SARS-CoV-2 phenotypic trajectory and the Omicron Paradox. Discussion of the following pre-print and the complex intertwined pathways that have dictated phenotypic outcomes for SARS-CoV-2. This can be otherwise termed the pandemic to endemic trajectory: biorxiv.org/content/10.64898…
5
41
139
24,752
1- For a long time I have been talking at length about ACE2 & SARS-CoV-2 and how ACE2 dictates complex outcomes in various tissues based on its conformation when engaged as a solute carrier chaperone or in the angiotensin system. Now we see this is 1/2 the story.
1
3
37
1,593
2- So lets recap. What is the Omicron paradox? We define this as Omicron attenuation we observe in vitro as Omicron becomes fitter in transmitting. We observe this in Vero based cells that co-express high levels of ACE2 and TMPRSS2 (see panel C).
1
1
29
1,360
3- The best way to see panel C is the further the curves shift to the right, the more attenuated that particular variant is. What you see here is as Omicrons have become fitter in the real-world, they step to the right and become more attenuated in vitro...thats the paradox.
1
2
28
1,010
4- But as one wise reviewer pointed out, we have become somewhat obsessed about ACE2 conformations at the cell surface and hadnt looked closely enough at the virus itself. So what do we see when we do that?
1
28
963
5- Here we saw something very interesting. So lets take our cells (Veros) with high ACE2 and TMPRSS2. Infect them with representative variants that span the pandemic and then look at the Spike protein on viral particles and in infected cells.
2
29
963
6- So what are we seeing here... Two spike pools with what appears to be two different outcomes. On the virion pool of Spike, it is vanishing. The virus has literally lost its "crown" of Spike. Then looking at infected cells, we see Spike persisting but only in the uncleaved form
1
1
31
870
7. So what is happening here?...to figure this out we did a brief set of passages using Omicron KP.3 (a JN.1 lineage) using Veros and Veros with high ACE2 and TMPRSS2. What happened? Within only a few passages a mutant appeared with only one Spike change. R685P in the FCS.
1
25
847
9-.... You rescue Spike on the viral membrane (albeit uncleaved) on Omicron viral particles and through literally 1 amino acid change in Spike at FCS position R685P, you increase virion infectivity by several orders of magnitude. But this sounds odd. Why would the virus do this?
Sep 25, 2026 · 10:27 AM UTC
3
1
24
782
10- Lets look at a recent paper by the Menachery lab with Ancestral and the Omicron specific change N679K: pubmed.ncbi.nlm.nih.gov/4271…
They see three things: 1. Better FCS cleavage. 2. Spike depletion on virions when grown in Veros and lung derived Calu3s. 3. .....
1
23
743
11. ... The characteristic tropic shift of infection from the lower respiratory tract to upper respiratory tract that we see in Omicron lineages.... But Spike is being depleted on viral particles... this seems odd....
1
21
732
12. So lets look at spike on virions in various settings. Nasal epithelia, lung derived Calu3, Veros & a cell line identified by the Swartz lab as a really good cell line to grow Omicron lineages (IGROV-1 Ovarian cell line).
1
21
726
13. What do we see... (note A.2.2 = Ancestral)... Omicron KP.3 is obliterated only in Vero Calu3 lines but in cells where it thrives Spike is left alone. As with the Menachery lab we see across all cells that Spike is cleaved far better than Ancestral...BUT....
1
21
680
14. In certain cells (Veros and Calu3s).... it appears cleavage of the FCS has crossed a line and likely is too much. This we define as Spike hyper-cleavage, where the FCS is cleaved so well it depletes from Virions. Why doesn't it do this in other cells?
1
24
675
15. Key here is likely the very different glycosylation pathways of Spike across different cells types. Cells of the upper respiratory tract are characterized by abundant and heavily sialylated mucin-type O-glycosylation, which may shield S and lower the activity towards the FCS.
2
24
677
16. Here the FCS has reached a "goldilocks" point in one tissue but crossed a line in another. So thats it right... where done...not quite... thats the first rabbit hole and regulation of Spike on Viral particles. Now lets look at the pool on the infected cell membrane.
1
1
22
637
17. Whilst we often see viral life cycles as cell free entities that come and go, there is another layer that people often don't apprepriate. The viral dissemination power of an infected cell. The path we often refer to as cell-cell viral spread.
