Widespread ncaas Imprints in the Serum Proteome of COVID-19 Convalescents Uncovering Immune System Sequelae, 2026, Liu et al.

[SNT Gatchaman]

Widespread ncaas Imprints in the Serum Proteome of COVID-19 Convalescents Uncovering Immune System Sequelae

Spoiler: Authors

Kun Liu; Zhigang Ren; Bowen Dong; Wenli Liu; Yuyuan Gao; Li Zhang; Jingyi Li; Zhao Sun; Hongyi Li; Qian Zhao; Xinchao Hu; Jinfeng Chen; Yuanyuan Wang; Yang Yang; Lei Zhang; Xinli Xue; Aiguo Xu; Zujiang Yu; Jing-Hua Yang

Post-COVID-19 sequelae has become an emerging global health issue, but the mechanisms for the sustained susceptibility of convalescents to the sequelae remain poorly understood.

Here we report the use of a restricted open-search approach to explore the molecular imprints of SARS-CoV-2 infection left on the proteome of 412 COVID-19 patients and convalescences. A total of 827 non-standard amino acid variations, chemically modified residues as well as post-translational modifications, termed non-coded amino acids (ncAAs), are found spreading over 29,814 sites in patients serum proteins.

Markedly, widespread ncAAs are induced and sustainedly imprinted on the serum proteome predominately perturbing the immunoglobulin-mediated immune response, complement activation and coagulation regulation even 12 months after recovery.

Sustained amino acid variations and chemical modifications are found in the CDRs of the variation region of immunoglobulin contributing to the interactions between the emerging antibody and antigens; durable chemical amino acid modifications found in the HMRs of the constant region of immunoglobulin important for the interaction with the complement and regulatory receptors. In the complement system, inducible ncAAs are memorized in the components essential for the complement activation, amplification cascades and membrane attack processes.

Thus, the workflow described in this study can be used to identify the molecular imprints of viral infection at the proteomic scale, particularly the specific antibodies and the immune targets left in COVID-19 patients and convalescents.

HIGHLIGHTS
• Multiple non-coded amino acids (ncAAs) are found spreading on COVID_19 patients serum proteome

• Sustained amino acid variations and chemical modifications are found in immunoglobulins, complement and coagulation factors even 12 months after COVID_19 recovery

Web | DOI | PDF | Molecular & Cellular Proteomics | Open Access

 

[SNT Gatchaman]

I think this is worth read-through. Chinese team looking at the recovery (or not) through the lens of protein post-translational modifications. I think the acute patients were hospitalised but not ICU — I couldn't see this in supplementary data. They highlight immunoglobulins and complement cascade factors in particular, and coagulation to a lesser extent, with some high AUCs.

I don't want to get side-tracked but will note that the dates given for the enrolment might be wrong.

Spoiler: Enrolment Feb 2019 to March 2020

This is a prepress, not the final manuscript. They said —

A total of 412 samples between February, 2019 and March, 2020 were collected in this research. Among them, 73 acute COVID-19 patients, 21 convalescents in 1 month, 58 convalescents in 3 months, and 96 convalescents in 12 months were from Xinyang, Henan; 164 health people were from the First Affiliated Hospital of Zhengzhou University.

(Unless Covid-19 started a lot earlier than we knew!!)

The clinical acute phase patients of COVID-19 underwent antiviral therapy with nirmatrelvir/ritonavir, anti-inflammatory treatment with glucocorticoids/tocilizumab, or anticoagulation/antiplatelet therapy.

Those treatments weren't standard (or potentially available) in March 2020. Eg Paxlovid wasn't available until late 2021. Wikipedia says —

The research that led to nirmatrelvir began in March 2020, when Pfizer formally launched a project at its Cambridge, Massachusetts site to develop antiviral drugs for treating COVID-19. In July 2020, Pfizer chemists were able to synthesize nirmatrelvir for the first time. In September 2020, Pfizer completed a pharmacokinetic study in rats which suggested that nirmatrelvir could be administered orally.

In February 2021, Pfizer launched the company's first phase I trial of PF-07321332 (nirmatrelvir) at its clinical research unit in New Haven, Connecticut.

It's possible the early-late 2019 timeframe was for a cohort of prepandemic healthy controls only, but maybe this should have read February 2020 to March 2021 or even 2022?

Anyway on to the findings.

 

[SNT Gatchaman]

Consistently, the dysregulated and sustained ncAAs in COVID-19 convalescents were functionally correlated with the pathways relevant to immune responses, complement activation, and coagulation, suggesting the biological impacts of the molecular imprints.

