C–C Bond formation catalyzed by natural gelatin and collagen proteins

Summary The activity of gelatin and collagen proteins towards C–C bond formation via Henry (nitroaldol) reaction between aldehydes and nitroalkanes is demonstrated for the first time. Among other variables, protein source, physical state and chemical modification influence product yield and kinetics, affording the nitroaldol products in both aqueous and organic media under mild conditions. Significantly, the scale-up of the process between 4-nitrobenzaldehyde and nitromethane is successfully achieved at 1 g scale and in good yield. A comparative kinetic study with other biocatalysts shows an increase of the first-order rate constant in the order chitosan < gelatin < bovine serum albumin (BSA) < collagen. The results of this study indicate that simple edible gelatin can promote C–C bond forming reactions under physiological conditions, which may have important implications from a metabolic perspective.


Expanded introduction • Henry reaction
The Henry (nitroaldol) reaction, discovered more than 100 years ago [1], is one of the most important CC bond-forming reactions used to generate -nitroalcohols, which constitute versatile organic synthons for the synthesis of nitroalkenes, 2-nitroketones, 2-aminoalcohols, -hydroxy carboxylic acids and bioactive compounds including, among others, antibiotics, fungicides and insecticides [2]. Due to its practical importance, a vast number of publications have focused on the development of new catalysts for this transformation, including amines, alkali-metal bases, ammonium salts, guanidine derivatives, organo-phosphorous derivatives, Lewis acids or electrophilic activators [3].
Nevertheless, the major limitation of conventional catalysis has been the formation of several byproducts, including those originated from aldol self-condensation, polymerization, Cannizzaro and water-elimination reactions. On the other hand, the reaction usually requires organic solvents, a large excess of nitroalkanes, long reaction times, and challenging timeand labor-intensive isolation techniques [4]. Hence, the development of new sustainable, ecofriendly and selective catalysts for CC bond formation also in aqueous media continues to be an important task [5]. In this sense, microwave techniques, solvent-free conditions and biocatalytic processes represent attractive alternatives [3,6]. In addition, substituent-and catalyst-dependent selectivity toward aldol or nitrostyrene products in a heterogeneous basecatalyzed Henry reaction has been also reported [7].

• Gelatin and collagen features
Although gelatin has been known to mankind for over 3000 years, it continues to find new uses in additional research areas, as we have demonstrated in this paper. Looking at its properties, gelatin has a molecular weight distribution between 50100 kD. Type-A gelatin has ca. 7880 millimoles of free carboxyl groups per 100 g of protein and a broad iso-electric point (IEP) [8] between 6.09.0. In contrast, type-B gelatin has 100115 millimoles of free carboxyl groups per 100 g of protein and a rather narrow IEP range of 4.55.5. protein content [10]. Table S1 outlines the general amino acid composition of gelatin, sorted in increasing content [11]. Several of these amino acids have been proven to be active in amine-based catalysis, including the Henry reaction. Thus, gelatin seems to be a natural choice as a polyvalent catalyst for this reaction.
On the other hand, gelatin forms thermoreversible hydrogels in water, with melting temperatures below body temperature (< 35 °C). From a rheological point of view, both the storage and loss shear modulus G' and G'' exhibit a power-law dependence on the frequency at the sol-gel transition point (rheological exponent n = 0.58 ± 0.05) [12].
For potential alternative experiments dealing with gelatin catalysis, it should be also considered that gelatin proteins in aqueous media are prone to degradation on incubation at elevated temperatures (> 4050 °C), where the salt content has an influence on both the unfolding of the polypeptide chains and the hydrolytic stability [13,14]. In addition, gelatin hydrogel (prepared without any additional hardener) was found to be fairly stable at rt in water, hexanes, Et 2 O, EtOAc and cyclohexane, but undergoes shrinking and clear-to-opaque transition in MeCN, THF, acetone and MeOH. Clear-to-opaque transition, but no shrinking, was observed in Et 3 N, CH 2 Cl 2 and CHCl 3 .
In the case of collagen, triple-stranded helical collagen molecules are stabilized by numerous hydrogen bonds and further laterally associated to form collagen fibrils of 10300 nm diameter. Figure S1 shows the typical procedure for the extraction of different types of gelatin from collagen. Figure S1: Representative preparative process for acidic and basic gelatins from collagen. A typical structure of gelatin is -Ala-Gly-Pro-Arg-Gly-Glu-4Hyp-Gly-Pro-.

