Comment

Selective protection of functional groups is one of the main issues in organic chemistry (Greene & Wuts, 1999). The strategy of protection is vital in peptide synthesis, especially when peptides are used as scaffolds for complex bioconjugates (e.g. Alkhader et al., 2010). The tert-butoxycarbonyl (Boc) and bis(tert-butoxycarbonyl) (bis-Boc) groups are frequently used to protect amino groups. While Boc-protected organic structures are fairly well represented (452 hits) in the Cambridge Structural Database (CSD, Version 5.33; Allen, 2002), to the best of our knowledge only nine organic structures with bis-Boc protection on the N atom have been deposited to date, and the structures of only two sets of compounds with Boc and bis-Boc protection are available in the CSD [CSD refcodes MUTTOK and MUTTUQ (Macleod et al., 2003), and XUDQUJ and XUDQOD (Ikonen et al., 2009)]. During the synthesis of some novel nonproteinaceous amino acids, we obtained two derivatives of aspartic acid, and their structures form a third pair of this type. We thus present here the crystal structures of dimethyl (2S)-2-[(tert-butoxycarbonyl)amino]butanedioate, (I), and dimethyl (2S)-2-{bis[(tert-butoxycarbonyl]amino}butanedioate, (II). We compare the geometry at the N atom in each structure, as well as in the other two Boc-protected pairs mentioned above and in structures protected by related groups. It is worth noting that (II) is the first chiral and second non-aromatic example of a bis-Boc compound, opening the possibility of observing the arrangement of bulky substituents.

The asymmetric units of (I) and (II) are shown in Fig. 1. The absolute configuration at atom C3 is S in both structures, as the applied synthetic procedure preserves the configuration of the substrate. All bond lengths and angles are typical, except for those in the urethane group, which will be discussed in detail (Tables 1 and 3). In (I), the orientation of the carbonyl O atom in the Boc substituent is antiperiplanar relative to atom H1N [torsion angle O5-C7-N1-H1N = -175.9 (14)°], stabilized by two very weak intramolecular hydrogen-bond contacts of C-HO type (C10-H10BO5 and C11-H11AO5; Table 2). Therefore, rotation around the C8-O6 single bond is most probably hindered. In (II), carbonyl atom O5 is synperiplanar to atom C3 and carbonyl atom O7 is antiperiplanar to atom C3. In the other two bis-Boc compounds, the carbonyl O atoms are oriented in the same direction, both antiperiplanar to the C-N bond, but these structures are much less crowded (Macleod et al., 2003; Ikonen et al., 2009). Four very weak intramolecular hydrogen-bond contacts of C-HO type (C9-H9AO5, C11-H11CO5, C15-H15CO7 and C16-H16AO7; Table 4) stabilize the conformation of both Boc groups in (II) (Fig. 3a) by hindering rotation around the C8-O6 and C13-O8 single bonds.

In (I), intermolecular hydrogen bonds, utilizing mainly the NH group and both ester groups, stabilize infinite chains running along the b axis (Fig. 2a). Two adjacent chains (related by a 2₁ symmetry operation) are joined to each other via C1-H1CO5ⁱⁱ hydrogen bonds to form helical ribbons down the b axis (Fig. 2b and Table 2; symmetry code as in Table 2). The same atom C1 and urethane atom O5 are involved in weaker C1-H1BO5ⁱⁱⁱ contacts, giving rise to layers parallel to the (001) plane (Table 2). In contrast, the crystal structure of (II) has a three-dimensional architecture, although chains of molecules (related by a direct a-axis translation), joined via C1-H1BO1ⁱⁱⁱ and C3-H3O4ⁱⁱⁱ hydrogen bonds, may be distinguished (Fig. 3a and Table 4; symmetry code in Table 4). Nonetheless, each chain interacts with five others (related by a b-axis translation and symmetries of two 2₁ axes), utilizing ester, tert-butyl and urethane groups to give a three-dimensional network of hydrogen bonds, as shown in Fig. 3(b).

