Diversity-generating retroelements

Abstract

Parasite adaptation to dynamic host characteristics is a recurrent theme in biology. Diversity-generating retroelements (DGRs) are a newly discovered family of genetic elements that function to diversify DNA sequences and the proteins they encode. The prototype DGR was identified in a temperate bacteriophage, BPP-1, on the basis of its ability to generate variability in a gene that specifies tropism for receptor molecules on host Bordetella species. Tropism switching is a template-dependent, reverse transcriptase mediated process that introduces nucleotide substitutions at defined locations within a target gene. This cassette-based mechanism is theoretically capable of generating trillions of different amino acid sequences in a distal tail fiber protein, providing a vast repertoire of potential ligand-receptor interactions. Variable residues are displayed in the context of a specialized C-type lectin fold, which has evolved a unique solution for balancing protein diversity against structural stability. Homologous DGRs have been identified in the chromosomes of diverse bacterial species. These unique genetic elements have the potential to confer powerful selective advantages to their hosts, and their ability to generate novel binding specificities and dynamic antimicrobial agents suggests numerous applications.

Figure 1

DGR mediated tropism switching by Bordetella bacteriophage. (1) The BPP-1 Mtd tail fiber protein binds to pertactin on the surface of Bvg⁺ phase Bordetella with subsequent injection of phage DNA. (2) Following phage genome replication and DGR function, ~1% of progeny phage contain a variant mtd (colored DNA). Since the frequency of diversification is low, parental genomes are expected to be in the vast majority in phage-producing cells, and variant genomes are likely to be packaged in virions containing the parental Mtd [31]. (3) Irrespective of genotype, the tropism specificity of the parent phage is retained for the second round of infection. (4) In the second round of infection, genomes diversified in the first round are finally packaged into virions with the variant Mtd molecules they encode. Approximately 1 in 1,000,000 of these progeny will express a novel Mtd that recognizes a receptor expressed on the surface of Bvg- phase Bordetella. (5) Ensuing phage infection and replication cycles will continue to generate Mtd variants.

Figure 2

Potential mechanisms of DGR retrotransposition. Four potential mechanisms for cDNA priming and mutagenic homing are shown. (1) Replication Fork. An Okazaki fragment (green arrow) serves as the initiating primer as the TR RNA transcript anneals to the recently synthesized antisense IMH complementary strand [32]. (2) Single Stranded Nick. A nick in the antisense strand occurs at IMH by a single-strand endonuclease that has yet to be identified. The resulting 3′ hydroxyl serves as the initiating primer. (3) Double Stranded Break. A yet to be determined factor creates a double strand break at the IMH. Progressive exonuclease degradation of the sense strand is accompanied by TR RNA transcript priming by the free 3′ hydroxyl on the antisense strand [33,34]. (4) cDNA Recombination. Brt reverse transcription of the TR RNA transcript is initiated by an unknown primer. VR variants are created by RecA independent homologous recombination of TR derived cDNA with the parental VR [35,36]. In all cases, DGR homing is followed by replication that produces mosaic VRs with patches of TR-derived variable sequence [••17].

Figure 3

Structure of Mtd [••19]. (a) Left: BPP-1 with bi-lobed globular structures at the distal ends of each tail fiber; Right: Each globular structure corresponds to a single Mtd trimer with three VR regions present on the bottom face. (b) Left: Mtd monomer; Right: Mtd CLec domain, VR sequences in red. The β5 strand, located at the very C-terminus of Mtd, is encoded by the 21 bp IMH sequence that sets the directionality of information transfer. Positioned in the central core of the trimer, β5 makes close intra- and inter-molecular contacts that would be disrupted by variation and, despite having adenine-encoded amino acids, IMH remains invariant. (c) Ligand-binding surface of an Mtd variant that binds pertactin. Variable side chains (red) are solvent exposed. Although the Mtd variant shown has a preponderance of hydrophobic variable residues, highly hydrophilic binding sites can also be generated. (d) Superimposed VR regions from five Mtd variants that bind different ligands. The main chain conformation of the Clec domain is remarkably invariant despite large differences in the chemical nature of variant side chains.

Figure 4

Conserved and variable features of DGRs (Xu M et al., unpublished data, Gingery M et al., unpublished data) [••17]. (a) The prototype BPP-1 DGR is shown at the top. Orthologous reverse transcriptases (RTase) are located near predicted VRs and TRs. VRs are predominantly located within CLec domains at the C-terminal end of VR-containing ORFs. VR and TR pairs differ almost exclusively at positions correlating to adenines within the TR. Atd homologs are more divergent than the reverse transcriptase genes and they are present in a subset of DGRs. Hrdc loci, predicted to encode proteins with 80 amino acid helicase and RNAseD C-terminal (HRDC) domains, overlap RTase loci in a subset of DGRs [37]. (b) The short sequence represented by Gln151-Pro157 is highly conserved in HIV-1, LINE-1 and other “canonical” RTs, playing a pivotal role in dNTP and template recognition [38]. In particular, Q151 is located in the dNTP binding pocket and it directly interacts with the sugar moieties of incoming dNTPs. Mutations at this site (e.g. Q151M) confer resistance to nucleoside inhibitors such as AZT and they alter both mutational specificity and fidelity [38,39]. Although other shared unique features are evident, it may be significant that DGR-associated RTases share nearly identical polymorphisms (Q151 to I/L181, P157 to Q187) in a region known to play such a major role in controlling RT fidelity.