absorption spectra is in line with literature and they are assigned to S$_0$-S$_2$ and S$_0$-S$_1$ transition, respectively [21]. In ethanol, when the concentration of Zn$^{2+}$ is increased, the 376 nm absorption peak gradually shifts to red-side. At the same time, a new absorption peak with maximum at $\sim$705 nm appears and the 670 nm peak becomes a shoulder peak. Fluorescence emission of BVE itself is very weak (FIG. 2(B)). However, fluorescence intensity is enhanced by about 60$-$80 times after adding Zn$^{2+}$ in BVE solution and the fluorescence peak is centered at 740 nm, which is consistent with literature reports about most linear tetrapyrrole molecules when binding to protein or other materials [4]. These results suggest the existence of BVE-Zn$^{2+}$ complex. The Zn$^{2+}$ can complex with the pyrrole nitrogen atoms in BVE to form an annular planar structure (FIG. 1(B) ) [22]. This complex has a more rigid molecular structure than that of BVE and it can lead to higher fluorescence QY similar to the case that fluorescence of BV shows great enhancement at low temperatures [23]. FIG. S1 in supplementary materials shows that fluorescence intensity of this complex increase linearly until the BVE:Zn ratio over 1:1, indicating that the new complex is formed by complex BVE and Zn with 1:1 stoichiometry in ethanol.

 FIG. 2 (A, C) Steady-state absorption and (B, D) fluorescence spectra of BVE-Zn$^{2+}$ complex in ethanol and DMSO.

Similar absorption and fluorescence results (FIG. 2 (C) and (D)) were observed in BVE-Zn$^{2+}$ complex DMSO solution. However, when excess Zn$^{2+}$ is added into the solution, an 805 nm absorption peak appears while the fluorescence is quenched. From FIG. S2 in supplementary materials, we can see that BVE and Zn$^{2+}$ can form a complex with 1:2 ratio in DMSO, but this complex shows very low fluorescence emission, we will not study it in detail in this work. The complex with 1:1 ratio in DMSO will be also kept in the time-resolved spectroscopy study.

B. Fluorescence lifetime and QY of BVE-Zn$^{\textbf{2+}}$ complex

To further explore the mechanism of enhancement in BVE-Zn$^{2+}$ complex, we measured fluorescence lifetimes and QYs of BVE-Zn$^{2+}$ complex in ethanol, $n$-propanol and DMSO. As shown in FIG. 3, fluorescence decay at 750 nm from TCSPC measurements with 80 ps time resolution reveals a multi-exponential decay of BVE-Zn$^{2+}$ complex and the lifetime is obviously longer in DMSO. Table Ⅰ summarized the best-fit parameters for the fluorescence kinetics of BVE-Zn$^{2+}$ complex. In ethanol and $n$-propanol, the three time constants are similar and the average lifetime is 0.4 ns for the complex. However, the amplitude of the 0.79 ns component increases dramatically in DMSO (from $\sim$2% to $\sim$62%) and it leads to a 0.8 ns average lifetime for the complex.

 FIG. 3 Time resolved fluorescence of BVE-Zn$^{2+}$ complex from TCSPC measurement (650 nm excitation and 750 nm detection)
Table Ⅰ Fluorescence lifetime $\tau$ (in ns) with $A$ (%) in parenthesis for BVE-Zn$^{2+}$ complex in ethanol, $n$-propanol, and DMSO

Fluorescence QYs of BVE-Zn$^{2+}$ complex in three solvents are measured by using Nile Blue A as reference and the results are listed in Table Ⅱ. As mentioned above, fluorescence QY of BVE is less than 0.01%. However, after complexing with Zn$^{2+}$, the QY increased to 2.7%, 3.3%, and 7.4% in ethanol, $n$-propanol, and DMSO, respectively. With fluorescence lifetimes and QYs, radiative and non-radiative rate constants can be determined according to Eq.(5) and Eq.(6), and the results are listed in Table Ⅱ. Notably, radiative rate constant of BVE-Zn$^{2+}$ complex in DMSO only increases a little (1.3 times) compared with that in ethanol and $n$-propanol, but the non-radiative rate constant decreases to only half of its value in ethanol. This is the main factor that leads to the $\sim$3 times increasing of fluorescence QY in DMSO.