The limited data available on mono- and bis-Boc substituted compounds prompted us to take a closer look at the N-atom geometry. We found almost no information on the values of C-N bond lengths in such urethane-type structures in International Tables for Crystallography, Vol. C (Prince, 2004). The only available structural information is the H-N distance in X₂-N-H of 1.01 Å (Prince, 2004) and the Csp³-Nsp³ distance in C*-NH-C=O (acyclic amides) of 1.45 Å (Prince, 2004), which is the best available representation of a urethane group. We found 874 urethane-type organic structures in the CSD [tert-butoxycarbonyl-, benzyloxycarbonyl-, ethoxycarbonyl-, methoxycarbonyl-, (9H-fluoren-9-ylmethoxy)carbonyl-, and others] and only 11 bis-urethane structures. The mean values of selected geometric parameters, defined in Fig. 4, are presented in Table 5 for comparison with (I) and (II), and with acetyl-substituted compounds (non-urethane derivatives).

All bonds at the N atom in (II) are longer than those in (I) due to the introduction of a second bulky tert-butoxycarbonyl substituent. Elongation of these bonds should be expected from resonance theory, as the introduction of a second substituent provides a decrease in the double-bond character of the N-C bonds, while the elongation of the N1-C3 bond (DIST3) from 1.4438 (16) to 1.4669 (13) Å seems to be caused by steric effects. It is worth noting that the C=O and C-O bonds in the Boc substituent in (II) are shorter [C=O = 1.2067 (14) and 1.2057 (13) Å, and C-O = 1.3241 (13) and 1.3312 (13) Å] than the respective bonds in (I) [C=O = 1.2169 (14) Å and C-O = 1.3455 (15) Å]. The same is true for the other two sets of Boc and bis-Boc structures (Macleod et al., 2003; Ikonen et al., 2009).

The hybridization of atom N1 in both structures is sp², with angles tending towards 120°. The urethane group in (I) is almost planar and atom N1 is out-of-plane by 0.031 (2) Å. In (II), atom N1 is out of the plane of the neighbouring atoms (C3, C7 and C12) by 0.184 (2) Å, suggesting a partial sp³ character. The shift of electron density around atoms C7 and C12 towards the O atoms in (II) may be a consequence of this change, probably related to the limited delocalization of electrons in a bis-protected amino acid derivative. It is also possible that the steric repulsion between two bulky Boc substituents also affects the N1 out-of-plane distance. In (II), the C7-N1-C3 angle (ANG1) is much smaller than the C12-N1-C3 angle (ANG2). As atoms C7 and C12 should create the same chemical environment, this is the effect of stereochemistry and resulting steric correlations with the C-C(O)OMe group of aspartic acid. Only a single H atom (H3) points towards atom C7 (of the Boc group), while the rest of the amino acid chain points towards atom C12 from the other Boc group and therefore decreases the value of ANG1, as the result of steric repulsion between two bulky Boc groups and the side chain. It is of interest that the side-chain ester group has dramatically changed its conformation, as shown in Fig. 5. It seems that this effect was forced by the introduction of the second Boc group and provides the best spatial location of the substituents on the N atom: two Boc groups and the ester side chain of aspartic acid.

The elongation of bonds in (II) is similar to two other Boc and bis-Boc structures (see Scheme 2), as shown in Table 5. The N-atom out-of-plane value is much greater in (II) than in XUDQOD or MUTTUQ. In (II), DIST3 is longer than in the two bis-Boc aromatic structures, while the other two distances (DIST1 and DIST2) are similar. The differences in the angle values are greater in the two aromatic CSD structures than in (II), as the Boc substituents are rotated out of the plane of the aromatic ring, whereas in (II) the side chain of the aspartic acid limits the available space, even with the conformational change of the ester side chain (C2-C3-C4-C5 torsion angle).

