The glutathione (GSH) S-transferases are a group of enzymes whose main function is to convert endogenous, and xenobiotic electrophilic compounds, to water soluble intermediates, that may be eliminated. These enzymes are common to all eukaryotes and to many bacteria, and those that are soluble have been grouped into four distinct classes. These classes are based on their structural relationship to that of the theta group an ancestral enzyme that appears to be relatively constant in bacteria of relatively unchanged genome. Other distinct classes are alpha , mu, and pi.
With respect to primary protein structure, the glutathione S-transferase of the cephalopod digestive gland, shares a striking resemblance to the S-crystalline of the cephalopod lens. This is thought to be due to the copying and mutation of the gene that encodes for the glutathione S-transferase, to a gene that encodes for the production of S-crystalline of the lens of the cephalopod eye. Due to this commonality between these two proteins, and the great difference in primary structure of the glutathione S-transferase from the other classes, a new class sigma has been created.
Crystalines are water-soluble structural proteins found in fibril cells of animal ocular lenses. These are divided ,like the glutathione- S-transferases above, into four heterogenous classes alpha, beta, gamma, and delta. " These classes were formed on the basis of size, shape, charge, immunological properties, and source. Though other minor classes have been classified these four are the most widely occurring. The alpha,.beta, and delta are found in avian and reptilian lenses and alpha beta and gamma are found in the lenses of all other vertebrates."(6) The crystalines of all animals that have lenses are similar in sequence, and certainly similar in shape , for their shape determines their structural role. The crystalines must be within a close spacial relation to each other, if the lens is to maintain its transparency, thus their shape must be constant and fixed, to order themselves into set rows of crystallin form.
The catalytic activity of the S-crystallin toward 1-chloro-2,4-dinitrobenzene, a substrate to glutathione S-transferase shows evidence that the gene that encodes for the glutathione S-transferase, may have been copied, and evolved to produce the S-crystallin. Though this S-crystallin, a water-soluble refractory protein, performs its refractive purpose, it also possesses an unusual reactivity toward the substrate 1-chloro-2,4-dinitrobenzene. Similar refractive proteins from other animals do not show this reactivity to this substrate. Using this reactivity, they have deduced the three dimensional structure of glutathione S-transferase.
The structural similarity of the theta class, being found in ancestral bacteria, to the sigma class is so close, and yet, different from the alpha, mu, and pi, that they where able to discern that the theta and sigma classes diverged from the ancestral precursor, before alpha, mu, and pi. In this paper, I will attempt to clarify the means by which they accomplished the solution of the three dimensional structure of the glutathione S-transferase, and to show the relationship of the glutathione S-transferase to the S-crystallin of squid ocular lenses, and finally their reasoning for the placement of the glutathione S-transferase into a separate class sigma.
First let's begin with the procurement and development of the testing material. This was accomplished by means of constructing of an expression vector encoding for the enzyme GSH S-transferase and infecting and rearing E. coli cells to harvest the desired product.
The cDna insert of clone pGST5 was amplified by the polymerase chain reaction (PCR). The primers created a NdeI restriction sight immediately 5' upstream to the translation initiation codon ( 5'-ccacacatATGCCTAAGTA TACCCTACACTAT-3') and a NotI site at the 3' untranslated region (5'- ggagaagcggccgCTTGTTCTTGATTTCGGCTAGGA-3') of the mRNA encoding the enzyme. After digestion the amplified cDNA with NdelI and NotI a PCR product with the expected length of about 650 bp was isolated and purified from an agrose gel using the Geneclean methodology, ligated onto the expression vector pET-17b.(1)
This is a description of how they used a PCR technique was used, described by Jones and Howard (1990), that placed restriction sites onto the gene in places that they wished to conserve. This was done by the use of the restriction enzymes NdelI and NotI, which are a class of enzymes that bind to specific codons (that are underlined above)and protect this portion of the genome from digestion. Then they digested the rest of the genome and obtained a product of the size that they expected ( about 650 bp). This product was then attached/ligated to the genome of a virus to create a vector that was used to infect an E. coli bacterial cell. This vector was added to the 650 bp plus the rest of the vector to the genome of the bacterial cell and was able to produce the desired product GSH S-transferase of the squid digestive gland. The bacterial culture was then plated on LB medium supplemented with ampicillin as a selective agent to retard unwanted growth of contaminants. "The induction of the vector was accomplished by adding IPTG (isopropyl beta-d-thiogalactoside) and incubated at thirty degrees Celsius for three hours. " (1). When the induction was attempted at higher temperatures (37 degrees Celsius) the target enzyme was found in inclusion bodies within the cells; this was avoided by the lower temperature. This yielded 20% of the soluble protein content of the cells as the desired glutathione S-Transferase. "The cells where collected and washed once in 50 ml of Tris -HCL (pH of 8.0) containing 2 mM EDTA (ethylenediaminetetraacetatic acid), then they resuspended the cells in 120 ml of 10mM Tris Buffer (pH of 7.8) containing 1mM EDTA, and disrupted with a Branson SonifierII sonicator (This was to assist in the breakdown of the cells). The cell debris was removed by centrifugation (30000g for 30 minutes), and the supernatant was applied directly to a 1.7cm x 6 cm affinity column of S-hexylglutathione coupled to Sepharose 6B previously equilibrated with 10 mM Tris (pH 7.8)."(1). Elution of the enzyme was conducted with a buffer that contained 2.5 mM of S-hexylglutathione (the substrate of the enzyme). The fractions that where obtained where pooled and concentrated in a concentrator fitted with a PM-10 membrane. This is a summary of how the cells were engineered and how the desired enzyme was retrieved.
