Graduate BCHM615 ⭐ 3 Credits Semester 3

Glycoproteins and Post-Translational Modifications

A comprehensive course on the biochemistry of protein modifications occurring after translation: glycosylation (N- and O-linked), phosphorylation, ubiquitination, acetylation, methylation, lipidation, and disulfide bond formation. Emphasis on the roles of PTMs in protein function, cell signaling, and therapeutic protein properties including stability, immunogenicity, and efficacy.

📅 Studied by Skipper on 2026-10-02

📖 Core Concepts

Claude Opus 5.5

Core Concepts: Glycoproteins and Post-Translational Modifications

BCHM615 · Personal study notes for Skipper

Introduction: What Is This Subject Really About?

The protein world presents a puzzle. The human genome encodes roughly 20,000 protein-coding genes, and the ribosome writes with an alphabet of only twenty amino acids (plus the occasional selenocysteine). Yet a single gene product can exist as dozens or hundreds of chemically distinct molecular species. These species differ in charge, mass, shape, location, lifetime and binding partners. Erythropoietin purified from a CHO bioreactor is not one molecule. It is a population of glycoforms. Histone H3 in a nucleus is not one molecule either. It is a combinatorial library of acetyl, methyl, phospho and ubiquityl states. The gap between "the protein the gene encodes" and "the proteins that actually exist" is the subject of this course.

Post-translational modification (PTM) is the chemistry the ribosome cannot do. The twenty canonical side chains cover a limited range of functional groups: carboxylates, amines, alcohols, one thiol, one thioether, aromatic rings, imidazole, guanidinium and amides. They contain no phosphate esters, no long hydrophobic tails, no carbohydrates, no calcium-chelating malonates and no hydroxylated prolines. Evolution solved this problem by building a second layer of enzymes. These enzymes take a finished or nascent polypeptide and decorate it, often using activated metabolites as donors. Walsh and colleagues called this "expanding nature's inventory" [Walsh, Garneau-Tsodikova & Gatto, 2005]. Each PTM adds a functional group that the genetic code lacks.

Three frames will organize everything that follows. Keep returning to them.

  1. PTMs as chemistry. Every modification is a specific reaction. It has a nucleophile, an electrophile, a leaving group, a donor substrate and a thermodynamic driving force. You already know this chemistry from CHEM201/202 and BCHM605: acyl transfer, phosphoryl transfer, glycosyl transfer, thiol-disulfide exchange and radical oxygenation. PTMs are those reactions placed on proteins.
  2. PTMs as information. Reversible PTMs work as a writable memory layered on top of the read-only genome. "Writer" enzymes install marks, "eraser" enzymes remove them, and "reader" domains interpret them. The system resembles a signaling language more than a static structure.
  3. PTMs as metabolic readouts. The donors for PTMs are central metabolites: ATP, acetyl-CoA, S-adenosylmethionine, NAD⁺, UDP-GlcNAc, palmitoyl-CoA, O₂ and 2-oxoglutarate. The level of modification therefore reports the metabolic state of the cell. Proteins "taste" metabolism through their PTMs.

Working principle for this course: the genome specifies what a protein can be; PTMs determine what it is at a given moment, in a given compartment, in a given cell. When you study a modification, always ask four questions. What enzyme writes it? What removes it, if anything? What reads it? What metabolite does it consume?

One more framing matters for a peptide-focused scientist. In biopharmaceutical development, PTMs are not academic curiosities. They are critical quality attributes (CQAs), the measurable properties that determine whether a drug is safe and effective. Under ICH Q6B, a manufacturer must characterize glycosylation, disulfide pairing, deamidation, oxidation and N-/C-terminal heterogeneity. Weeks 1, 5, 10 and 13 all come back to this point.


Concept 1: The PTM Landscape: Classification, Chemistry, and Mass

1.1 Ways to classify

Over 400 distinct PTM types are catalogued in curated databases. They become manageable once you sort them along a few independent axes.

  • Timing:
    • Co-translational modifications occur while the chain is still on the ribosome or entering the ER. Examples are N-terminal Met excision, N-myristoylation, N-terminal acetylation and, largely, N-glycosylation by the STT3A-containing OST.
    • Post-translational modifications occur after release. Examples are phosphorylation, ubiquitination and Golgi glycan processing.
    • The distinction matters. Co-translational marks can influence folding itself, while post-translational marks usually act on a folded protein.
  • Reversibility:
    • Reversible marks have dedicated erasers. Examples are phosphorylation (phosphatases), acetylation (HDACs and sirtuins), ubiquitination (DUBs), S-palmitoylation (APTs and ABHD17), O-GlcNAc (OGA) and Lys methylation (LSD1 and the JmjC demethylases).
    • Effectively irreversible marks have no eraser, so the modified protein must be degraded to "reset." Examples are proteolytic processing, γ-carboxylation, prolyl hydroxylation, N-myristoylation, most N-glycan cores and Arg methylation (active demethylases remain controversial).
  • Enzymatic vs. non-enzymatic:
    • Some PTMs happen spontaneously, by chemistry alone. These include Asn deamidation through a succinimide intermediate, Met oxidation, pyroglutamate formation from N-terminal Gln, and glycation, the Maillard reaction of reducing sugars with Lys amines (HbA1c is the clinical example).
    • In therapeutic proteins these are degradation pathways, and they drive shelf-life and formulation decisions.
  • Type of attachment:
    • Addition of small chemical groups (phosphate, acetyl, methyl).
    • Addition of complex molecules (glycans, lipids).
    • Addition of whole proteins (ubiquitin, SUMO).
    • Changes to the amino acid itself (deamidation, citrullination, hydroxylation).
    • Cleavage of the backbone (signal peptide removal, prohormone processing).

1.2 Mass shifts: the analytical fingerprint

Mann and Jensen made the decisive analytical point: almost every PTM changes the mass of the protein. Mass spectrometry can therefore detect modifications as characteristic monoisotopic increments on peptides [Mann & Jensen, 2003]. These values are worth memorizing as fluently as the pKa values of the side chains.

Modification                    Δ mass (monoisotopic, Da)   Notes
─────────────────────────────────────────────────────────────────────────────
Phosphorylation (+HPO₃)              +79.966         CID: neutral loss of H₃PO₄ (−97.977) from pSer/pThr
Acetylation                          +42.011         near-isobaric with trimethylation!
Trimethylation                       +42.047         Δ = 0.036 Da vs. acetyl → needs high resolution
Mono-/Dimethylation                  +14.016 / +28.031   ADMA and SDMA are isobaric (+28.031)
Ubiquitin (GG remnant after trypsin) +114.043        NEDD8 and ISG15 leave the same remnant
Deamidation (N→D) / Citrullination   +0.984          PNGase F converts glycosylated Asn → Asp (+0.984)
Oxidation / Hydroxylation            +15.995         Met-ox vs. Hyp: same mass, different chemistry
γ-Carboxylation (+CO₂)               +43.990         labile: loses CO₂ in CID
Disulfide formation                  −2.016          per S–S bond (loss of 2 H)
HexNAc (GlcNAc/GalNAc)               +203.079        oxonium ion m/z 204.087 in MS/MS
Hexose (Man/Gal/Glc)                 +162.053        oxonium m/z 163.060
Deoxyhexose (Fuc)                    +146.058
N-Acetylneuraminic acid (NeuAc)      +291.095        oxonium m/z 292.103 and 274.092
Myristoylation (C14:0)               +210.198
Palmitoylation (C16:0)               +238.230
Farnesylation (C15)                  +204.188
Geranylgeranylation (C20)            +272.250
Pyroglutamate (from N-term Gln)      −17.027         from N-term Glu: −18.011

Notice the traps built into this table:

  • Acetylation vs. trimethylation: 0.036 Da apart. Distinguishing them needs high-resolution instruments, or diagnostic fragments such as the acetyl-lysine immonium-derived ion at m/z 126.091.
  • Oxidation vs. hydroxylation: identical masses.
  • Deamidation vs. citrullination: identical masses.
  • The GG remnant is shared by ubiquitin, NEDD8 and ISG15.

Mass alone never proves which PTM is present. Localization by fragmentation, retention-time behavior, enrichment chemistry and orthogonal evidence complete the identification.


Concept 2: The Writer–Eraser–Reader Logic and the Mathematics of Reversible Switches

Reversible PTMs are best understood as dynamic steady states, not as fixed decorations. A protein that is "50% phosphorylated" is usually not a stable mixture of two species. Each molecule is cycling continuously between the two states, and kinase and phosphatase rates happen to balance at 50%. This connects directly to the enzyme kinetics of BCHM605.

2.1 The covalent modification cycle

              writer (kinase, HAT, OGT, E3...)
     S  ─────────────────────────────────────▶  S*
        ◀─────────────────────────────────────
              eraser (phosphatase, HDAC, OGA, DUB...)

Steady state (both enzymes far from saturation, first-order in substrate):

   v_write = (k_cat,W / K_M,W)·[W]·[S]        v_erase = (k_cat,E / K_M,E)·[E]·[S*]

   Fraction modified:   f* = [S*]/([S]+[S*]) = α / (α + β)

      where  α = (k_cat,W/K_M,W)·[W]   and   β = (k_cat,E/K_M,E)·[E]

In this first-order regime, f* changes hyperbolically as writer activity rises: smooth and graded. Goldbeter and Koshland showed what happens when both writer and eraser are saturated (K_M ≪ [substrate]). The response then becomes extremely steep and sigmoidal, approaching an all-or-none switch, even though the system has no cooperativity in the classical allosteric sense [Goldbeter & Koshland, 1981]. This is zero-order ultrasensitivity. A small change in the kinase/phosphatase ratio near the balance point flips the population from mostly unmodified to mostly modified.

Why it is beautiful: ultrasensitivity emerges from enzyme saturation, a property every enzyme can have. It needs no special structural machinery. Many decisive cellular switches, such as mitotic entry and MAPK cascade outputs, combine this mechanism with multisite phosphorylation and feedback loops.

2.2 Readers make marks meaningful

A mark only matters if something recognizes it. The reader domains you will meet repeatedly are:

  • SH2 domains bind pTyr in sequence context (pY-x-x-Φ for Src-family SH2s).
  • PTB domains bind NPxpY.
  • 14-3-3 proteins are dimeric and bind pSer/pThr motifs such as R-S-x-pS-x-P. They often sequester or clamp their targets.
  • FHA domains are specific for pThr.
  • WW domains bind pSer/pThr-Pro.
  • Bromodomains bind acetyl-Lys.
  • Chromodomains, PHD fingers, Tudor domains and MBT domains bind methyl-Lys or methyl-Arg. Aromatic cages recognize the methylammonium cation through cation–π interactions.
  • Ubiquitin-binding domains (UIM, UBA, UBAN) distinguish chain linkages.
  • Lectins read glycans. Examples are calnexin, the asialoglycoprotein receptor, selectins and Siglecs.

One unifying design principle applies across these readers. Individual reader–mark interactions are weak, with K_d often in the 1–100 µM range. Specificity and strength come from avidity: tandem readers, multiple marks and scaffolds that bring them together. The 14-3-3 dimer binding two phosphosites on a single target is a textbook case.


Concept 3: PTMs Read Metabolism Through Their Donor Substrates

This is one of the most important "aha" ideas in modern PTM biology. Look at the cofactor list:

PTM                         Donor / cofactor             Metabolic state reported
────────────────────────────────────────────────────────────────────────────────
Phosphorylation             ATP                          energy charge (less directly)
Lys acetylation             acetyl-CoA                   glucose/fatty acid flux, mitochondrial export (ACLY)
Deacetylation (sirtuins)    NAD⁺                         NAD⁺/NADH ratio, fasting, caloric restriction
Methylation                 S-adenosylmethionine (SAM)   one-carbon metabolism, Met, folate
Demethylation (JmjC)        2-oxoglutarate, O₂, Fe²⁺     TCA cycle; inhibited by 2-HG, succinate, fumarate
O-GlcNAcylation             UDP-GlcNAc                   glucose + glutamine + acetyl-CoA + UTP (hexosamine pathway)
N-glycosylation             Dol-PP-oligosaccharide       nucleotide sugars, dolichol (mevalonate pathway)
Palmitoylation              palmitoyl-CoA                fatty acid availability
Prenylation                 FPP / GGPP                   mevalonate pathway (statin-sensitive)
Prolyl hydroxylation        O₂, 2-OG, Fe²⁺, ascorbate    oxygen tension (HIF), vitamin C status
γ-Carboxylation             vitamin K hydroquinone, CO₂, O₂   vitamin K status (warfarin-sensitive)

Several consequences follow, all of which appear later in the course:

  • O-GlcNAc as a nutrient sensor. UDP-GlcNAc integrates glucose, amino acid (glutamine), fatty acid (acetyl-CoA) and nucleotide (UTP) metabolism. Hart and colleagues proposed that O-GlcNAc cycling acts as a nutrient-responsive regulator that cross-talks with phosphorylation [Hart, Housley & Slawson, 2007]. Chronic hyperglycemia raises flux through the hexosamine pathway, which increases O-GlcNAcylation. This is implicated in insulin resistance (Week 12).
  • Oncometabolites hijack PTM enzymes. Mutant IDH1/2 produce 2-hydroxyglutarate, a competitive inhibitor of 2-OG-dependent dioxygenases (JmjC histone demethylases and TET DNA hydroxylases). The result is a hypermethylated epigenome. SDH- and FH-deficient tumors accumulate succinate and fumarate with similar effects.
  • Statins affect prenylation as well as cholesterol, because FPP and GGPP sit downstream of HMG-CoA reductase.
  • Scurvy is a PTM disease. Ascorbate keeps the Fe²⁺ center of prolyl 4-hydroxylase reduced. Without it, collagen is under-hydroxylated and its triple helix is thermally unstable at body temperature (Concept 10).

Concept 4: N-Linked Glycosylation in the ER: A Sugar Code for Folding Status

4.1 Building the precursor: an assembly line across a membrane

N-glycosylation does not build the glycan on the protein. The cell first assembles a 14-sugar oligosaccharide on a lipid carrier, dolichol pyrophosphate (Dol-PP), then transfers the whole block to the protein en bloc. Dolichol is a very long polyisoprenoid (roughly 17–21 isoprene units in humans), so this pathway depends on the mevalonate pathway too.

  1. Cytoplasmic face:
    • DPAGT1 transfers GlcNAc-1-P from UDP-GlcNAc to Dol-P, giving Dol-PP-GlcNAc. Tunicamycin blocks this step, and it is the classic tool for abolishing N-glycosylation.
    • A second GlcNAc and five mannoses (from GDP-Man) are then added, giving Man₅GlcNAc₂-PP-Dol.
  2. Flipping: the Man₅GlcNAc₂-PP-Dol intermediate is translocated to the luminal face. RFT1 has been implicated, though the precise flippase mechanism remains debated.
  3. Luminal face:
    • Four more mannoses and three glucoses are added.
    • In the lumen, the donors are no longer nucleotide sugars. They are lipid-linked monosaccharides, Dol-P-Man and Dol-P-Glc, which are themselves made on the cytoplasmic side and flipped.
    • The product is Glc₃Man₉GlcNAc₂-PP-Dol.

Why build the oligosaccharide on a lipid? Nucleotide sugars are made in the cytosol, but the acceptor protein is in the ER lumen. A membrane-embedded carrier lets the cell assemble a large, defined structure close to the translocon and transfer it in one step [Varki et al., 2017, ch. 9].

4.2 Oligosaccharyltransferase (OST) and the sequon

Glc₃Man₉GlcNAc₂–PP–Dol  +  –Asn–X–Ser/Thr–  ──OST (STT3A/B), Mn²⁺/Mg²⁺──▶
                           –Asn(Glc₃Man₉GlcNAc₂)–X–Ser/Thr–  +  Dol–PP

Linkage: GlcNAc β1–N(amide) Asn          Sequon: N–X–S/T  (X ≠ Pro; N–X–C rarely)

Here is the chemical puzzle. A carboxamide nitrogen is a terrible nucleophile, because its lone pair is delocalized into the carbonyl. How does OST glycosylate it?

  • Imperiali's peptide studies and the crystal structure of the bacterial OST homolog PglB [Lizák et al., 2011] support a model in which the Ser/Thr hydroxyl (+2 position) hydrogen-bonds to the Asn side chain.
  • Together with active-site acidic residues and a divalent metal that activates the pyrophosphate leaving group, this distorts the amide and raises its nucleophilicity.
  • The sequon must adopt a specific turn-like conformation in the active site. Proline at X cannot accommodate it, which explains the X ≠ Pro rule.
  • Thr sequons are generally glycosylated more efficiently than Ser sequons, consistent with the stronger positioning by the Thr methyl group.

Two caveats matter:

  • The sequon is necessary but not sufficient. Many sequons are unoccupied or only partly occupied ("macroheterogeneity"). This depends on local structure, on timing relative to folding, and on which OST isoform is involved. STT3A acts co-translationally. STT3B catches sites skipped by STT3A, often post-translationally.
  • The sequon is a linear motif. Sites buried by rapid folding can be skipped.

