The Mechanism of Cyclization in Chromophore Maturation of Green

Jul 1, 2010 - An intriguing aspect of the green fluorescent protein (GFP) is the ... 2005, 280, 26248−26255] are constructed on the basis of the X-r...
0 downloads 0 Views 3MB Size
Download PDF
9698

J. Phys. Chem. B 2010, 114, 9698–9705

The Mechanism of Cyclization in Chromophore Maturation of Green Fluorescent Protein: A Theoretical Study Yingying Ma,†,‡ Qiao Sun,‡ Hong Zhang,‡ Liang Peng,† Jian-Guo Yu,*,† and Sean C. Smith*,‡ College of Chemistry, Beijing Normal UniVersity, Beijing 100875, People’s Republic of China, and Centre for Computational Molecular Science, Australian Institute for Bioengineering and Nanotechnology, The UniVersity of Queensland, Qld 4072, Brisbane, Australia ReceiVed: May 1, 2010

An intriguing aspect of the green fluorescent protein (GFP) is the autocatalytic post-translational modification that results in the formation of its chromophore. Numerous experimental and theoretical studies indicate that cyclization is the first and the most important step in the maturation process. In this work, two proposed mechanisms for the cyclization were investigated by using the hybrid density functional theory method B3LYP. Cluster models corresponding to the two mechanisms proposed by Wachter et al. [J. Biol. Chem. 2005, 280, 26248-26255] are constructed on the basis of the X-ray crystal structure (PDB entry 2AWJ) and corresponding reaction path potential energy profiles for the two cyclization mechanisms are presented. Our results suggest that the backbone condensation initiated by deprotonation of the Gly67 amide nitrogen is easier than deprotonation of the Tyr66 R-carbon. Moreover, Arg96 fulfills the role of stabilizing the enolate moiety, and Glu222 plays the role of a general base. The formation of the cyclized product is found to be 16.0 and 18.6 kcal/mol endothermic with respect to the two models, which is in agreement with experimental observation. 1. Introduction The 2008 Nobel Prize in Chemistry honors three scientists for the discovery and development of the green fluorescent protein (GFP), which has become a ubiquitous tool for tagging and observing gene expression, protein localization, and cell developments.1-5 The GFP possesses a barrel-like structure, the wall of the barrel consisting of 11 β-sheets. The chromophore of the GFP is in the center of the barrel and is linked by the R-helix which runs through the center of the barrel. It was reported that wild-type GFP absorbs light mainly at 395 nm, with a smaller peak at 475 nm, due to the neutral and anionic chromophores, respectively.1,6 Upon irradiation, the neutral chromophore undergoes excited state proton transfer, generating the anionic chromophore. The mechanism of proton transfer in the GFP, as well as mechanistic and structural aspects of other fluorescent proteins, have been studied by our group among a number of others.7-25 The GFP proton transfer studies not only point out that the proton transfer is likely to be a single, concerted kinetic step but also reveals clear evidence of a sequential ordering of the movement of the protons in the chain. Moreover, GFPs and their mutants are particularly useful not only due to the fluorescence they can display, but also to their stabilities and the fact that the chromophores are formed in an autocatalytic cyclization of a peptide from their own backbone structure that does not require a cofactor (only molecular oxygen).26 The autocatalytic GFP cyclization is not unique; e.g., the histidine ammonia lyase and the phenylalanine ammonia lyase also are able to form a ring through the attack of the protein backbone on themselves.27,28 It is well-known that the autocatalytic cyclization of wild-type GFP is completed by its own three amino acids Ser65, Tyr66, and Gly67 (see Figure * To whom correspondence should be addressed. E-mail: jianguo_yu@ bnu.edu.cn and [email protected]. † Beijing Normal University. ‡ The University of Queensland.

1c), and the GFP fluorescence is observed at about 90 min to 4 h after protein synthesis.1,29 Despite the investigations summarized above, the detailed mechanism for the formation of the chromophore in GFP is not yet clear. In 1994, Tsien et al.1,2 studied the mechanism of chromophore formation based on expression of GFP in a wide variety of organisms and mutation experiments, proposing that the first step in chromophore formation is Gly67 amide nitrogen attack on Ser65 carbonyl carbon, followed by dehydration and then a final oxidation step. In 1997, Reid and Flynn30 observed the S65T-GFP chromophore maturation process in vitro, proposing three distinct kinetic steps: (1) protein folding, which precedes any chromophore modification and occurs fairly slowly (kf ) 2.44 × 10-3 s-1); (2) cyclization of the tripeptide with a rate constant of 3.8 × 10-3 s-1; and (3) oxidation of the cyclized chromophore (kox ) 1.51 × 10-4 s-1). On the basis of the density functional calculations, Siegbahn et al.31 subsequently proposed that dehydration of residue Tyr66 prior to cyclization (oxidized mechanism) is more favorable in energy than the commonly accepted reduced mechanism (wherein cyclization precedes dehydrogenation). However, it has been suggested that their calculations are problematic because parts of the models do not depend on the reliable experiment data, for example, the position and charge of Arg96.26 Recently, the mechanism of GFP chromophore maturation has continued to attract attention. Two mechanisms have been proposed respectively by Getzoff et al.32-37 and Wachter et al.38-45 based on their experiments. The first mechanism was called cyclization-dehydration-oxidation, proposed by Getzoff and her co-workers,32-37 in which the first step is the cyclization in chromophore maturation, followed dehydration of the Ser65 carbonyl group and finally oxidation of the CR-Cβ bond in Tyr66. Their explanation is based on the crystal data,36 which is an enolate moiety, cyclized and dehydrated but not oxidized. The experiment was carried out under anaerobic conditions and with dithionite reducing GFPsol (F64L S65T F99S M153T

