Human Cytomegalovirus Induces the Expression of the AMPKa2 Subunit to Drive Glycolytic Activation and Support Productive Viral Infection

Abstract

Human Cytomegalovirus (HCMV) infection modulates cellular metabolism to support viral replication. Calcium/calmodulin-dependent kinase kinase (CaMKK) and AMP-activated protein kinase (AMPK) regulate metabolic activation and have been found to be important for successful HCMV infection. Here, we explored the contributions that specific CaMKK isoforms and AMPK subunit isoforms make toward HCMV infection. Our results indicate that various CaMKK and AMPK isoforms contribute to infection in unique ways. For example, CaMKK1 is important for HCMV infection at a low multiplicity of infection, but is dispensable for AMPK activation at the earliest times of infection, which our data suggest is more reliant on CaMKK2. Our results also indicate that HCMV specifically induces the expression of the non-ubiquitous AMPKa2 catalytic subunit, found to be important for both HCMV-mediated glycolytic activation and high titer infection. Further, we find that AMPK-mediated glycolytic activation is important for infection, as overexpression of GLUT4, the high capacity glucose transporter, partially rescues viral replication in the face of AMPK inhibition. Collectively, our data indicate that HCMV infection selectively induces the expression of specific metabolic regulatory kinases, relying on their activity to support glycolytic activation and productive infection.IMPORTANCE Viruses are obligate parasites that depend on the host cell to provide the energy and molecular building blocks to mass produce infectious viral progeny. The processes that govern viral modulation of cellular resources have emerged as critical for successful infection. Here, we find that HCMV depends on two kinase isoforms to support infection, CaMKK1 and AMPKa2. We find that HCMV specifically induces expression of the AMPKa2 subunit to induce metabolic activation and drive robust viral replication. These results suggest that HCMV has evolved mechanisms to target specific metabolic regulatory kinase subunits to support productive infection, thereby providing insight into how HCMV hijacks cellular metabolism for its replication, and sheds light on potential viral therapeutic vulnerabilities.

FIG 1

CaMKK2 activity is important for early activation of AMPK during HCMV infection. (A) CaMKK induces AMPK activity, leading to successful infection and HCMV-mediated glycolytic activation. The entire process can be halted with the addition of CaMKK inhibitor, STO-609. (B and C) Clonal CaMKK1 and CaMKK2 knockout (KO) cell lines were generated in fibroblast cells via CRISPR genomic editing. (B) Western blot showing CaMKK1 protein expression in parental (par), nontargeting control (NT), CaMKK1 KO, and CaMKK2 KO cells that were mock (M) or HCMV (H) infected. (C) Insertions and deletions introduced are indicated. Nomenclature follows (60). Cells were M or H infected (MOI = 3.0) at 24 (D), 48 (E), and 72 (F) h postinfection (hpi). Expression of P-ACC (Ser79), total ACC, P-AMPKa1/a2 (Thr172), and total AMPKa1/a2 was assessed by Western blotting. GAPDH was used as a loading control. Densitometry analysis was used to quantify the average ratios from independent blots (n = 3 for 24- and 72-hpi blots, n = 2 for 48-hpi blot), which were combined to determine the P-AMPK-to-total AMPK and P-ACC-to-total ACC values shown in panels D to F. BJhT parental cells were used for the generation of all cells used for panels B to F.

FIG 2

CaMKK1 is important for HCMV infection at a low MOI. (A to C) Parental (par), nontargeting control (NT), CaMKK1 knockout (KO), and CaMKK2 KO cells were infected, and viral titer was assessed by TCID₅₀ at indicated times postinfection. Cells were infected at an MOI of 3.0 (A) or an MOI of 0.01 (B), and virus was harvested at 5 days postinfection (dpi). (C) Cells were infected at an MOI of 0.01, and virus was harvested at 5, 10, 15, 20, and 25 dpi. Multiple statistical comparisons were made by one-way ANOVA followed by Student’s t tests in panels A to C. **, P < 0.01; ***, P < 0.001; ns, not significant. All comparisons are made to nontargeting control (average [avg] ± standard error of the mean [SEM], n = 3). BJhT parental cells were used for the generation of KO cells.

FIG 3

HCMV infection induces expression of the AMPKa2 catalytic subunit. (A and B) MRC5hT cells were mock (M) or HCMV (H) infected (MOI = 3.0) for indicated hours postinfection (hpi). (A) At 48 hpi, RNA was harvested for qPCR analysis for the AMPK subunit isoforms. Signals were normalized to GAPDH and to the mock-infected sample (avg ± SEM, n ≥ 3). Statistical comparisons were made using a Student’s t test. *, P < 0.05; **, P < 0.01; ns, not significant. (B) AMPK subunit isoform protein expression in M- or H-infected parental cells at 24, 48, 72, 96, and 120 hpi, representative of two separate infections.

FIG 4

Knockout of AMPK catalytic subunits via CRISPR engineering. (A) Polyclonal AMPKa1 knockout (KO) and AMPKa2 KO MRC5hT fibroblast cell lines were generated. Expression of total AMPKa1/a2 and P-AMPKa1/a2 (Thr172) was assessed by Western blotting and compared to parental (par) cell line expression. (B to D) Clonal AMPKa1 KO and AMPKa2 KO cell lines were generated from AMPKa1-1 and AMPKa2-3 polyclonal cell lines shown in panel A. (B) Clonal AMPKa1 KO and AMPKa2 KO cell lines were generated in fibroblast cells via CRISPR genomic editing followed by monoclonal selection. Insertions and deletions introduced are indicated. Nomenclature follows (60). Expression in KO cells was compared to that in par using antibodies specific to AMPKa1 (C) or AMPKa2 (D). GAPDH was used as a loading control.

