A transient initiator for polypeptoids post-polymerization α-functionalization via activation of thioester group

Here we introduce a post-polymerization modification method of an α-terminal functionalized poly(N-methyl-glycine), also known as polysarcosine. We utilized 4-(methylthio)phenyl piperidine-4carboxylate as an initiator for the ring-opening polymerization of N-methyl-glycine-Ncarboxyanhydride followed by oxidation of the thioester group to yield an α-terminal reactive 4(methylsulfonyl)phenyl piperidine-4-carboxylate polymer. This represents an activated carboxylic acid terminus, allowing straightforward modification with nucleophiles under mild reaction conditions and provides the possibility to introduce a wide variety of nucleophiles as exemplified using small molecules, fluorescent dyes and model proteins. The new initiator yielded polymers with well-defined molar mass, low dispersity and high end-group fidelity, as observed by gel permeation chromatography (GPC), nuclear magnetic resonance (NMR) spectroscopy and matrixassisted laser desorption/ionization time-of-flight (MALDI-ToF) mass spectroscopy. The introduced method could be of great interest for bioconjugation, but requires optimization, especially for protein conjugation. Borova et al. A transient initiator for polypeptoids post-polymerization α-functionalization via activation of thioester group

Polypeptides and polypeptoids have had significant attention for biomedical application for decades and saw an additional boost in the last few years [68][69][70][71][72]. The absence of hydrogen bonds [73,74] in polypeptoids gives them additional advantages like better solubility in various solvents and the formation of stable structures for broad biomedical application [68,75,76].
PSar is most commonly prepared via nucleophilic living ring-opening polymerization (NuLROP) of N-substituted α-amino acid-N-carboxyanhydrides (NNCAs) [88,89], which can be initiated with a variety of nucleophiles. Most commonly, amines are employed [90][91][92] , but water [92], alcohol [93][94][95], thiols [96] and others [70,92,[97][98][99][100] have been reported. Notably, PSar and other polypeptoids can also be obtained by NuLROP from the more stable N-substituted αamino acid-N-thiocarboxyanhydrides (NNTAs) [101,102]. Apart from determining the polymer chain length via [M]0/[I]0, an initiator can be chosen to introduce specific functional groups in the α-terminus of the polymers. However, it is important to consider that the introduced functionalities must be compatible with the polymerization process. Only a few papers introduced functional initiators for NCAs polymerization. Tao et al. reported on ROP of Nsubstituted glycine N-thiocarboxyanhydride with cysteamine with a further application for thiolene and thiol-yne click chemistry [103]. Later, Johann and co-workers introduced the aminofunctional trans-cyclooctenes (TCO) and 6-methyl-tetrazine (mTz) as initiators to obtain functional polypeptoids and performed their post-polymerization modification [104]. Postpolymerization modification (PPM) can help to overcome limitations that occur during polymerization and introduce responsive, structural and functional properties into polymers, which are otherwise incompatible with the polymerization process [36,[105][106][107][108]. PPM can offer several advantages. On one hand, a variety of different functionalities can be introduced into the side chain or termini of a polymer. On the other hand, the resulting polymer will have the same degree of polymerization and chain length distribution as the original polymer. PPM also allows the synthesis of polymers with functionalities that can not or only inconveniently be introduced directly via the polymerization. Different post-polymerization concepts like modification of polymeric active esters, anhydride, isocyanates, oxazolones, epoxides, Michael-type addition reactions, modification by thiol exchange, etc., have been reported for functionalization of polymers [36]. PPM used for bioconjugation, in particular, should offer high efficacy under mild conditions. In the last two decades, click-chemistries have had a major impact in the field of bioconjugation [109].
Despite the preference for PPM utilizing click chemistry, the activated ester exchange reaction still has some value.
