Tải bản đầy đủ - 0 (trang)
2 Primary Structure, X-Ray Diffraction and Homology Models of Secondary and Tertiary Structure, Domain Organization

2 Primary Structure, X-Ray Diffraction and Homology Models of Secondary and Tertiary Structure, Domain Organization

Tải bản đầy đủ - 0trang

1 Structure of Transglutaminases: Unique Features Serve Diverse Functions


complex with four different irreversible inhibitors (PDB IDs:2Q3Z; 3S3J; 3S3P;

3S3S) (Liu et al. 2002; Pinkas et al. 2007; Han et al. 2010; Jang et al. 2014). The

crystals have been resolved at 2.00–3.14 Å. Human TGM3 was crystallized as a

zymogen in complex with a single calcium ion (PDB ID: 1L9M) and in the partially

cleaved form with two additional calcium ions (PDB ID: 1L9N) (Ahvazi

et al. 2002).

To construct a homology model for human TGM1, crystal structures of TGM2

and 3, and the model of TGM5 were used (Boeshans et al. 2007). As the N-terminus

could not be modeled from other Tgases, the segment Pro24-Ala55 was modeled on

a segment of Cd-6 metallothionenin (PDB ID:1DMF) and residues Asp66-Leu109

on human fascin (PDB ID: 1DFC). For Met1-Ser23 no template was found. The

isolated second β-barrel of human TGM1 comprising residues 693–787 has been

also crystallized and resolved at 2.3 Å (PDB ID:2XZZ). An illustration made of a

homology model of TGM5 was presented in Pietroni et al.; however, no technical

detail was made public (Pietroni and von Hippel 2008). For TGM6 two homology

models were generated, one based on the activated and calcium-bound form of

TGM3 (PDB ID:1L9N), the other based on the GDP-bound TGM2 to infer the

structures of both the calcium binding sites and the guanine nucleotide binding

pocket (PDB ID:1KV3) (Thomas et al. 2013). The total energies calculated from

these models were À32,300 and 17,400 kJ/mol, respectively, reflecting, possibly,

the stabilizing effect of calcium binding and the significantly higher level (50 %) of

sequence conservation between TGM6 and TGM3. No model has been published of

the 3D structures of TGM4 and 7. The crystal of red sea bream (Pargus major) liver

transglutaminase (PDB ID:1G0D) was resolved at 2.50 Å resolution (Noguchi

et al. 2001). The models of all vertebrate Tgases reveal the same four folded

domains (Fig. 1.1) which are similar in organization and structure: an N-terminal

β-sandwich domain, a catalytic core domain, and two C-terminal β-barrel domains

(Kiraly et al. 2011). These domains are also conserved in TGM1 with the exception

that according to its predicted model it contains an additional fifth domain at its

N-terminus (Boeshans et al. 2007). All except FXIIIa are monomeric. FXIIIa works

as a dimer. As will be discussed later the enzymes are thought to exist in closed and

open forms.

The structure of Streptoverticillium mobaraense transglutaminase was determined at 2.4 Å resolution (PDB ID:1 IU4) (Kashiwagi et al. 2002). As can be

expected from the sequence dissimilarity and the different sizes of the molecules

the overall structure of the Streptoverticillium mobaraense transglutaminase is

completely different from that of the vertebrate transglutaminases. The microbial

Tgase consists of a single domain that folds into a plate-like shape, and has one deep

cleft at the edge, which contains the active site (Kashiwagi et al. 2002).

The shape of the closed forms of the vertebrate enzymes is reminiscent of a

flattened triangle and has roughly the dimensions of 110x65x50 Å. Secondary

structural elements are conserved enough so that a description of TGM2 will give

an idea about all isoforms. The N-terminal β-sandwich domain is 130–140 amino

acids long and is folded into 9 β-strands and an α-helix. The catalytic core domain


´ . Deme´ny et al.