10.1021/jp1039817  2010 American Chemical Society Published on Web 07/01/2010

Chromophore Maturation of Green Fluorescent Protein

J. Phys. Chem. B, Vol. 114, No. 29, 2010 9699

Figure 1. Optimized reactant of model A. For clarity, some hydrogen atoms are omitted, which is also applied to other figures below (a). Optimized reactant of model B (b). The crystal structure of the precyclized chromophore active site in GFP (coordinates taken from PDB 2AWJ32); the residues 96 and 65 have been changed into Arginine and Serine, respectively, with DeepView55 software (c).

SCHEME 1: Backbone Condensation Initiated by Deprotonation of the Tyr66 r-Carbon

9700

J. Phys. Chem. B, Vol. 114, No. 29, 2010

Ma et al.

SCHEME 2: Backbone Condensation Initiated by Deprotonation of Gly67 Amide Nitrogen

V163A);35 however, this might not translate to the real case because the whole process of chromophore formation in vivo is under aerobic conditions. The second mechanism proposed by Wachter et al.38-45 is cyclization-oxidation-dehydration, in which the cyclized intermediate is trapped by oxidation, followed by the dehydration involving the loss of hydroxyl on the five-membered heterocycle and a proton from Cβ in Tyr66. Both the crystal structure (PDB entry 1S6Z)38 and mass spectroscopy data45 support Wachter et al.’s mechanism. Maybe the strongest evidence for this mechanism is the kinetic isotope effect studies44 which indicate that the proton on Cβ in Tyr66 is not removed in the oxidation step. Meanwhile, Wachter et al. suggest two mechanisms40 by which cyclization is mediated: namely, backbone condensation initiated by either deprotonation of the Tyr66 R-carbon (Scheme 1) or deprotonation of the Gly67 amide nitrogen (Scheme 2). In the formation of the chromophore, whether the second step is dehydration or oxidation may depend on the oxygen concentration and the efficiency of ring dehydration for the particular fluorescent protein variant.46 However, the great majority of current evidence suggests that the first step is cyclization, i.e., the Gly67 amide nitrogen (NG) attacks on the Ser65 carbonyl carbon (CS), forming an imidazolone ring. There are two possible mechanisms for the cyclization40 in which backbone condensation is initiated either by deprotonation of the Tyr66 R-carbon (Scheme 1) or by deprotonation of Gly67 amide nitrogen (Scheme 2). Zimmer et al.26 pointed out in their studies that no hydrogen bonding network was found between Glu222 and the amide nitrogen of Gly67. Obviously, when the proton on the nitrogen of Gly67 is abstracted, the close proximity alignment of the Gly67 amide lone pair with the π* orbital of the residue 65 carbonyl is easily formed. Wachter et al. pointed out that deprotonation from Gly67 amide nitrogen should be easier than deprotonation from Tyr66 R-carbon.40 Thus, the mechanism study of cyclization in chromophore maturation of GFP is still not clear.

In this paper, the two proposed mechanisms for backbone condensation in the cyclization process of GFP chromophore maturation are investigated and compared. The study will gain insight into the detailed mechanism of cyclization in chromophore maturation of GFP. A better understanding of the mechanism of the autocatalytic post-translation reaction of fluorescent proteins will be beneficial in designing brighter and faster mutants, thereby increasing their utility in biotechnology and cell biology.46 2. Computational Details and Models 2.1. Computational Details. All calculations presented here were performed by means of the density functional theory (DFT) functional B3LYP,47,48 which is composed of Becke’s three-parameter hybrid exchange functional (B3) and the correlation functional of Lee, Yang, and Parr (LYP). This approach has previously been successfully applied to the study of a number of enzyme mechanisms.49-53 All geometry optimizations in our calculations were carried out with the double-ζ plus polarization basis set 6-31G(d,p). On the basis of these geometries, more accurate energies were obtained by performing single-point energy calculations with the larger basis set 6-311+G(2d,2p). Frequency calculations were performed at the same theory level as the optimizations to obtain zero-point energies (ZPE) and to confirm the nature of the stationary points. The latter implies no negative eigenvalues for minima and only one negative eigenvalue for transition states. Freezing some atoms to their crystallographic positions gives rise to a few small negative eigenvalues for the optimized structures; however, these are only in the order of 10 cm-1 and do not affect the obtained energetic results significantly. The energies reported here are corrected for zero-point vibrational effects. All calculations were performed with the Gaussian 03 program package.54 2.2. Our Model. The chromophore of wild-type GFP is autocatalytically generated by a multistep chemical reaction of