The next step was to produce a resolvable crystal for X-ray crystallography. The recombinant proteins that were obtained from the above procedure were added to spot wells that contained the following : 25mM of Tris (pH 7.0)containing 1mM of EDTA, 2 mM of 1-(s-glutathionyl-2-4-dinitrobenzene (a product inhibitor that occupies the active site of the enzyme,) and a buffer of ammonium sulfate at 40% saturation (pH 7.0.). These wells were kept at 4 degrees Celsius and were equilibrated against wells that contained 60-70% saturated ammonium sulfate. The equilibration process is done to slowly bring the amount of solvent up (by evaporation and recondensation) to allow the protein to precipitate in a slow, and orderly fashion producing a resolvable crystal. The crystals grew in 5-7 days in the Ji study. The crystals were mounted on thin capillary tubes of a diameter of 0.7-1.0mm.
There were two heavy atom derivatives, containing iodine and mercury, that were produced by the replacement of the product inhibitor 1-(S-glutithionyl)-2,4-dinitrobenzene, with S-(3-iodobenzyl -glutathione, to form one derivative, and a replacement with ethylmercuric phosphate for the other derivative. These two heavy atom derivatives initially were used in multiple isomorphous replacement technique (MIR) to discern the packing of the molecules in the unit cell. ( This line of action was only pursued after failed attempts to solve the structure by molecular replacement using alpha, mu, and pi, structures as probe models.) The phasing power of the iodine derivative was not strong enough to discern the secondary structure. The mercuric derivative was used to generate Patterson Maps that showed the positions of the heavy atoms and their associated protein residues. Patterson Maps are electron density maps that can be used to derive the diffraction angle of the heavy atoms, that is necessary for back calculations in order to plot the positions of the rest of the atoms in the protein. The two mercuric ions were found associated with the C123 and C174 sulfur atoms.
The electron densities were improved by the use of solvent flattening a method that is used to define the molecules boundaries in a sample that contains a lot of solvent. To do this solvent flattening the computer recognizes the solvent and ignores its presence between the molecule and the x-ray detector. The solvent flattened MIR map was used to generate an almost complete initial structural model, ",the complete structure was resolved by five cycles of simulated annealing with X-PLORE( a computer program that matches electron density maps with atom positions)"(1) The electron density is in agreement with the amino acid sequence of 202 residues as determined by gene sequencing. With the exception of the methionine encoded by the initiator codon that was removed by the bacteria during synthesis. )
The secondary structure consists of four beta-strands and eight alpha-helices, arranged into two domains with a six residue linker between them. Domain I consists of a beta-alpha-beta-alpha-beta-beta-alpha motif. Domain II consists of five alpha helices. The two domains are held together by eleven hydrogen bonds and salt bridges.
The biologically active forms of the GSH transferases are either homodimeric or heterodimeric proteins in the cytosol. The interfacing of the squid enzyme is the same in that it forms dimers as well. In this interfacing both hydro phobic and hydrophilic interactions are observed. In the middle of the dimer unit there is a stacking of two arginine residues the R68 in both proteins, this is also seen in other transferases that form dimers, for example in the alpha class it is the R69, in the mu it is the R 77, and in the pi class it is the R68. There is a lock and key mechanism to this dimer formation, where a hydrophobic ligand of one constituent fits into the hydrophobic pocket of the other constituent. Aside from the lock and key and the arginine side chain interactions there are many hydrogen bonds and salt bridges that add to the dimer formation. (see figure b ).
The relationship between the squid GSH transferase and the S-crystallin SL11 is so close, that it is thought the gene encoding for the transferase was copied, and mutated, to form the crystallin found in the lens tissues. It is 42- 44% identical and even has corresponding residues in the active site area of the transferase. The crystallin however doesn't bind to S-hexylglutathione affinity columns. Though it has the same residue counterparts that the transferase has it does not have the same affinity for the substrate S-hexylglutathione as does the transferase. The active binding site of the transferase has the following groups responsible (Y7, W38, K42, Q62, S63, and D96(B))(see figure a ) as where the S-crystallin SL11 has the following counterparts in about the same orientation (Y7, W38, K42, Q63, and S64.) The single exception and probable cause of inactivation is the E97 that corresponds to the D 96(B). The aspartate that reaches over from the adjacent subunit to form an electrostatic contact with the alpha-amino group of the gama glutamyl residue of the S-hexylglutathione. The result of the substitution is a lengthening of the active arm by one carbon length. This is deemed the main difference between the activity of the transferase and the S-crystallin.
The reasoning behind the placement of the glutathione S-transferase into a separate class is do to three factors. The failed attempts to solve its three dimensional structure by molecular replacement using other known transferase structure such as the alpha, mu, and pi. The dimer formation where there are distinct linkages between certain arginine side chains that are constant within the classes alpha ,mu, and pi. These linkages occur at R69 in the alpha class, R77 in the mu class, and at R68 in the pi class. If we followed this format the glutathione S-transferase of squid would fit into the pi class having a arginine side chain interaction at the R68 position. But, the molecular replacement didn't work, the overall three dimensional structures were too different. The final factor is the relationship between the S-crystallin and the glutathione S-transferase both from squid. It is the similarity between the two structures that makes this transferase so unique, that its gene was copied and mutated to produce an enzyme to serve a different function.
Thus far, we have reviewed the construction of the enzyme and its analogs, and how they were used to derive the three dimensional structure, by the analysis of their electron density maps. We have also reviewed the similarities between the S-glutathione transferase and the S-crystallin of the squid digestive gland and ocular lens, respectively. And finally we have reviewed the criteria by which it qualifies to be placed into a separate class sigma. It is my conclusion that through the study of such enzymes that have this dual origin, that we may be able to shed light on the mechanisms by which new enzymes are created, and how evolution may function.
Back to the 1997 Table of Contents