Sequencing of Kaposi's Sarcoma Herpesvirus (KSHV) genomes from persons of diverse ethnicities and provenances with KSHV-associated diseases demonstrate multiple infections, novel polymorphisms, and low intra-host variance

Abstract

Recently published near full-length KSHV genomes from a Cameroon Kaposi sarcoma case-control study showed strong evidence of viral recombination and mixed infections, but no sequence variations associated with disease. Using the same methodology, an additional 102 KSHV genomes from 76 individuals with KSHV-associated diseases have been sequenced. Diagnoses comprise all KSHV-associated diseases (KAD): Kaposi sarcoma (KS), primary effusion lymphoma (PEL), KSHV-associated large cell lymphoma (KSHV-LCL), a type of multicentric Castleman disease (KSHV-MCD), and KSHV inflammatory cytokine syndrome (KICS). Participants originated from 22 different countries, providing the opportunity to obtain new near full-length sequences of a wide diversity of KSHV genomes. These include near full-length sequence of genomes with KSHV K1 subtypes A, B, C, and F as well as subtype E, for which no full sequence was previously available. High levels of recombination were observed. Fourteen individuals (18%) showed evidence of infection with multiple KSHV variants (from two to four unique genomes). Twenty-six comparisons of sequences, obtained from various sampling sites including PBMC, tissue biopsies, oral fluids, and effusions in the same participants, identified near complete genome conservation between different biological compartments. Polymorphisms were identified in coding and non-coding regions, including indels in the K3 and K15 genes and sequence inversions here reported for the first time. One such polymorphism in KSHV ORF46, specific to the KSHV K1 subtype E2, encoded a mutation in the leucine loop extension of the uracil DNA glycosylase that results in alteration of biochemical functions of this protein. This confirms that KSHV sequence variations can have functional consequences warranting further investigation. This study represents the largest and most diverse analysis of KSHV genome sequences to date among individuals with KAD and provides important new information on global KSHV genomics.

Fig 1. K1 Phylogenetic Analysis and SplitsTree Phylogenetic Analysis of KSHV Whole Genomes.

(A) KSHV subtypes determined by the K1 gene amino acid (aa) sequence performed using IQTree with 1000 bootstrap replicates [44]. The analysis included 120 K1 amino acid sequences (344 aa) from current study individuals including multiple infections that were generated using de novo K1 sub-assembly. 52 published KSHV K1 aa sequences representing all subtypes A-F define the branches as indicated (shaded in grey). Current study samples were KSHV K1 subtypes A, B, C, E, and F. Study sequences are labeled as FNL and bolded in black, those samples with multiple infections are colored in blue, and multiple infections identified in more than one tissue from the same individual are red. (B) Near-full genome splits network for 85 current study KSHV and 32 published genomes. A total of approximately 130,000 nucleotide positions were included in the final data set excluding 5 repetitive regions. Analysis included KSHV genomes assembled from longitudinal PBMC and multiple tissue comparisons. The KSHV genomes of 17 of 18 comparisons determined to not have significant sequence variations, highlighted in blue. FNL002, shown in red, was published in 2020 as a K1 A2 subtype. The subtype of the PBMC genome is consistent with K1 A4 as shown in Fig 1A. Bootstrap values >80 are as indicated. A phi test for recombination performed using SplitsTree indicated strong evidence of viral recombination (P-value < 0.001). Published genomes used for phylogenetic analysis: GK18 (NC_009333.1), BC-1 (U75698.1), Japan/Miyako sequences (LC200587.1-LC200589.1), BCBL-1 (HQ404500.1), JSC-1 (MK143395.1), BrK.219 (KF588566.1), Zambian sequences ZM007-ZM130 (KT271453-KT271468), P044 (MK876733.1), P100 (MK876737.1), P133 (MK876738.1), P030 (MK876732.1), P075 (MK876735.1), P076 (MK876736.1), P072 (MK876734.1), UNC_KICS009 (MK733606.1), ZM004 (KT271453), ZM091 (KT271455.1), ZM095 (KT271456.1), ZM128 (KT271467.1), ZM130 (KT271468.1), ZM121 (KT271464), UG130 (SAMEA103926549), UG145 (SAMEA103926607), UG151 (SAMEA103926625), UG160 (SAMEA103926554), UG128 (SAMEA103926544), UG146 (SAMEA103926610), UG163 (SAMEA103926564), DG1 (JQ619843.1), SPEL (AP017458), and BC-2 (AF133042.1).