Chromophore Maturation of Green Fluorescent Protein

J. Phys. Chem. B, Vol. 114, No. 29, 2010 9701

a 3-amino acid sequence (S65-Y66-G67). Besides these three amino acids, Arg96 and Glu222 also fulfill important roles in chromophore formation, especially the positive charge of Arg96 plays a pivotal role in accelerating the GFP chromophore maturation.32 So, in order to get the precyclized structure experimentally, the Arg96 is usually mutated. For example, in the high resolution crystal structure (PDB entry 2AWJ),32 the 96th residue Arginine is mutated to Methionine. However, since our interest is to investigate the chromophore formation process, the 96th residue should be Arginine. Therefore, to study the chromophore maturation of wild-type GFP, rather than S65TGFP, the 96th and 65th residues in crystal structure 2AWJ are mutated to Arginine and Serine respectively with DeepView software,55 which can place the best rotamer of the new amino acid into the old amino acid position. Models A and B (Figure 1a,b) were constructed according to the two cyclization mechanisms suggested by Wachter et al.40 on the basis of the crystal structure. Respectively, they correspond to the backbone condensation initiated by deprotonation of the Tyr66 R-carbon and Gly67 amide nitrogen. The two models all contain Ser65, Tyr66, Gly67, Arg96, Glu222, H2O1042, and H2O1158. The amino group of the Ser65, the carboxyl group of the Gly67, and the amino group and carboxyl group of the Glu222 and Arg96 are truncated and saturated with H as usual but all atoms of Tyr66 were kept in the model. The resulting models include 78 atoms and the total charge is zero. The only difference between the two models is the position in which the two water molecules appear. The two water molecules in model B are situated between the Gly67 and Glu222. To keep the calculated structures close to those obtained experimentally, some atoms were fixed to their X-ray positions during the geometry optimizations. The fixed atoms are marked with asterisks in the figures below. 3. Results and Discussions 3.1. Cyclization Initiated by Deprotonation of the Tyr66 r-Carbon. The optimized structure of the reactant in model A is shown in Figure 1a. Figure 1c is the crystal structure of precyclized chromophore active site in GFP (coordinates taken from PDB 2AWJ32) with the Met96 and Thr65 mutated into Arg96 and Ser65, respectively, using DeepView55 software. The overall geometric parameters obtained from optimization are basically in agreement with experimental structure. The distance between the guanidinium group of Arg96 and the carbonyl of Tyr66 in our optimized reactant is shorter than the corresponding distance in the crystal structure. The reason is that there is electrostatic interaction between the guanidinium (positive charge) and the carbonyl oxygen (negative charge), so they get close during the optimization. Accordingly, in the crystal structure, the distance between the R-carbon of Tyr66 and the oxygen atom of H2O1158, the oxygen atom of H2O1042, and the oxygen of Glu222 acetate is 5.07, 2.88, and 3.20 Å, respectively, and the corresponding distances are 3.24, 2.70, and 2.65 Å, respectively, in our optimized structure. Generally, the results show that the position of two water molecules in the optimized reactant is closer to Tyr66 and Glu222 than those in the crystal structure. 3.1.1. Proton Transfer from Tyr66 r-Carbon to Glu222. The optimized transition state for proton transfer (A-TS1) and the resulting intermediate (A-Int1) are shown in Figure 2. Our calculational results show that the proton transfer from R-carbon (CR) to oxygen atom of Glu222 acetate (OE) via two water molecules is a concerted process. The barrier is calculated to be 15.6 kcal/mol, and A-Int1 is found to lie 0.7 kcal/mol higher than A-Re. The A-TS1 is confirmed to be the first-order saddle

Figure 2. Optimized structures of the transition state of proton transfer (A-TS1), product of proton transfer (A-Int1), the transition state of nucleophilic attack (A-TS2), and the product of cyclization (A-pro) of model A. The atoms marked with asterisks were fixed during the calculations.

point with only the imaginary frequency (817i cm-1), which corresponds to the stretch mode of CR-HR, O1-H1, and O2-H2 bond simultaneously. The CR-HR, O1-H1, O2-H2, and H2-OE bond distances in A-Re are 1.10, 0.99, 1.01, and 1.65 Å, respectively (see Table 1). The four values in A-TS1 are 1.30, 1.44, 1.46, and 1.05 Å, respectively. While in A-Int1, the four bond distances are 2.50, 1.78, 1.52, and 1.03 Å, respectively. These data show that the proton on CR transferred to OE via the transition state A-TS1. The distance between CR and carbonyl carbon of Tyr66 (CY) is 1.55 Å in A-Re to 1.50 Å at A-TS1 and 1.38 Å at A-Int1, and the distance between CY and the carbonyl oxygen of Tyr66 (OY) is 1.26 Å in A-Re to 1.28 Å at A-TS1 and 1.35 Å at A-Int1. All these show that the carbonyl has changed into enolate accompanying the process of proton transfer. We further notice that the distance between the

9702

J. Phys. Chem. B, Vol. 114, No. 29, 2010

Ma et al.