The commonly induced modifications by coronaviral infection include deamination, oxidation, hydroxylation, dehydrogenation and sulfation. It is worthy to emphasize that deamidations of asparagine (N) and glutamine (Q) are among the top chemical modifications in the host cells, resulting in the addition of negative charges at the original positions (Fig. 2). Deamidation is a spontaneous and irreversible posttranslational modification that are shown to affect protein function and stability.

In our study, deamidations of asparagine are induced in immunoglobulins and complement components in the acute patients, which are not reversed in the convalescents even 12 months after COVID-19 recovery. Therefore. deamination is a hallmark event of coronaviral infection and a persistent molecular imprint in COVID-19 convalescents.

Focusing on some summary quotes re: immunoglobulins —

Notably, a great portion of AA-substitutions were found within immunoglobulins, particularly for tyrosine and serine. For immunoglobulin, three hyper ncAA-modified regions (HMRs) in the variable regions of λ, κ and heavy chains were identified that matched the complementarity-determining regions known as CDR1, CDR2, and CDR3. Notably, the most ncAAs within HMRs (>75%) were AA substitutions in agreement with antibody somatic (V(D)J) recombination.

… suggesting that, on top of somatic (V(D)J) recombination, ncAA-modifications provided an additional level of regulation for antigen-antibody interaction in immunoglobulin responses.

Within the constant heavy (CH) chain, however, the hotspots HMR1-8 were clustered differently in the IgM, IgG, and IgA chains, except IgD and IgE. HMR1-2 was found in the CH1 region, HMR4-5 in the CH2 region, HMR7 in the CH3 region, and HMR3/6 near the transitional regions between CH1, CH2 and CH3. Notably, IgG and IgA were extensively modified across HMR1-7, suggesting the importance of ncAA-modifications in regulating immunoglobulin-directed immune activation.

Notably, these COVID-19-associated ncAAs were identified in the hotspots (HMR3-5) important for the complement and the Fc receptor binding sites, suggesting potential effects on the downstream of immune responses associated with COVID-19.

Thus, COVID-19-associated ncAAs as molecular imprints were reasoned to affect antibody-antigen recognition, complement activation and Fc receptor interaction in the acute phase and COVID-19 convalescents.

on top of somatic mutations, post-translational protein modifications are important factors to regulate the specific recognition between the newly emerged antigens and antibodies. […] Markedly, more ncAAs are found in the constant regions, particularly near the motifs important for immune receptor recognitions, including the complement-binding sites of IgG and the pentamer-forming sites of IgM.

As many of the ncAAs within FcRn are induced in the acute phase and sustained in the convalescents even 12 months after recovery, persistent complement activation and immune responses are reasoned to last for a long period of time. […] Therefore, the omics of ncAAs represents a characteristic pattern of molecular imprints that are responsive for the persistent complement activation and immune dysregulation in COVID-19 convalescents.

 

[Murph]

To what extent were we previously aware that viral infection could cause changes to amino acids? I feel surprised to learn this.

• I threw the term "non coding amino acid" into PubMed, there's relatively few hits.
• Seaching "non-canonical amino acids" shows far more hits, but they are mostly seen as a protein engineering tool rather than a natural phenomenon.
• Searching for "deamidation of asparagine" shows thousands of hits, that seems to be a well-described phenomenon.

I remain a bit unsure whether this paper describes something new.

 

[SNT Gatchaman]

I think the term they use is "non-coded amino acid" rather than "non-coding" ie the genes would have specified one AA but a post-translational modification or other process leads to something other than gene-specified. (In contradistinction they do use the term "coding amino acids" but I think they intend "coded amino acids"?)

(PubMed search for non-coded amino acids)

The abstract introduces as "chemically modified residues as well as post-translational modifications, termed non-coded amino acids (ncAAs)" so I think it's a superset that includes (active/deliberate) post-translational modifications. Maybe alternate processes would include passive phenomena similar to glycation in the context of hyperglycaemia, forming AGEs, although focused on AA changes.

In the first part of results they say —

The delta masses were clustered with Gaussian regression, filtered with the known modification databases, and used as a pre-selected delta mass list for a restricted search to generate the final list of non-coded amino acids and derivatives, termed non-coded amino acids (ncAAs). Following this workflow, 827 different ncAAs were identified from the patient’s serum, theoretically representing structural variations of proteins by amino acid substitutions, post-translational modifications (PTMs), and unknown amino acid derivatives. These ncAAs were spreading over 29,814 protein sites, suggesting widespread amino acid polymorphisms or modified proteins on top of the proteomic expression. Accounting for the incidents of ncAAs, 207,345 (7.6%) matched amino acid substitutions, 365,383 (13.4%) were known post-translational modifications (PTMs), and 1,946,090 (71.3%) were annotated as chemical derivatives, demonstrating widespread protein sequence and structural variations.