Fig. 1.1 General domain organization of vertebrate Tgases illustrated by human FXIIIa and

TGM2. (a) Folded 3D organization of the FXIIIa (left) and TGM2 (right) isoforms (PDB 1FIE

and 1KV3) with the characteristic four structural domains. (b) The four domains of the TGM2

protein are represented with rectangles. Exon boundaries within the coding sequence of the TGM2

gene are marked with arrowheads. Functional regions and amino acid positions are shown as

follows: red, amino acids of the catalytic triad; yellow, key residues involved in GTP/GDPbinding; FN, khaki, fibronectin binding site, WTATVVDQQDCTLSLQLTT(88–106) (Hang

et al. 2005); NLS1 and NLS2, violet, putative nuclear localization signals (Peng et al. 1999); S1–

S5, magenta, calcium binding regions (Kiraly et al. 2009); dashed brown, matrix

metalloproteinase 2 cleavage sites C-terminal to Pro375, Arg458 and His461 (Belkin

et al. 2004); PL, green, phospholipid binding motif KIRILGEPKQRKK(590–602) (Zemskov

et al. 2011); blue, Cys230, Cys370 and Cys371 are capable of forming disulfide bonds in the

closed and open conformations, respectively (Stamnaes et al. 2010)

could be encased within a 35 Â 35 Å box with a large triangular protrusion that has

been called the ‘flap’. It contains ~330 amino acids and consists of 12 or 13 β-sheets

interspersed with 9 or 10 α-helices and a 310 helix. (Secondary structure assignment

can be subjective in borderline cases. The description given here follows the

1 Structure of Transglutaminases: Unique Features Serve Diverse Functions


predictions in the DSSP database (Kabsch and Sander 1983).) The longest α-helix is

located in the center of the molecule and carries the catalytic cysteine toward its

end. β14, β16 and β17-18 seem to be the remainder of the ancient β-barrel

characteristic of the so called ‘transglutaminase fold’ with the catalytic histidine

and aspartate sitting on the adjacent β-strands, β16 and β17. If one compares it with

the blueprint given by Anantharaman, half of this barrel appears to have been

disrupted by insertions of helical segments in the connecting loops (Anantharaman

and Aravind 2003). β19-20 were pushed away with respect to the rest of the barrel.

The base of the hairpin made of β21 and 22 being dragged with it, the hairpin

twisted and turned downward changing the orientation of the strands to the opposite

with respect to the rest. The hairpin containing β15, which constitutes the ‘flap’

together with β19–20, seems to be a later evolutionary addition (Makarova

et al. 1999). The active site is shielded from contact with the solvent because it is

buried within a tunnel that is covered by the side chains of two conserved tryptophans, one of which is part of the same β-sheet as the catalytic histidine and the

other sitting in a loop of the α-helical bundle. The residues connecting the last

α-helical segment of the core domain to the β1-strand of the first β-barrel form a

flexible solvent-exposed loop that is the site of proteolytic cleavage in TGM1, 3, 5

and 6. At the same time, this connecting region constitutes the hinge around which

the domain movements opening the enzymes are thought to occur, as has been

demonstrated for TGM2 and FXIIIa (Pinkas et al. 2007; Stieler et al. 2013). The two

β-barrels are composed of ~100–110 amino acids, each folded into β-sheets that are

organized into two four- and three- antiparallel stranded plates. The first barrel

contains a short α-helical segment between β-sheets 5–6. In the closed form of the

enzymes the two β-barrels are closely appositioned to the surface of the catalytic

domain in the region of the active site. The loop between β-sheets 3 and 4 in barrel

1 harbors a conserved tyrosine residue that protrudes into the catalytic site area in

the closed form and is hydrogen bonded to the active site cysteine completing the

occlusion of the catalytic residue.

There are two non-proline and one proline conserved cis peptide bonds in the red

sea bream and human TGM2, the zymogen and active crystals of human TGM3 and

FXIIIa (Noguchi et al. 2001; Ahvazi et al. 2002; Weiss et al. 1998). The bonds are

present in the same locations within conserved sequence stretches in the isoforms

and much heed has been formerly paid to their relevance for stabilization, activation

or mechanism of action of the enzymes. Cis peptide bonds are energetically not

favorable and are found only in 0.03 % of proteins. Their stabilization is attributed

to extensive hydrogen bonding with neighbor atoms (Jabs et al. 1999). This network

of bonds may weave the cis peptide bond containing regions together, potentially

holding the active site containing residues in the proper orientation (Ahvazi

et al. 2002). Alternatively, they may isomerize to trans in association with the

conformational change and activation of the enzymes. However, the cis bonds were

found in the substrate-bound, open forms of both TGM2 and FXIIIa, which

supposedly represent the active enzymes and the proteolytically activated, nevertheless, closed form of TGM3, likely dismissing the hypothesis that the energy


´ . Deme´ny et al.


stored in these bonds is necessary for activation and supporting the idea that they

are important for stability. The question remains unresolved.