TABLE 1: Important Distances (Å) for the Various Stationary Points along the Reaction Pathways of Proton Transfer from the r-Carbon of Tyr66 (Cr) to the Oxygen Atom of Glu222 Carboxylate (OE) in Models A and B, Respectively CR-HR

HR-O1

O1-H1

H1-O2

O2-H2

H2-OE

CR-CY

CY-OY

OY-HR

1.10 1.30 2.50 1.10 1.33 2.06

2.17 1.35 0.97 2.33 1.32 0.99

0.99 1.44 1.78 0.98 1.56 1.93

1.74 1.06 0.98 1.86 1.04 0.98

1.01 1.46 1.52 1.02 1.65 1.69

1.65 1.05 1.03 1.57 1.02 1.00

1.55 1.50 1.38 1.53 1.46 1.38

1.26 1.28 1.35 1.26 1.28 1.30

1.79 1.71 1.01 1.81 1.73 1.70

A-Re A-TS1 A-Int1 B-Int2b B-TS3 B-Pr

TABLE 2: Important Distances (Å) for the Various Stationary Points along the Reaction Pathways of Nucleophilic Attack in Model A A-Int1 A-TS2 A-Pr

NG-CS

HG-OS

CS-OS

NG-HG

NG-CY

3.08 2.08 1.46

1.90 1.01 0.97

1.25 1.38 1.43

1.02 1.68 2.34

1.37 1.36 1.41

hydrogen atom on guanidinium of Arg96 (HR) and the OY is 1.79 Å in A-Re to 1.71 Å at A-TS and 1.01 Å at A-Int1. This shows that the Arg96 provides electrostatic stabilization to the transition state and the enolate intermediate, thereby lowering the barrier for the proton transfer. The Glu222 plays the role of a general base, facilitating proton abstraction from the Tyr66 R-carbon. So, our results are in line with the proposed mechanism based on the experiment,35,40 and Zimmer et al.’s recent theoretical results26 also support the above role of Arg96 and Glu222. 3.1.2. Nucleophilic Attack. The second part of the reaction for model A is that the amide nitrogen atom of Gly67 (NG) attacks the carbonyl carbon of Ser65 (CS), resulting in the formation of the five-membered heterocycle. The transition state for this (called A-TS2) is also shown in Figure 2. It was found that, simultaneously with the NG-CS bond formed, a proton transferred from the NG to the OS. The transition state is characterized by an imaginary frequency of 301i cm-1, which mainly corresponds to the bond of NG-CS and NG-HG simultaneous stretch mode. The calculated energy of A-TS2 is 42.2 kcal/mol higher than that of A-Int1 (i.e., 42.9 kcal/mol higher than that of A-Re), and the resulting structure corresponds to the five-membered heterocycle with the hydroxyl attached to it (A-Pr, Figure 2). The energy of this structure is calculated to be 16.0 kcal/mol higher than that of A-Re. The energy barrier is high, because the four-membered-ring NG-CSOS-HG is formed at A-TS2. The four-membered ring has a big tensile force and needs high energy to form. At A-TS2, the critical NG-CS distance is 2.08 Å, and the NG-HG, CS-OS, and OS-HG distance is 1.68, 1.38, and 1.01 Å, respectively. As seen in Table 2, the corresponding four distances are 1.46, 2.34, 1.43, and 0.97 Å, respectively, at A-Pr, which show that the five-membered heterocycle has formed and the HG has completely transferred to the OS. At A-TS2, the phenol ring has rotated and is nearly perpendicular to the five-membered heterocycle. There is evidence for this in the crystal structure (PDB entry 2FZU),36 in which the phenol ring is nearly perpendicular to the five-membered heterocycle. The calculated energy profile for the reaction mechanism in model A is displayed in Figure 3. It can be seen that the ratelimiting step is nucleophilic attack, which is the amide nitrogen of Gly67 attack on the carbonyl carbon of Ser65. The relative energy of the rate-limiting step is 42.9 kcal/mol, which is too high for the reaction to occur in vivo. The energy of final product A-Pr is 16.0 kcal/mol relative to the A-Re. This shows that the process of cyclization is endothermic, which is consistent with the experiment results.