The PPM via polymeric active esters became an attractive tool after the first introduction by Jatzkewitz [110][111][112] and later picked up by Ferruti et al. [113] and Ringsdorf et al. [114]. The reaction of activated esters with amines leads to the formation of stable amide bonds which are of course most common in biological systems. Additionally, there is no need to use of potentially toxic (metal) catalysts or other chemical reagents making them an attractive material in biomedical research [115]. Apart from alkyne-azide click chemistry [116][117][118][119][120][121], sulfur-based chemistries have been widely exploited [122][123][124][125][126]. Thiol moieties exhibit a favorable low pKa [127,128], have ability to form disulfide bonds upon oxidation [129,130], show a good reactivity profile [131] and are abundant in biological systems [132]. The versatility of sulfur-based chemistry and the ease of modification makes it an attractive choice in organic chemistry and life science [133][134][135]. A multitude of chemoselective reactions utilizes sulfur chemistry [135][136][137][138] with application in many fields, including self-healing materials or drug delivery systems [122,139]. One particular thiol-based reagent, 4-(methylthio)phenol (4MTP), has been described long ago for peptide coupling, but to date, has not been investigated for PPM bioconjugation. Johnson et al. reported on the use of 4MTP esters as a carboxyprotecting group during polypeptide synthesis in 1968. When 4MTP is oxidized to yield 4-(methylsulfonyl)phenyl (4MTO2P), it transforms a protected ester into an activated activate ester, allowing further aminolysis and amide formation [140][141][142]. Chen introduced 4MTP esters for quinoxaline antibiotic synthesis [143]. Later, Siemens applied activated 4MTP esters during peptide solid-phase synthesis [144]. Cho also reported on the application of 4MTP moieties as a safety-catch protecting group during peptide coupling or as an active ester that can act with other N-free peptide fragments to form new bonds [145]. More recently, Popovic' et al. reported on peptide 4MTP ester synthesis suitable for peptide segment coupling or direct amidation with peptide N-termini [146]. However, to the best of our knowledge, 4MTP has not been utilized or suggested for any PPMs.
Detection was typically set from 1000 m/z to 7000 m/z. After parameter optimization, the instrument was calibrated with CsI3 or PEG standards depending on the m/z range of the individual sample. Samples were prepared with α-cyano-4-hydroxycinnamic acid or with sinapinic acid as matrices, using the dried-droplet spotting technique (0.5-1.5 µL). Exemplarily, samples (1 g/L) were dissolved in MeOH (supplemented with 1.0 % TFA) or in MeCN. Laser power was set slightly above the threshold, typically at 50% -70%. Poisson distributions were calculated using eq 1: The Mn, Mw were calculated using eq. 2 from peak analysis obtained from MALDI-ToF experimental data: where Ni and Mi are the abundance and mass of the i-th oligomer, respectively.
Ultraviolet-visible (UV-Vis) spectrum was taken on a Jasco V-630 spectrometer in MeOH in a quartz cuvette (0.1 or 1.0 cm). The concentration of the dye was calculated by UV-Vis spectroscopy in an absorption maximum of 325 nm.
The polymers were purified via dialysis. Dialysis was performed using Spectra/Por membranes with a molecular weight cutoff (MWCO) of 1 and 10 kDa (material: cellulose acetate) obtained from neoLab (Heidelberg, Germany) against water (Millipore).
The Sarcosine NCA was synthesized as was described previously [149]. Sarcosine 5.00 g (0.056 mol, 1.7 eq.) pre-dried by azeotropic distillation with toluene was placed in 250 mL dried
Diphenylcarbonat (120 g, (0.56 mol, 1.0 eq.) was dissolved in 200 mL of THF, and the rest of the DPC was washed with an additional 200 mL of THF. The reaction flask was left to react for 7-9 days at room temperature. Subsequently, the reaction mixture was filtered, and volatiles was removed under reduced pressure. The residue was dissolved in 400 mL of 5 % NaHCO3 and extracted three times with 300 mL of ethyl acetate (EtOAc). The water phase was adjusted to pH 3-4 with concentrated HCl and extracted three times with 300 mL of EtOAc. The organic phase was dried under Na2SO4 overnight. The solvent was removed under reduced pressure, and the high viscous gel was obtained as the raw product. The yield of the obtained product is 37.6 % (44.1 g).
The monomer solution was prepared in a dried Schlenk flask under an inert atmosphere by dissolving Sar-NCA (1.47 g 13.03 mmol; 96.5 eq.) and 4MTPPC (0.034 g 0.135 mmol, 1.0 eq) in 5 mL of PhCN and 3 mL of MeCN. The reaction solution was stirred under an inert and dry atmosphere at room temperature for 24 h. In the first 2.0 h, a vial was opened to reduce CO2.
PSar was purified by precipitation (3x) from cold diethyl ether (re-dissolved in DCM) and lyophilized.
Polymerization of Sar-NCA from Poc-Sar.
After the initiator and monomer were dissolved, 0.4 mL (0.29 g 2.87 mmol, 50 eq.) of TEA was added, and the reaction solution was heated up to 70°C and left to react for 24h. The pure polymer was obtained by precipitation (3x) from cold diethyl ether (re-dissolved in DCM).

Oxidation of PSar.
The oxidation of PSar was performed according to the procedure described [96]. PSar

Functionalization of PSar.
