Why DEL Building Blocks Self-Cleave During Boc Deprotection — and How to Prevent It

What self-cleavage looks like on LC-MS

The clearest illustration in our data set is an oxetane-pyrrolidine diamine building block (internally, BB070). Immediately after conjugation, HPLC-MS of the crude reaction showed a single dominant, correctly-massed species: the intact oligonucleotide-PEG12-BB-Boc conjugate (calculated 16,639.9 Da, found 16,640.0 Da). After subjecting that same conjugate to standard Boc deprotection conditions (borate buffer, 80 °C, 48 h), the chromatogram changed completely: the major species shifted to an earlier retention time and a lower mass, matching bare oligonucleotide-PEG12-NH₂ with no building block attached at all (calculated 16,357.7 Da, found 16,357.2 Da) — the properly deprotected, BB-intact conjugate was only a minor peak. In other words, the deprotection conditions didn’t just remove the Boc group; for this building block, they removed the entire building block.

Trigger 1: a free amine positioned to attack the linker

The oxetane-diamine case above is the clean example of the first mechanism: once Boc deprotection liberates a free primary amine somewhere on the building block, if that amine sits in a favorable geometry relative to the amide bond linking the BB to the PEG12 linker, it can attack that amide intramolecularly and cleave the building block off — effectively an intramolecular transamidation. This isn’t universal even within a structurally similar series: of two enantiomers of the same oxetane scaffold tested, this cleavage was significant for one but not observed for the related oxetanes elsewhere in the same series, underscoring how geometry-dependent the effect is.

The same pattern showed up in a 4-carboxypyrrazole subseries substituted with pyrrolidine: the pyrrolidine nitrogen, once freed, was positioned for the same kind of intramolecular attack, and Boc deprotection gave low yields. The corresponding piperidinyl-substituted version (one ring size larger) showed the same effect much less strongly — piperidine’s larger ring puts the nitrogen in a less favorable geometry for attack than pyrrolidine’s does.

Trigger 2: oxazolone formation from amide-substituted building blocks

A second, distinct mechanism appeared in two unrelated scaffold series — azetidines and piperidines, each substituted at the ring nitrogen with an amide. In both cases, conjugation to the oligonucleotide went well, but Boc deprotection caused extensive or complete cleavage of the building block. The proposed mechanism is internal oxazolone formation: at the elevated temperature required for Boc removal, the amide substituent can cyclize onto the adjacent linker amide, displacing it and releasing the building block. We didn’t isolate the oxazolone intermediate directly, but the diagnostic signature was consistent with this pathway — LC-MS after deprotection showed only bare oligonucleotide, with no trace of the intended BB-conjugate.

Trigger 3: electron-poor heterocycles weaken the linker amide itself

A third pattern doesn’t require a newly liberated amine at all. 5-carboxypyrimidines substituted with pyrrolidine or piperidine conjugated cleanly as the Boc derivative, but the amide bond to the linker was completely cleaved during Boc deprotection anyway. The pyrimidine ring’s electron-poor character makes the adjacent amide bond intrinsically more labile — more so than the already-sensitive 4-carboxypyrazoles — so it doesn’t survive 48 hours at 80 °C in basic buffer regardless of geometry.

The fix is the same in every case: switch to pNs

For all three failure modes, the solution we used was the same — synthesize the corresponding pNs-protected building block instead of the Boc version. Because pNs deprotection runs at room temperature for one hour (aqueous NAC/DBU) instead of 80 °C for 48 hours, it doesn’t give these cleavage pathways time or thermal energy to compete. On the pyrimidine scaffold specifically, switching to pNs kept the linker amide fully intact and recovered good conjugate yields on a building block that Boc deprotection destroyed completely. The same recovery held for the pyrrolidine-substituted 4-carboxypyrrazoles: pNs derivatives deprotected cleanly with no BB loss where Boc had given low yields.

A practical checklist

Before committing a new building block to Boc deprotection at scale, it’s worth checking the structure against these three patterns:

• Does deprotection free a primary or secondary amine anywhere near the linker amide? If geometry allows a 5- or 6-membered transition state for intramolecular attack, expect cleavage risk — pyrrolidine-type geometries are higher risk than piperidine-type ones.
• Is there an amide substituent on the ring bearing the Boc group? Oxazolone-forming cleavage is a risk under prolonged heat, independent of the core scaffold.
• Is the acid attached to an electron-poor heterocycle (pyrimidines, certain carboxypyrazole regiochemistries)? The linker amide itself may be too labile for 80 °C/48 h regardless of what else is on the molecule.

Any building block that trips one of these flags is a strong candidate to make as the pNs derivative from the start, rather than discovering the cleavage after a failed Boc deprotection.