3.2. Cyclization Initiated by Deprotonation of the Gly67 Amide Nitrogen. The optimized structure of reactant in model B is shown in Figure 1b. Comparing with the reactant of model A, the main difference is that there are two water molecules between the Gly67 amide nitrogen (NG) and the Glu222 acetate oxygen (OE), which means there is a hydrogen bond network formed between the residues of Gly67 and Glu222. Thus it is easier for the hydrogen on the NG (HG) to transfer to the OE by the hydrogen bond network. The fixed atoms in model B are the same as those of model A. 3.2.1. Proton Transfer from Gly67 Amide Nitrogen to Glu222. The first step of model B is the HG atom of Gly67 transfers to the atom OS of Glu222, via the transition state of B-TS1, which is confirmed with one imaginary frequency (830i cm-1). This frequency corresponds to NG-HG, O1-H1, and O2-H2 simultaneous with the same direction stretch mode. There is no barrier at the level of B3LYP/6-311+G(2d,2p) with zero-point energy correction (Figure 5), and the barrier is 5.5 kcal/mol without zero-point correction at the same level. Large effects from zero-point vibration are expected for these types of reactions where there are partly broken X-H bonds in the transition states.31 The firm conclusion, however, is that this step is very fast. Comparing to deprotonation from the R-carbon of Tyr66 (the energy needed is 15.6 kcal/mol, see section 3.1.1), it can be seen that deprotonation from the amide nitrogen of Gly67 is much easier than deprotonation from the R-carbon of Tyr66. This is in agreement with Wachter’s suggestion.40 The resulting B-Int1 is calculated to be 8.2 kcal/mol lower than that of B-Re. The NG-HG, O1-H1, O2-H2, and H2-OE distances are 1.05, 1.00, 1.00, and 1.73 Å, respectively, in B-Re to 1.10, 1.18, 1.34, and 1.13, Å respectively, at B-TS1, and 1.93, 1.72, 1.71, and 1.01 Å, respectively, at B-Int1 (see Table 3). This shows that the proton on amide nitrogen of Gly67 transfers to the oxygen of Glu222 acetate. During the proton transfer, the CY-NG distance decreases from 1.33 Å in B-Re to 1.32 Å at B-TS1, and 1.30 Å in B-Int1, while the CY-OY distance is elongated from 1.27 Å in B-Re to 1.28 Å at B-TS1 to 1.31 Å at B-Int1. In particular, the HR-OY distance is decreased from 1.81 Å in B-Re to 1.78 Å at B-TS1 to 1.67 Å at B-Int1. This shows that the Arg96 positive charge has electrostatic interaction with the Tyr66 carbonyl, favoring Gly67 nitrogen deprotonation,

Figure 3. Potential energy profile for the cyclization reaction of model A. Digits in black are the energies of B3LYP/6-31G(d,p); digits in blue are the energies at the B3LYP/6-311+G(2d,2p)//B3LYP/6-31G(d,p) level with zero-point energies included.

Chromophore Maturation of Green Fluorescent Protein

J. Phys. Chem. B, Vol. 114, No. 29, 2010 9703

Figure 4. Optimized structure of the transition states and intermediates of model B. The atoms marked with asterisks were fixed during the calculations.

thereby stabilizing the enloate intermediate and accelerating the GFP post-translational modification. Experiment obversation32,37,40 also support our conclusion. 3.2.2. Nucleophilic Attack. The optimized transition state for nucleophilic attack (B-TS2) and the resulting five-membered heterocycle intermediate (B-Int2a) are displayed in Figure 4. It was found that, simultaneously with the NG attacking on the carbonyl carbon of Ser65 (CS), a proton moves from the oxygen of Glu222 acetate (OE) to the carbonyl oxygen of Ser65 (OS). So this is a concerted process. The transition state is confirmed to be the first-order saddle point with only one imaginary frequency (141i cm-1), which corresponds to NG-CS, H1-OS, H2-O1, and HE-O2 simultaneous stretch mode. The calculated energetic barrier for this step is 20.4 kcal/mol, and B-Int2a is calculated to be 1.9 kcal/mol higher than that of B-Re. At B-TS2, the critical NG-CS distance is 2.16 Å, and there is one

Figure 5. Potential energy profile for the cyclization reaction of model B. Digits in black are the energies of B3LYP/6-31G(d,p); digits in blue are the energies at the B3LYP/6-311+G(2d,2p)//B3LYP/6-31G(d,p) level with zero-point energies included.

9704

J. Phys. Chem. B, Vol. 114, No. 29, 2010

Ma et al.