1
2
24
637
18. Here is HIV engaged in cell-cell spread courtesy of work from the Chen Lab: youtu.be/5TC35-ssd_0 science.org/doi/10.1126/scie…
This mechanism can drive viral replication and also better navigate neutralising antibodies.
2
1
21
927
19. So lets go back to what we originally observed in our cells with high ACE2 and TMPRSS2. Maybe this is influencing the other pool of Spike on the infected cell surface?
1
18
587
21. Here we started to look at known ACE2 domains that TMPRSS2 interacted with. The ACE2 Collectrin like domain (CLD). When we mutated the CLD we rescued Omicron infection in vitro and in vivo.
2
18
577
22. Curiously the cells we engineered were Vero based. Some how by altering the ACE2 CLD, we could rescue infection on Omicron's but in an environment where the Spike on virions was being depleted....basically here we were seeing viral replication sustained by cell-cell spread.
2
18
562
23. But ACE2 mutants like this dont exist in the real world. What else could play with ACE2 conformation and around the ACE2-CLD? In pink we have a solute carrier (SLC6A19/SLC6A20) that forms a heterodimer of dimers with ACE2. Does this structure help Omicrons?
1
1
17
574
24. Turns out if you co-express solute carriers SLC6A19 or SLC6A20 with ACE2 and TMPRSS2, you see Omicrons replicate even better than ACE2 CLD mutants + TMPRSS2. And once again, its primarily during cell-cell spread as virion Spike is depleted.
2
1
22
539
25. So lets look at Spike specifically on cells where Omicrons are rescued and happily spreading between cells. Here we see Spike being cleaved at the FCS in cells that express Solute carriers+ACE2+TMPRSS2....mmm whats going on here?
3
19
543
26. To cut a long story short... TMPRSS2 is cleaving the FCS of Spike at the infected cell membrane. When ACE2 is co-expressed with TMPRSS2 it blocks this cleavage event. Solute carriers work by quarantining ACE2 from TMPRSS2 & in doing so free TMPRSS2 to cut the FCS.
2
2
20
526
27. It impacts Omicrons far greater than pre-Omicrons, as the Spike has evolved primary to use TMPRSS2 at the infected cell membrane to target the FCS. How would this help? TMPRSS2 remember works on the S2 part of Spike but only when the FCS is cleaved by other proteases.
1
1
20
525
28. If the virus uses the same protease (i.e. TMPRSS2), it gates itself by FCS cleavage to rapidly reach the second cleavage target S2. We see this in vitro as sustaining very rapid cell-cell spread with infections peaking at 24 to 48 hours.
1
2
19
539
29. But where is the evidence this is happening in vivo? This would proceed primarily in the upper respiratory tract and help transmission in theory. But what is the evidence that a solute carrier known to complex with ACE2 does this in the real world?
1
1
17
549
30. Enter one of two large GWAS studies. What do we see? The second hit is SLC6A20 and when you resolve the data a little more you see it increases transmission: nature.com/articles/s41588-0…
2
1
22
952
31. So if we think about the Spike on the infected cell membrane, this then makes sense and finally mechanistically links a pathway now resolved to increase infection in vitro (through increased cell-cell spread) to transmission in vivo.
2
1
13
320
32. But what is also interesting about SARS GWAS studies is the more recent data highlighting a very high Omicron specific signal.... ST6GAL1 aka SIAT1. What does that do? It decorates cell-surface glycoproteins with sialic acid in an α-2,6 linkage to galactose units....
1
14
327
33. Whilst this acts primarily on N-linked glycans, it may provide the key difference from the upper and lower respiratory tract that may have led to outcomes we see in Spikes that embed in viral membranes and suseptible or resistance to FCS hyper cleavage.
1
14
303
34. We need to also recognise that SIAT1 is a key factor for influenza infection and its role in promoting infection (albeit via another mechanism) for that respiratory virus has been known for some time.
1
1
17
311
35. If I have lost you along this thread, the take home message is the phenotypic shift of SARS-CoV-2 is complex and is the culmination of two Spike pools being differentially regulated that end up on viral membranes vs infected cell membranes. Both pathways are tissue specific
1
5
21
545
36. In one tissue setting, the virus has optimised both pathways to its advantage (e.g. Upper respiratory tract). In other settings (e.g. lung) it represents additive attenuation through Spike loss and the inability of TMPRSS2 to activate Spike during cell-cell spread.
1
2
24
517