A substitution was performed as follows. Oxidized PSar was dissolved in 5 mL DCM (or any other suitable solvent) and stirred with 3.0 eq. of nucleophile overnight. A pure product was obtained by precipitation from cold diethyl ether followed by dialysis and lyophilization.

Synthesis of 4MTPPC-functionalized polysarcosine
This work presents a synthetic pathway to obtain 4MTPPC-functionalised polypeptoids via NuLROP using a new 4MTP-functionalized initiator. As was already mentioned, the presence In CHCl3, monomer consumption was fastest and a markedly higher proportion of cyclic dimer was observed compared to the reactions carried out in MeCN and PhCN. MALDI-ToF mass spectrometry of the product obtained from CHCl3 confirmed low molar mass. Unfortunately, the desired product (MTP-initiated PSar) and the macrocyclic PSar have an identical m/z ratio and virtually the same isotope pattern, and therefore we cannot distinguish the two by mass spectrometry (supporting information, Figure S1 and S2). The signal attributed to PSar initiated with 4MTP was observed. However, the presence of low molecular weights cyclic polypeptoids or dimers could not be determined. Interestingly, the obtained signals could also be assigned to oxidized or partly oxidized PSar initiated with 4MTP even though any oxidation was performed. In addition, MALDI-ToF MS data corroborate the molar mass values obtained by GPC (supporting information, Figure S3) Therefore, we investigated a different strategy, i.e., synthesis of an initiator that contains 4MTP but initiates the ROP through a more reliable nitrogen nucleophile ( Figure 3a) and prevents dimer formation at the same time.
While commonly secondary amines are not considered good nucleophiles (they are more basic in character), piperidine and its derivatives are good nucleophiles due to cyclic restraint.
Accordingly, 4-(methylthio)phenyl piperidine-4-carboxylate (4MTPPC) was successfully synthesized in two steps. In the first step, we performed the esterification of Boc-Inp-OH with 4MTP, followed by TFA treatment to remove the Boc group in a second step (Figure 3a). The product was obtained as a white powder with a good yield (93 %) and satisfactory purity. 1   4MTP, (Figure 3b). The signal attributed to the 4MTP methyl group (signal 1, blue) is shifted slightly to higher ppm values due to the introduced ester group. Signals 4 and 6 (blue and green, respectively) are represented with two peaks that most likely arise from the axial and equatorial chair conformation of the piperidine group [150,151]. The synthesis of polysarcosine was performed by ROP of Sar-NCA monomer according to a well-known method [148,149]. In contrast to initiation using 4MPT initiation, this approach   with the system peak and residual solvent. We also assume that the particularly low yield for PSar with DP = 10 is due to significant loss during work-up rather than due to problems with initiation/polymerization. Optimization of this work-up for the low-molar mass PSar was outside the scope of this contribution. Importantly, successful initiation with 4MTPPC was further  (Figure 5d). The broad signal could be attributed to either a partially hydrolyzed α-terminus or a fragmented 4MTPPC α-termini bearing potassium(α) and sodium (β) ion doping, respectively. The obtained signals could also be attributed to the desired PSar with a carbamic acid terminus. However, the latter species seems to be improbable, considering that this rather unstable species have not been reported before for PSar MALDI-ToF mass spectra.  [66,154]. Treatment with 3.0 eq. of m-chloroperoxybenzoic acid (m-CPBA) in dioxane for 4 h was later found also to convert MTP ester into the activated form [141]. Recently

Activation of the 4MTPPC-functionalised polymers and their further functionalization
Popovic and colleagues introduced oxone as the oxidizing agent for 4MTP activation [146]. In a slightly different but related approach, Wu et al. reported a very interesting PPM of thiocarbamate derivatives of poly(2-oxazoline)s, which were also oxidized using m-CPBA to make them amendable for nucleophilic substitution [155]. Accordingly, we chose to utilize the same strategy to prevent polymer degradation and confirmed that the sulfone is generated under mild conditions with 3.0 eq. m-CPBA ( Figure 6). 1 H NMR spectra of the activated polymer confirm the high efficiency of the oxidation. We observe a significant shift of the aromatic signals 2 and 3 (at 7.86 ppm and 7.46 ppm, respectively) before and after oxidation (Figure 7a, red) due to an increased electronegativity of the sulfonyl group. Likewise, signal 1, attributed to the methyl group next to sulfide/sulfonyl, is shifted and overlaps with the polymer backbone after oxidation, as was verified by oxidizing the initiator 4-MTPPC as a model compound (see supporting information, Figure S4). As the MTP ester is prone to hydrolysis, particularly after oxidation, we also find signals that can be attributed to a hydrolyzed terminal group. In particular, signals 2' and 3' (at 7.71 ppm and 7.13 ppm, respectively) (Figure 7a, red) are attributed to 4-(methylsulfonyl)phenol, cleaved off the polymer via hydrolysis. The signal 1'' (at 2.64 ppm) is attributed to the methyl group next to the sulfoxide (partially oxidized MTP), even though we did not observe the other signal expected for the partly oxidized phenyl ring.