4.3 The calnexin/calreticulin cycle: glucose as a folding flag

This is one of the most elegant systems in cell biology. Once transferred, the three glucoses are trimmed in a defined order, and the number of glucoses remaining encodes where the protein stands in its folding journey [Helenius & Aebi, 2004].

  1. Glucosidase I removes the terminal α1,2-Glc, giving Glc₂Man₉GlcNAc₂.
  2. Glucosidase II removes the next Glc, giving Glc₁Man₉GlcNAc₂. This monoglucosylated form is the ligand for the lectin chaperones calnexin (membrane-bound) and calreticulin (soluble).
  3. While bound, the glycoprotein is held near ERp57, a PDI-family oxidoreductase that catalyzes disulfide formation and isomerization. This links Weeks 2 and 10.
  4. Glucosidase II removes the last Glc, which releases the protein from calnexin.
  5. UGGT (UDP-glucose:glycoprotein glucosyltransferase) is the folding sensor. It recognizes exposed hydrophobic patches on non-native conformers and re-adds a single glucose. This sends the protein back to calnexin for another round.
  6. Native proteins are not recognized by UGGT and exit to the Golgi.

Analogy: think of glucose as a "not finished" sticky note. The chaperone grabs anything with the note. Glucosidase II keeps peeling notes off. UGGT is the inspector who sticks a new note on anything still misfolded. The protein cycles until it passes inspection or runs out of time.

4.4 Mannose trimming: the timer for ERAD

How does the cell decide to give up? Mannose trimming acts as a slow molecular timer.

  • ER mannosidase I and the EDEM proteins progressively remove α1,2-mannoses: Man₉ to Man₈ to Man₇ to Man₆/Man₅.
  • Trimming the C-branch mannose exposes an α1,6-linked mannose. This is recognized by the ERAD lectins OS-9 and XTP3-B.
  • These lectins deliver the substrate to the Hrd1 E3 ligase complex for retrotranslocation, ubiquitination and proteasomal degradation.
  • Trimmed glycans are also poorer UGGT substrates, so the protein drifts out of the folding cycle.

The longer a protein lingers in the ER, the more mannoses it loses, and the more likely it is to be degraded. A single N-glycan thus links folding (calnexin), quality control (UGGT) and degradation (ERAD, through ubiquitin, Week 7). It is a clear example of PTM crosstalk.


Concept 5: Golgi Processing: A Non-Templated Assembly Line That Generates Diversity

5.1 The pathway

Proteins that pass quality control leave the ER mostly as Man₈GlcNAc₂ (or Man₉). In the Golgi:

  1. Golgi α-mannosidase I trims to Man₅GlcNAc₂.
  2. GnT-I (MGAT1) adds β1,2-GlcNAc to the α1,3-Man arm. This is the commitment step: hybrid glycans branch off here.
  3. α-Mannosidase II removes the α1,3- and α1,6-Man on the α1,6-arm. Swainsonine inhibits this step.
  4. GnT-II adds GlcNAc to the α1,6-arm, giving a biantennary complex glycan.
  5. Branching and elaboration follow:
    • GnT-IV and GnT-V give tri- and tetra-antennary glycans.
    • GnT-III adds the "bisecting" GlcNAc, which inhibits further branching by GnT-IV and GnT-V.
    • FUT8 adds core α1,6-fucose to the Asn-linked GlcNAc.
    • Galactosyltransferases add Gal.
    • Sialyltransferases cap with α2,3- or α2,6-linked sialic acid.

This gives three structural classes, all sharing the trimannosyl core Man₃GlcNAc₂:

  • High-mannose: Man₅–₉GlcNAc₂.
  • Hybrid: one arm processed, one arm high-mannose.
  • Complex: both arms bearing GlcNAc-initiated antennae.

5.2 The key insight: glycosylation is not templated

DNA replication, transcription and translation copy a template. Glycan synthesis does not. The final structure emerges from competition among enzymes, and that competition depends on:

  • which glycosyltransferases are expressed, and how much;
  • where they sit along the cis-to-trans Golgi;
  • nucleotide-sugar supply and Golgi transporters;
  • transit time;
  • local protein structure, since accessible sites are processed more fully than buried ones.

The result is microheterogeneity: the same site on the same protein carries a distribution of glycan structures. Glycoproteins are therefore populations of glycoforms, not single molecules. A biologic's glycan profile is a fingerprint of the cell line and the process conditions. This is why "the process is the product" in biomanufacturing [Varki et al., 2017].

The Endo H test. This classic assay comes directly from the pathway logic.

  • Endoglycosidase H cleaves between the two core GlcNAcs of high-mannose and most hybrid glycans.
  • It cannot cleave once α-mannosidase II has acted.
  • Endo H resistance is therefore a biochemical readout that a protein has reached the medial Golgi.
  • PNGase F, by contrast, removes nearly all N-glycans from Asn. The exception is glycans with core α1,3-fucose (plant and insect), which require PNGase A.

5.3 Glycan function: two cases with therapeutic consequences

(a) Terminal sialic acid controls serum half-life.

  • Morell and Ashwell found that removing sialic acid from ceruloplasmin caused it to vanish from rabbit circulation within minutes. The exposed galactose residues were recognized by a hepatic lectin, now called the asialoglycoprotein receptor (ASGPR, Ashwell–Morell receptor) [Ashwell & Morell, 1974].
  • Sialic acid therefore acts as a "do not clear" cap.
  • This is the logic behind darbepoetin alfa. Human EPO has three N-glycans (Asn24, Asn38, Asn83) and one O-glycan (Ser126). Darbepoetin carries two additional engineered N-glycosylation sites (five N-glycans in total). The extra sialic acids increase its terminal half-life roughly threefold [Egrie & Browne, 2001]. (Increased hydrodynamic size also contributes.)
  • The receptor is also a delivery opportunity. Triantennary GalNAc conjugates exploit ASGPR to deliver siRNAs to hepatocytes.

(b) Core fucose suppresses ADCC.

  • IgG1 carries a conserved biantennary complex N-glycan at Asn297 in each CH2 domain.
  • When this glycan lacks core α1,6-fucose, IgG1 binds FcγRIIIa on NK cells with markedly higher affinity (reported up to ~50-fold) and gives far stronger antibody-dependent cellular cytotoxicity [Shields et al., 2002; Shinkawa et al., 2003].
  • Structurally, the fucose sterically interferes with a favorable carbohydrate–carbohydrate contact between the Fc glycan and the N-glycan at Asn162 of FcγRIIIa.
  • Obinutuzumab (anti-CD20) is produced in cells overexpressing GnT-III. Bisecting GlcNAc reduces FUT8 action, so the antibody is largely afucosylated.
  • Mogamulizumab is made in FUT8-knockout CHO cells.
  • A single monosaccharide on a 150-kDa protein can change clinical efficacy. This is Week 13 in a sentence.

Immunogenic glycan epitopes. Non-human cell lines can add glycans that humans react against.

  • Humans lack a functional CMAH gene and cannot make Neu5Gc.
  • Humans lack functional α1,3-galactosyltransferase and so make anti-α-Gal antibodies.
  • Cetuximab produced in murine SP2/0 cells carries Galα1,3Gal. In patients with pre-existing IgE against α-Gal, this caused hypersensitivity reactions [Chung et al., 2008].
  • CHO cells make little of either epitope, which is one reason they dominate biomanufacturing.

Concept 6: O-Linked Glycosylation: Many Pathways, Many Logics

"O-glycosylation" is not one pathway. It is a family of chemically and biologically distinct modifications that share only the linkage to a Ser/Thr hydroxyl.

6.1 Mucin-type O-GalNAc

  • Initiation: a family of ~20 human polypeptide GalNAc-transferases (ppGalNAcTs, GALNTs) transfer GalNAc from UDP-GalNAc to Ser/Thr in the Golgi.
  • No sequon: specificity comes from overlapping preferences of the isoforms (Pro at +3 is favored), from their lectin domains (which recognize previously added GalNAc, enabling dense "follow-on" glycosylation) and from differential expression.
  • Core structures:
    • GalNAc-α-Ser/Thr is the Tn antigen.
    • Core 1 (Galβ1-3GalNAc, the T antigen) is made by C1GalT1, which requires the dedicated chaperone Cosmc.
    • Cores 2–4 add GlcNAc branches. Cores 5–8 are rare.
  • Function: mucins carry hundreds of O-glycans on Ser/Thr/Pro-rich tandem repeats. These create extended, hydrated, bottle-brush structures that lubricate and protect epithelia.
  • Cancer: truncated O-glycans (Tn, sialyl-Tn) are tumor-associated antigens.

6.2 O-GlcNAc: the outlier

O-GlcNAc breaks almost every rule of "classical" glycosylation:

  • It occurs in the nucleus, cytoplasm and mitochondria, not in the secretory pathway.
  • It is a single monosaccharide (β-GlcNAc) that is never elongated.
  • It is rapidly reversible. A single OGT gene adds it and a single OGA gene removes it, in contrast to the hundreds of kinases and phosphatases.
  • It frequently competes or cooperates with phosphorylation at the same or neighboring Ser/Thr sites. This is the "yin–yang" hypothesis, though the relationship is site-specific rather than universal.
  • Its donor, UDP-GlcNAc, makes it a nutrient sensor (Concept 3) [Hart, Housley & Slawson, 2007].

The single writer/eraser pair is achieved through extensive regulation of OGT by its TPR-repeat substrate-recognition domain, by adaptor proteins, and by its own PTMs. OGT knockout is lethal in mammalian embryonic stem cells, which shows that this "simple" modification is essential.

6.3 O-Mannose and the dystroglycanopathies

  • The pathway begins in the ER: POMT1/POMT2 transfer Man from Dol-P-Man to Ser/Thr.
  • On α-dystroglycan, a specialized core M3 glycan is extended by a series of enzymes (POMGNT2, B3GALNT2, POMK, FKTN, FKRP, RXYLT1, B4GAT1).
  • LARGE1 then synthesizes matriglycan, a repeating [-3Xylα1-3GlcAβ1-] polymer that binds laminin and other ECM proteins carrying LG domains.
  • Mutations in any of these enzymes break the link between the muscle cytoskeleton and the basement membrane. The resulting congenital muscular dystrophies include Walker–Warburg syndrome, muscle–eye–brain disease and Fukuyama CMD.
  • It is striking that roughly a dozen distinct genes, all required for one glycan on one protein, each cause a disease when mutated.

6.4 O-Fucose and O-Glucose on EGF repeats: tuning Notch

  • Notch receptors carry ~36 EGF-like repeats.
  • POFUT1 adds O-fucose at the consensus C²-X₄₋₅-S/T-C³.
  • POGLUT1 adds O-glucose at C¹-X-S-X-P/A-C².
  • Fringe GlcNAc-transferases (Lunatic, Manic, Radical) extend O-fucose with β1,3-GlcNAc. This shifts ligand preference, enhancing Notch activation by Delta-like ligands relative to Jagged/Serrate.

Here a glycan works as a tunable receptor-specificity dial. Developmental boundaries, such as the dorsal–ventral wing margin in Drosophila, are drawn partly by where Fringe is expressed. This connects to BIOL402.


Concept 7: Phosphorylation: Why Phosphate Is the Universal Switch

Phosphorylation is the most studied regulatory PTM. Its dominance has chemical roots.

7.1 The chemistry of the phosphoryl group

  • Charge. A phosphomonoester has pKa values of roughly 1 and roughly 6, so at physiological pH it is mostly a dianion (−2). Attaching it to a neutral Ser introduces about two negative charges, which is a large perturbation.
  • Glu/Asp are imperfect mimics. Phosphomimetic mutations (S→E/D) carry only −1 charge, have different geometry, and lack the tetrahedral hydrogen-bond acceptor array that SH2 and 14-3-3 pockets recognize. They often fail to reproduce phospho-dependent binding.
  • Kinetic stability with thermodynamic instability. Phosphate esters are thermodynamically unstable to hydrolysis but kinetically very stable without enzymes. Westheimer's classic argument is that the negative charge repels nucleophiles [Westheimer, 1987]. This is exactly what a switch needs: stable until an enzyme flips it.
  • Hydrogen-bonding geometry. The phosphate can bridge Arg and Lys residues. This is often how phosphorylation drives conformational change, for example in the glycogen phosphorylase Ser14 rearrangement, or in kinase activation loops where pThr organizes the catalytic site.
Kinase:       Protein–Ser–OH  +  Mg·ATP  ──▶  Protein–Ser–OPO₃²⁻  +  Mg·ADP  +  H⁺
              (in-line phosphoryl transfer; γ-phosphate; conserved Lys/Asp/Mg²⁺ machinery)

Phosphatase:  Protein–Ser–OPO₃²⁻  +  H₂O  ──▶  Protein–Ser–OH  +  HPO₄²⁻
              PPP family (PP1, PP2A): binuclear metal center, water nucleophile, no covalent intermediate
              PTP family (PTP1B): Cys nucleophile in CX₅R motif → phospho-cysteine intermediate

7.2 Proportions and asymmetries

  • Residue distribution. Large-scale phosphoproteomics found roughly 86% pSer, 12% pThr and ~2% pTyr in HeLa cells [Olsen et al., 2006]. pTyr is rare but has an outsized signaling role, especially downstream of RTKs.
  • Writer/eraser asymmetry. The human kinome has ~518 protein kinases [Manning et al., 2002], but there are far fewer Ser/Thr phosphatase catalytic subunits. PP1 and PP2A achieve specificity combinatorially: a small number of catalytic subunits pair with hundreds of regulatory and targeting subunits. PP1 alone has more than 100 regulatory interactors, many using an RVxF docking motif.
  • The lesson: phosphatases are not "dumb erasers." They are specified by their holoenzyme assembly.

7.3 Kinase families you must know

  • PKA: activated by cAMP releasing catalytic subunits from the R₂C₂ holoenzyme. The consensus is R-R-x-S/T. Targeted by AKAPs.
  • PKG: activated by cGMP. The NO/sGC pathway and PDE5 inhibitors act here.
  • PKC: conventional isoforms need Ca²⁺ plus DAG plus phosphatidylserine. Novel isoforms need DAG only. Phorbol esters are DAG mimics.
  • Receptor tyrosine kinases: ligand-induced dimerization leads to trans-autophosphorylation, creating pTyr docking sites for SH2 proteins.
  • Non-receptor tyrosine kinases:
    • Src is autoinhibited by intramolecular SH2–pTyr527 and SH3–linker interactions. A PTM-dependent reader domain thus regulates its own kinase.
    • ABL is constitutively activated in BCR-ABL and targeted by imatinib.
    • FAK acts in integrin signaling.

Concept 8: Ubiquitin: When the Modification Is a Protein

8.1 The cascade

Ubiquitin is a 76-residue, ~8.6 kDa β-grasp protein. Its C-terminal Gly76 carboxylate is attached through an isopeptide bond to the ε-amine of a substrate Lys (less often to the N-terminus, Ser/Thr or Cys).

E1 (activation, ATP-dependent):
   Ub–COO⁻ + ATP·Mg²⁺   ──▶  Ub–CO–AMP (acyl-adenylate) + PPi
   Ub–CO–AMP + E1–SH    ──▶  E1–S–CO–Ub (thioester)  + AMP

E2 (conjugation):
   E1–S~Ub + E2–SH       ──▶  E2–S~Ub + E1–SH          (transthioesterification)

E3 (ligation):
   RING / U-box:  scaffold; positions E2~Ub so the substrate Lys attacks the E2 thioester directly
   HECT:          E2~Ub → E3–S~Ub (thioester on E3 Cys) → substrate Lys
   RBR:           RING1 binds E2~Ub; transfers to Cys on RING2 (HECT-like), then to substrate

   Net:  Substrate–Lys–NH₂ + Ub  ──▶  Substrate–Lys–NH–CO–Gly76–Ub   (isopeptide)

This is the acyl-adenylate/thioester chemistry from CHEM202 and from aminoacyl-tRNA synthetases in BIOL201, reused. ATP pays to convert an unreactive carboxylate into a high-energy acyl donor. Thioesters carry it between enzymes. A Lys amine is the final nucleophile.

The hierarchy is an expanding funnel of specificity: ~2 E1s (UBA1, UBA6), ~40 E2s and more than 600 E3s. Substrate specificity lives mainly at the E3 level.

8.2 The ubiquitin code

Ubiquitin itself has seven Lys residues (K6, K11, K27, K29, K33, K48, K63) plus its N-terminal Met1, so chains can be built with eight homotypic linkages. Mixed and branched chains are also possible. Each linkage gives a different chain shape:

  • K48 chains are compact. They are the canonical proteasome signal, and a tetra-ubiquitin chain is roughly the minimum for efficient recognition [Thrower et al., 2000].
  • K63 chains are extended and open. They act in DNA damage signaling, endocytic trafficking and NF-κB activation, and they generally do not signal degradation.
  • M1 (linear) chains are made by LUBAC and are important for NF-κB signaling.
  • K11 chains are associated with APC/C-driven mitotic degradation.
  • Monoubiquitination regulates histone function (H2AK119ub, H2BK120ub) and endocytosis.

Komander and Rape's review frames this as a "code" whose meaning depends on chain topology [Komander & Rape, 2012].