Table Ⅱ Fluorescence QY of BVE-Zn2+ complex and radiative and non-radiative rate constants.

In combination with the steady-state spectra, we speculate that it is structure difference of BVE-Zn$^{2+}$ complex in ethanol and DMSO that leads to the changes seen in fluorescence QY and lifetime. The structure of BVE in ethanol and DMSO has been studied by NMR and the results suggest that they actually have similar helical conformations in solution [24]. However, it was shown that the C = O group on ring D could form hydrogen bond with ethanol while the N$-$H at ring D could form hydrogen bond with S = O group in DMSO [25]. The strength of hydrogen bond in DMSO is stronger than that in ethanol. Therefore, structure of BVE itself is more rigid in DMSO. When Zn$^{2+}$ is added, the complex is proposed to have a structure similar to heme [22] and it can also stabilize the structure of BVE and restrain the out-of-plate motion of pyrrole as well as C15 = C16 double bond twisting [26]. In this case, the non-radiative decay pathway related with BVE structural deformation is greatly affected and this can lead to fluorescence QY enhancement as observed in our experiments. Our results are in line with literature that BVE shows high fluorescence QY when it is in protein environment [27]. We believe that due to stronger hydrogen bonding in DMSO, BVE-Zn$^{2+}$ complex should have a more rigid structure compared with that in ethanol and this is the key to its higher fluorescence QY and longer lifetime.

C. Transient absorption spectra of BVE-Zn$^{\textbf{2+}}$ complex

In order to understand excited state dynamics, TA measurements were further conducted in BVE-Zn$^{2+}$ complex solutions and the results are shown in FIG. 4. In all solvents, there are three main TA signals. The first one is ground state bleach (GSB) signal ranging from 350 nm to 420 nm, which corresponds well with the S$_0$-S$_2$ absorption; the second is excited state absorption (ESA) band ranging from 420 nm to 600 nm; and the last one is another GSB signal centered around 725 nm. No obvious stimulated emission (SE) signal is found in the TA data possibly due to the low fluorescence quantum yields and overlapping with GSB. There is no obvious shift of the ESA and GSB signal and they decay back to base line in $\sim$2$-$3 ns after excitation. TA kinetics for BVE-Zn$^{2+}$ complex in three solutions are displayed in FIG. 5. In each solvent, the ESA and GSB kinetics are mirror image-like, suggesting that decay of the excited state leads to direct repopulation of the ground state. As shown in FIG. 5, the excited state lifetimes are somehow similar in ethanol and $n$-propanol, but it is obviously longer in DMSO.

 FIG. 4 TA spectra of BVE-Zn$^{2+}$ complex in (A) ethanol, (B) $n$-propanol, and (C) DMSO. (D$-$F) are the evolution-associated difference spectra (EADS) extracted from global fitting.
 FIG. 5 TA kinetics for BVE-Zn$^{2+}$ complex in ethanol (left panel), $n$-propanol (middle panel) and DMSO (right panel).

Global analysis was carried out on the TA spectra of BVE-Zn$^{2+}$ complex. We extracted decay associated spectra (DAS) (FIG. S4 in supplementary materials) and evolution-associated difference spectra (EADS) to visualize the evolution of the excited states. In ethanol, three components are required in the global analysis, with lifetime of 5.2, 57.5, and 400 ps (Table Ⅲ), and the corresponding EADS is shown in FIG. 4(C). The first EADS (red) with 5.2 ps shows an ESA band around 475 nm, two GSB bands around 380 and 720 nm. It converts into the second EADS (green), which has similar shape but slightly lower amplitude of the GSB. The second EADS has a lifetime of 57.5 ps and then it decays into the third EADS (blue). Once again, the third EADS losses some amplitude in the GSB. Similar EADS and spectra evolution are seen in $n$-propanol. The third EADS has a lifetime of 400 ps, which closely corresponds to $\tau_2$ from TCSPC experiment, and is the result of main emissive deactivation channel. The loss of GSB amplitude during the first two EADS suggests depopulation of excited state can also take place and certain non-radiative pathway should exist in BVE-Zn$^{2+}$ complex when dissolved in ethanol and $n$-propanol. On the other hand, no loss of GSB signal during EADS evolution in DMSO and longer lifetime of the third EADS are observed. The 716 ps lifetime of the third EADS in DMSO matches with $\tau_3$ from TCSPC experiment. These results indicate that the non-radiative pathway for the first two EADS is prohibited and the third EADS should represent the main emissive decay channel of BVE-Zn$^{2+}$ complex in DMSO.