Perhaps @MelbME might have capacity to offer a comment on whether this could be of relevance.

I remain a bit unsure whether this paper describes something new.

If it's valid I think the new bit would be the sheer scale of proteome disturbance ("widespread"). This seems to be being described as ongoing (at least out to 12 months) in post Covid convalescents (without ongoing symptoms?). That suggests it's happening in nearly everyone infected, so something else would have to explain the ME/LC symptomatology if this were relevant to the story, but that might be where genetic risk signals could come in.

Tagging @mariovitali as you expressed interest in post-translational modifications in the recent PTM-tagged preprint thread Post-translational modifications within fibrinaloid microclot complexes distinguish Pre-COVID-19 Postural Orthostatic Tachycardia Syndrome, Long COVID, and Long COVID-POTS and reveal disease-specific molecular pathways (2025)

Just to add a short comment, PTMs were indeed a signal in my analyses and I believe there are plausible hypotheses as to why PTMs may be disrupted.

 

[SNT Gatchaman]

I don't think we had posted a thread at the time but there is a German study from 2023 —

Amino acids, post-translational modifications, nitric oxide, and oxidative stress in serum and urine of long COVID and ex COVID human subjects (2023)

Spoiler: Abstract

In this study, we investigated the status of amino acids, their post-translational modifications (PTM), major nitric oxide (NO) metabolites and of malondialdehyde (MDA) as a biomarker of oxidative stress in serum and urine samples of long COVID (LoCo, n = 124) and ex COVID (ExCo, n = 24) human subjects collected in 2022.

Amino acids and metabolites were measured by gas chromatography–mass spectrometry (GC–MS) methods using stable-isotope labelled analogs as internal standards. There were no differences with respect to circulating and excretory arginine and asymmetric dimethylarginine (ADMA). LoCo participants excreted higher amounts of guanidino acetate than ExCo participants (17.8 ± 10.4 µM/mM vs. 12.6 ± 8.86 µM/mM, P = 0.005).

By contrast, LoCo participants excreted lower amounts of the advanced glycation end-product (AGE) NG-carboxyethylarginine (CEA) than ExCo participants did (0.675 ± 0.781 µM/mM vs. 1.16 ± 2.04 µM/mM, P = 0.0326). The serum concentrations of MDA did not differ between the groups, indicating no elevated oxidative stress in LoCo or ExCo. The serum concentration of nitrite was lower in LoCo compared to ExCo (1.96 ± 0.92 µM vs. 2.56 ± 1.08 µM; AUC, 0.718), suggesting altered NO synthesis in the endothelium. The serum concentration of nitrite correlated inversely with the symptom anxiety (r = − 0.293, P = 0.0003). The creatinine-corrected urinary excretion of Lys and its metabolite L-5-hydroxy-Lys correlated positively with COVID toes (r = 0.306, P = 0.00027) and sore throat (r = 0.302, P = 0.0003).

Our results suggest that amino acid metabolism, PTM and oxidative stress are not severely affected in long COVID. LoCo participants may have a lower circulating NO reservoir than ExCo.

Web | PDF | Amino Acids | Open Access

They concluded post-translational modifications were not severely affected in LC, but they were comparing to post-infection convalescent HCs and had no uninfected HCs.

A potential limitation of our study might be the absence of a healthy control group that was not affected at all by COVID-19 and our sampling method (convenience sampling). However, the primary aim of our study was to study potential differences between LoCo and ExCo subjects.

So if there are widespread PTM effects as suggested by this thread's paper, they may have missed that signal.

 

[mariovitali]

I haven't heard of non-coding amino acids before so I enetered this term to the Information retrieval system I use to find out more. Here are the results :


So the most hits are found in research about aminoacid conjugation then we have asparagine. Recall that asparagine was found to be disrupted in the paper by Li et al and is a key part of N-Linked glycosylation :

https://pubmed.ncbi.nlm.nih.gov/40649860/

Regarding aminoacid conjugation, from AI

Amino acid conjugation is highly important in both health and disease, acting as a crucial Phase II metabolic pathway for detoxifying drugs, environmental chemicals, and bile acids by increasing their solubility for excretion. This process involves attaching amino acids like glycine or taurine to hydrophobic compounds, which also serves to enhance the stability, permeability, and target specificity of pharmaceuticals.