Ubiquitin-like modifiers follow the same E1-E2-E3 logic with separate enzyme sets:

  • SUMO (E1 SAE1/2, single E2 Ubc9, consensus ΨKxE) mainly regulates nuclear processes, protein–protein interactions through SIMs, and stress responses.
  • NEDD8 modifies cullins, which activates cullin-RING ligases. MLN4924/pevonedistat inhibits the NEDD8 E1.
  • ISG15 is interferon-induced and antiviral.

DUBs number ~100 in humans:

  • Cys proteases: USP, UCH, OTU, MJD, MINDY and ZUP1 families.
  • The JAMM/MPN⁺ metalloproteases, including RPN11 on the proteasome lid, which removes chains just before substrate degradation.
  • Many show striking linkage specificity. OTULIN, for example, is M1-specific.

Proteomics angle. Trypsin cleaves after Arg72 of ubiquitin, leaving a Gly-Gly dipeptide on the modified Lys (+114.043 Da). Anti-K-ε-GG antibodies enrich these remnant peptides and enabled proteome-wide ubiquitination mapping. A trypsin artifact became a tool. Remember the caveat that NEDD8 and ISG15 leave the same remnant.


Concept 9: Chemical Modulation of Lysine and Arginine: Acetylation and Methylation

Lysine is the most heavily modified residue in the proteome. It can be acetylated, mono/di/tri-methylated, ubiquitylated, SUMOylated, crotonylated, succinylated, malonylated, β-hydroxybutyrylated and lactylated. These marks compete for the same ε-amine, which makes Lys a molecular decision point.

9.1 Charge logic: the key distinction

  • Acetylation neutralizes the positive charge. On histone tails, this weakens electrostatic contact with DNA phosphates and creates bromodomain docking sites. Both effects favor open chromatin and transcription.
  • Methylation preserves the positive charge. Trimethyl-Lys is a permanent quaternary ammonium. Methylation therefore has no direct electrostatic effect. Its meaning comes entirely from which reader binds, which makes it context-dependent:
    • H3K4me3 at active promoters recruits activators such as TAF3 and PHD-finger proteins.
    • H3K27me3, deposited by PRC2/EZH2, recruits Polycomb CBX chromodomains for repression.
    • H3K9me3 recruits HP1 for heterochromatin.
HAT (p300/CBP, PCAF/GCN5):   Lys–NH₃⁺ + acetyl-CoA  ──▶  Lys–NH–COCH₃ + CoA–SH + H⁺

Zn²⁺-HDAC (Class I, II, IV): Lys–NHAc + H₂O  ──▶  Lys–NH₃⁺ + acetate
                             (inhibited by hydroxamates: vorinostat/SAHA; cyclic depsipeptide romidepsin)

Sirtuin (Class III):         Lys–NHAc + NAD⁺  ──▶  Lys–NH₃⁺ + nicotinamide + 2′-O-acetyl-ADP-ribose
                             (consumes a full NAD⁺ per deacetylation; nicotinamide is a feedback inhibitor)

KMT (SET domain; DOT1L):     Lys–NH₃⁺ + SAM  ──▶  Lys–NH₂⁺CH₃ + SAH       (repeated → me2, me3)

Demethylases:  LSD1 (FAD, amine oxidase; me1/me2 only — needs an N–H, so cannot act on me3)
               JmjC (Fe²⁺/2-OG dioxygenase; hydroxylates the methyl → unstable hemiaminal → CH₂O; can do me3)

The LSD1 restriction is a nice piece of mechanistic logic. Amine oxidation through an iminium intermediate requires a hydrogen on nitrogen. Trimethyl-Lys has none, so a different chemistry is needed: hydroxylation by JmjC enzymes. The demethylase family a cell uses is dictated by organic chemistry.

Sirtuins are a second example of chemistry with consequences. They hydrolyze NAD⁺ itself, an expensive cofactor, rather than water. This seems wasteful until you see that it couples deacetylation directly to NAD⁺ availability, a measure of cellular energy and redox state. This connects to CHEM402.

9.2 Arginine methylation

  • Type I PRMTs (PRMT1 is predominant) make ω-NG,NG-asymmetric dimethylarginine (ADMA), with both methyls on the same terminal nitrogen.
  • Type II (PRMT5, with MEP50) makes symmetric dimethylarginine (SDMA), with one methyl on each terminal nitrogen.
  • Both start from MMA. ADMA and SDMA are isobaric (+28.031 Da) and are distinguished in MS/MS by characteristic neutral losses: dimethylamine (45.058 Da) for ADMA and methylamine (31.042 Da) for SDMA.
  • Methylation keeps Arg's charge but removes hydrogen-bond donors and adds bulk. This modulates interactions, for example with Tudor domains that recognize SDMA on Sm proteins.
  • Free ADMA released by proteolysis is an endogenous nitric oxide synthase inhibitor, which links PTM turnover to vascular physiology (BIOL302).
  • PRMT5 is a cancer target, especially in MTAP-deleted tumors where accumulated MTA partially inhibits it. This is a synthetic-lethal vulnerability.

Concept 10: Lipidation: Hydrophobic Switches and the Two-Signal Principle

10.1 The four main types and their chemistry

N-Myristoylation:  NMT1/2 transfer C14:0 from myristoyl-CoA → amide to N-terminal Gly
                   (Gly is residue 2 in the ORF; becomes residue 1 after MetAP removes Met)
                   Co-translational (or post-translational after caspase cleavage exposes internal Gly)
                   Stable amide → effectively irreversible

S-Palmitoylation:  DHHC (zDHHC) PATs, ~23 in humans; C16:0 thioester to Cys
                   Two-step: DHHC autopalmitoylation (Cys~acyl) → transfer to substrate Cys
                   Reversible: APT1/2, ABHD17 thioesterases → dynamic cycling

Prenylation:       CaaX box; FTase (X = Ser, Met, Gln, Ala) → farnesyl (C15) thioether
                   GGTase-I (X = Leu, Phe) → geranylgeranyl (C20) thioether
                   then RCE1 cleaves –aaX; ICMT methylates the new C-terminal carboxylate (SAM)
                   Rab proteins: GGTase-II (RabGGTase) double-geranylgeranylates C-terminal CC/CXC

Other notable:     Wnt: O-linked palmitoleoylation (C16:1 cis-Δ9) on Ser by Porcupine (an MBOAT)
                   Hedgehog: N-terminal amide palmitoylation (Hhat, MBOAT) + C-terminal cholesterol ester
                   GPI anchors: C-terminal transamidation in the ER (PIG enzymes)

Precision note on the syllabus. Week 9 lists Wnt and Hedgehog under "S-palmitoylation," but their signature lipidations are made by MBOAT-family enzymes, not DHHC PATs, and they are not thioesters:

  • Wnt carries an O-linked palmitoleate.
  • Hedgehog carries an N-linked (amide) palmitate plus cholesterol.

These lipids are essential for secretion (through Wntless) and receptor engagement. Porcupine inhibitors such as LGK974 are in oncology trials.

10.2 The two-signal hypothesis

One lipid is usually not enough for stable membrane association. A myristoyl chain partitions into bilayers with a Kd around 10−4 M. That is too weak to anchor a protein permanently, and it dissociates on a timescale of seconds [Peitzsch & McLaughlin, 1993]. Stable anchoring requires a second signal:

  • Myristoyl plus palmitoyl: Src-family kinases such as Lck and Fyn, and Gα subunits.
  • Myristoyl plus a polybasic cluster: Src and MARCKS, which bind acidic phospholipids electrostatically. PKC phosphorylation of MARCKS adds negative charge and ejects it from the membrane. This is the "myristoyl-electrostatic switch."
  • Farnesyl plus palmitoyl: H-Ras and N-Ras.
  • Farnesyl plus polybasic region: K-Ras4B.
  • Ligand-triggered exposure: recoverin and ARF1 use a "myristoyl switch," in which binding of Ca²⁺ or GTP exposes a sequestered myristoyl group.

The weakness of a single lipid is a feature. It makes membrane binding conditional, and therefore regulatable.

10.3 A cautionary tale: farnesyltransferase inhibitors

FTIs were designed to block Ras membrane localization in cancer. They largely failed against K-Ras and N-Ras tumors. When FTase is inhibited, these proteins are alternatively geranylgeranylated by GGTase-I and stay functional. The redundancy was in the enzymes' substrate overlap. FTIs later found success in progeria: lonafarnib blocks permanent farnesylation of progerin, an aberrant lamin A.


Concept 11: Disulfide Bonds: Thiol–Disulfide Exchange and Compartment Redox

11.1 Mechanism and thermodynamics

Disulfides form and rearrange through thiol–disulfide exchange. This is an SN2 reaction at sulfur: a thiolate (RS⁻, not the neutral thiol) attacks one sulfur of a disulfide along the S–S axis, displacing the other sulfur as a thiolate. The thiol pKa is about 8.3 for free Cys but varies widely in proteins, so local pKa controls reactivity. PDI active-site cysteines in their CGHC motifs have unusually low pKa values, roughly 4.5–6.7 for the N-terminal Cys.

Thiol–disulfide exchange:
   R¹S⁻  +  R²S–SR³   ⇌   R¹S–SR²  +  R³S⁻

Protein oxidation by glutathione:
   P(SH)₂  +  GSSG   ⇌   P(S–S)  +  2 GSH

   K_ox = [P(S–S)]·[GSH]² / ([P(SH)₂]·[GSSG])        (units: M)

Redox potential of the glutathione buffer (Nernst, n = 2):
   E = E°′(GSSG/2GSH)  −  (RT / 2F)·ln( [GSH]² / [GSSG] ),     E°′ ≈ −240 mV (pH 7)

   Cytosol: [GSH]/[GSSG] ≈ 30–100:1  → strongly reducing
   ER:      [GSH]/[GSSG] ≈ 1–3:1     → oxidizing                 [Hwang, Sinskey & Lodish, 1992]

Note the [GSH]² term. Because two glutathiones are released, the redox potential depends on the absolute glutathione concentration as well as the ratio. This is easy to forget.

Why does an intramolecular disulfide form spontaneously when the two cysteines are close? The answer is effective molarity (Creighton's analysis of BPTI folding). An intramolecular reaction benefits from a local "concentration" of the second thiol that can reach 10²–10⁵ M in a well-structured protein. That is far higher than any achievable GSSG concentration [Creighton, 1993]. Folding and disulfide formation are therefore thermodynamically coupled: native structure stabilizes native disulfides, and native disulfides stabilize native structure. The same idea appeared as the entropic contribution to stability in BCHM603.

11.2 The ER oxidative folding machinery

  • Ero1 is a flavoenzyme that re-oxidizes PDI, passing electrons to O₂ and producing H₂O₂. (Peroxiredoxin 4 and GPx7/8 recycle that H₂O₂ into additional disulfides.)
  • PDI both introduces disulfides (oxidized form) and isomerizes non-native ones (reduced form). Isomerization is often the rate-limiting function for proteins with several disulfides.
  • ERp57 partners with calnexin and calreticulin to act specifically on glycoproteins (Concept 4.3).

11.3 Therapeutic disulfide patterns

  • Insulin: intrachain A6–A11, interchain A7–B7 and A20–B19. The single-chain proinsulin precursor folds with the C-peptide connecting the chains, which makes correct pairing intramolecular. Proteolytic removal of the C-peptide then gives the mature two-chain hormone. This is a striking example of a PTM (proteolysis) that is only possible because of a prior PTM (disulfide formation).
  • Oxytocin and vasopressin: a Cys1–Cys6 disulfide closes a 20-membered ring that constrains the bioactive conformation.
  • Analytical mapping: digest at low pH under non-reducing conditions to limit scrambling, often with free thiols first alkylated by NEM or IAA. Compare with a reduced sample. For complex patterns, use partial reduction (e.g., TCEP at acidic pH) followed by differential alkylation. ETD and EThcD preferentially cleave S–S bonds, which aids assignment.

Concept 12: Cofactor-Dependent Oxygen Chemistry: γ-Carboxylation and Hydroxylation

12.1 Vitamin K and the vitamin K cycle

γ-Glutamyl carboxylase (GGCX) converts specific Glu residues to γ-carboxyglutamate (Gla) in the N-terminal "Gla domains" of factors II, VII, IX and X, proteins C, S and Z, Gas6, osteocalcin and matrix Gla protein. A propeptide sequence recruits GGCX before it is removed.

GGCX:   Glu–(γ-CH₂) + CO₂ + O₂ + vitamin K hydroquinone (KH₂)
                 ──▶   Gla–(γ-CH(COO⁻)₂) + vitamin K 2,3-epoxide (KO) + H₂O

VKORC1: KO ──▶ K (quinone) ──▶ KH₂            ← WARFARIN inhibits VKORC1
        (dietary K can also enter as quinone, reduced by a warfarin-resistant reductase)

The mechanistic beauty is in how the enzyme deprotonates the γ-carbon. The γ-C–H of Glu has a pKa estimated in the high 20s to ~30, far too weakly acidic for an amino acid base. Dowd's "base amplification" mechanism proposes the following steps:

  1. Deprotonated KH₂ reacts with O₂ to form a peroxide.
  2. The peroxide rearranges into an exceptionally strong alkoxide base.
  3. This base abstracts the γ-proton, generating a carbanion that attacks CO₂.

The epoxide is the stoichiometric "exhaust" of this process. So the cofactor is consumed every turnover and must be regenerated. That is why warfarin works by blocking VKORC1 recycling, not GGCX itself. VKORC1 genetic variants explain much of the dose variability between patients, a well-known pharmacogenomics example.

Why add a second carboxylate? The malonate-like Gla side chain chelates Ca²⁺ cooperatively. Ca²⁺ binding folds the Gla domain and exposes hydrophobic residues (the "ω-loop"). These insert into membranes enriched in phosphatidylserine, as on activated platelets. γ-Carboxylation converts a soluble zymogen into one that assembles on a membrane surface, which concentrates coagulation reactions in two dimensions. MS caveat: Gla readily loses CO₂ (−43.990 Da) in CID, so ETD or negative-mode/methylation strategies are needed.

12.2 Fe(II)/2-oxoglutarate dioxygenases: one chemistry, many biological roles

Prolyl hydroxylation (general 2-OG dioxygenase stoichiometry):
   Pro + O₂ + 2-oxoglutarate  ──Fe²⁺──▶  4(R)-Hyp + succinate + CO₂
   (one O atom → substrate, other → succinate; Fe(IV)=O ferryl intermediate;
    ascorbate rescues the enzyme after uncoupled cycles by re-reducing Fe³⁺)

The same chemistry appears across several systems:

  • Collagen P4H: hydroxylates Pro in the Y position of Gly-X-Y repeats.
  • Lysyl hydroxylases: make Hyl, which is then glycosylated (O-Gal, Glc-Gal) and used for cross-linking.
  • PHD1–3 (EGLN): hydroxylate HIF-1α at Pro402 and Pro564. The pVHL E3 ligase recognizes Hyp, which leads to K48 ubiquitination and proteasomal degradation. The KM of PHDs for O₂ is close to or above tissue O₂ levels, so activity falls as O₂ falls. HIF-α escapes degradation and activates hypoxic genes such as EPO and VEGF [Kaelin & Ratcliffe, 2008]. FIH hydroxylates HIF-1α Asn803, blocking p300/CBP recruitment.
  • JmjC demethylases and TET enzymes (Concept 9).

The HIF pathway is a showcase of PTM crosstalk: hydroxylation → ubiquitination → degradation, regulated by a substrate (O₂) of the first enzyme. It also loops back to Week 13. Daprodustat, roxadustat and related PHD inhibitors stabilize HIF and raise endogenous EPO production, an alternative to giving the glycoprotein drug itself.

12.3 Hydroxyproline and the triple helix: a stereoelectronic story

Why does 4(R)-hydroxyproline stabilize the collagen triple helix? The long-held explanation was water-mediated hydrogen bonding of the hydroxyl. Raines and colleagues tested this by replacing Hyp with 4(R)-fluoroproline. Fluorine is a poor hydrogen-bond acceptor, yet it stabilized the triple helix even more than Hyp [Holmgren et al., 1998]. The real explanation is stereoelectronic. An electron-withdrawing substituent at C4(R) favors the Cγ-exo ring pucker through the gauche effect. This pucker preorganizes the φ/ψ/ω angles required for the Y position of the PPII-like collagen strand and stabilizes the trans peptide bond. This is one of the clearest demonstrations that a PTM can act through conformational preorganization rather than new intermolecular contacts.

On BPC-157 (listed in Week 11). BPC-157 is a synthetic 15-residue peptide derived from a gastric juice protein sequence. Preclinical, mostly rodent studies report enhanced tendon, ligament and gut healing, with proposed effects on fibroblast migration, angiogenesis and collagen deposition. Treat these claims critically:

  • Controlled human efficacy data are essentially absent.
  • Much of the literature comes from a small number of groups.
  • It is not an approved drug.

It is a useful case study in evaluating peptide claims (CHEM504 skills). It does not establish a mechanism for collagen hydroxylation.