TABLE 3: Important Distances (Å) for the Various Stationary Points along the Reaction Pathways of Proton Transfer from the Amide Nitrogen of Gly67 (NG) to the Oxygen Atom of Glu222 Carboxylate (OE) in Model B B-Re B-TS1 B-Int1

NG-HG

HG-O1

O1-H1

H1-O2

O2-H2

H2-OE

CY-NG

CY-OY

OY-HR

1.05 1.10 1.93

1.74 1.49 0.99

1.00 1.18 1.72

1.76 1.26 1.00

1.00 1.34 1.71

1.73 1.13 1.01

1.33 1.32 1.30

1.27 1.28 1.31

1.81 1.78 1.67

TABLE 4: Important Distances (Å) for the Various Stationary Points along the Reaction Pathways of Nucleophilic Attack in Model B B-Int1 B-TS2 B-Int2a

NG-CS

CS-OS

OS-H1

H1-O1

O1-H2

H2-O2

O2-HE

CY-NG

CY-OY

OY-HR

3.83 2.16 1.51

1.23 1.31 1.39

1.90 1.05 1.01

0.98 1.45 1.66

1.72 1.06 1.01

1.00 1.45 1.64

1.71 1.03 1.01

1.30 1.29 1.33

1.31 1.30 1.26

1.67 1.70 1.83

hydrogen bond network between OS and OE. The CS-OS distance is 1.23 Å in B-Int1 to 1.31 Å at B-TS2 and 1.39 Å at B-Int2a. The OS-H1, O1-H2, and O2-HE distances are 1.05, 1.06, and 1.03 Å, respectively, at B-TS2 and 1.01, 1.01, and 1.01 Å, respectively, at B-Int2a (see Table 4). All these show that while NG attacks on the CS, a proton transfers from the OE to the OS. 3.2.3. Proton Transfer from Tyr66 r-Carbon to Glu222. The final step of model B is proton transfer, in which the hydrogen on the R-carbon of Tyr66 transfers to the oxygen of Glu222 acetate via the two water molecules. The only difference between B-Int2b and B-Int2a is the position of one water molecule, as is shown in Figure 4. The two water molecules in B-Int2b form a hydrogen bonding network between R-carbon of Tyr66 and the oxygen of Glu222 acetate. The energy of B-Int2b is 4.0 kcal/mol higher than that of B-Int2a. The optimized transition state B-TS3 for this proton transfer and resulting B-Pr are shown in Figure 4. The transition state B-TS3 is confirmed to be the first-order saddle point with only one imaginary frequency (990i cm-1), which corresponds to CR-HR, O1-H1, and O2-H2 simultaneously with the same direction stretch mode. The energetic barrier is 18.0 kcal/mol relative to the B-Int2b. The energy of the resulting B-Pr is calculated to be 18.6 kcal/mol relative to the B-Re. The key distances of CR-HR, O1-H1, O2-H2, and H2-OE are 1.10, 0.98, 1.02, and 1.57 Å, respectively, in B-Int2b to 1.33, 1.56, 1.65, and 1.02 Å, respectively, at B-TS3, and 2.06, 1.93, 1.69, and 1.00 Å at B-Pr (see Table 1). This shows that the proton on the Tyr66 R-carbon has transferred to the Glu222. The CR-CY, and CY-OY distances are 1.53 and 1.26 Å, respectively, in B-Int2b to 1.46 and 1.28 Å, respectively, at B-TS3 and 1.38 and 1.30 Å, respectively, at B-Pr. This shows that when proton transfers from CR to OE, the enolate is formed in B-Pr. The OY-HR distance is 1.81 Å in B-Int2b to 1.73 Å at B-TS3 and 1.70 Å at B-Pr, which show that during the proton transfer the interaction between the carbonyl of Tyr66 and the guanidinium of Arg96 becomes stronger. The calculated energy profile for the reaction mechanism is displayed in Figure 5. The rate-limiting step is the proton on the R-carbon of Tyr66 via two water molecules transfers to the oxygen of Glu222 acetate. The relative energy of the ratelimiting step is 23.9 kcal/mol, which is feasible in vivo. The relative energy of the final product B-Pr is 18.6 kcal/mol higher relative to B-Re, which shows that the cyclization in model B is endothermic. Comparing Figures 3 and 5, it can be seen that the backbone condensation initiated by deprotonation of Gly67 amide nitrogen is easier than deprotonation of the Tyr66 R-carbon. The reason is that the barrier of the rate-determining step in model B is 18.0 kcal/mol, whereas the barrier of the rate-determining step