GPC traces of the oxidized polymer remained monomodal, narrow and Mn remained essentially unchanged (Figure 7b), which indicates that polymer degradation or cross-linking does not occur [66,154]. It should be noted, when we increased the oxidation time up to 24 h, GPC analysis suggests polymer degradation as evidenced by a decrease of the molar mass, as well as the presence of the high molecular weight shoulder point on the interaction of macro chains, ends with each other (Supporting information Figure S5).  (Figure 8c), while signals 9-11 were not distinguishable in the 1 H NMR spectra. However, we still observed signals at 7.5 -7.8 ppm (Figure 8ab, red frame) possibly attributed to unoxidized PSar (however, the oxidative analysis shows that MALDI-ToF MS data corroborates the functionalization of PSar with the low molecular weight nucleophiles ( Figure 9). In the mass spectra of the α-benzylamine-PSar , signals attributed to the desired product featuring an amine ω-terminus and sodium (γ) and potassium (α) ion doping, respectively (Figure 9a), are observed. We also observed signals that can be attributed to oxidized PSar with proton (δ) ion doping. Signals at m/z = 3711 could be attributed to the PSar starting material that underwent fragmentation and sodium (β) ion doping. Signals suggesting the presence of PSar with a hydrolyzed MTP ester and sodium ion doping (ε) were also detected.
An analysis of α-neopentylamine-PSar showed the presence of signals that can be attributed to the product bearing proton (ω), sodium (δ) and potassium (α) ion doping with amine ωtermini ( Figure 9b). However, we also observed signals that could be attributed to unoxidized PSar (γ) or PSaroxid bearing sodium (α) or potassium (γ) ion doping with an amine ω-terminus.
Signals with lower intensity can be assigned to PSar with hydrolyzed 4MTPPC-α-termini bearing potassium (ε) ion doping. The weaker δ distribution may be attributed to fragmented PSar chains.
Successful functionalization with dansylcadaverine is evidenced by the two most intense signals in the MS at m/z = 3668 and 3683, which can be attributed to the desired product ionized with Na + (α) and K + (β), respectively, featuring an amine ω-terminus. The weaker γ and δ distributions can be attributed to residual PSar and PSaroxid bearing potassium ion doping, respectively ( Figure 9c).
Additionally, the UV-Vis spectra of PSarDansyl show a strong absorbance at 335 nm attributed to the presence of dansylcadaverine, which was not observed in the precursor polymers ( Figure S6). Quantification of the polymer conjugated dye was in excellent correlation with expected values, corroborating the potential to apply the proposed method for PPM.
As a next step, we tested the more challenging conjugation with bio(macro)molecules such as peptides and proteins. The activated polymer was incubated with glutathione as a model peptide, red fluorescent protein (RFP) and bovine serum albumin (BSA) as model proteins.
The conjugation was performed in phosphate buffer solution (PBS) at pH = 8 for 4-5 h.
Glutathione (GSH) was chosen due to its antioxidant properties and the presence of two available nucleophilic groups that can, in theory, react with the activated ester group of PSaroxid.
As expected, GPC traces of the resulting product (PSarGHS) after incubation demonstrate a slight but noticeable difference between the respective elution volumes. The minor difference in the apparent molar mass suggests that substitution occurs only with one available nucleophilic group (Figure 10a). The presence of a new signal at 3.80 ppm in the 1 H NMR spectra also substantiates a successful reaction (Figure 10b). An overlay of the 1 H NMR spectra of PSarGHS and GHS showed that the shifting of the signals 1 and 5 depends on the used buffer solution. After incubation in the TRIS-HCl buffer, signal 1 shifted towards lower In contrast, incubation in PBS leads to a shift of the signals 1 and 5 to lower ppm values and both signals appear to overlap. Such low ppm shifting can result from the hydrogen bonds in the water solution [156]. Unfortunately, signals 4 and 6 could not be unambiguously assigned in NMR spectra, either due to low intensity or overlap with the broad and intense signal of the polymer side chain. Important to note, the same peak shifting is observed after the reaction between oxidized 4MTPPC (4MPTO2PC) and GHS (Supporting information Figure S7). Signal 5 shifts completely after incubation in PBS but not TRIS-HCl, In addition, after conjugation in TRIS-HCl, signal 6 remains largely unchanged, it essentially disappears after incubation in PBS.