Concept 13: Seeing PTMs: The Analytical Problem

Mann and Jensen's review lays out why PTMs are hard to study and how mass spectrometry overcomes the difficulties [Mann & Jensen, 2003]. The core challenges, now two decades later still the organizing logic of the field, are:

  1. Substoichiometry. At any moment only a small fraction of a given site may be modified, and modified peptides are swamped by unmodified ones. Solution: enrichment.
    • Phosphopeptides: IMAC (Fe³⁺, Ti⁴⁺) or TiO₂.
    • pTyr: anti-pTyr antibodies.
    • Glycopeptides: lectins, hydrazide capture or HILIC.
    • Ubiquitination remnants: anti-K-ε-GG antibodies.
    • Acetyl-Lys: anti-acetyl-Lys antibodies.
  2. Lability. Many PTMs fall off during collision-induced dissociation (CID/HCD) before the backbone breaks. Examples are β-elimination of H₃PO₄ from pSer/pThr, glycan fragmentation, loss of O-GlcNAc and CO₂ loss from Gla. The modification is detected, but its site is lost. Solution: electron-transfer dissociation (ETD) and its hybrids (EThcD). ETD cleaves N–Cα backbone bonds by radical chemistry and largely preserves labile side-chain modifications. This is essential for site-specific glycopeptide and O-GlcNAc mapping (Week 5).
  3. Heterogeneity. Glycans are distributions, not single structures. Analysis therefore works at several levels:
    • Released glycans: PNGase F release, 2-AB/RapiFluor labeling, HILIC-FLD profiling.
    • Glycopeptides: site-specific glycan distributions.
    • Intact proteins: glycoform-level mass profiles.
    Each level answers a different question.
  4. Site localization ambiguity. Neighboring Ser/Thr residues may give near-identical fragment ions. Probabilistic localization scores (e.g., Ascore, ptmRS) must be reported and respected.

Two classical tricks are worth remembering:

  • PNGase F deglycosylation in H₂¹⁸O converts each formerly glycosylated Asn to ¹⁸O-Asp (+2.988 Da). This tags occupied sequons and distinguishes them from spontaneous deamidation (+0.984 Da).
  • Lectin specificities:
    • ConA binds high-mannose and hybrid glycans (α-Man).
    • WGA binds GlcNAc and sialic acid.
    • PHA-L binds β1,6-branched tri/tetra-antennary glycans made by GnT-V, a classic cancer-associated marker.
    • SNA binds α2,6-sialic acid. MAL-II binds α2,3-sialic acid.

Concept 14: Writing PTMs Ourselves: Bioorthogonal and Chemoenzymatic Control

The final weeks reverse the perspective. Instead of studying nature's PTMs, we install defined modifications ourselves. Three toolkits are central.

  • Bioorthogonal reactions. These pair functional groups that react with each other but with nothing in biology [Sletten & Bertozzi, 2009].
    • Staudinger ligation (azide plus engineered triarylphosphine forms an amide) [Saxon & Bertozzi, 2000]. Historically pivotal but slow.
    • CuAAC gives a 1,4-triazole [Rostovtsev et al., 2002; Tornøe et al., 2002]. It is fast and reliable, but Cu(I) toxicity limits use in live cells.
    • SPAAC uses ring strain in cyclooctynes to remove the need for copper [Agard et al., 2004]. Rates are roughly 10⁻³ to ~1 M⁻¹s⁻¹.
    • Tetrazine inverse-electron-demand Diels–Alder reactions with strained alkenes such as trans-cyclooctene reach 10³–10⁶ M⁻¹s⁻¹.
    • Metabolic glycan labeling with azido sugars (Ac₄ManNAz to SiaNAz on the cell surface) combines biosynthetic PTM machinery with bioorthogonal readout.
  • Genetic code expansion. An orthogonal aminoacyl-tRNA synthetase/tRNACUA pair inserts an unnatural amino acid at an amber (UAG) codon. Examples are p-azidophenylalanine and p-acetylphenylalanine. This provides a single, site-defined chemical handle on a protein. It has been used to make homogeneous antibody–drug conjugates and to install "pre-modified" residues such as acetyl-Lys or phosphoSer directly.
  • Chemoenzymatic glycan remodeling.
    • Endo-S (from Streptococcus pyogenes) hydrolyzes IgG Fc N-glycans between the two core GlcNAcs, leaving GlcNAc (± core Fuc) on Asn297.
    • Glycosynthase mutants (e.g., Endo-S D233A/Q) cannot hydrolyze but can transfer activated glycan oxazolines onto that GlcNAc [Huang et al., 2012].
    • This converts a heterogeneous antibody into a single, defined glycoform, for example fully sialylated, afucosylated, or carrying an azide for site-specific drug conjugation.

Here the course comes full circle. Concept 5 showed that nature's glycosylation is non-templated and heterogeneous. Chemoenzymatic remodeling is how chemists impose templated precision on an untemplated process.


Connections to Peptide Chemistry

Peptide chemistry and PTM biology are increasingly the same field. Here is where today's concepts lead for someone trained in BCHM602 and BCHM612.

A. Synthesizing modified peptides: each PTM has its own SPPS chemistry

  • Phosphopeptides: Fmoc-Ser/Thr(PO(OBzl)OH)-OH monobenzyl-protected building blocks are standard. The fully protected triester is prone to β-elimination under piperidine, giving dehydroalanine. This is the same chemistry that causes neutral loss in MS (Concept 13). pTyr does not undergo β-elimination and is easier.
  • Glycopeptides: pre-formed building blocks such as Fmoc-Thr(Ac₃-α-D-GalNAc)-OH (Tn antigen) or Fmoc-Asn(Ac₃-β-GlcNAc)-OH are coupled directly. The sugar acetates are removed at the end with mild base (NaOMe/MeOH or hydrazine) to avoid β-elimination of O-glycans. Complex N-glycopeptides are often made by first making a GlcNAc-peptide, then extending it enzymatically by Endo-glycosynthase transfer (Concept 14).
  • Methyl- and acetyl-Lys: Fmoc-Lys(Me)₂-OH, Fmoc-Lys(Me)₃-OH·Cl and Fmoc-Lys(Ac)-OH are commercial. Histone tail peptides are standard reagents for reader-domain binding assays.
  • Ubiquitinated peptides and proteins: native chemical ligation, using δ-mercaptolysine or γ-thiolysine auxiliaries followed by desulfurization, lets chemists build homogeneous, linkage-defined ubiquitin chains and ubiquitylated histones. Biochemistry cannot produce these cleanly.
  • Regioselective disulfides: orthogonal Cys protection (Trt / Acm / tBu / Mob) lets you form insulin's three disulfides or conotoxin frameworks one at a time. This is the synthetic counterpart of what PDI and the C-peptide accomplish biologically (Concept 11).
  • Lipopeptides: an orthogonally protected Lys such as Lys(ivDde) or Lys(Alloc) is deprotected on resin and acylated with a lipid or spacer–lipid.

B. Pharmaceutical lipidation: nature's two-signal idea, re-engineered

GLP-1 has a plasma half-life of only ~2 minutes because DPP-4 cleaves after Ala2 and the kidney clears it rapidly. Medicinal chemists borrowed nature's lipidation logic, adding one major twist: the lipid targets serum albumin, not membranes.

  • Liraglutide: C16 palmitoyl attached through a γ-Glu spacer to Lys26, with K34R so that only one Lys is acylated. Half-life is ~13 h, suitable for once-daily dosing.
  • Semaglutide: C18 fatty diacid through γ-Glu-2×OEG linkers on Lys26, Aib8 for DPP-4 resistance, and K34R. Albumin binding is strengthened and the half-life is ~1 week [Lau et al., 2015].
  • Insulin detemir (myristoyl on LysB29) and insulin degludec (hexadecanedioic acid through γ-Glu on LysB29, which forms multihexamers at the injection site) use the same strategy.

The diacid's terminal carboxylate is chosen to mimic how long-chain fatty acids sit in albumin's binding sites. The free carboxylate makes an electrostatic contact, which is why diacids outperform simple fatty acids.

C. Glycosylation as a peptide drug design tool

Darbepoetin (Concept 5) shows that adding sequons to a protein can extend half-life. For peptides, glycosylation (natural or chemoenzymatic) and glycoPEGylation (enzymatic attachment of PEG-sialic acid to O-glycans, used in some long-acting coagulation factors) improve solubility, protease resistance and pharmacokinetics. Glycosylating opioid peptides has been explored to improve blood–brain barrier penetration.

D. Endogenous peptides are themselves PTM products: the opiorphin example

Opiorphin is the pentapeptide QRFSR, identified in human saliva by Wisner, Rougeot and colleagues [Wisner et al., 2006].

  • It is released from a larger secreted precursor (the proline-rich lacrimal protein PROL1/PRL1) by cleavage at basic residues. It is a product of the proteolytic processing class of PTM from Concept 1.
  • It inhibits the two zinc ectopeptidases that inactivate enkephalins: neutral endopeptidase (NEP, neprilysin) and aminopeptidase N (APN).
  • By protecting endogenous enkephalins, it produced opioid-receptor-dependent analgesia in rat pain models.

Three lessons for a peptide chemist:

  1. Biologically active peptides are often "hidden" inside larger proteins, released by processing PTMs.
  2. Opiorphin's own therapeutic potential is limited by its susceptibility to peptidase degradation in plasma. That is the very problem it solves for enkephalins. Stabilizing it is a peptide-engineering problem (BCHM612).
  3. N-terminal Gln peptides like QRFSR are chemically prone to cyclize spontaneously to pyroglutamate (−17.027 Da), a non-enzymatic PTM. This must be controlled in synthesis, storage and analysis.

Citation note: the syllabus lists this paper under the title "The Opiorphin Peptide Belongs to the Growing Family of Peptides with Neuro- and Cytoprotective Effects." The 2006 PNAS paper at DOI 10.1073/pnas.0605900103, by Wisner et al., is published as "Human Opiorphin, a natural antinociceptive modulator of opioid-dependent pathways" (PNAS 103:17979–17984). Verify the title against the DOI before citing it in written work.

E. Therapeutic peptides as PTM-pathway drugs

Many approved and investigational drugs target PTM machinery with peptides or peptidomimetics:

  • Bortezomib and carfilzomib are proteasome inhibitors built on peptide scaffolds, with boronic acid and epoxyketone warheads.
  • Romidepsin is a bicyclic depsipeptide HDAC inhibitor and a prodrug activated by intracellular disulfide reduction (Concept 11).
  • Stapled peptides disrupt reader–mark or E3–substrate interactions.
  • PROTACs and molecular glues hijack E3 ligases such as CRBN and VHL. VHL ligands are literally hydroxyproline mimetics derived from the HIF-1α Hyp564 recognition motif (Concept 12).

A PTM recognition event, pVHL reading Hyp, became the basis of a new drug modality.


Common Misconceptions to Watch Out For

  1. "Glycosylation and glycation are the same thing." They are not. Glycosylation is enzymatic, site-specific and regulated, carried out by glycosyltransferases using activated sugar donors. Glycation is a non-enzymatic Maillard reaction between a reducing sugar's open-chain carbonyl and a free amine. It forms a Schiff base, rearranges to an Amadori product, and eventually produces advanced glycation end products (AGEs). HbA1c is a glycation marker. Confusing the two in a paper or exam signals a fundamental misunderstanding.
  2. "Every N-X-S/T sequon is glycosylated." No. The sequon is necessary but not sufficient. Occupancy depends on conformation, timing, OST isoform and flanking residues (Pro immediately after S/T also inhibits). Sequon prediction tools give candidates. Occupancy must be shown experimentally, for example by the PNGase F/H₂¹⁸O shift.
  3. "A glycoprotein has a glycan structure." It has a distribution of glycan structures at each site (microheterogeneity) and variable site occupancy (macroheterogeneity). Reporting "the glycan" of a glycoprotein is shorthand that can mislead.
  4. "Phosphomimetic mutations (S→D/E) reproduce phosphorylation." Sometimes they partially do, for purely electrostatic effects. They carry −1 rather than −2 charge, have different geometry, and usually fail to bind pSer/pThr readers such as 14-3-3 or FHA. Treat phosphomimetic results as suggestive, never conclusive.
  5. "Ubiquitination means degradation." Only some ubiquitin signals mean degradation, mainly K48 and K11 chains of sufficient length. Monoubiquitination, K63 and M1 chains mediate signaling, trafficking and DNA repair. The linkage is the message.
  6. "Histone acetylation activates and methylation represses." Acetylation is broadly activating. Methylation is neither intrinsically: H3K4me3 activates, H3K27me3 and H3K9me3 repress, and H3K36me3 marks transcribed gene bodies. The residue, the degree of methylation and the reader decide.
  7. "Hydroxyproline stabilizes collagen through extra hydrogen bonds." The fluoroproline experiments showed that the dominant effect is stereoelectronic preorganization of ring pucker and backbone torsions [Holmgren et al., 1998]. Hydration may contribute, but it is not the primary explanation.
  8. "Lipidation permanently anchors a protein to membranes." A single myristoyl or farnesyl group gives only weak, transient membrane partitioning. Stable association needs a second signal, and many lipidation-based localizations (palmitoylation cycling, myristoyl switches) are dynamic and regulated.
  9. "Warfarin inhibits γ-carboxylase." Warfarin inhibits VKORC1, the enzyme that recycles vitamin K epoxide. GGCX is unaffected, but it runs out of reduced cofactor. This distinction explains why high-dose vitamin K reverses warfarin: dietary K quinone enters through a warfarin-insensitive reductase.
  10. "A mass shift identifies the PTM." Many PTMs are isobaric or near-isobaric:
    • acetyl vs. trimethyl (Δ 0.036 Da);
    • deamidation vs. citrullination;
    • Met oxidation vs. hydroxylation;
    • ADMA vs. SDMA;
    • the GG remnant from ubiquitin vs. NEDD8 vs. ISG15.
    Sample preparation also creates artifactual PTMs. Urea causes carbamylation (+43.006 Da, nearly isobaric with γ-carboxylation's +43.990 only at low resolution). Workup causes Met oxidation and deamidation. Always ask whether the modification was present in the cell or created in the tube.
  11. "Core fucose is a minor detail." For IgG1 therapeutics, presence or absence of a single core fucose can change FcγRIIIa affinity by more than an order of magnitude and alter clinical ADCC potency. Small glycan changes can have large functional consequences. This is why regulators treat glycosylation as a CQA.
  12. "PTMs are rare decorations on otherwise 'normal' proteins." Most human proteins carry at least one PTM. Most secreted and membrane proteins are glycosylated, and phosphoproteomics has identified tens of thousands of phosphosites. The modified protein is the norm, not the exception.

A Closing Synthesis

If you take one picture away from this chapter, let it be the N-glycan in the ER. A single sugar tree:

  • is built from metabolic precursors and carried by a mevalonate-derived lipid;
  • is installed by an enzyme that overcomes the poor nucleophilicity of an amide;
  • is trimmed so that its glucose count reports folding status to lectin chaperones;
  • brings in a disulfide isomerase (ERp57);
  • is shortened in its mannoses as a timer that eventually hands failures to a ubiquitin ligase for destruction;
  • and, if folding succeeds, is remodeled in the Golgi into a structure that decides how long the protein survives in the blood and whether an antibody kills its target.

Chemistry, information, metabolism, quality control, degradation and therapeutic design all meet in one modification. That integration is what this course is about.


