Identification of critical residues of the serotype modifying O-acetyltransferase of Shigella flexneri

Thirteen serotypes of S. flexneri have been recognised, all of which are capable of causing bacillary dysentery or shigellosis. The difference in the various serotypes arises from variations in the nature of their lipopolysaccharide (LPS) O-antigen structure. All serotypes (except serotype 6) share a common O-antigen backbone structure consisting of repeating units of a tetrasaccharide made up of an N-acetylglucosamine residue linked to three rhamnose residues [1]. Glucosyl and/or O-acetyl groups added to the backbone allow diversification of the parent serotype into different serotypes. The factor responsible for adding an acetyl group to the Rhamnose 3 molecule on the O-antigen backbone has been identified to be an integral membrane protein known as O-acetyltransferase or Oac encoded by the temperate bacteriophage Sf6 [2]. Oac is capable of acetylating the O-antigen of serotypes Y, X, 1a and 4a and changing the antigenic signature of the parent S. flexneri strain to generate a new serotype (serotypes 3b, 3a, 1b and 4b, respectively).

We previously reported the solved two-dimensional membrane topology of Oac [3]. Using a gene-fusion approach with dual reporter genes pho A-lac Zα, Oac’s topology with reference to the cytoplasmic membrane was proposed. Oac consists of ten α-helical membrane spanning regions with both the N and C-termini located in the cytoplasm.

Although knowledge about integral membrane acetyltransferases belonging to S. flexneri and other Gram-negative bacteria is mounting, the actual mechanism of serotype conversion mediated by these proteins remains largely unknown. An understanding of the key structural and functional components involved in O- acetylation is thus required to give an insight into the overall mechanism. Several studies on acetyltransferases belonging to different organisms have been carried out to identify residues playing critical roles [4–7]. Critical residues have been identified in acetyltransferases in these studies either by the isolation of spontaneous phenotype-modifying mutants or by targeted mutagenesis of residues.

In this study, Oac amino acids which showed a high level of conservation with those of other homologous acetyltransferases were identified by BLASTp searches and subsequent alignment of the protein sequences having the highest hits with S. flexneri Oac. This information was coupled with Oac’s topology data to obtain a visual representation of the location of the conserved residues in Oac. The identified conserved residues were then subjected to mutagenesis in order to study their effects on Oac. Based on the findings, roles for the different amino acids in context to their location are proposed.

This study, aimed at investigating amino acid residues critical to Oac, forms an important milestone towards understanding the mechanistics of acetyltransferases in general. An alignment of proteins homologous to Oac revealed several conserved amino acids mainly present in the TM helical regions. Accordingly, these residues were targeted for mutagenesis. All residues mutated were then assessed for their effect on Oac’s function and assembly in the cytoplasmic membrane. In order to assign specific roles for all the other critical amino acids identified, the effects of the mutations on protein assembly in the membrane has been used as an initial criteria followed by the effect on function, since naturally under in vivo conditions, protein (synthesis and) assembly pre-determine functional capabilities.

From these studies three non-essential amino acids of Oac have been identified. Amino acids R62, C84 and S114 could be replaced without having any adverse effects on either the function or the assembly of Oac. Among the three amino acids only the location of R62 appears fairly conserved (in four of the seven homologous proteins used for the alignment); however this positional conservation does not appear to offer any importance to Oac.

Arginine residues have been frequently implicated to play a structural role [5, 6] and also to function as active-site residues in acetyltransferases of different organisms [10–12]. We previously reported the identification of critical arginines (R73 and RR 75–76) [3]. In this study the inclusion of the quantitative assessment of protein assembly has provided a better insight into the role played by these residues. It appears that although these residues are in close proximity to one another, they perform different roles in Oac. RR 75–76 appear to disrupt the assembly of the protein suggesting they have a structural role, while the absence of function in R73 despite protein assembly (as seen in [3]) suggests it either causes the assembly of an inefficient form of the protein or it directly relates to the specific absence of catalytic residue.

The replacement of the two basic residues, RK110-111, with alanines increases protein assembly and function. The specific location of this residue pair could be a contributing factor to the very different role these residues play. On the basis of the activities observed in this study, it could be hypothesised that the RK 110–111 motif assists in regulating the kinetics of protein assembly and folding at an optimal rate (and quantity) suited for the overall purpose of Oac in O-acetylation. Similar effects were shown by Loladze et al. [13] with ubiquitin, where the substitution of basic amino acids Lys, Arg and His, with both neutral and acidic amino acids increased overall protein stability.