FIG 5

AMPK subunit expression and activity in CRISPR knockout cell lines. (A to D) Cells were mock (M) or HCMV (H) infected (MOI = 3.0, n = 1), and 72 h postinfection (hpi), P-AMPKa1/a2 (Thr172), AMPKa1, AMPKa2, P-ACC (Ser79), and total ACC levels were assessed in the AMPKa1 knockout (KO) (A) and AMPKa2 KO (B) clonal cell lines compared to that in the nontargeting control (NT) cell line via immunoblotting. NS, nonspecific band. Densitometry analysis was used to quantify the ratios of P-AMPK to total AMPK and P-ACC to total ACC in panels A and B. Total AMPKa was calculated by the addition of AMPKa1 and AMPKa2 intensities prior to calculating the ratio. NT cells and AMPKa2 clone no. 1 KO (C), AMPKa2 clone no. 2 KO (D), or AMPKa1 KO (E) cells were infected over the course of 96 h, and AMPKa1/a2 phosphorylation was accessed by Western blotting. Expression of pp28 indicates successful infection; GAPDH was used as a loading control. MRC5hT parental cells were used for the generation of KO cells.

FIG 6

AMPKa2 catalytic subunits contribute to viral production. Parental (par), nontargeting control (NT), AMPKa1 knockout (KO), and AMPKa2 KO cells were infected (MOI = 3.0) for 96 h. Viral titer was quantified by plaque assay. AMPK inhibitor, 5 μM compound C (CC), was used a control for plaque formation in the par and NT cells (avg ± SEM, n ≥ 4). Multiple statistical comparisons were made by one-way ANOVA followed by Tukey post hoc test. **, P < 0.01; ns, not significant versus NT. MRC5hT parental cells were used for the generation of KO cells.

FIG 7

AMPK knockout reduces lactate secretion during HCMV infection. Nontargeting control (NT) and AMPKa2 knockout (KO) clonal cell lines were mock (M) or HCMV (H) infected (MOI = 3.0). At 48 h postinfection (hpi), medium was changed for 1 h and then harvested, and lactate secretion was measured by LC-MS/MS; 5 μM compound C (CC) was used as a control. Samples were normalized to 10⁶ cells/h (avg ± SEM, n = 4). Multiple statistical comparisons were made by one-way ANOVA followed by Tukey post hoc test. *, P < 0.05; **, P < 0.01; ns, not significant. Comparisons were between M and H infection samples unless indicated by a line between sample comparisons. MRC5hT parental cells were used for the generation of KO cells.

FIG 8

GLUT4 overexpression cannot rescue viral infection during CaMKK inhibition. (A to C) MRC5hT fibroblast cells overexpressing doxycycline (Dox)-inducible empty vector (EV) control or GLUT4 were induced with Dox (1 μg/ml) for 48 h and then mock (M) or HCMV (H) infected (MOI = 3.0). CaMKK was inactivated with its inhibitor, 30 µM STO-609 (STO) when indicated. (A) Ninety-six hours postinfection (hpi), RNA was harvested for qPCR analysis of GLUT4 RNA levels and normalized to GAPDH (avg ± SEM, n = 3). (B) Conditioned medium was harvested for analysis of glucose consumption. Measurements were normalized to total cell counts (avg ± SEM, n = 3). (C) Virus was harvested 96 hpi and quantified by TCID₅₀ (avg ± SEM, n = 3). Multiple comparisons were made by one-way ANOVA followed by Student’s t test in panels A and C. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

FIG 9

GLUT4 overexpression increases HCMV replication in the face of AMPK inhibition. (A to G) MRC5hT fibroblast cells overexpressing doxycycline (Dox)-inducible empty vector (EV) control or GLUT4 were induced with Dox (1 μg/ml) for 48 h and then mock (M) or HCMV (H) infected (MOI = 3.0). AMPK was inactivated with its inhibitor, 5 μM compound C (CC) when indicated. (A) Ninety-six hours postinfection (hpi), RNA was harvested for qPCR analysis of GLUT4 RNA levels and normalized to GAPDH (avg ± SEM, n = 5). (B) Ninety-six hours postinfection (hpi), protein was harvested and GLUT4 levels assessed by Western blotting. GAPDH was used as a loading control. All samples were normalized to GAPDH and the M-infected EV sample via densitometry analysis; averages from two blots (n = 1, run on two separate occasions). Conditioned medium was harvested for analysis of glucose consumption (C) and lactate secretion (D). Measurements were normalized to total cell counts (avg ± SEM, n = 3). (E) Representative images of cells at 72 hpi, infected with HCMV-GFP. (F) Individual cell GFP intensities were analyzed from images shown in panel E and are presented in a box and whisker plot. (G) Virus was harvested 96 hpi and quantified by TCID₅₀ (avg ± SEM, n = 4). Multiple comparisons were made by one-way ANOVA followed by Student’s t test in panels A and G or Tukey post hoc test in panel F. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.