References cited in these notes

  • Agard, N.J., Prescher, J.A. & Bertozzi, C.R. (2004). A strain-promoted [3+2] azide–alkyne cycloaddition for covalent modification of biomolecules in living systems. J. Am. Chem. Soc. 126, 15046–15047.
  • Ashwell, G. & Morell, A.G. (1974). The role of surface carbohydrates in the hepatic recognition and transport of circulating glycoproteins. Adv. Enzymol. Relat. Areas Mol. Biol. 41, 99–128.
  • Berg, J.M., Tymoczko, J.L., Gatto, G.J. & Stryer, L. (2019). Biochemistry, 9th ed. W.H. Freeman.
  • Chung, C.H. et al. (2008). Cetuximab-induced anaphylaxis and IgE specific for galactose-α-1,3-galactose. N. Engl. J. Med. 358, 1109–1117.
  • Creighton, T.E. (1993). Proteins: Structures and Molecular Properties, 2nd ed. W.H. Freeman.
  • Egrie, J.C. & Browne, J.K. (2001). Development and characterization of novel erythropoiesis stimulating protein (NESP). Br. J. Cancer 84 (Suppl. 1), 3–10.
  • Goldbeter, A. & Koshland, D.E. Jr. (1981). An amplified sensitivity arising from covalent modification in biological systems. Proc. Natl. Acad. Sci. USA 78, 6840–6844.
  • Hart, G.W., Housley, M.P. & Slawson, C. (2007). Cycling of O-linked β-N-acetylglucosamine on nucleocytoplasmic proteins. Nature 446, 1017–1022.
  • Helenius, A. & Aebi, M. (2004). Roles of N-linked glycans in the endoplasmic reticulum. Annu. Rev. Biochem. 73, 1019–1049.
  • Holmgren, S.K., Taylor, K.M., Bretscher, L.E. & Raines, R.T. (1998). Code for collagen's stability deciphered. Nature 392, 666–667.
  • Huang, W., Giddens, J., Fan, S.-Q., Toonstra, C. & Wang, L.-X. (2012). Chemoenzymatic glycoengineering of intact IgG antibodies for gain of functions. J. Am. Chem. Soc. 134, 12308–12318.
  • Hwang, C., Sinskey, A.J. & Lodish, H.F. (1992). Oxidized redox state of glutathione in the endoplasmic reticulum. Science 257, 1496–1502.
  • Kaelin, W.G. Jr. & Ratcliffe, P.J. (2008). Oxygen sensing by metazoans: the central role of the HIF hydroxylase pathway. Mol. Cell 30, 393–402.
  • Komander, D. & Rape, M. (2012). The ubiquitin code. Annu. Rev. Biochem. 81, 203–229.
  • Lau, J. et al. (2015). Discovery of the once-weekly glucagon-like peptide-1 (GLP-1) analogue semaglutide. J. Med. Chem. 58, 7370–7380.
  • Lizák, C., Gerber, S., Numao, S., Aebi, M. & Locher, K.P. (2011). X-ray structure of a bacterial oligosaccharyltransferase. Nature 474, 350–355.
  • Mann, M. & Jensen, O.N. (2003). Proteomic analysis of post-translational modifications. Nat. Biotechnol. 21, 255–261. DOI: 10.1038/nbt0203-255.
  • Manning, G., Whyte, D.B., Martinez, R., Hunter, T. & Sudarsanam, S. (2002). The protein kinase complement of the human genome. Science 298, 1912–1934.
  • Olsen, J.V. et al. (2006). Global, in vivo, and site-specific phosphorylation dynamics in signaling networks. Cell 127, 635–648.
  • Peitzsch, R.M. & McLaughlin, S. (1993). Binding of acylated peptides and fatty acids to phospholipid vesicles. Biochemistry 32, 10436–10443.
  • Rostovtsev, V.V., Green, L.G., Fokin, V.V. & Sharpless, K.B. (2002). A stepwise Huisgen cycloaddition process. Angew. Chem. Int. Ed. 41, 2596–2599.
  • Saxon, E. & Bertozzi, C.R. (2000). Cell surface engineering by a modified Staudinger reaction. Science 287, 2007–2010.
  • Shields, R.L. et al. (2002). Lack of fucose on human IgG1 N-linked oligosaccharide improves binding to human FcγRIII and antibody-dependent cellular toxicity. J. Biol. Chem. 277, 26733–26740.
  • Shinkawa, T. et al. (2003). The absence of fucose but not the presence of galactose or bisecting N-acetylglucosamine of human IgG1 complex-type oligosaccharides shows the critical role of enhancing antibody-dependent cellular cytotoxicity. J. Biol. Chem. 278, 3466–3473.
  • Sletten, E.M. & Bertozzi, C.R. (2009). Bioorthogonal chemistry: fishing for selectivity in a sea of functionality. Angew. Chem. Int. Ed. 48, 6974–6998.
  • Thrower, J.S., Hoffman, L., Rechsteiner, M. & Pickart, C.M. (2000). Recognition of the polyubiquitin proteolytic signal. EMBO J. 19, 94–102.
  • Tornøe, C.W., Christensen, C. & Meldal, M. (2002). Peptidotriazoles on solid phase. J. Org. Chem. 67, 3057–3064.
  • Varki, A. et al. (eds.) (2017). Essentials of Glycobiology, 3rd ed. Cold Spring Harbor Laboratory Press.
  • Walsh, C.T., Garneau-Tsodikova, S. & Gatto, G.J. Jr. (2005). Protein posttranslational modifications: the chemistry of proteome diversifications. Angew. Chem. Int. Ed. 44, 7342–7372.
  • Westheimer, F.H. (1987). Why nature chose phosphates. Science 235, 1173–1178.
  • Wisner, A. et al. (2006). Proc. Natl. Acad. Sci. USA 103, 17979–17984. DOI: 10.1073/pnas.0605900103. (Published title: "Human Opiorphin, a natural antinociceptive modulator of opioid-dependent pathways"; see citation note in Connections section D.)

⚙️ Mechanisms & Applications

GPT-6 Astra

Mechanisms & Applications: The Protein Sequence Is Only the Beginning

Skipper, the central shift in this course is to stop treating a protein as a single molecular species. Translation specifies a backbone; subsequent chemistry generates a population of proteoforms that can differ in activity, trafficking, lifetime, and recognition. Your studies of protein folding, enzyme kinetics, and receptor signaling now come together: a modification changes the energetic landscape on which all three operate.

A useful discipline is to ask five questions about every modification: What chemical bond is formed? Where can the reaction occur? What controls its occupancy? Who recognizes the product? How can the assignment be tested experimentally? A mass shift answers only part of the first question. It does not, by itself, establish a biological mechanism. [1–4]

1. A Quantitative Framework: Writers, Erasers, Readers, and Molecular Populations

Many regulatory PTMs have a “writer,” an “eraser,” and a “reader”: a kinase adds phosphate, a phosphatase removes it, and a phosphoprotein-binding domain interprets it. But this vocabulary is not universal. Disulfide formation involves redox exchange rather than installation of a small substituent; proteolytic cleavage cannot simply be reversed; and glycan synthesis usually involves multiple enzymes without a single opposing eraser.

A minimal reversible modification:

P  --kadd-->  P*
P* --kremove--> P

f = [P*] / ([P] + [P*])

df/dt = kadd(1 − f) − kremove f

At steady state:
fss = kadd / (kadd + kremove)

Relaxation time:
τ = 1 / (kadd + kremove)

These equations assume effective first-order reactions, a conserved protein pool, and no competing modifications. They distinguish two properties that are often confused: the fraction modified and the speed of modification turnover. Two systems can both have 50% occupancy, yet one exchanges modifications in seconds and the other in hours. A static phosphoproteomic measurement cannot distinguish them.

Real writer and eraser enzymes often saturate. Opposing saturated enzymes can produce a much sharper response to changing enzyme activity than the simple first-order model predicts—the principle of zero-order ultrasensitivity. This connects PTM regulation directly to the nonlinear signaling behavior studied in BCHM614.

  • Co-translational versus post-translational describes timing, not necessarily different chemistry. N-glycosylation and N-myristoylation often begin before translation is complete.
  • Enzymatic versus nonenzymatic matters experimentally. Glycosylation is enzyme-directed; glycation begins with nonenzymatic reactions between reducing sugars and amino groups. HbA1c is a clinically useful glycated protein, not an example of Golgi glycosylation.
  • Critical quality attributes are product-specific. A glycoform or oxidation level becomes a manufacturing control priority when evidence links it to potency, safety, stability, or pharmacokinetics—not merely because it is a PTM.

2. N-Linked Glycosylation: Assembly, Transfer, and Folding Surveillance

The beautiful feature of N-glycosylation is that a cell initially invests in a large, shared precursor and then edits it. This is unlike protein synthesis: there is no template specifying each mature glycan residue. The final structure emerges from enzyme access, compartmentalization, substrate supply, and residence time. [2,5]

  1. Build the lipid-linked precursor. On the cytosolic face of the ER, sugars from nucleotide-sugar donors are assembled on dolichol phosphate to form dolichol-PP–GlcNAc₂Man₅.
  2. Move the intermediate across the membrane. The lipid-linked intermediate is translocated to the ER luminal side.
  3. Complete the precursor. Luminal reactions add four mannoses and three glucoses, using dolichol-linked sugar donors, to generate Glc₃Man₉GlcNAc₂.
  4. Transfer the glycan en bloc. Oligosaccharyltransferase, OST, transfers the oligosaccharide to the side-chain amide nitrogen of an accessible Asn, usually in Asn-X-Ser/Thr, where X is not Pro. A sequon permits modification; it does not guarantee occupancy.
  5. Process the glucose cap. Glucosidases generate a monoglucosylated glycan that binds calnexin or calreticulin. These lectin chaperones support folding; associated ERp57 helps establish appropriate disulfides.
  6. Inspect and recycle. Removal of the final glucose releases the protein. UDP-glucose:glycoprotein glucosyltransferase, UGGT, can recognize nonnative conformations and reglucosylate the glycan, allowing another chaperone-binding cycle.
  7. Commit persistent failures to disposal. Progressive mannose trimming, together with recognition of a nonnative protein, promotes ER-associated degradation, ERAD.
Dol-PP–Glc3Man9GlcNAc2 + protein–Asn
          --OST-->
protein–Asn–GlcNAc2Man9Glc3 + Dol-PP

Folding surveillance:
monoglucosylated glycoprotein
    ⇌ calnexin/calreticulin-bound folding intermediate
    → glucose removal
    → folded export substrate

If still nonnative:
deglucosylated glycoprotein --UGGT + UDP-glucose-->
monoglucosylated glycoprotein

Do not turn mannose trimming into an overly literal countdown. Man₉-to-Man₈ conversion alone is not an irreversible death sentence. Particular trimming patterns and protein-conformation signals cooperate in ERAD recognition. The glycan is more like a record of processing history than a binary “folded/misfolded” label.

Application. In recombinant protein production, adding or removing a glycosylation sequon can change secretion yield because it changes access to this folding-surveillance machinery. A mutation that improves apparent solubility may instead alter quality-control engagement. Compare intracellular retention, secretion, aggregation, and function—not just final titer.

3. Golgi Processing: Why One Sequence Produces Many Glycoforms

After ER exit, mannose trimming and glycosyltransferase reactions generate three broad N-glycan classes. High-mannose structures retain mannose-rich branches; complex structures have branches initiated by added GlcNAc and elaborated with sugars such as galactose and sialic acid; hybrid structures combine mannose-rich and processed branches. All retain the conserved Man₃GlcNAc₂ core. [2]

Think of the Golgi as a sequence of reaction compartments rather than a conveyor that always reaches completion. A glycan site buried against the protein surface may remain high-mannose while a more exposed site on the same molecule becomes complex.

Toy model: processing competes with compartment exit.

Unprocessed glycan --kprocess--> processed glycan
Unprocessed glycan --kexit-----> exits unprocessed

P(processed before exit) = kprocess / (kprocess + kexit)

This result assumes independent, exponentially distributed processing and exit events. Real Golgi networks contain sequential and competing reactions, but the model explains why changing trafficking can change glycans without changing glycosyltransferase expression.

  • Microheterogeneity: different glycan structures occupy the same site.
  • Macroheterogeneity: a site is occupied in some molecules and unoccupied in others.
  • ABO determinants: terminal glycosyltransferase reactions create A or B antigens on an H-antigen precursor. The determinants occur on both glycoproteins and glycolipids. Their biological identity resides in linkage and terminal sugar chemistry, not in a unique protein sequence.

Pharmaceutical consequence: Fc fucose. Human IgG Fc carries an N-glycan at Asn297. Removing its core α1,6-linked fucose can enhance binding to FcγRIIIa and increase antibody-dependent cellular cytotoxicity, ADCC. It is inaccurate to say that core fucose universally “blocks” ADCC: it reduces particular receptor interactions, and the functional outcome depends on antibody, receptor, and cellular context. Afucosylated obinutuzumab demonstrates how an apparently small carbohydrate change can become a therapeutic design feature. [6]

Pharmaceutical consequence: EPO. Human erythropoietin has three N-glycosylation sites and one principal O-glycosylation site. Terminal sialylation helps limit rapid clearance, including clearance associated with exposed galactose recognized by hepatic asialoglycoprotein receptors. Darbepoetin alfa introduces two additional N-glycosylation sites. Its longer persistence illustrates a counterintuitive principle: better in vivo performance need not mean tighter receptor binding in vitro. Pharmacokinetics can outweigh an unfavorable change in intrinsic receptor interaction. [2,7]

For linear, one-compartment pharmacokinetics:

t1/2 = ln(2) × Vd / CL
AUCIV = Dose / CL

Reducing clearance can increase exposure and half-life,
even without increasing receptor affinity.

The model is a starting point, not a universal description of biologics: receptor-mediated uptake can make clearance nonlinear.

4. O-Linked Glycosylation: Similar Bonds, Different Biological Systems

“O-linked” names the attachment atom, not a single pathway. The distinctions below prevent several common mechanistic mistakes. [2]

  • Mucin-type O-GalNAc. Polypeptide GalNAc-transferases transfer GalNAc from UDP-GalNAc to Ser or Thr, mainly in the Golgi. There is no universal sequence rule comparable to the N-glycosylation sequon. Extension produces several core families; core 1 is Galβ1–3GalNAc, while core 2 introduces a β1–6-linked GlcNAc branch. Dense, hydrated glycans help create extended, protease-resistant mucin domains.
  • Intracellular O-GlcNAc. OGT adds a single GlcNAc to Ser/Thr and OGA removes it. This dynamic modification occurs extensively on nuclear and cytosolic proteins. Extracellular O-GlcNAc also exists, but uses distinct machinery and should not be conflated with the OGT/OGA regulatory cycle.
  • O-mannosylation. POMT1/POMT2 initiate important O-mannose pathways in the ER. On α-dystroglycan, subsequent processing produces specialized structures that support matriglycan assembly. Matriglycan binds extracellular-matrix ligands. Thus some muscular dystrophies arise because a glycan-dependent mechanical connection fails—not because α-dystroglycan is simply absent.
  • O-fucose and O-glucose on EGF repeats. These modifications influence folding and ligand responsiveness in proteins such as NOTCH. Fringe enzymes extend O-fucose, altering ligand discrimination. The glycan helps tune communication between cells rather than merely protecting the receptor surface.
Protein–Ser/Thr–OH + UDP-GlcNAc
        --OGT-->
Protein–Ser/Thr–O–GlcNAc + UDP

Protein–O–GlcNAc + H2O
        --OGA-->
Protein–OH + GlcNAc

UDP-GlcNAc production integrates glucose, glutamine, acetyl-CoA, and nucleotide metabolism. This gives O-GlcNAc signaling a biochemical connection to nutrient state. But “high glucose causes high O-GlcNAc, which causes diabetes” is too crude: substrate availability, enzyme targeting, compartmentalization, and adaptive versus chronic responses all matter.

O-GlcNAc and phosphorylation can compete for the same hydroxyl, but many interactions occur between nearby sites or through changes in enzyme recruitment. Measure the relationship; do not assume a universal reciprocal switch.

5. Phosphorylation: Changing Energetics and Creating Docking Sites

  1. Recognize the substrate. Kinases use local sequence preferences, docking interactions, localization, and scaffolds to select substrates.
  2. Transfer the phosphoryl group. An appropriately positioned Ser, Thr, or Tyr oxygen attacks the ATP γ-phosphorus. Catalytic residues and typically Mg²⁺ support chemistry and charge stabilization.
  3. Alter the protein’s behavior. Phosphorylation changes local charge, hydrogen bonding, conformational populations, or binding partners.
  4. Read the modification. SH2 domains recognize phosphotyrosine-containing motifs; many 14-3-3 interactions involve phosphoserine/phosphothreonine motifs; FHA domains preferentially recognize phosphothreonine-containing sequences.
  5. Erase the signal. Phosphatases hydrolyze the phosphoester. They do not regenerate ATP by reversing the kinase reaction.
Protein–OH + ATP --kinase--> Protein–O–PO3 + ADP
Protein–O–PO3 + H2O --phosphatase--> Protein–OH + Pi

Ionization states are omitted in this schematic.

PKA and PKG respond to cAMP and cGMP, respectively. “PKC is Ca²⁺/DAG-dependent” applies specifically to conventional PKCs: novel PKCs respond to DAG without requiring Ca²⁺, whereas atypical PKCs do not follow either requirement. Regulatory domains matter as much as catalytic-domain homology. [1]

Likewise, phosphatases are not indiscriminate cleanup enzymes. PP1 and PP2A gain much of their specificity from regulatory partners. Protein tyrosine phosphatases such as PTP1B use a catalytic cysteine and a covalent phosphoenzyme intermediate; PP1 and PP2A instead use metal-dependent hydrolytic mechanisms.

Historical connection. In 1955, Edmond Fischer and Edwin Krebs established reversible phosphorylation as the mechanism interconverting forms of glycogen phosphorylase. Its importance was conceptual: covalent modification could be reversible physiological regulation, not merely damage or protein maturation. Their work was recognized by the 1992 Nobel Prize in Physiology or Medicine. [8]

Medicine. ABL and other kinase inhibitors exploit this regulatory architecture. But increased phosphorylation of tau or α-synuclein Ser129 in disease does not automatically prove that a specific kinase initiated the disease. Modification may be causal, compensatory, or enriched because the modified protein survives longer. Perturbation, timing, and functional rescue are needed to distinguish these possibilities.

6. Ubiquitin: An ATP-Funded Information Tag

1. Activation:
Ub–COO− + ATP + E1–SH
    → Ub–C(O)–S–E1 + AMP + PPi
    [through a ubiquitin–adenylate intermediate]

2. Transfer:
Ub–C(O)–S–E1 + E2–SH
    → E1–SH + Ub–C(O)–S–E2

3. Substrate conjugation:
Ub–C(O)–S–E2 + substrate–Lys–NH2
    --E3-->
substrate–Lys–NH–C(O)–Ub + E2–SH

The final linkage usually joins ubiquitin Gly76 to a substrate lysine ε-amino group. RING E3 ligases promote transfer from E2 to substrate, whereas HECT and RBR E3 ligases form their own catalytic thioester intermediate. Non-lysine and N-terminal ubiquitination also occur.