The presence of side reactions such as cyclization [157] or rapid hydrolysis of resulting thioester bonds cannot be excluded. Mitamura et al. reported on the rapid hydrolysis of the thioester bond in GHS conjugated products and reports the presence of a thiocarboxylic acid and an elimination product of the thioester in mass spectra [158]. Here, MALDI-ToF mass spectrometry was performed to confirm the successful GHS conjugation. Interestingly, compared to PSaroxid, the MALDI-ToF MS of PSarGHS is much better resolved again gives a clear and narrow m/z distribution. While we could not attribute signals of the fully desired GHS conjugate PSarGHS, the most intense distribution can be attributed to a thiocarboxylic acid fragment with a K + (α) ion doping (Figure 10c). We suppose that resulting PSarGHS is not ionized efficiently itself but rather degrades during MALDI-ToF analysis. However, also the presence of the signals attributed to partly oxidized (ε) and oxidized (δ) PSar with amine ωtermini as well as partly hydrolyzed PSar bearing sodium (γ) and potassium (β) ion doping is observed (Figure 10c). The presence of the PSaroxid clearly shows that the conversion is not the complete under the chosen conditions.
Results obtained after conjugation with a red fluorescent protein (RFP) were found to be less successful. RFP was incubated with PSaroxid (with 1:5 ratio, respectively) in PBS (pH = 8.0) for 4 h. The resulting product (PSarRFP) was purified with a centrifugal filter and analyzed with FPLC ( Figure 11a) and PAGE techniques (Figure 11b).
Judging from FPLC traces before and after conjugation (Figure 11a), only a minor fraction of RFP was conjugated with PSar, as evidenced by a minor shoulder at lower elution volumes.
The PSarRFP reaction mixture is represented with two clear signals attributed to excessed PSaroxid and conjugated RFP (Figure 11a, dark blue). This unexpected observation could be related either to inferior conjugation degree or decreased PSarRFP hydrodynamic radius after functionalization. A noticeable decrease of the second peak and increased molecular weight shoulder after centrifugation (most molecules lower than 10.0 kg/mol were removed) are observed (Figure 11a, blue). Side reactions could also explain the presence of the high molecular shoulder in PSarRFP traces during conjugation. PAGE analysis shows that electrophoretic patterns of RFP alone exhibit a characteristic band at ca. 25 kg/mol while the conjugated PSarRFP before centrifugation shows three prominent bands that could be attributed to free RFP and the desired PSarRFP that appeared at lower Mw, which suggesting that presence of PBS increases the migration rate of PSarRFP (Figure 11b). PAGE of the purified PSarRFP shows only one band at the same region that native RFP and the presence of minor smear at higher Mw was not detected (Figure 11b). We expect that reaction conditions critically influence the degree of conjugation and that further improvements are needed. BSA was selected as another model protein for conjugation due to its ready availability, robust nature, and the presence of an accessible free thiol at Cys-34 and 60 free amine groups available for conjugation [159][160][161]. The conjugation method uses a large excess of PSaroxid (Mw = 3.5 kg/mol). The experiment was carried out for 5 h and 72 h in PBS at pH = 8.0. Subsequently, the reaction mixture was dialyzed against water for 4 h to remove unreacted polymer before lyophilization. GPC analysis shows a minor shift after 5 h and a significant tailing at higher molar mass after 72 h (Figure 12a). The negligible molar mass difference between PSarBSA and native BSA suggests that the conjugation via the amine group is avoided under these pH conditions, and only cysteine group is available for interaction. PSarBSA incubated for 72 h shows higher molecular weight tailing compared with native BSA and, also, presence of low molecular weight tailing for signal attributed to PSaroxid (Figure 12a). The presence of tailing points that increased incubation time leads to increased side reactions with PSaroxid like chain coupling or degradation and increases the degree of conjugation (Figure 12a). The SDS-PAGE analysis of PSarBSA performed after coupling in PBS at pH = 8.5 (PSarBSA1) and 11 (PSarBSA2) demonstrate a successful mere successful result. A small but distinct increase in molar mass was found between BSA and its corresponding conjugates (PSarBSA1 and PSarBSA2), respectively. The average molar mass was estimated to be ca. 75-80 kg/mol (supporting information Figure SI). Our results suggest that the degree of conjugation depends on the pH of the reaction media (Figure 12), which is well known for the reaction of activated carboxylic acids. The difference in band intensity points that significant excess of unreacted PSar is present even after 24h of dialysis (membrane pore size is 10 kDA).