SARS-CoV-2 peptide fragments selectively dysregulate specific immune cell populations via Gaussian curvature targeting
Immune cell populations are dysregulated in COVID-19 for currently unknown reasons: Plasmacytoid dendritic cell (pDC) populations are reduced, thus hampering antiviral responses. CD8+ T cell populations are reduced, the level of which has emerged as an index of disease severity. Recent work has shown that the proteome of SARS-CoV-2 is a rich reservoir of antimicrobial peptide-like sequence motifs (xenoAMPs) which can chaperone and organize dsRNA for amplified Toll-Like Receptor 3 (TLR3)-mediated inflammation in vitro and in vivo.
Here, we demonstrate that proteolytic digestion of the SARS-CoV-2 spike protein by host trypsin-like serine proteases directly produces xenoAMPs. Synchrotron Small Angle X-ray Scattering, mass spectrometry, and a theoretical analysis based on continuum membrane elasticity show that proteolytically generated xenoAMPs from SARS-CoV-2 proteins in vitro and machine learning-predicted high-scoring xenoAMPs all induce negative Gaussian curvature (NGC) necessary for pore formation in membranes.
We find that xenoAMPs alone as well as xenoAMPs synergistically with endogenous AMP LL-37 can induce NGC in membranes. A computational analysis of immune cells with morphologically complex shapes (e.g., pDC, CD8+, and CD4+ T cells) suggests that surfaces with high local NGC can concentrate AMP-like sequences and promote selective membrane disruption.
Consistent with this hypothesis, experiments with freshly isolated human peripheral blood mononuclear cells confirm that viable pDCs, DCs, and T cells are significantly depleted after xenoAMP exposure, in contrast to monocytes and neutrophils, the immune cell subsets with spheroidal morphology. Structural data from Omicron variant xenoAMP homologs indicate reduced pore formation, consistent with clinical observations of reduced T cell cytopenia in Omicron variant infections.
SIGNIFICANCE
Previous work has demonstrated that the proteome of SARS-CoV-2 can potentially be a rich source of AMP-like viral fragments, exemplars of which are associated with severe COVID-like inflammation in vitro and in vivo. Here, we demonstrate that direct proteolytic processing of SARS-CoV-2 proteins can yield xenoAMPs, and that the full heterogeneous ensemble of resultant fragments can collectively exert AMP-like pore-forming activity. We describe an unanticipated general mechanism of host cellular targeting for viral AMP-like pore forming peptides, based on local Gaussian curvatures of the host cell membrane, and show that this mechanism can selectively target and deplete specific immune cell types in a manner consistent with clinical observations for severe COVID-19 patients.
Web | DOI | Proceedings of the National Academy of Sciences | Open Access
Zhang, Yue; Silvestre-Roig, Carlos; Fu, Han; Alimohamadi, Haleh; Mandal, Taraknath; Chen, Jonathan W; Luo, Elizabeth Wei-Chia; Anda, Jaime de; Linard Matos, Anna Lívia; Richter, Mathis; Mennella, Anna; Lee, HongKyu; Chan, Liana C; Wang, Yingrui; Wang, Naixin; Wang, Hongyu; Wang, Xiaohan; Lee, Calvin K; Ghosh, Susmita; Matsui, Tsutomu; Weiss, Thomas M; Guo, Tiannan; Zhang, Maomao; Li, Dapeng; Wolfgang, Matthew C; Hagan, Robert S; Li, Melody M H; Gunzer, Matthias; Sickmann, Albert; Frasca, Loredana; Yeaman, Michael R; Lande, Roberto; Cui, Qiang; Soehnlein, Oliver; Wong, Gerard C L
Immune cell populations are dysregulated in COVID-19 for currently unknown reasons: Plasmacytoid dendritic cell (pDC) populations are reduced, thus hampering antiviral responses. CD8+ T cell populations are reduced, the level of which has emerged as an index of disease severity. Recent work has shown that the proteome of SARS-CoV-2 is a rich reservoir of antimicrobial peptide-like sequence motifs (xenoAMPs) which can chaperone and organize dsRNA for amplified Toll-Like Receptor 3 (TLR3)-mediated inflammation in vitro and in vivo.
Here, we demonstrate that proteolytic digestion of the SARS-CoV-2 spike protein by host trypsin-like serine proteases directly produces xenoAMPs. Synchrotron Small Angle X-ray Scattering, mass spectrometry, and a theoretical analysis based on continuum membrane elasticity show that proteolytically generated xenoAMPs from SARS-CoV-2 proteins in vitro and machine learning-predicted high-scoring xenoAMPs all induce negative Gaussian curvature (NGC) necessary for pore formation in membranes.
We find that xenoAMPs alone as well as xenoAMPs synergistically with endogenous AMP LL-37 can induce NGC in membranes. A computational analysis of immune cells with morphologically complex shapes (e.g., pDC, CD8+, and CD4+ T cells) suggests that surfaces with high local NGC can concentrate AMP-like sequences and promote selective membrane disruption.
Consistent with this hypothesis, experiments with freshly isolated human peripheral blood mononuclear cells confirm that viable pDCs, DCs, and T cells are significantly depleted after xenoAMP exposure, in contrast to monocytes and neutrophils, the immune cell subsets with spheroidal morphology. Structural data from Omicron variant xenoAMP homologs indicate reduced pore formation, consistent with clinical observations of reduced T cell cytopenia in Omicron variant infections.
SIGNIFICANCE
Previous work has demonstrated that the proteome of SARS-CoV-2 can potentially be a rich source of AMP-like viral fragments, exemplars of which are associated with severe COVID-like inflammation in vitro and in vivo. Here, we demonstrate that direct proteolytic processing of SARS-CoV-2 proteins can yield xenoAMPs, and that the full heterogeneous ensemble of resultant fragments can collectively exert AMP-like pore-forming activity. We describe an unanticipated general mechanism of host cellular targeting for viral AMP-like pore forming peptides, based on local Gaussian curvatures of the host cell membrane, and show that this mechanism can selectively target and deplete specific immune cell types in a manner consistent with clinical observations for severe COVID-19 patients.
Web | DOI | Proceedings of the National Academy of Sciences | Open Access