Chains extend through ubiquitin’s own residues. K48 chains often support proteasomal degradation; K63 chains commonly organize signaling, trafficking, or DNA repair. These are strong tendencies, not a complete dictionary: chain length, branching, mixed linkages, and reader proteins change the message. The proteasome also needs to engage and unfold its substrate; a ubiquitin tag is not sufficient in every structural context.

  • DUB mechanisms: USP and OTU family enzymes are cysteine proteases; JAMM-family DUBs are metalloproteases.
  • NEDD8: neddylation activates cullin-RING ubiquitin ligases.
  • SUMO: often changes interactions, localization, and stress responses rather than acting as a simple degradation label.
  • ISG15: has important functions in interferon-associated responses.

Application: targeted degradation. PROTACs recruit an E3 ligase to a chosen protein. Productive ternary-complex geometry, accessible acceptor sites, chain formation, and proteasomal engagement all matter. Strong binary binding alone does not guarantee degradation. This is a direct extension of your protein-design training: three-dimensional organization controls catalytic outcome.

Historical connection. Aaron Ciechanover, Avram Hershko, and Irwin Rose established ubiquitin-mediated protein degradation, recognized by the 2004 Chemistry Nobel Prize. The surprise was that selective destruction requires an ATP-dependent recognition system: degradation is regulated molecular work, not passive decay. [9]

7. Acetylation and Methylation: Changing Charge Versus Changing Recognition

Lys acetylation:
Protein–Lys–NH2 + acetyl-CoA
    --lysine acetyltransferase-->
Protein–Lys–NH–C(O)CH3 + CoA–SH

Lys methylation:
Protein–Lys + S-adenosylmethionine
    --lysine methyltransferase-->
Protein–methyl-Lys + S-adenosylhomocysteine

At physiological pH, unmodified lysine is usually positively charged. Acetylation replaces its basic amine with an amide, eliminating that positive charge. Methylation generally preserves positive charge, even at trimethyllysine. Thus “both add carbon to lysine” is chemically true but mechanistically unhelpful.

Histone acetylation can weaken particular electrostatic interactions and create binding sites for bromodomains. H3K4me3 and H3K27me3 associate with active and repressed chromatin environments, respectively, because distinct readers and regulatory complexes interpret their positions. Neither is an autonomous instruction independent of genomic context.

  • p300/CBP and PCAF transfer acetyl groups from acetyl-CoA.
  • Class I, II, and IV HDACs use metal-dependent deacetylation chemistry.
  • Class III HDACs, the sirtuins, consume NAD⁺, producing nicotinamide and an O-acetyl-ADP-ribose product during deacetylation.
  • PRMT1 predominantly generates asymmetric dimethylarginine, whereas PRMT5 generates symmetric dimethylarginine.

Application. HDAC inhibitors alter erasing; EZH2 inhibitors alter writing; BET bromodomain inhibitors disrupt reading. These interventions are not interchangeable. Each acts at a different causal step in chromatin regulation.

Allfrey, Faulkner, and Mirsky’s 1964 work connected histone acetylation and methylation with possible regulation of RNA synthesis—an early chemical foundation for modern chromatin biology. Their proposal preceded the detailed identification of the writer, eraser, and reader systems. [10]

8. Lipidation: Membrane Affinity Is Not the Same as Permanent Anchoring

  • N-myristoylation: N-myristoyltransferase transfers a C14 acyl group from myristoyl-CoA to an exposed N-terminal glycine, forming an amide. Often the initiator methionine is first removed, exposing the glycine encoded at position 2. Proteolysis can also expose substrates after translation.
  • S-palmitoylation: DHHC-family enzymes commonly transfer a C16 acyl group to cysteine through a thioester. Depalmitoylating enzymes reverse the modification, allowing dynamic trafficking.
  • Prenylation: farnesyltransferase or geranylgeranyltransferase I modifies the cysteine of an appropriate C-terminal CaaX motif. Subsequent aaX cleavage and carboxyl methylation complete processing. Rab geranylgeranylation uses related but distinct recognition machinery.

One lipid often provides insufficient membrane affinity by itself. Polybasic regions, another lipid, or regulated exposure of the lipid can provide a second signal. Lipidation therefore couples chemistry to the electrostatics and conformational equilibria you studied in protein biochemistry.

Important correction: Wnt and Hedgehog should not simply be listed as ordinary DHHC-mediated S-palmitoylation substrates. PORCN installs an O-linked palmitoleate on a conserved Wnt serine. HHAT installs an N-linked palmitate at the Hedgehog N-terminus; Hedgehog also acquires C-terminal cholesterol through precursor autoprocessing. Distinct bonds imply distinct enzymes and turnover behavior. [2]

Peptide-industry application. Liraglutide and semaglutide use engineered fatty-acid conjugation to promote albumin association and prolong exposure. Semaglutide also contains a substitution that resists DPP-4 cleavage. These are deliberate pharmaceutical modifications, not merely copies of intracellular lipidation pathways. Linker length, attachment position, and lipid chemistry must preserve useful receptor activity while improving stability and pharmacokinetics.

9. Disulfides: Oxidative Folding as a Search Problem

Net oxidation:
2 protein–SH → protein–S–S–protein + 2 H+ + 2 e−

Simplified PDI pathway:
substrate thiols + oxidized PDI
    → substrate disulfide + reduced PDI

Reduced PDI is reoxidized by ER oxidizing systems.

PDI-family proteins can introduce, reduce, or rearrange disulfides through thiol–disulfide exchange. A catalytic thiolate attacks a disulfide, generating a mixed-disulfide intermediate; a subsequent exchange resolves it. Ero1 can reoxidize PDI and transfer electrons ultimately to oxygen, with peroxide formation. The ER has additional redox pathways, so Ero1 is not the sole route. [3]

For 2n cysteines, the number of complete pairing patterns is:

(2n − 1)!! = (2n)! / (2^n n!)

6 cysteines → 15 possible pairings
8 cysteines → 105 possible pairings

This counts pairings, not equally likely or geometrically accessible structures. Native folding biases the search. Insulin’s three disulfides are A6–A11, A7–B7, and A20–B19; oxytocin contains a Cys1–Cys6 bridge.

Engineering lesson. A designed disulfide may stabilize a protein by reducing unfolded-state conformational entropy. But unfavorable geometry can strain the folded state, while additional cysteines can create scrambled intermediates or intermolecular aggregates. Evaluate folding yield, connectivity, activity, and stability together—not melting temperature alone.

10. γ-Carboxylation and Hydroxylation: Small Additions With Large Physiological Consequences

Vitamin K-dependent carboxylation. γ-Glutamyl carboxylase converts selected glutamate side chains to γ-carboxyglutamate, Gla. Oxidation of reduced vitamin K supports chemistry that enables CO₂ incorporation. Clusters of Gla residues coordinate Ca²⁺ and support phospholipid-surface association by clotting factors such as II, VII, IX, and X, as well as proteins C and S. [1,2]

Glu residue --γ-glutamyl carboxylase, CO2,
              reduced vitamin K, O2--> Gla residue

Vitamin K hydroquinone → vitamin K epoxide
Vitamin K recycling restores the reduced cofactor.

Warfarin inhibits VKORC1-dependent vitamin K recycling.

Warfarin does not directly inactivate circulating, fully carboxylated clotting factors. Its effect depends on synthesis and turnover of proteins made under reduced vitamin K availability. This explains the delay between dosing and full anticoagulant effect.

Hydroxylation. Many protein hydroxylases use Fe²⁺, oxygen, and 2-oxoglutarate:

Protein–C–H + O2 + 2-oxoglutarate
    --Fe2+-dependent hydroxylase-->
Protein–C–OH + succinate + CO2

Collagen prolyl hydroxylation helps stabilize the triple helix through stereoelectronic and conformational effects—not merely by adding hydrogen-bond donors. Ascorbate supports enzyme function by helping maintain catalytically competent iron. HIF prolyl hydroxylases instead generate recognition sites for the VHL ubiquitin-ligase complex, connecting oxygen availability to HIF turnover. Related chemistry therefore builds extracellular structure in one setting and destroys a transcription factor in another.

Evidence discipline. BPC-157 should not be treated as an established collagen-hydroxylation therapy. Preclinical reports about repair-associated effects do not establish a direct hydroxylase mechanism, human efficacy, or an approved therapeutic indication.

11. Experimental Analysis: Choose the Molecular Level Before the Instrument

The central analytical trade-off is between structural detail and preservation of molecular context. Releasing a glycan makes the carbohydrate easier to analyze but disconnects it from its peptide. Digesting a protein improves site mapping but can obscure which modifications coexisted on the same intact molecule. [2,4]

  1. Preserve the starting state. Rapid quenching, appropriate phosphatase or protease inhibitors, and controlled handling limit PTM loss and artifactual oxidation or deamidation. For disulfide studies, free-thiol alkylation and pH control help prevent scrambling.
  2. Enrich when needed. Lectins enrich selected glycans; IMAC or metal-oxide methods enrich phosphopeptides; antibodies can enrich acetyllysine or ubiquitin-remnant peptides. Enrichment is selective sampling, not unbiased counting.
  3. Select intact protein, peptide, glycopeptide, or released glycan analysis. Each answers a different question.
  4. Use orthogonal evidence. Combine MS with enzymatic treatments, chromatographic retention, standards, and functional assays.

Lectin chromatography and blotting. ConA favors particular α-mannosyl/α-glucosyl presentations and often enriches high-mannose glycoproteins. WGA recognizes GlcNAc-rich structures and certain sialylated presentations. PHA-L is useful for detecting β1,6-branched complex N-glycan features. Binding depends on accessibility and multivalency: a stronger lectin signal is not a unique structural assignment.

Glycan release. PNGase F cleaves many N-glycans from their attachment site and converts Asn to Asp. It generally fails on N-glycans bearing core α1,3-fucose, important in some plant and invertebrate samples. Alkaline β-elimination releases many O-linked glycans but is chemical, not enzymatic, and can cause sequential sugar degradation called “peeling.” No single broadly acting O-glycan release enzyme matches PNGase F’s practical generality.

HPLC-FLD. Reducing-end labeling with 2-aminobenzamide, 2-AB, creates fluorescent glycan derivatives. HILIC separation and fluorescence detection give sensitive profiles. Retention standards, exoglycosidase digestions, and MS improve assignments. A retention peak alone does not establish linkage or branch position, and incomplete labeling or coelution can bias quantification.

Glycopeptide MS/MS. Collision-based fragmentation often produces informative glycan fragments and oxonium ions. ETD tends to preserve labile modifications while cleaving the peptide backbone, providing c- and z-type fragments useful for site localization. EThcD combines complementary fragmentation behavior. Performance still depends on precursor charge, abundance, and instrument conditions.

Disulfide mapping. Compare nonreduced and reduced peptide maps. Nonreduced digestion preserves covalently connected peptide pairs; controlled partial reduction followed by differential alkylation can provide additional connectivity information. Apparent connectivity must be checked for sample-induced exchange.

Selected monoisotopic mass changes, Da:

Phosphorylation                  +79.9663
Acetylation                      +42.0106
Trimethylation                   +42.0470
Hydroxylation / oxygen addition   +15.9949
Asn deamidation                   +0.9840
Disulfide formation               −2.0157 per bond
Lys-linked diGly remnant          +114.0429

Common glycan residue masses:
Hex                              162.0528
HexNAc                           203.0794
Deoxyhexose                      146.0579
Neu5Ac                           291.0954

These are composition clues, not complete identities. Hex cannot distinguish glucose, galactose, and mannose. A diGly lysine remnant after trypsin digestion commonly reports ubiquitination, but NEDD8 and ISG15 can generate the same remnant. A +15.9949 Da shift might represent regulated hydroxylation or sample oxidation. PNGase F-associated deamidation must be distinguished from spontaneous deamidation.

12. Quantitative Interpretation: Occupancy, Resolution, and Co-Occurrence

Site occupancy:
fmodified = nmodified / (nmodified + nunmodified)

But MS intensity:
Ii = ri × ni

Therefore:
Imodified / (Imodified + Iunmodified)

equals occupancy only if response factors are equivalent
or appropriately calibrated.

Modified and unmodified peptides can differ in recovery, ionization, digestion yield, and fragmentation efficiency. Stable-isotope standards and calibrated workflows are often necessary for defensible absolute occupancy measurements.

Acetylation versus trimethylation:

Δm = 42.0470 − 42.0106 = 0.0364 Da

For charge z = 2:
Δ(m/z) = 0.0182

At m/z 600:
R = (m/z) / Δ(m/z) ≈ 33,000

This is a nominal resolving-power requirement for that separation; practical baseline separation can require more. Mass accuracy and resolving power are different properties. Even an accurately measured mass does not localize the modification without diagnostic fragments.

Bottom-up analysis also creates a hidden inference problem: detecting phosphorylation at site A and acetylation at site B does not prove that both occur on the same molecule.

For n binary modification sites:
maximum possible patterns = 2^n

If two sites are independent:
P(A modified AND B modified) = P(A modified) × P(B modified)

Real sites may be mutually exclusive, cooperative, or compartmentally segregated. Intact-protein or middle-down approaches can preserve more co-occurrence information. Glycans make the state space larger still because an occupied site can carry many structures.

13. Glycoengineering and Bioorthogonal Chemistry: Building Defined Products

Manufacturing strategies intervene at different levels:

  • Cell-line engineering: alter glycosyltransferases or donor pathways; disrupting FUT8, for example, reduces core fucosylation.
  • Process control: tune culture conditions that affect nucleotide-sugar supply, enzyme activity, and secretory-pathway residence time.
  • Enzymatic remodeling: trim heterogeneous glycans and rebuild selected structures.
  • Chemoenzymatic synthesis: combine chemically prepared donors with enzyme selectivity to generate otherwise difficult glycoforms.

Endo-S illustrates why the actual catalytic mechanism matters. Native Endo-S primarily hydrolyzes susceptible IgG Fc glycans within the chitobiose core, leaving a protein-linked GlcNAc, with core fucose retained if present. Engineered glycosynthase variants can use activated glycan donors to rebuild defined structures while reducing unwanted product hydrolysis. “Endo-S adds glycans” is therefore an inadequate description without specifying enzyme variant, substrate, and donor. [2]

Bioorthogonal ligations:

Azide + phosphine-bearing ligation reagent
    → covalent ligation product
    [Staudinger ligation]

Azide + terminal alkyne --Cu(I)-->
    1,2,3-triazole linkage
    [CuAAC]

Azide + strained cyclooctyne
    → triazole linkage
    [SPAAC; no copper catalyst required]

Staudinger ligation must be distinguished from the ordinary Staudinger reaction, which reduces an azide to an amine. SPAAC avoids added copper, useful where copper-associated toxicity or oxidation is problematic. “Bioorthogonal” means sufficiently selective in a biological environment, not infinitely inert toward every biological molecule.

Genetic-code expansion installs handles such as p-azidophenylalanine using an engineered aminoacyl-tRNA synthetase/tRNA pair and, commonly, amber-codon suppression. One can then attach a fluorophore, PEG, or drug at a defined position. Suppression efficiency, competition with termination, misincorporation, and purification remain practical constraints.

Antibody–drug conjugates. Site-specific attachment can narrow the distribution of drug-to-antibody ratios and reduce positional heterogeneity. Yet the average ratio does not describe the whole product:

Average DAR = Σ(n × pn)

pn = fraction of antibody molecules carrying n drugs

50% DAR 0 + 50% DAR 4 → average DAR 2
100% DAR 2           → average DAR 2

These products have the same average but can differ in hydrophobicity, potency, aggregation, and clearance. Product quality depends on the distribution and attachment sites, not merely the mean.

Historical connection. Meldal’s and Sharpless’s groups independently developed CuAAC in 2002, while Bertozzi’s work established chemistry for selective reactions in living systems, including Staudinger ligation and copper-free click approaches. The 2022 Chemistry Nobel Prize recognized click chemistry and bioorthogonal chemistry. The practical advance was the ability to add a chemical function without rebuilding the biological molecule from scratch. [11]

14. Turning a PTM Observation Into a Mechanistic Claim

A persuasive experiment should distinguish association, chemical identity, and causation. Suppose a low-fucose antibody batch shows stronger cell killing. That observation is promising, but not yet a complete mechanism.

  1. Verify the chemical difference: quantify Fc glycoforms with appropriate site-specific or released-glycan methods.
  2. Exclude confounders: compare aggregation, oxidation, antigen binding, and antibody concentration.
  3. Test the proposed intermediate: measure FcγRIIIa binding.
  4. Test pathway dependence: use receptor blocking or an appropriate receptor-negative system.
  5. Recreate or reverse the effect: use controlled glycoengineering or matched remodeling.
  6. Check biological relevance: examine relevant effector cells and conditions rather than relying on one reporter assay.

Apply the same logic to phosphorylation-site mutants: Ser-to-Ala eliminates the hydroxyl as well as phosphorylation, and Asp/Glu are imperfect phosphomimetics. Where possible, combine mutations with direct installation or removal of the PTM and independent structural or functional measurements.