in model A is much higher, with the value of 42.2 kcal/mol, although the energy of B-Re is about 9 kcal/mol higher than that of A-Re. Moreover, the energy of nucleophilic attack in model B is much lower than that in model A. We can rationalize the quite different barriers involved for these two mechanisms with the observations that when the hydrogen on the nitrogen of Gly67 is abstracted, it is more favorable for the close alignment of the Gly67 amide lone pair with the π* orbital of the residue 65 carbonyl. In contrast, for the mechanism of model A, when the nitrogen atom of Gly67 attacks the carbonyl carbon of Ser65, the hydrogen transfers from the amide nitrogen of Gly67 to the carbonyl oxygen of Ser65. During this process a four-membered cycle N-C-O-H involving significant strain is formed;leading to a much higher barrier. Thus, the mechanism of backbone condensation initiated by deprotonation of Gly67 amide nitrogen (Scheme 2) is demonstrated to be the more feasible mechanism. 4. Conclusions In this paper, we investigated the mechanism of cyclization in the maturation of the GFP using the hybrid density functional theory method B3LYP. Our calculations are based on the two cyclization mechanisms proposed by Wachter et al., in which the backbone condensation may be initiated either by deprotonation of Gly67 amide nitrogen or by deprotonation of the Tyr66 R-carbon. For the first time, the models were built on the precyclized chromophore crystal structure (PDB entry 2AWJ), which gives reasonable starting points for the calculations. The calculated results are conducive toward explanation of experimental observations and illuminate for the first time that backbone condensation initiated by deprotonation of Gly67 amide nitrogen (Scheme 2) is easier than deprotonation of the Tyr66 R-carbon (Scheme 1) in the cyclizaton process. These results are consistent with Getzoff and Wachter’s experimental observations, showing, e.g., that Arg96 fulfills the role of stabilizing the enolate moiety, that Glu222 plays the role of a general base, and that the cyclization process is endothermic. Finally, it is important to note that the models presented herein shed light only on the intrinsic mechanism of cyclization in chromophore maturation in the absence of the protein matrix. Clearly, when reliable QM/MM calculations become available for the reaction path, it will be very interesting to find out whether or not the influence of the other surrounding amino acid residues in the GFP chromophore cavity changes the picture obtained from these cluster models for the intrinsic mechanism of cyclization in the GFP chromophore maturation process. Acknowledgment. Y.Y.M. thanks the Australian Institute for Bioengineering and Nanotechnology at The University of

Chromophore Maturation of Green Fluorescent Protein Queensland for facilitating a four-month research visit (supported in part by Australian Research Council Discovery Project DP0771867) and the provision of computing resources at the Centre for Computational Molecular Science (funded in part by ARC LIEF project LE0882357). We thank Dr. Sven de Marothy (from Stockholm University) for providing XYZViewer to draw most of the figures of the molecule models. This work was also supported by grants from the National Natural Science Foundation of China (Grant Nos. 20733002 and 20873008). References and Notes (1) Heim, R.; Prasher, D. C.; Tsien, R. Y. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 12501. (2) Cubitt, A. B.; Heim, R.; Adams, S. R.; Boyd, A. E.; Gross, L. A.; Tsien, R. Y. Trends Biochem. Sci. 1995, 20, 448. (3) Orm, M.; Cubitt, A. B.; Kallio, K.; Gross, L. A.; Tsien, R. Y.; Remington, S. J. Science 1996, 273, 1392. (4) Zimmer, M. Chem. ReV. 2002, 102, 759. (5) Giepmans, B. N. G.; Adams, S. R.; Ellisman, M. H.; Tsien, R. Y. Science 2006, 312, 217. (6) Niwa, H.; Inouye, S.; Hirano, T.; Matsuno, T.; Kojima, S.; Kubota, M.; Ohashi, M.; Tsuji, F. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 13617. (7) Wang, S. F.; Smith, S. C. Chem. Phys. 2006, 326, 204. (8) Wang, S. F.; Smith, S. C. J. Phys. Chem. B 2006, 110, 5084. (9) Wang, S. F.; Smith, S. C. Phys. Chem. Chem. Phys. 2007, 9, 452. (10) Zhang, H.; Smith, S. C. J. Theor. Comput. Chem. 2007, 6, 789. (11) Zhang, H.; Wang, S. F.; Sun, Q.; Smith, S. C. Phys. Chem. Chem. Phys. 2009, 11, 8422. (12) Sun, Q.; Doerr, M.; Li, Z.; Smith, S. C.; Thiel, W. Phys. Chem. Chem. Phys. 2010, 12, 2450. (13) Vendrell, O.; Gelabert, R.; Moreno, M.; Lluch, J. M. J. Chem. Theory Comput. 2008, 4, 1138. (14) Vendrell, O.; Gelabert, R.; Moreno, M.; Lluch, J. M. J. Phys. Chem. B 2008, 112, 5500. (15) Vendrell, O.; Gelabert, R.; Moreno, M.; Lluch, J. M. J. Am. Chem. Soc. 2006, 128, 3564. (16) Olsen, S.; Smith, S. C. J. Am. Chem. Soc. 2008, 130, 8677. (17) Olsen, S.; Smith, S. C. J. Am. Chem. Soc. 2007, 129, 2054. (18) Olsen, S.; Smith, S. C. Chem. Phys. Lett. 2006, 426, 159. (19) Olsen, S.; Prescott, M.; Wilmann, P.; Battad, J.; Rossjohn, J.; Smith, S. C. Chem. Phys. Lett. 2006, 420, 507. (20) Reuter, N.; Lin, H.; Thiel, W. J. Phys. Chem. B 2002, 106, 6310. (21) Scha¨fer, L. V.; Groenhof, G.; Boggio-Pasqua, M.; Robb, M. A.; Grubmu¨ller, H. PLoS Comput. Biol. 2008, 4, e1000034. (22) Scha¨fer, L. V.; Gerrit, G.; Klingen, A. R.; Ullmann, G. M.; Martial, B.-P.; Robb, M. A.; Helmut, G. Angew. Chem., Int. Ed. 2007, 46, 530. (23) Zhang, H. S. Q.; Wang, S.; Olsen, S.; Smith, S. C. Theoretical studies of Green and Red Fluorescent Proteins; John Wiley & Sons, Inc, 2010; in press. (24) Sun, Q.; Wang, S. F.; Zhang, H.; Li, Z.; Pifisterer, C.; Fischer, S.; Nanbu, S.; Smith, S. C. Aust. J. Chem. 2010, 63, 363. (25) Andresen, M.; Wahl, M. C.; Stiel, A. C.; Grater, F.; Schafer, L. V.; Trowitzsch, S.; Weber, G.; Eggeling, C.; Grubmu¨ller, H.; Hell, S. W.; Jakobs, S. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 13070. (26) Lemay, N. P.; Morgan, A. L.; Archer, E. J.; Dickson, L. A.; Megley, C. M.; Zimmer, M. Chem. Phys. 2008, 348, 152. (27) Schwede, T. F.; Retey, J.; Schulz, G. E. Biochemistry 1999, 38, 5355. (28) Donnelly, M.; Fedeles, F.; Wirstam, M.; Siegbahn, P. E.; Zimmer, M. J. Am. Chem. Soc. 2001, 123, 4679.

