A Distinctive Structural Twist in the Aminoimidazole Alkaloids from a Calcareous Marine Sponge: Isolation and Characterization of Leucosolenamines A and B

Abstract

Bioassay-guided fractionation of Papua New Guinea collections of Leucosolenia sp. resulted in the isolation of the novel compounds leucosolenamines A (5) and B (6) and the known compound thymidine (7). Compound 5 contains a 2-aminoimidazole core substituted at C-4 and C-5 by an N,N-dimethyl-5,6-diaminopyrimidine-2,4-dione and a benzyl group, respectively. Compound 6 retains the same core structure; however C-4 is substituted by a 5,6-diamino-1,3-dimethyl-4-(methylimino)-3,4-dihydropyrimidin-2(1H)-one moiety. This substitution pattern is unique and has never been observed in calcareous imidazole alkaloid chemistry. Leucosolenamine A (5) was mildly cytotoxic against the murine colon adenocarcinoma C-38 cell line.

The sustained position of marine sponges as a source of structurally diverse and biologically active natural products was emphasized in a recent review article covering the 2003 literature.¹ Complex sponge natural products have also been the basis for the development of several clinical leads.² At the top of the list are natural products containing multiple chiral centers such as the bengamides,³ fijianolides,⁴ the halichondrin analogue E7389,⁵ and the discodermolides⁶ or sponge metabolites with exotic functionality, as in the psammaplins.⁷ The research described in this paper began with the goal of adding additional unique structures to the rapidly growing list of sponge-derived alkaloids. In this regard, the Calcarea class of sponges is now recognized as a reliable source of richly bioactive metabolites possessing a 2-aminoimidazole⁸ core that is often further substituted at the C-4 and C-5 positions.⁹ The range of substitution patterns observed to date at these positions include (a) benzyl side chains, (b) oxo groups at C-4, and (c) annulation by five- or six-membered rings. The structures of naamine C (1),¹⁰ leucettamine B (2),¹¹ spirocalcaridine A (3),^8c and 2-amino-2-deoxykealiiquinone (4)¹² provide illustrations of these structural features.

Investigations to date of Calcarea sponges have been restricted to the genera Leucetta (syn. Pericharax) and Clathrina (both of the order Clathrinida, family Leucettidae). Yet, in view of the rich chemistry discovered, it is surprising that very little chemical work has been conducted on other members of this taxonomic class comprised of sponges in five orders spanning more than 75 genera.¹³ The major impediment to further studies of this group is that they appear to be less abundant and poorly described in comparison to the class Demospongiae. During recent expeditions to Papua New Guinea (PNG), specimens of the genus Leucosolenia (order Leucosolenida, family Lecucosoleniidae) were obtained and became high priority for further investigation because the crude MeOH extract showed positive activity in our primary screen for in vitro solid tumor selectivity.¹⁴ Disclosed below are the results of this study culminating in the characterization two new compounds, leucosolenamines A (5) and B (6), possessing a 2-amino imidazole core substituted at C-4 and C-5, and the isolation of thymidine (7).

Results and Discussion

The PNG sponge Leucosolenia sp. (coll. no. 03528) yielded three extracts using an accelerated solvent extraction process. The methanol-soluble fraction, coded XFM, had selective bioactivity (for murine adenocarcinoma C-38 cell line) in the disk diffusion solid tumor whole cell assay with an inhibitory zone differential of 9 mm between murine C-38 and CFU-GM cells at 180 µg per disk.¹⁴ Because of this selective bioactivity, the MeOH fraction was subjected to HPLC purification to afford five fractions (Figure S2). Fraction 3 was pure leucosolenamine A (5) (5.6 mg) with an m/z of 384.2 [M + H]⁺ (ESI-TOF) and exhibited inhibition zones of 10.5 mm (at 180 µg per disk) against murine colon 38 (vs 0 mm against CFU-GM). Fraction 1 (coded XFMP1) was also active and showed an inhibition zone of 21 mm (at 180 µg per disk) against C-38 and was further purified by HPLC to yield thymidine (7) (12.2 mg). Additional samples of Leucosolenia sp. (coll. no. 05413 and 06231) were also examined to isolate more of 5. To our delight, these two collections (vide infra) yielded 5 accompanied by a new analogue, leucosolenamine B (6).