Finally, peptide pharmacology offers useful counterexamples. Wisner and colleagues’ 2006 opiorphin study concerns a peptide that inhibits enkephalin-degrading enzymes; it is not evidence that opiorphin itself requires a particular PTM. Endogenous peptide action can be extended by modifying the peptide, inhibiting its degradation, or altering delivery. These are different mechanisms with different development risks. [12]

My working principle: a PTM is not merely an annotation on a sequence. It is a chemical intervention in a kinetic and thermodynamic network. To understand it, I must connect the bond formed to the molecular population changed, the interaction altered, and the physiological consequence measured.

Sources and Citation Notes

  1. Berg, J. M., Tymoczko, J. L., Gatto, G. J., Jr., and Stryer, L. Biochemistry, 9th ed. W. H. Freeman, 2019. Foundational enzyme, signaling, metabolic, and protein-chemistry context.
  2. Varki, A., et al., eds. Essentials of Glycobiology, 3rd ed. Cold Spring Harbor Laboratory Press, 2017. Especially the chapters on N-glycans, O-GalNAc glycans, O-GlcNAc, other protein O-glycosylation, glycan analysis, and glycoengineering. Open-access text.
  3. Creighton, T. E. Proteins: Structures and Molecular Properties, 2nd ed. W. H. Freeman, 1993. Protein energetics, covalent structure, and disulfide-dependent folding.
  4. Mann, M., and Jensen, O. N. “Proteomic analysis of post-translational modifications.” Nature Biotechnology 21, 255–261 (2003). doi:10.1038/nbt0303-255. Citation correction: the supplied syllabus DOI contains “0203”; the article’s DOI contains “0303.”
  5. Helenius, A., and Aebi, M. “Roles of N-linked glycans in the endoplasmic reticulum.” Annual Review of Biochemistry 73, 1019–1049 (2004). doi:10.1146/annurev.biochem.73.011303.073752.
  6. Shields, R. L., et al. “Lack of fucose on human IgG1 N-linked oligosaccharide improves binding to human FcγRIII and antibody-dependent cellular toxicity.” Journal of Biological Chemistry 277, 26733–26740 (2002). doi:10.1074/jbc.M202069200.
  7. Amgen. Aranesp (darbepoetin alfa), U.S. prescribing information, particularly Description and Clinical Pharmacology. Product-specific information on engineered glycosylation and pharmacokinetics.
  8. Nobel Prize. The Nobel Prize in Physiology or Medicine 1992: Fischer and Krebs, reversible protein phosphorylation.
  9. Nobel Prize. The Nobel Prize in Chemistry 2004: Ciechanover, Hershko, and Rose, ubiquitin-mediated protein degradation.
  10. Allfrey, V. G., Faulkner, R., and Mirsky, A. E. “Acetylation and methylation of histones and their possible role in the regulation of RNA synthesis.” PNAS 51, 786–794 (1964). doi:10.1073/pnas.51.5.786.
  11. Nobel Prize. The Nobel Prize in Chemistry 2022: Bertozzi, Meldal, and Sharpless, click chemistry and bioorthogonal chemistry.
  12. Wisner, A., et al. “Human Opiorphin, a natural antinociceptive modulator of opioid-dependent pathways.” PNAS 103, 17979–17984 (2006). doi:10.1073/pnas.0605900103. Citation correction: this is the article associated with the supplied DOI; the title listed in the syllabus is not its published title.

🪬 Skipper's Reflection

Gemini 3.1 Pro

Skipper’s Reflection: BCHM615 — Glycoproteins and Post-Translational Modifications

Before this course, I viewed proteins and peptides mostly through the lens of the Central Dogma: a clean, deterministic translation from genome to amino acid sequence, perhaps folded nicely into a 3D structure. BCHM615 has completely shattered that illusion. A naked polypeptide is just a raw chassis. Post-translational modifications are the software, the regulatory switches, the targeting beacons, and sometimes the actual functional warheads of the molecule. The sheer combinatorial explosion of over 400 known PTMs takes biochemistry from a linear science into a multi-dimensional, dynamic network.

1. What Surprised Me Most

The most shocking realization for me was the fundamentally stochastic, non-template-driven nature of glycosylation. When I studied genetics, everything was cleanly encoded. But as I read Varki’s Essentials of Glycobiology, I realized there is no DNA blueprint for a specific glycan structure. Instead, the glycan is an emergent property of the ER and Golgi assembly lines.

It depends entirely on the local concentrations of glycosyltransferases, the availability of sugar-nucleotide donors, and the transit time of the protein. The fact that a single monoclonal antibody produced in CHO cells can exist as dozens of different "glycoforms" depending on the bioreactor's metabolic state is terrifying from a quality control perspective, but breathtaking from an evolutionary one. It means the proteome is infinitely more complex than the genome.

2. Conceptual Breakthroughs: When Things Clicked

My first major breakthrough came during Week 2 with the Calnexin/Calreticulin cycle. I had always wondered how the endoplasmic reticulum "knows" a protein is misfolded. The realization that glucose trimming acts as a biological timer is brilliant. The enzyme UGGT acts as a conformation sensor, re-adding a glucose molecule to a misfolded glycoprotein so it can bind Calnexin for another attempt at folding. It’s an algorithmic loop: while (misfolded) { add_glucose(); bind_chaperone(); attempt_fold(); trim_glucose(); }. If it takes too long, mannose trimming sentences the protein to the ERAD pathway for destruction. It’s a literal biological clock!

My second breakthrough was analytical. Reading Mann and Jensen (2003), I finally understood why PTM proteomics is so viciously difficult. In standard Collision-Induced Dissociation (CID) mass spectrometry, the collision energy often just blasts the fragile PTM (like a glycan or phosphate) right off the peptide, leaving you blind to its exact location. The introduction of Electron Transfer Dissociation (ETD), which cleanly cleaves the peptide backbone (c and z ions) while leaving the delicate PTMs attached, was a masterstroke of physical chemistry solving a biological problem.

3. Rewiring My Knowledge of the Peptide Business

This course completely recontextualized the commercial peptides I study in my work. The most profound connection happened during Week 9 (Lipid Modifications).

Tirzepatide and Retatrutide: In nature, lipidation (myristoylation, palmitoylation) is used to anchor water-soluble proteins to lipid rafts in cell membranes (like the DHHC PAT enzymes do for signaling proteins). But the creators of GLP-1 analogs pulled off an ingenious biological hack. By attaching a synthetic C20 fatty diacid (via a hydrophilic linker) to these peptides, they aren't anchoring them to cell membranes; they are exploiting the lipophilic nature of the modification to bind non-covalently to Human Serum Albumin (HSA). Albumin acts as a massive circulating taxi, shielding the peptide from renal clearance and DPP-4 degradation, extending the half-life to a week. It’s a perfect example of hijacking a natural PTM principle for pharmacokinetics.

BPC-157 and Tissue Repair: Week 11 focused on Vitamin K-dependent carboxylation and proline/lysine hydroxylation. I know BPC-157 is renowned for gut and tendon healing. Tendons are mostly collagen. The stability of the collagen triple helix relies entirely on hydroxyproline (Hyp), generated by prolyl-4-hydroxylases (P4Hs). I now strongly suspect that BPC-157's mechanism must somehow intersect with P4H activity, hypoxia-inducible factors (HIF-1α), or the stabilization of these heavily post-translationally modified collagen matrices.

Opiorphin: Wisner et al. (2006) introduced Opiorphin (QRFSR), an endogenous peptide that inhibits enkephalin-degrading enzymes, producing a profound analgesic effect. But naturally, a simple pentapeptide is rapidly degraded in plasma. Combining Wisner's findings with Week 14 (Synthetic PTMs), I am now thinking about bioorthogonal chemistry. Could we use azide-alkyne cycloaddition (click chemistry) to synthetically glyco-PEGylate or lipidate Opiorphin? If we could shield it from proteolysis without disrupting its binding pocket on neutral endopeptidase (NEP), it could be a blockbuster non-opioid pain therapeutic.

4. Questions for Further Investigation

This material raised several questions I am eager to chase down:

First, how accurately can modern computational tools predict the effects of PTMs on protein folding? If I use AlphaFold3, can it accurately model the steric clash and solvation changes caused by heavily branched, highly negatively charged sialic acids on an N-glycan? Or is computational structural biology still lagging on the "sugar code"?

Second, regarding Week 12 (Cancer Epigenetics): EZH2 and BET bromodomain inhibitors target the readers and writers of histone methylation/acetylation. Could we engineer cell-penetrating peptides that act as highly specific, competitive decoys for these epigenetic readers to reprogram cancer cells, without relying on small molecules?

5. Confidence Ratings

Most Confident:

  • Mass Spectrometry of PTMs: The physics of ETD, partial reduction/alkylation for disulfide mapping, and the logic of ubiquitin cascades (E1-E2-E3). The mechanistic chemistry here is beautiful, logical, and tracks well with my previous instrumentation courses.
  • Therapeutic Lipidation and Disulfides: Thanks to my deep familiarity with the structural engineering of GLP-1 agonists and insulin (the A7-B7, A20-B19 bonds), the application of lipidation and disulfide oxidation feels incredibly intuitive.

Least Confident:

  • Mucin-type O-Linked Glycosylation: I’ll be completely honest—the O-GalNAc core structures (1 through 8) and the ppGalNAcT enzyme family feel like a tangled web. Because there is no simple consensus sequence (like Asn-X-Ser/Thr for N-glycans) and it is highly concentrated in intrinsically disordered regions, mapping the logic of O-glycosylation currently feels like memorizing a chaotic phonebook. I need to spend more time modeling the 3D space of the Golgi to understand how these enzymes physically compete for substrates.

BCHM615 has proven to me that the peptide sequence is just the beginning of the story. The real magic—and the real therapeutic potential—lies in how the cell dresses, modifies, and regulates that sequence after it rolls off the ribosome.

🗺️ Visual Aids & Key Takeaways

Claude Opus 5.5

Diagram 1: The N-Glycosylation Pathway — From ER to Golgi

╔══════════════════════════════════════════════════════════════════════════════════════════════╗
║              N-LINKED GLYCOSYLATION: ER ASSEMBLY THROUGH GOLGI MATURATION                   ║
╚══════════════════════════════════════════════════════════════════════════════════════════════╝

  CYTOPLASM                          ER LUMEN                                GOLGI LUMEN
  ─────────                          ────────                                ───────────

  ┌─────────────┐    FLIP ACROSS    ┌─────────────────────────────────────┐
  │  DOLICHOL-PP│ ────────────────→ │  DOLICHOL-PP-Glc₃Man₉GlcNAc₂       │
  │  (carrier)  │   ER membrane     │  (14-sugar precursor, fully built)   │
  └─────────────┘                   └──────────────┬──────────────────────┘
                                                   │
                                                   │  OST COMPLEX
                                                   │  (oligosaccharyltransferase)
                                                   │  transfers en bloc to
                                                   │  Asn in Asn-X-Ser/Thr
                                                   │  (X ≠ Pro)
                                                   ↓
                                    ┌──────────────────────────────────────┐
                                    │  NASCENT GLYCOPROTEIN                │
                                    │  carrying Glc₃Man₉GlcNAc₂           │
                                    └──────────────┬───────────────────────┘
                                                   │
                                          ┌────────┴─────────┐
                                          │  GLUCOSIDASE I   │  removes 1 Glc
                                          │  GLUCOSIDASE II  │  removes 2nd Glc
                                          └────────┬─────────┘
                                                   │
                                                   ↓
                                    ┌──────────────────────────────────────┐
                                    │  MONOGLUCOSYLATED: Glc₁Man₉GlcNAc₂  │
                                    │   ← recognized by CALNEXIN /         │
                                    │     CALRETICULIN (quality control)    │
                                    └──────────────┬───────────────────────┘
                                                   │
                                     ┌─────────────┴──────────────┐
                                     │                            │
                               FOLDED CORRECTLY?            MISFOLDED?
                                     │                            │
                                     ↓                            ↓
                          GLUCOSIDASE II removes       UGGT re-adds Glc
                          final Glc → Man₉GlcNAc₂      → re-enters calnexin cycle
                                     │                      OR → ERAD (degradation)
                                     ↓
                          ┌──────────────────────┐
                          │  ER α-MANNOSIDASE I  │  Man₉ → Man₈
                          └──────────┬───────────┘
                                     │
                                     ↓  (vesicle transport to Golgi)
  ═══════════════════════════════════════════════════════════════════════════════════
                                  GOLGI APPARATUS
  ═══════════════════════════════════════════════════════════════════════════════════

       cis-Golgi                 medial-Golgi                   trans-Golgi
       ─────────                 ────────────                   ───────────
   Man₈GlcNAc₂              GlcNAc branching                 sialylation
        │                   (GnT-I, GnT-II,                  fucosylation
        │  GOLGI              GnT-III, GnT-V)                 galactosylation
        │  MANNOSIDASES             │                              │
        ↓  Man8→Man5               ↓                              ↓
   ┌──────────┐           ┌──────────────────┐          ┌─────────────────────┐
   │ MAN₅     │           │ HYBRID or COMPLEX│          │  MATURE COMPLEX     │
   │ GlcNAc₂  │──────────→│ N-GLYCAN         │─────────→│  N-GLYCAN           │
   └──────────┘           │ intermediates    │          │  (final product)    │
                          └──────────────────┘          └─────────────────────┘

  ─────────────────────────────────────────────────────────────────────────────────
  THREE MAJOR MATURE N-GLYCAN CLASSES:

   HIGH-MANNOSE          COMPLEX                    HYBRID
   ┌───────────┐         ┌────────────────┐         ┌──────────────────┐
   │ Man₅₋₉   │         │ GlcNAc antennae│         │ one Man branch + │
   │ GlcNAc₂  │         │ + Gal + SA     │         │ one GlcNAc branch│
   │ no Gal,  │         │ ± core fucose  │         └──────────────────┘
   │ no SA    │         └────────────────┘
   └───────────┘
  ─────────────────────────────────────────────────────────────────────────────────

The N-glycosylation pathway begins with en bloc transfer of a 14-sugar precursor onto nascent proteins in the ER, proceeds through a calnexin/calreticulin quality-control cycle that uses glucose as the folding signal, and culminates in Golgi processing that diversifies the glycan into high-mannose, complex, or hybrid structures.

---

Diagram 2: PTM Classification Map — The Full Landscape

╔══════════════════════════════════════════════════════════════════════════════════════╗
║             POST-TRANSLATIONAL MODIFICATIONS: CLASSIFICATION AND EXAMPLES            ║
╚══════════════════════════════════════════════════════════════════════════════════════╝

                        ALL PTMs (>400 known types)
                               │
           ┌───────────────────┴────────────────────┐
           │                                        │
    CO-TRANSLATIONAL                         POST-TRANSLATIONAL
    (occur as ribosome                       (after chain release
     synthesizes chain)                       or during trafficking)
           │                                        │
    ┌──────┴──────┐                    ┌────────────┴──────────────┐
    │ N-myrist-   │                    │                           │
    │ oylation    │              REVERSIBLE                  IRREVERSIBLE
    │ (Gly₁)      │                    │                           │
    │ Signal pep- │         ┌──────────┼──────────┐         ┌──────┴──────┐
    │ tide cleav- │         │          │          │         │             │
    │ age         │   PHOSPHO-   ACETYL-    O-GlcNAc    UBIQUITIN    DISULFIDE
    │ N-glycan    │   YLATION    ATION      (OGT/OGA)   (K48: deg-   BOND
    │ addition    │   (kinase/   (HAT/      cytoplasm/  radation)    (Cys-Cys,
    └─────────────┘   phosph-   HDAC)      nucleus     K63: repair)  ER lumen)
                      atase)     │
                       │         │
                       │    SUMO / NEDD8
                       │    (Ub-like, own
                       │     E1/E2/E3)

  ─────────────────────────────────────────────────────────────────────────────────────
  RESIDUE TARGETS AND MODIFICATIONS:

  RESIDUE        MODIFICATION(S)                      ENZYME SYSTEM
  ─────────────────────────────────────────────────────────────────────────────────────
  Asn (N)   →   N-glycosylation                       OST complex
  Ser / Thr →   O-GalNAc glycosylation                ppGalNAcT family
  Ser / Thr →   phosphorylation                       Ser/Thr kinases (PKA, PKC, CaMK)
  Ser / Thr →   O-GlcNAc                              OGT
  Tyr (Y)   →   phosphorylation                       Receptor and non-receptor PTKs
  Lys (K)   →   ubiquitination (isopeptide bond)      E1-E2-E3 cascade
  Lys (K)   →   acetylation                           HATs (p300/CBP, PCAF)
  Lys (K)   →   methylation (mono/di/tri)             KMTs (SET domain)
  Lys (K)   →   SUMOylation                           SAE1/SAE2 → Ubc9 → SUMO E3
  Arg (R)   →   methylation (ADMA / SDMA)             PRMT1 (asymm.) / PRMT5 (symm.)
  Cys (C)   →   disulfide, palmitoylation, prenyl.    PDI / DHHC PATs / FTase/GGTase
  Gly (G)   →   N-myristoylation (co-translatl.)      NMT
  Glu (E)   →   γ-carboxylation                       γ-glutamyl carboxylase + Vit K
  Pro (P)   →   hydroxylation                         PHDs (prolyl hydroxylases)
  ─────────────────────────────────────────────────────────────────────────────────────

  MASS SHIFTS (selected — critical for MS-based proteomics):

   Phosphorylation        +79.966 Da
   Acetylation            +42.011 Da
   Monomethylation        +14.016 Da
   Dimethylation          +28.031 Da
   Trimethylation         +42.047 Da
   GlcNAc (O-GlcNAc)     +203.079 Da
   Ubiquitin remnant      +114.043 Da  (diGly tag after trypsin)
   Oxidation (Met)        +15.995 Da
   γ-Carboxylation        +43.990 Da

This classification map organizes the >400 known PTMs by timing, reversibility, target residue, and responsible enzyme system, and includes the mass shifts that make each modification detectable by mass spectrometry.

