Design and synthesis of digitally encoded polymers that can be decoded and erased

Biopolymers such as DNA store information in their chains using controlled sequences of monomers. Here we describe a non-natural information-containing macromolecule that can store and retrieve digital information. Monodisperse sequence-encoded poly(alkoxyamine amide)s were synthesized using an iterative strategy employing two chemoselective steps: the reaction of a primary amine with an acid anhydride and the radical coupling of a carbon-centred radical with a nitroxide. A binary code was implemented in the polymer chains using three monomers: one nitroxide spacer and two interchangeable anhydrides defined as 0-bit and 1-bit. This methodology allows encryption of any desired sequence in the chains. Moreover, the formed sequences are easy to decode using tandem mass spectrometry. Indeed, these polymers follow predictable fragmentation pathways that can be easily deciphered. Moreover, poly(alkoxyamine amide)s are thermolabile. Thus, the digital information encrypted in the chains can be erased by heating the polymers in the solid state or in solution.

iterative steps on the solid support S1 (purple line) or on the soluble support S2 (dark-blue line). These data correspond to Entries 13-14 in Supplementary Table 1 a The letter T denotes the TEMPO spacer. The number 1 corresponds to the building-blocks defined as 1-bit. b Values measured by size exclusion chromatography. M p is the peak molecular weight. c ∆M p = M p, Entry n -M p, Entry n-1 .

Supplementary methods.
Synthesis of 2-bromopropionic anhydride (a-0). This procedure was adapted from the literature. 1 2-Bromopropionic acid (5.0 g, 32.7 mmol) was dissolved in CH 2 Cl 2 (40 mL). N,N-Dicyclohexylcarbodiimide (3.745 g, 18.1 mmol) was added, and the milky mixture was stirred overnight at room temperature. After reaction, the precipitate was filtered off and the filtrate was concentrated using rotary evaporation. The concentrated solution was poured in npentane, thus resulting in the precipitation of insoluble species. The residue was filtered, and the filtered solution was concentrated in vacuo. The resulting slightly greenish liquid was Removal of the Fmoc-group. 0.6 g of S2 was dissolved in a mixture of piperidine/DCM (1/1, 6 mL) and the solution was stirred for 1.5h at room temperature. The reaction mixture was concentrated under reduced pressure and precipitated in methanol. The precipitate was collected by filtration, washed with methanol and dried in vacuum.

Attachment of the 1-motif to the soluble support S2. The amine terminated polymer S2
(0.50 g, 1 Eq.) along with a-1 (0.166 g, 6 Eq.) and K 2 CO 3 (0.181 g, 15 Eq.) were dissolved in 5mL of THF. The solution was stirred for 50 min at room temperature. After reaction, the solvent was removed under reduced pressure. The resulting white solid was dissolved in THF and filtered. The filtrate was concentrated under reduced pressure and precipitated in MeOH.
The precipitate was filtered, washed with MeOH and dried. Wesdemiotis et al. as briefly described hereafter. 2 Protonated nitroxides are named "y" since they contain the original -end-group (hence designated by a letter from the end of the alphabet) and they are formed after the cleavage of the second bond in the monomer (when counting bonds in the skeleton from right to left). In contrast, protonated carbon-centered radicals contain the original α-end-group (hence designated by a letter from the beginning of the alphabet) and are formed after the cleavage of the third bond in the monomer (when counting bonds in the skeleton from left to right): they are then named "c". For both product ions, a superscripted "•+" is added to indicate that they are radical cations, and the subscripted "i" value corresponds to the number of partial or entire motifs (0-T or 1-T) that they contain.

Attachment of amino-TEMPO
Using this nomenclature to annotate peaks in MS/MS spectra, interpretation of CID data can readily be performed based on relative locations of 1 and 0 in the precursor ion, according to the rules described below: Rule 1. The largest congener in the c i •+ product ion series (i.e., c n •+ ) is formed after the precursor ion has eliminated either a 305 Da radical ( • T-0-) when containing the 0- moiety, or a 319 Da radical ( • T-1-) when containing the 1- moiety.
Rule 3. The additional coding unit present in c i •+ compared to c i-1 •+ (as well as in y i •+ compared to y i-1 •+ ) is revealed by the m/z difference (Δm/z) between these two product ions and is a 0 if Δm/z = 226 Da or a 1 if Δm/z = 240 Da.
As a result, measuring the mass of the two smallest neutrals released from the precursor ions allows the highest congeners of each product ion series to be identified, and so the coding unit linked to each termination. Then, measuring the distance between consecutive peaks from c n •+ down to c 2 •+ and from y n •+ down to y 2 •+ allows the binary (1, 0) sequence in the precursor ion to be reconstructed, starting from the  or αchain-end, respectively. However, it is important to note that the smallest congeners in each series (i.e., c 1 •+ and y 1 •+ ) were never observed. As shown in Figure 3b-c, isomers composed of one 0 and two 1 coding units can readily be distinguished when applying these sequencing rules. Longer sequences can also easily be deciphered as shown in Figure 1e.  (2011)