Table Ⅲ Global fit parameters of the TA spectra of BVE-Zn$^{2+}$ complex

According to our TA results we proposed the excited state decay mechanism of BVE-Zn$^{2+}$ complex in ethanol and DMSO. As shown in FIG. 6, in ethanol, A is Frank-Condon region of BVE-Zn$^{2+}$ complex in its S$_1$ state and it can either decay back to ground state or evolve to B with a lifetime of 5.2 ps. B is more like a vibrational relaxed S$_1$ state and it can also decay back to ground state or further decay to C. C is the main emission decay channel in BVE-Zn$^{2+}$ complex. C has a lifetime of 400 ps in ethanol and it is the main excited state decay channel of BVE-Zn$^{2+}$ complex. On the other hand, due to structure rigidity in DMSO, the non-radiative decay channels in A and B are restrained and BVE-Zn$^{2+}$ complex can only evolve to C and then decay back to ground state.

 FIG. 6 Proposed excited state decay mechanism of BVE-Zn$^{2+}$ complex in (A) ethanol and (B) DMSO
Ⅳ. CONCLUSION

In this work, BVE-Zn$^{2+}$ complex with 1:1 stoichiometry is studied by using both steady-state and time-resolve spectroscopy techniques. BVE is believed to have a heme-like structure after complexing with Zn$^{2+}$ and the structure rigidity is much higher than molecule itself. The fluorescence QY of complex is 60$-$80 times higher than that of BVE monomer. TCSPC results suggest that fluorescence enhancement is mainly due to suppressing the non-radiative decay rate constant, especially when dissolving the complex in DMSO due to stronger hydrogen bonding network. TA spectroscopy reveals that non-radiative decay channel is limited for the complex in DMSO and it is the key to the higher fluorescence QY and longer lifetime. Our results provide a guidance for the study about the emission mechanism of bacteriophytochrome where BV/BVE acts as chromophores.

Supplementary materials: Data analysis method for TCSPC and transient absorption spectroscopy; additional fluorescence spectra for determining BVE-Zn$^{2+}$ complex stoichiometry in ethanol and DMSO; fluorescence quantum yield measurements of BVE-Zn$^{2+}$ complex in ethanol, $n$-propanol and DMSO; and decay associated spectra of BVE-Zn$^{2+}$ complex are available.

Ⅴ. ACKNOWLEDGMENTS

This work was supported by the National Nature Science Foundation of China (No.11674101, No.21873030 and No.91850202)

Data analysis Methods

TCSPC data. The fluorescence lifetimes were determined by fitting fluorescence decay profiles to a decay model based on the nonlinear least-squares method with convolution of the IRF (FWHM= 80 ps).

Transient absorption data. Singular value decomposition (SVD) and decay-associated spectra (DAS) were generated from global fitting with Surface Xplorer 4.1.0 software (Ultrafast System). Evolution-associated difference spectra (EADS) were further implemented in Igor Pro. Software with the sequential model. The time dependent concentrations of the species used for the global fittings in EADS are as follows:

 $A(t) = {e^{ - {k_1}t}}$
 $B(t) = \frac{{{k_1}}}{{ - {k_1} + {k_2}}}{e^{ - {k_1}t}} - \frac{{{k_1}}}{{ - {k_1} + {k_2}}}{e^{ - {k_2}t}}$
 $C(t) = \frac{{{k_1}{k_2}}}{{\left( { - {k_1} + {k_2}} \right)\left( { - {k_1} + {k_3}} \right)}}{e^{ - {k_1}t}} - \frac{{{k_1}{k_2}}}{{\left( { - {k_1} + {k_2}} \right)\left( { - {k_2} + {k_3}} \right)}}{e^{ - {k_2}t}} + \frac{{{k_1}{k_2}}}{{\left( { - {k_1} + {k_3}} \right)\left( { - {k_2} + {k_3}} \right)}}{e^{ - {k_3}t}}$