Scaling relation between genome length and particle size of viruses provides insights into viral life history

Abstract

In terms of genome and particle sizes, viruses exhibit great diversity. With the discovery of several nucleocytoplasmic large DNA viruses (NCLDVs) and jumbo phages, the relationship between particle and genome sizes has emerged as an important criterion for understanding virus evolution. We use allometric scaling of capsid volume with the genome length of different groups of viruses to shed light on its relationship with virus life history. The allometric exponents for icosahedral dsDNA bacteriophages and NCDLVs were found to be 1 and 2, respectively, indicating that with increasing capsid size DNA packaging density remains the same in bacteriophages but decreases for NCLDVs. We argue that the exponents are largely shaped by their entry mechanism and capsid mechanical stability. We further show that these allometric size parameters are also intricately linked to the relative energy costs of translation and replication in viruses and can have further implications on viral life history.

Keywords: Biocomputational method; Genomics; Virology.

Graphical abstract

Figure 1

The log-log plot of outer capsid volume as a function of genome length for viruses infecting different hosts Size of the data points indicate the number of genes while shapes are in accordance with the genetic material (filled circle-dsDNA viruses, empty circle-RNA viruses, filled triangle-ssDNA viruses). A power law $y=a{x}^m$ is used as a fitting expression for the entire data and appears as a straight line on the log-log plot. A linear regression fit of the form $Y=mX+A$ to the data, where $Y=\log y$, $X=\log x$, and $A=\log a$, gives $A=3.64$ and $m=1.13$ (p value < 2.2 × 10⁻¹⁶ and R² = 0.67). All logs are to the base 10. Formulas to calculate capsid volume are described in STAR methods and data are available in the Table S1. See also Figure S1 and S2 and Table S2.

Figure 2

The log-log plot of inner capsid volume versus genome length for icosahedral dsDNA bacteriophages (including jumbo phages) and NCLDVs (algal and protozoan) Inner radius estimated by subtracting capsid thickness, 3 nm for bacteriophages and 10 nm for NCLDVs, from the outer radius. A power law $y=a{x}^m$ appears as a straight line on the log-log plot and is used to independently fit the data for bacteriophages and NCLDVs. For that linear regression fit of the form $Y=mX+A$ is performed over the data, where $Y=\log y$, $X=\log x$, and $A=\log a$. Parameters for phage are $A=3.49,m=0.95$ (p value < 2.2 × 10⁻¹⁶ and R² = 0.70) and for NCLDVs are $A=1.38,m=2.00$ (p value = 4.87 × 10⁻⁹.and R² = 0.73). All logs are to the base 10. Data and formulas are presented in the Table S1.

Figure 3

The log-log plot of translation and replication energy cost of icosahedral dsDNA bacteriophages and NCLDVs as a function of their inner radius The replication cost is obtained using genome length (Equation 1) and capsid radius (Equation 2). The translation cost for capsid protein molecules is obtained from Equation 3. The gap in the energy costs of NCLDVs and bacteriophages is because of the difference in the thickness t of their viral capsids (Equation 3). Energy costs are reported in terms of the number of ATP hydrolysis events (Mahmoudabadi et al., 2017). As discussed in Mahmoudabadi et al. (2017) translation ($\sim r_c^2$) and replication ($\sim r_c^3$ ) rates dominate at lower and higher capsid sizes, respectively. This trend works well for bacteriophages (data points for replication cost using actual length). For NCLDVs, however, the translation cost always dominates because $L_g\sim r_c^{3/2}$. See text for more detail. All logs are to the base 10. The entire data are represented in the Table S1.