The interatomic distances in (I) and (II) are within the range of other tert-butoxycarbonyl- and bis(tert-butoxycarbonyl) structures, and those of general urethane and bis-urethane structures. They are also close to those in acyl-substituted species. Resonance theory allows the participation of the second noncarbonyl O atom in resonance structures. Urethane groups are known for their racemization-protecting properties during peptide synthesis, as the resonance of the urethane group lowers the rate of cyclic transient product formation, which is associated with decreased acidity of urethane NH (R-O-CO-NH) compared with amide NH (R-CO-NH) (Sewald & Jakubke, 2002; Goodman, 2004). As can be estimated from the Boc and acetyl derivatives, the difference in bond length is rather insignificant, which supports the theory of acidity.

In conclusion, we have observed the elongation of all bonds formed by the N atom in bis-substituted urethane organic compounds. This is probably caused by steric repulsion of bulky substituents and is enhanced by delocalization of the electron pair of the N atom. The nature of the changes depends on the structure of the main part of the molecule and leads to the optimum spatial accommodation of the Boc substituents.

Experimental

Dimethyl (2S)-2-aminobutanedioate hydrochloride was obtained as a colourless oil according to a modified procedure (Schröder & Lübke, 1965) from L-aspartic acid (Reanal). The oil was diluted with MeCN, neutralized with N,N,N-triethylamine (Ubichem Ltd) and subjected to further reaction with di-tert-butyl dicarbonate (Boc₂O, GL Biochem) for 9 h, according to Bodanszky & Bodanszky (1994). The resulting oil was treated with n-hexane (oil-n-hexane 1:6 v/v), yielding long colourless block-shaped crystals of (I) (yield 89%; m.p. 335.7-337.1 K). ¹H NMR (500 MHz, CD₃OD): 4.55 (t, J₁ = 6.0 Hz, 1H), 3.77 (s, 3H), 3.73 (s, 3H), 2.89 (dd, J₁ = 6.0 Hz, J₂ = 16.5 Hz, 1H), 2.81 (dd, J₁ = 6.0 Hz, J₂ = 16.5 Hz, 1H), 1.48 (s, 9H). MS (ESI), m/z, calculated for C₁₁H₂₀NO₆ [M + H]⁺: 262.128; found: 262.129.

Compound (II) was obtained from (I) according to a modified method (Englund et al., 2004) by overnight reaction with Boc₂O in MeCN in the presence of 4-(dimethylamino)pyridine (Fluka) at room temperature. Thin colourless needle-shaped crystals of (II) (yield 88%; 325.2-326.0 K) were obtained by slow evaporation of AcOEt. ¹H NMR (500 MHz, CD₃OD): 5.44 (t, J₁ = 6.5 Hz, 1H), 3.77 (s, 3H), 3.73 (s, 3H), 3.21 (dd, J₁ = 6.5 Hz, J₂ = 16.3 Hz, 1H), 2.83 (dd, J₁ = 6.5 Hz, J₂ = 16.3 Hz, 1H), 1.54 (s, 18H). MS (ESI), m/z, calculated for C₁₆H₂₇NNaO₈ [M + Na]⁺: 384.163; found: 384.163.

All H atoms were found in difference Fourier maps. In the final refinement cycles, C-bonded H atoms were positioned geometrically and treated as riding atoms, with C-H = 0.98-1.00 Å and U_iso(H) = 1.2U_eq(CH and CH₂) or 1.5U_eq(CH₃). Atom H1N in (I) was refined isotropically.

For both compounds, data collection: CrysAlis CCD (Oxford Diffraction, 2009); cell refinement: CrysAlis RED (Oxford Diffraction, 2009); data reduction: CrysAlis RED; program(s) used to solve structure: SHELXS97 (Sheldrick, 2008); program(s) used to refine structure: SHELXL97 (Sheldrick, 2008); molecular graphics: DIAMOND (Brandenburg, 2012); software used to prepare material for publication: SHELXL97.

---

Supplementary data for this paper are available from the IUCr electronic archives (Reference: QS3019 ). Services for accessing these data are described at the back of the journal.

---