Genome-Wide Sequence Analysis of Kaposi Sarcoma-Associated Herpesvirus Shows Diversification Driven by Recombination

Abstract

Background: Kaposi sarcoma-associated herpesvirus (KSHV) establishes lifelong infection in the human host and has been associated with a variety of malignancies. KSHV displays striking geographic variation in prevalence, which is highest in sub-Saharan Africa. The current KSHV genome sequences available are all tumor cell line-derived or primary tumor-associated viruses, which have provided valuable insights into KSHV genetic diversity.

Methods: Here, we sequenced 45 KSHV genomes from a Ugandan population cohort in which KSHV is endemic; these are the only genome sequences obtained from nondiseased individuals and of KSHV DNA isolated from saliva.

Results: Population structure analysis, along with the 25 published genome sequences from other parts of the world, showed whole-genome variation, separating sequences and variation within the central genome contributing to clustering of genomes by geography. We reveal new evidence for the presence of intragenic recombination and multiple recombination events contributing to the divergence of genomes into at least 5 distinct types.

Discussion: This study shows that large-scale genome-wide sequencing from clinical and epidemiological samples is necessary to capture the full extent of genetic diversity of KSHV, including recombination, and provides evidence to suggest a revision of KSHV genotype nomenclature.

Figure 1.

Whole-genome variation across 70 genomes. Line graph plotted across the genome showing the proportion of variant bases in a 1000-nucleotide (nt) sliding window where at least 1 Kaposi sarcoma-associated herpesvirus genome sequence has a single nucleotide polymorphism relative to the GK18 reference sequence. Grey bars indicate masked-out repeat regions.

Figure 2.

Principal component analysis (PCA) separates strains by type and geographic origin. PCA of all Kaposi sarcoma-associated herpesvirus (KSHV) strains (45 new strains and 25 published strains) based on single-nucleotide polymorphisms relative to the GK18 reference sequence in a full-genome multiple-sequence alignment. A, Principal component 1 (PC 1) separates all strains based on type P (blue dotted circle) and type M (green dotted circle). B, PCA of central genome minus variable K1 and K15 genes shows some geographic clustering of KSHV strains, with separation of the African strains (black and red) compared to Western (blue) and Japanese (green).

Figure 3.

Neighbor-net phylogenetic network based on complete genomes of all 70 Kaposi sarcoma-associated herpesvirus samples. Parallel edges depict conflicting phylogenetic signals. The samples clustered into 5 distinct types (dotted circles) designated types P1, P2, M1, M2, and N, and 1 outlier ZM004, which did not cluster well with any type and thus may be a putative intertype recombinant or a novel type. The tips are labeled to represent sample geographic origin: black (Uganda), red, (Zambia), blue (Western), and green (Japan).

Figure 4.

Evidence for recombination between different Kaposi sarcoma-associated herpesvirus types. A, Intertypic recombinant type M2. Bootscan plot comparing type M2 query with consensus sequence of types P2 (red), P1(green), M1(yellow), N (purple), and ZM004 (orange). B, Intratypic recombinant ZM091. Bootscan plot comparing ZM091 with consensus sequences of types P2 (red), P1(green), and ZM004 (orange). Shifting phylogenetic signals and fragmented genomes are indicative of the presence of recombination.