J. Phys. Chem. B, Vol. 114, No. 29, 2010 9705 (29) Crameri, A.; Whitehorn, E. A.; Tate, E.; Stemmer, W. P. C. Nat. Biotechnol. 1996, 14, 315. (30) Reid, B. G.; Flynn, G. C. Biochemistry 1997, 36, 6786. (31) Per, E. M. S.; Maria, W.; Marc, Z. Int. J. Quantum Chem. 2001, 81, 169. (32) Barondeau, D. P.; Hitomi, C.; Kassmann, C. J.; Tainer, J. A.; Getzoff, E. D. Biochemistry 2005, 44, 16211. (33) Barondeau, D. P.; Kassmann, C. J.; Tainer, J. A.; Getzoff, E. D. J. Am. Chem. Soc. 2007, 129, 3118. (34) Barondeau, D. P.; Kassmann, C. J.; Tainer, J. A.; Getzoff, E. D. J. Am. Chem. Soc. 2006, 128, 4685. (35) Barondeau, D. P.; Putnam, C. D.; Kassmann, C. J.; Tainer, J. A.; Getzoff, E. D. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 12111. (36) Barondeau, D. P.; Tainer, J. A.; Getzoff, E. D. J. Am. Chem. Soc. 2006, 128, 3166. (37) Kassmann, C. J.; Tainer, J. A.; Getzoff, E. D. Biochemistry 2005, 44, 1960. (38) Huffman, H. A.; Phail, M. E.; Wachter, R. M. Biochemistry 2004, 43, 4464. (39) Patel, H. N.; Wachter, R. M. Biochemistry 2005, 44, 8303. (40) Sniegowski, J. A.; Lappe, J. W.; Patel, H. N.; Huffman, H. A.; Wachter, R. M. J. Biol. Chem. 2005, 280, 26248. (41) Sniegowski, J. A.; Phail, M. E.; Wachter, R. M. Biochem. Biophys. Res. Commun. 2005, 332, 657. (42) Wachter, R. M. Acc. Chem. Res. 2007, 40, 120. (43) Wachter, R. M. Mechanistic aspects of GFP chromophore biogenesis; Genetically Engineered Probes for Biomedical Applications, San Jose, CA, 2006. (44) Zhang, L.; Chan, N. H.; Dorrestein, P. C.; Wachter, R. M. Biochemistry 2008, 47, 10111. (45) Zhang, L.; Patel, H. N.; Lappe, J. W.; Wachter, R. M. J. Am. Chem. Soc. 2006, 128, 4766. (46) Craggs, T. D. Chem. Soc. ReV. 2009, 38, 2865. (47) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (48) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (49) Chen, S.-L.; Fang, W.-H.; Himo, F. J. Phys. Chem. B 2007, 111, 1253. (50) Chen, S.-L.; Marino, T.; Fang, W.-H.; Russo, N.; Himo, F. J. Phys. Chem. B 2008, 112, 2494. (51) Liao, R.-Z.; Yu, J.-G.; Raushel, F. M.; Himo, F. Chem.sEur. J. 2008, 14, 4287. (52) Liao, R.-Z.; Yu, J.-G.; Himo, F. Inorg. Chem. 2009, 48, 1442. (53) Yang, L.; Liao, R.-Z.; Yu, J.-G.; Liu, R.-Z. J. Phys. Chem. B 2009, 113, 6505. (54) Frisch, M. J.; Trucks G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision C.02; Gaussian, Inc., Wallingford, CT, 2004. (55) Guex, N.; Peitsch, M. C. Electrophoresis 1997, 18, 2714.

JP1039817