• Possible metabolic implications
It is remarkable the fact that edible gelatin can promote CC bond forming reactions under physiological conditions. In this regard, there are numerous aldehydes that are also present in a variety of foods (as well as cosmetic products) that could be, at least partially, transformed in the presence of gelatin proteins. Such process could influence to some extent the metabolic route of these compounds. In this sense, relevant aldehydes include, among others, (a) citronellal (the main component in the mixture of terpenoid compounds that provides citronella oil its distinctive lemon scent. It is isolated from lemon-scented gum and teatree), (b) citral or lemonal (present in the oil of several lemon plants widely used as a flavoring in the food industry and as an antimicrobial), (c) cinnamaldehyde (present in the bark of cinnamon trees and provides cinnamon its flavor and odor. It is widely use as a flavoring agent, agrochemical, antimicrobial and anticancer agent), (d) safranal (isolated from saffron and primarily responsible for this aroma in foods).

Preliminary optimization experiments
Preliminary screening of the reaction conditions by using 4-nitrobenzaldehyde (1a, 0.1 mmol) and nitromethane (2a) as a model reaction showed that 2 mg of PSTA gelatin afforded the corresponding nitroaldol product in good yields within 6 h at physiological temperature by using at least a 5-fold excess of 2a (Tables S2S3). For solvent screening, see discussion in the main paper.   Note that regarding the catalyst loading; the use of more than 10 mg of gelatin may make the work-up of the reaction more laborious.

Additional experiments
a) We carried out a series of experiments that confirmed no effect of possible metal impurities in the gelatin samples on the reaction conversion. Possible impurities of gelatin samples are moisture content (ca. 10%), heavy metals (e.g., Co, Cu, Fe, Ni, Pd) and ash [15]. Table S5 shows the results of this study.
Note that the direct use of nondiluted metal stock solutions (these are 100-fold more concentrated than in the experiments according to the published values [16]) yielded the nitroalcohol product in only 2%.  b) The contribution of the pH of the media was investigated by individual measurements (total volume of each sample = 10 mL) for the reaction between 1a and 2a in H 2 O/TBAB at 37 °C and concentrations as described in the main text: milliQ water: pH 7.02; 2a in milliQ water: pH 5.63; 1a in milliQ water: pH 4.89; PSTA gelatin in milliQ water: pH 5.31; TBAB in milliQ water: pH 6.11; all 4 components combined in milliQ water: pH 4.36; 1a + 2a in milliQ water: pH 3.44; 1a + 2a + PSTA gelatin in milliQ water: pH 3.80. c) Only in the cases of less reactive aldehydes, such as vanillin or iso-valeraldehyde, was it possible to detect by 1 H NMR also the formation of additional by-products (compounds 45, Figure S2), albeit in trace amounts (≤ 2%). In this sense, -nitroalkene 4 seemed to be the major product formed in the case of vanillin, whereas formation of -nitroalcohol 3 was evident in the case of iso-valeraldehyde. However, the complexity of the 1 H NMR spectra in these cases made difficult the unambiguous assignment of all signals.  e) Other heterocyclic aldehydes such as furfural were converted into the corresponding nitroaldol product in very low yield (ca. 4%). However, it should be considered that clean synthesis of such heteroarylnitroalcohols may require lower reaction temperatures [16]. Experiments in this regard were not pursued. Here, the use of a mixture DMSO:H 2 O 4:1 as solvent, and benzaldehyde as substrate, did not show significant improvement in comparison to pure DMSO. In MeOH:H 2 O 1:1 or MeCN:H 2 O 1:1, no reaction product was observed.
f) Both air-dried and freeze-dried (lyophilized) xerogel materials prepared from the corresponding hydrogels provided similar results in the model reaction. Hence, potential water uptake of the xerogel specimen did not play any critical role in the reaction.   Figure S4: Representative crude 1 H NMR spectra of the gelatin-catalyzed Henry reaction between 4nitrobenzaldehyde and nitromethane (5 equiv). Diphenylmethane was used as the internal standard. Figure S5: Representative crude 1 H NMR spectrum of the gelatin-catalyzed Henry reaction between 4-nitrobenzaldehyde and nitromethane (5 equiv). Dimethylacetamide was used as the internal standard (Note that methyl "a" of dimethylacetamide appears at ca. 2 ppm, overlapping with the signal from EtOAc). Figure S6: Additional FESEM images of different biocatalysts used in this work: a) esterificated PSTA gelatin; b) succinated PSTA gelatin; c-d) xerogel prepared by freeze-drying the hydrogel made of PSTA gelatin; e) cooked gelatin purchased from a supermarket; f) powdered chitosan; g) powdered BSA; h) powdered collagen.