Fig 2. Principal Component Analysis (PCA) of sequence variations and KSHV-associated disease.

Analysis included 1989 informative variants across 1788 positions that passed QC generated from a consensus alignment of 45 non-replicative genomes. KSHV K1 subtypes are colorized and associated diseases are indicated by symbols. No KSHV-associated diseases grouped together. Instead, the genomes segregated by subtype with associated diseases interspersed within each quadrant. Individual samples discussed in detail are highlighted by borders for emphasis.

Fig 3

(A) Example of Sanger Sequence Confirmation of KSHV Multiple Infection: Screen shot of reference-guided assembly of reads within the KSHV K1 gene region for FNL0067 which is a multiple infection of an A5 and C3 genome as indicated shaded in peach and blue. K1 gene subtype specific nodes generated by de novo are shown for the C3 and A5 subtypes in both the oral fluid and PBMC material used for NGS. Sanger sequence subtype-specific traces used for confirmation for each subtype are aligned above. The position of the forward primer is boxed upstream of the beginning of the trace data (S3 Table). (B) Examples of De Novo Subassemblies of KSHV Multiple Infections: Screen shots of KSHV K1 gene alignments made in Geneious 2022.0.2 for three study samples with 2, 3, and 4 multiple infections indicated by de novo subassembly of K1 region-specific reads. The subtypes are indicated in black circles and variations indicated are highlighted in color determined by comparison to the NC_009333.1 reference genome, GK18.

Fig 4. KSHV Percent Identification by Gene Region.

(A) Boxplots reflect the distribution of pairwise distances calculated for each coding regions across the KSHV genome using nucleotide variation. The distances are based upon a consensus alignment of 66 KSHV genomes from the current study and the NC_009333.1 reference. The highly variable K1 and K15 genes diverge approximately 15% and 50% respectively. The genes within the central portion of the genome are very conserved at 98.5–99.4% with notable exceptions as shown in panel 3B. (B) Los Alamos Highlighter amino acids plots illustrating the variations in three gene regions, K3, ORF47, and vIRF2. The plots show the predicted amino acids changes referencing NC_009333.1 shown at the top of the plots. Gaps in sequences are indicated by gray bars.

Fig 5. KSHV sequence inversions.

(A) Reference-guided sequence alignments for samples FNL0015_NA and FNL0052_CA illustrating the ORF9-ORF10 sequence inversion involving 80 base pairs. Sequence deletions occur in FNL0052 accompanying the inversion as shown. (B) Reference-guided sequence alignments for samples FNL0015_NA and FNL0021_NA_20050819 depicting the 26bp ORF8-ORF9 sequence inversion. 20X read depth coverage is included for each sample out of mean coverages of greater than 180X for each. Sequence FNL0015_NA is a K1 C3 subtype genome while FNL0052_CA is B1 and FNL0021_NA_20050819 is F2.

Fig 6. AP site binding and catalytic activity of KSHVUNG variant V9.

(A) Primary sequence alignment of KSHVUNG and V9 variants, five aa changes in V9 are noted. (B) Location of aa changes in the regions resolved for the KSHV vUNG structure (23) using UCSF Chimera. (C) Hyperbolic curve for KSHVUNG binding to THF:G-fluorescein-dT dsDNA to plot anisotropy data. (D) summary bar graph to compare dissociate constant (Kd) of KSHVUNG and V9 with Extra sum of squares F-test, p<0.001. (E) Representative denaturing urea PAGE resolving uracil containing oligo incubated with recombinant KSHVUNG or V9 over time. (F) summary bar graph representing specific activity of KSHVUNG and V9 for uracil in ssDNA and G:U mismatch in dsDNA. Unpaired student T-test was used for statistical analysis. NS denotes not significant. Three replicated experiments were performed for Kd and specific activity determination.