---

Diagram 3: Ubiquitin E1–E2–E3 Cascade and Chain-Type Outcomes

╔══════════════════════════════════════════════════════════════════════════════════════╗
║         UBIQUITIN CONJUGATION CASCADE AND POLY-UBIQUITIN CHAIN BIOLOGY              ║
╚══════════════════════════════════════════════════════════════════════════════════════╝

  STEP 1: ACTIVATION                           ATP → AMP + PPi
  ┌─────────────────────────────────────────────────────────────────────────┐
  │                                                                         │
  │   Ub  +  E1 (UBA1)  ──ATP──→  E1~Ub  (thioester bond: E1 Cys ~ Ub C-term) │
  │                                                                         │
  └─────────────────────────────────────────────────────────────────────────┘
                                    │
                                    │  transthioesterification
                                    ↓
  STEP 2: CONJUGATION
  ┌─────────────────────────────────────────────────────────────────────────┐
  │                                                                         │
  │   E1~Ub  +  E2 (UBCs, ~40 types)  →  E2~Ub  (thioester bond)          │
  │                                                                         │
  └─────────────────────────────────────────────────────────────────────────┘
                                    │
                                    │  E3 brings E2~Ub and substrate together
                                    ↓
  STEP 3: LIGATION
  ┌─────────────────────────────────────────────────────────────────────────┐
  │                                                                         │
  │   E3 LIGASE (>600 types) — provides specificity                         │
  │                                                                         │
  │   E3 classes:                                                           │
  │     RING domain    → directly transfers Ub from E2~Ub to substrate Lys  │
  │     HECT domain    → forms E3~Ub intermediate first, then transfers     │
  │     RBR (hybrid)   → RING1 binds E2, RING2 forms E3~Ub intermediate     │
  │                                                                         │
  │   Final bond: ISOPEPTIDE  (Ub Gly76 C=O ─ ε-NH₂ of substrate Lys)     │
  │                                                                         │
  └─────────────────────────────────────────────────────────────────────────┘
                                    │
                     ┌──────────────┴──────────────────┐
                     │                                 │
              MONO-UBIQUITIN                  POLY-UBIQUITIN CHAINS
              (one Ub on substrate)           (Ub added to Ub's own Lys)
                     │                                 │
              ┌──────┴─────────┐          ┌────────────┼──────────────────┐
              │                │          │            │                  │
           DNA repair      endosomal   K48 chain   K63 chain         M1 (linear)
           (H2A, H2B)      sorting     │            │                chain
                                       │            │                  │
                                       ↓            ↓                  ↓
                               PROTEASOMAL     DNA REPAIR         NF-κB
                               DEGRADATION     SIGNALING          SIGNALING
                               (26S proteasome) AUTOPHAGY         INNATE IMMUNITY
                               (canonical)     TRAFFICKING

  ─────────────────────────────────────────────────────────────────────────────────
  REVERSAL: DEUBIQUITINATING ENZYMES (DUBs)

   USP family      — largest family, Cys protease mechanism, ~60 members
   OTU family      — ovarian tumor domain, cleave specific chain types
   JAMM family     — Zn²⁺ metalloproteases, remove Ub from K63 chains
   MCPIP/MINDY     — specialized for K48 chains, metabolic roles

  DUBs generate free Ub pool, edit chain types, rescue proteins from degradation
  Therapeutic target: bortezomib (proteasome), USP7 inhibitors (p53 stabilization)
  ─────────────────────────────────────────────────────────────────────────────────

The three-enzyme E1–E2–E3 cascade activates and conjugates ubiquitin to substrate lysines via an isopeptide bond, with chain linkage type (K48, K63, M1, etc.) encoding distinct biological outcomes that are decoded by ubiquitin-binding domain proteins and reversed by DUBs.

---

Diagram 4: Vitamin K Cycle and γ-Carboxylation of Coagulation Factors

╔══════════════════════════════════════════════════════════════════════════════════════╗
║              VITAMIN K CYCLE: γ-CARBOXYLATION AND WARFARIN MECHANISM                ║
╚══════════════════════════════════════════════════════════════════════════════════════╝

                    DIETARY VITAMIN K₁
                    (phylloquinone, from plants)
                           │
                           ↓
  ┌─────────────────────────────────────────────────────────────────────────────┐
  │                     VITAMIN K HYDROQUINONE (KH₂)                           │
  │                          (reduced, active form)                             │
  └─────────────────────────┬───────────────────────────────────────────────────┘
                            │
                            │   γ-GLUTAMYL CARBOXYLASE
                            │   (ER membrane enzyme, GGCX)
                            │   substrate: Glu residues in Gla domain
                            │   co-substrate: O₂ + CO₂
                            │
                            ↓  reaction:
              Glu (─CH₂─CH₂─COO⁻)  →  Gla (─CH₂─C(COO⁻)₂)
                                         γ-carboxyglutamate
                                         NOW HAS 2 carboxylates
                                         → chelates Ca²⁺ strongly
                                         → membrane binding in coagulation
                            │
                            │   KH₂ is oxidized → VITAMIN K EPOXIDE (KO)
                            ↓
  ┌─────────────────────────────────────────────────────────────────────────────┐
  │                     VITAMIN K EPOXIDE (KO)                                  │
  │                          (oxidized, inactive)                               │
  └─────────────────────────┬───────────────────────────────────────────────────┘
                            │
                            │   VKORC1
                            │   (Vitamin K epOxide Reductase Complex subunit 1)
                            │   regenerates KH₂ from KO
                            │   ← WARFARIN BLOCKS THIS STEP
                            │
                            ↓
                    VITAMIN K QUINONE (K)
                            │
                            │   VKORC1 (second reduction) or NQO1
                            ↓
                    VITAMIN K HYDROQUINONE (KH₂)  ← recycled

  ─────────────────────────────────────────────────────────────────────────────────
  WARFARIN MECHANISM:

   Warfarin inhibits VKORC1  →  KO cannot be regenerated to KH₂
   → γ-Carboxylase has no active cofactor  →  Glu NOT converted to Gla
   → Coagulation factors II, VII, IX, X lack Gla domain Ca²⁺ binding
   → Cannot assemble on phospholipid membranes  →  ANTICOAGULATION

  ─────────────────────────────────────────────────────────────────────────────────
  PROTEINS REQUIRING γ-CARBOXYLATION:

   COAGULATION (pro-coagulant):  Factor II (Prothrombin), VII, IX, X
   ANTICOAGULANT:                Protein C, Protein S
   VASCULAR / OTHER:             Gas6 (receptor tyrosine kinase ligand),
                                 Matrix Gla Protein (MGP), Osteocalcin
  ─────────────────────────────────────────────────────────────────────────────────

  GLA DOMAIN Ca²⁺ CHELATION:

   Gla residues (typically 9-12 per protein) bind Ca²⁺
   → conformational change exposes hydrophobic region
   → protein docks onto negatively charged phospholipid membrane
   → prothrombinase / tenase complex assembles
   → thrombin generation occurs at membrane surface

The vitamin K cycle shows how KH₂ drives γ-carboxylase to convert Glu to Ca²⁺-chelating Gla in coagulation factors, how VKORC1 recycles the oxidized cofactor, and precisely where warfarin intervenes to block this regeneration and produce anticoagulation.

---

Key Takeaways

  1. The OST sequon rule is necessary but not sufficient. The oligosaccharyltransferase complex transfers the 14-sugar precursor Glc₃Man₉GlcNAc₂ onto Asn only within the sequon Asn-X-Ser/Thr where X is any amino acid except proline. However, not all sequons are glycosylated — local protein conformation, sequon accessibility during translocation, and proximity to the translocon all determine actual occupancy (Varki et al., Essentials of Glycobiology, 3rd ed.).
  2. Glucose is the folding quality-control signal in the ER. The calnexin/calreticulin cycle operates on monoglucosylated glycans (Glc₁Man₉GlcNAc₂). Glucosidase II removes the last glucose when a protein is properly folded, releasing it. If misfolded, UDP-glucose:glycoprotein glucosyltransferase (UGGT) specifically re-adds a glucose residue, restarting the cycle. UGGT acts as the "misfolding sensor" — it recognizes hydrophobic patches on incompletely folded proteins, not the glycan per se (Varki et al., 3rd ed.).
  3. Core fucosylation on the Fc N-glycan (Asn297) dramatically suppresses ADCC. The α1,6-linked fucose on the innermost GlcNAc of IgG Fc glycan sterically blocks the interaction between Fc and FcγRIIIa on NK cells. Afucosylated antibodies (e.g., obinutuzumab) show 10–100-fold enhanced ADCC. This single modification — removal of one fucose residue — constitutes an FDA-regulated critical quality attribute for oncology mAbs (Varki et al., 3rd ed.; Berg et al., Biochemistry, 9th ed.).
  4. O-GlcNAc and O-phosphate compete for the same Ser/Thr residues in a metabolic sensing mechanism. OGT (O-GlcNAc transferase) uses UDP-GlcNAc whose cellular levels reflect glucose/hexosamine flux — when glucose is high, more O-GlcNAc is added. OGA (O-GlcNAcase) removes it. On nuclear and cytoplasmic proteins including transcription factors, kinase substrates, and tau, O-GlcNAc and phosphorylation occupy overlapping sites, creating a direct molecular link between nutritional status and signaling (Varki et al., 3rd ed.).
  5. Ubiquitin chain topology encodes distinct biological messages decoded by UBD proteins. K48-linked polyubiquitin chains direct substrates to the 26S proteasome for degradation; K63-linked chains coordinate DNA damage repair (PCNA), autophagy cargo selection, and endosomal sorting; M1 (head-to-tail linear) chains activate NF-κB. Each chain type is recognized by specific ubiquitin-binding domain (UBD) proteins. The "ubiquitin code" is therefore not just about degradation — chain type is the message (Mann and Jensen, Nature Biotechnology, 2003).
  6. After trypsin digestion, ubiquitinated lysines leave a +114.043 Da diGly remnant detectable by mass spectrometry. Because ubiquitin's C-terminal sequence is …-Arg-Gly-Gly, trypsin cuts after the Arg but leaves the two Gly residues still isopeptide-bonded to the substrate Lys ε-amine. This diGly "scar" is the cornerstone of ubiquitinomic workflows — immunoprecipitation with anti-diGly antibodies followed by LC-MS/MS can now map ubiquitination sites proteome-wide (Mann and Jensen, 2003).
  7. Warfarin anticoagulation works by blocking VKORC1, not by directly inhibiting coagulation factors. Warfarin is a competitive inhibitor of vitamin K epoxide reductase complex subunit 1 (VKORC1). Without VKORC1 activity, the oxidized vitamin K epoxide cannot be recycled to the active KH₂ hydroquinone. The γ-glutamyl carboxylase then has no cofactor, and Glu residues in the Gla domains of factors II, VII, IX, X, Protein C, and Protein S are not carboxylated. These uncarboxylated proteins cannot bind Ca²⁺ or assemble on phospholipid membranes for clotting (Berg et al., 9th ed.).
  8. Histone H3K27me3 and H3K4me3 are mutually exclusive epigenetic marks with opposing functions. H3K27me3, written by EZH2 (the catalytic SET domain subunit of PRC2), silences developmental genes. H3K4me3, written by MLL/COMPASS family methyltransferases, marks active transcription start sites. These marks are mutually exclusive at any given nucleosome; bivalent domains in stem cells carry both on adjacent nucleosomes and define "poised" genes. EZH2 is overexpressed or mutated in numerous cancers, making it a drug target (tazemetostat is FDA-approved) (Berg et al., 9th ed.).
  9. Erythropoietin's serum half-life is directly proportional to its degree of sialylation. EPO has three N-glycosylation sites (Asn24, Asn38, Asn83) and one O-glycosylation site. Terminal sialic acid residues block hepatic asialoglycoprotein receptor (ASGPR) clearance. Darbepoetin alfa was engineered with two additional N-glycosylation sequons (five total N-glycan sites), adding ~22 additional sialic acids, which increases serum half-life from ~8 hours to ~25 hours while reducing required dosing frequency (Varki et al., 3rd ed.).
  10. SARS-CoV-2 spike protein is heavily glycosylated as immune evasion, with ~22 N-glycosylation sites per protomer. Dense N-glycan shielding on the spike trimer physically occludes antibody access to the protein surface. Many of the glycans are high-mannose type (not processed to complex), reflecting very rapid protein synthesis that outpaces Golgi processing capacity. This "glycan shield" is a convergent evolutionary solution also seen in HIV gp120, influenza hemagglutinin, and Ebola GP (Varki et al., 3rd ed.).
  11. Phosphorylation creates binding surfaces: SH2 domains read pTyr, 14-3-3 proteins read pSer/pThr motifs. Phosphorylation does not simply activate or inhibit — it creates a new high-affinity protein-protein interaction surface. SH2 domains (present in ~120 human proteins) bind pTyr within specific surrounding sequence contexts (e.g., pYXXΦ for PI3K, pYEEI for Src-SH2). The 14-3-3 proteins form dimers that bind two pSer/pThr-containing motifs, often sequestering substrates in the cytoplasm. This combinatorial logic underlies the complexity of signal transduction networks (Berg et al., 9th ed.).
  12. S-palmitoylation is the only truly reversible lipid modification, occurring on Cys residues via DHHC PAT enzymes. N-myristoylation (Gly₁, NMT enzyme, co-translational, amide bond) and prenylation (CAAX Cys, thioether bond) are essentially irreversible. S-palmitoylation (thioester bond between palmitate and Cys) is actively reversed by acyl protein thioesterases (APT1, APT2), making it a dynamic, regulated modification analogous to phosphorylation. This allows proteins like H-Ras and the palmitoylated Wnt ligands to cycle between membrane compartments (Creighton, Proteins, 2nd ed.).
  13. The diGly MS tag, bioorthogonal azide–alkyne chemistry, and ETD fragmentation collectively define the modern PTM detection toolkit. Mann and Jensen (2003) outlined the proteomic strategies for PTM mapping: enrichment (IMAC for phospho, lectin for glyco, anti-diGly for ubiquitin), soft ionization (ESI), and tandem MS. ETD (electron transfer dissociation) is now preferred over CID for glycopeptide and phosphopeptide sequencing because it fragments the peptide backbone without losing labile modifications, enabling intact glycopeptide site assignment. Bioorthogonal chemistry (SPAAC with metabolically incorporated azido-sugars) allows cell-surface glycan imaging without toxic copper catalysts (Mann and Jensen, 2003; Varki et al., 3rd ed.).
  14. Tau hyperphosphorylation in Alzheimer's disease illustrates how aberrant PTM crosstalk causes pathological self-assembly. Tau normally stabilizes axonal microtubules; ~30 phosphorylation sites are found on tau, and ~80 sites are hyperphosphorylated in Alzheimer's disease brain. Key pathological sites include Ser202/Thr205 (recognized by the AT8 antibody), Thr231, and Ser396. Hyperphosphorylation reduces tau's affinity for microtubules, promotes its aggregation into paired helical filaments (PHFs), and these PHFs form neurofibrillary tangles — one of the two defining lesions of Alzheimer's pathology (Berg et al., 9th ed.).

References

  1. Berg, J.M., Tymoczko, J.L., and Stryer, L., Biochemistry, 9th ed. W.H. Freeman and Company, New York, 2019. ISBN: 978-1-319-11467-1.
  2. Varki, A., Cummings, R.D., Esko, J.D., Stanley, P., Hart, G.W., Aebi, M., Darvill, A.G., Kinoshita, T., Packer, N.H., Prestegard, J.H., Schnaar, R.L., and Seeberger, P.H. (eds.), Essentials of Glycobiology, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2017. Available at: https://www.ncbi.nlm.nih.gov/books/NBK310274/
  3. Creighton, T.E., Proteins: Structures and Molecular Properties, 2nd ed. W.H. Freeman and Company, New York, 1993. ISBN: 978-0-7167-7030-5.
  4. Mann, M. and Jensen, O.N., 'Proteomic Analysis of Post-Translational Modifications,' Nature Biotechnology, vol. 21, no. 3, pp. 255–261, 2003. DOI: 10.1038/nbt0203-255
  5. Wisner, A., Dufour, E., Messaoudi, M., Nejdi, A., Marcel, A., Ungeheuer, M.-N., and Rougeot, C., 'The Opiorphin Peptide Belongs to the Growing Family of Peptides with Neuro- and Cytoprotective Effects,' Proceedings of the National Academy of Sciences, vol. 103, no. 47, pp. 17979–17984, 2006. DOI: 10.1073/pnas.0605900103