Catalytic Mechanism – Organization of the Active Site

The essence of the catalytic mechanism of Tgases was understood long before the

first atomic scale structural models became available thanks to classical enzymology studies, homology with papain-like proteases and experimental observations

obtained from site-directed mutagenesis and chemical compound screening. However, to grasp the nuances of the catalytic mechanism it was necessary to reveal and

derive the sense of the structural organization of the active site region of these

enzymes (Keillor et al. 2015).

The reaction catalyzed by the Tgases has been described as a modified ping-pong

mechanism (Folk 1969). The ping-pong mechanism is used by enzymes which

sequentially react with two substrates. They also exist in two states, E and E’, the

latter being a relatively stable reaction intermediate, a form chemically modified as

a result of interaction with the first reactant. When this substrate dissociates from

the modified enzyme, E’ can react with the second substrate. In the Tgase reaction

(Fig. 1.2) the first substrate is the protein- or peptide-bound glutamine donor and the

second substrate is a primary amine. The reaction may deviate from this scheme

and is called ‘modified’ inasmuch as there is an alternate second substrate, water,

which can hydrolyze the acyl-enzyme intermediate (E’) and direct the reaction to

deamidation. In the Tgase reaction the carboxamido group of the substrate glutamine is attacked by a nucleophilic cysteine (Cys314 in Hs FXIIIa/Cys277 in Hs

TGM2/Cys272 in HsTGM3) located in the active site, that is made reactive by

partial deprotonation via a charge relay system composed of a nearby histidine

(His373 in Hs FXIIIa/His335 in Hs TGM2/His330 in HsTGM3) and an aspartate

(Asp396 in Hs FXIIIa/Asp358 in Hs TGM2/Asp353 in HsTGM3). One of the

imidazole nitrogens acting as a base deprives the sulfhydryl residue of the cysteine

of a proton generating an imidazolium thiolate. The nucleophilic thiolate reacts

with the carbonyl leading to a tetrahedral intermediate. The histidine imidazolium

acts as a base catalyst of the decomposition of this intermediate to one molecule of

released ammonia and an enzyme-substrate thioester. The reactive enzyme can be

regenerated from this acylated complex by either amino- or hydrolysis. In the case

of aminolysis the histidine again functions as a base to deprotonate the neutral

amine, which will undertake a nucleophilic attack on the thiolester carbonyl

forming a second tetrahedral intermediate. The decomposition of this releases the

transacylated product and regenerates the catalytically competent thiolate cysteine.

If there is no suitable acyl-acceptor amine but water has access to the active site, the

acyl-enzyme intermediate is hydrolyzed to thiolate and glutamic acid. In this

scheme the histidine functions as a general base, which is rendered more electronegative by the interaction of the appropriately situated aspartate residue with the

other nitrogen of the imidazole ring.

1 Structure of Transglutaminases: Unique Features Serve Diverse Functions


Fig. 1.2 Unified mechanism of Tgase-mediated acyl-transfer reactions on the example of TGM2.

(Adapted from Keillor et al. 2014). The active form of the enzyme is the imidazolium thiolate

shown as form I. The side chain of a select Gln residue is bound in a tunnel leading down to active

site (see Michaelis complex II). Nucleophilic attack by the active site thiolate on the amide

substrate carbonyl leads to formation of tetrahedral intermediate III. The subsequent decomposition of III gives acyl enzyme IV and results in expulsion of one equivalent of product ammonia,

presumably with general acid assistance by imidazolium. Acyl enzyme IV now has two fates –

either aminolysis or hydrolysis, the selectivity of which is discussed below. If a suitable amine

acyl-acceptor substrate (such as the side chain of an accessible Lys residue) is present, it is bound

in a second binding tunnel to form Michaelis complex V. Two imidazole groups near the active site

organized in charge relays with acidic residues now function as general bases to deprotonate the

amine, which is positively charged at physiological pH, during its attack on the thiolester carbonyl.

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

2 Primary Structure, X-Ray Diffraction and Homology Models of Secondary and Tertiary Structure, Domain Organization

Tải bản đầy đủ ngay(0 tr)