The definite role of the GR 269–270 and D333 residues is not clear. Although not critical to function, the drastic impact on their assembly indicates that the localized charges of the amino acids may have effects which are too subtle to be picked up by phenotype evaluations. The effect caused by GR 269–270 could be due to the elimination of the positive charge brought by the arginine residue which might be required to maintain the short loop 9 in the cytoplasm. Moreover, glycine residues possess considerable conformational flexibility which might also be necessary to maintain loop structures. The elimination of the glycine residue might have, if not drastically, interfered with the assembly of the protein beyond the amino acid position 269.

There is an abundance of glycine and serine residues, either as a pair or in close proximity to one other in most of the helices of Oac, but the motifs targeted for mutagenesis (SG 52–53, GS 138–139 and SYG 274–276) show a high level of conservation among other homologous acetyltransferases. Luck et al. [4] reported the occurrence of a spontaneous mutation in the lag-1 gene which encodes an O-acetyltransferase in Legionella pneumophilia resulting in the loss of enzyme activity without affecting mRNA transcription levels. An analysis of the DNA sequence revealed that a single nucleotide substitution resulted in the serine residue of the FFWLSG motif being mutated to leucine. This motif corresponds to Oac’s FFxISG motif containing SG 52–53 and in the current study it has been established that SG 52–53 and the other SG motifs affect protein assembly, although levels of transcription have not been investigated. All three of Oac’s SG motifs appear to be important to maintain the structural integrity of Oac. Similar opinions on the importance of glycine and serine residues occurring together and providing folding stability to membrane proteins have been expressed by other research groups [14, 15].

The conserved motifs, FP 78–79 and FPV 282–84, are found in transmembrane segments III and IX, respectively. Alanine replacements of these residues affected Oac function in spite of the fact that the protein assembly appeared to be more than what was seen for the wild type Oac. The effects on assembly could partly be explained by the absence of proline residues. Proline residues are normally not favoured in α-helices because they are thought to produce ‘kinks’ in transmembrane helices, thereby introducing destabilising energy on the helix [16]. The loss of a proline could have thus provided additional stability to the protein thereby increasing its assembly. The effect on function could either suggest that a conformationally unstable protein was assembled or that it truly represented the specific absence of a catalytic residue. It is possible that the conserved proline residue had a far more critical purpose relating to function. A functional role for prolines has been shown in another class of enzymes – the phenylalanine-specific permease of E. coli [17]. The conserved phenylalanine or valine (in FPV) residues could also have an effect on function. With phenylalanine, continuing from the identification of critical Arg9 and Arg64 (by Delomenie et al. [6] in human arylamine N-acetyltransferases NAT1 and NAT2), Goodfellow et al. [7] identified a critical transmembrane phenylalanine in these enzymes. A Phe125 was found to be involved in determining substrate selectivity leading to the hypothesis that like the arginines, the phenylalanine residue might form part of the active site. A similar arrangement of critical residues appears to exist in Oac. Critical R73, RR 75–76 lie in close proximity to FP 78–79 and all may be involved in forming one of the active sites.

Like the phenylalanine and proline motifs described above, the tryptophan-threonine pair (WT141-142) in TM IV appears to be important to Oac’s function. The effect of these residues in acetyltransferases is largely unknown, although tryptophan residues have been found to be critical in membrane proteins belonging to transporter families [18, 19]. In this study, tryptophan was replaced with alanine. Both are neutrally charged residues, but due to the presence of two fused aromatic rings, tryptophan is a considerably larger molecule than alanine. If tryptophan was involved in some aspect of catalysis, like perhaps the interaction with another residue bearing the captured acetyl group, it is possible that the extra spatial volume it provides (compared with alanine) is necessary to make the two residues interact.

The conserved and non-conserved Oac amino acids mutated in this study were assessed for their effect on Oac’s function and assembly. Conserved arginine residues such as R73, RR75-76, serine-glycine motifs (SG 52–53, GS 138–139 and SYG 274–276), phenylalanine-proline motifs (FP 78–79 and FPV 282–84) and motifs WT141-142 and GR269-270 were found to be critical to Oac. Non-conserved RK110-111 and D333 were also found to be critical. While the identification of conserved essential amino acids could suggest a common requirement among other homologous acetyltransferase, the identification of the non-conserved but critical amino acids found uniquely in S. flexneri Oac suggests they play a specialised role. Based on their effects on Oac function and assembly, an insight into the possible roles of these residues has been gained. This could form the basis of future studies to understand the overall mechanism of action of Oac.