The structure elucidation of 5 began by dereplicating against known sponge-derived metabolites, and the only literature match, naamine C (1)¹⁰ [C₂₁H₂₅N₃O₄, 383 amu], was confidently ruled out from the atom count assessed by NMR. A partial formula for 5 of C₁₇H₁₃O₂N₃ was supported by (a) the ¹³C and DEPT NMR data (Table 1) showing 10 quaternary, three methine, two methylene, and two methyl carbons, (b) the ä_C 160.1 characteristic of a guanidine residue of a 2-amino imidazole group, and (c) the ä_C/ä_H 100.5/5.90 (s, 2H’s) indicative of the 1,3 dioxolane ring as seen in leucettamine B (2).¹¹

Table 1.

NMR Data of Leucosolenamine (5)^a and Leucettamine B (2)^b

	5 (in DMSO-d6)			2 (in Me2CO-d6)
C#	δ C	δH mult. J (Hz), int.	gHMBC	δ C	δH mult. J (Hz), int.
2	160.1			160.2
4	148.0			170.5
5	158.0			140.4
6	38.1	4.05 s, 2H	C-4, C-5, C-7, C-8, C-12	114.6	6.42 s, 1H
7	133.2			131.7
8	109.7	6.97 d, J 1.5, 1H	C-6, C-9, C-10, C-12	110.8	8.04 d, J 1.6, 1H
9	146.8			148.6
10	145.3			148.1
11	107.8	6.73 d, J 8.0, 1H	C-9, C-7	108.8	6.82 d, J 8.1, 1H
12	121.9	6.79 dd, J 1.5, 8.0, 1H	C-6, C-8, C-10	126.3	7.33 dd, J 1.6, 8.1, 1H
13	100.5	5.90 s, 2H	C-9, C-10	102.0	6.01 s, 2H
2′	150.8
4′	146.3
5′	113.7
6′	159.5
CH3-N-1′	28.0	3.24 s, 3H	C-2′, C-6′
CH3-N-3′	28.8	3.39 s, 3H	C-2′, C-4′

^aMeasured at 500 MHz (¹H) and 125 MHz (¹³C).

^bRef 11.

Successive sets of high-resolution mass spectrometry data were needed to finalize the molecular formula of 5 as C₁₇H₁₇O₄N₇. The HRESI-TOFMS gave an [M + H]⁺ m/z of 384.1776; however, more accurate data were obtained later by the FTICRMS¹⁵ [M + H]⁺ m/z of 384.14161. Initially we considered three formulas based on the ESI-TOF peak as follows: (i) C₁₇H₂₂O₂N₉ (¢12 mmu vs calc), (ii) C₁₇H₁₈O₄N₇ (¢36 mmu vs calc), and (iii) C₁₇H₁₈O₄N₇ (¢83 mmu vs calc). At this point in the analysis (i) was favored; however, the ultrahigh-resolution data indicated the best match was with (ii) because of the following pattern of ¢ mmu vs calc: (ii) = 0.04, (iii) = 47.2, and (i) = 48.0. Comparing the final formula to the partial formula derived by NMR as noted above indicated that four heteroatom-attached hydrogens were present and an unsaturation number of 13 was required.

The ¹H NMR spectrum of 5 (Figure S4) was not information rich, as it showed just seven types of signals. As a confounding complication, none of the four heteroatom protons were visible by ¹H NMR using a variety of solvents. There was an ABX spin system for a 1,2,4-trisubstituted benzene ring identified by a double doublet at δ 6.79 (J = 1.5, 8.0 Hz), a doublet at ä 6.97 (J = 1.5 Hz), and a doublet at δ 6.73 (J = 8.0 Hz). This diagnostic spin system, the methylene singlet at δ 4.05, and the additional methylene singlet (δ 5.90) noted above justified substructure A (Figure 1). This array was confirmed by gHMBC correlations illustrated in Figure 1 and further justified by comparison of NMR properties of A to the parallel peaks of leucettamine B (2).¹¹ The remaining observed protons were identified as N-methyl singlets at δ 3.39 and 3.24 that flanked the intense water peak.

Figure 1.

Substructures with observed gHMBC correlations of 5.

The process of accounting for the seven nitrogens and four hetero-attached protons of 5 began with the linking of substructure A to a 2N-imidazole ring, resulting in substructure B. This proposal was initially envisioned in order to account for the extraordinarily low field shift of the aliphatic benzylic hydrogens (δ 4.05). The gHMBC correlations shown in Figure 1 were in full agreement with attachment of this heterocyclic ring to C-6. A side-by-side comparison of the ¹³C and ¹H NMR data of leucosolenamine (5) and leucettamine B (2) at atom numbers 2 and 7-13 provided expected similarities, as shown in Table 1, which further documented their congruent structural features. The obvious differences between these two compounds were also reflected in differing shifts at atom numbers 4-6.

The extremely difficult final task involved assembling one or more substructures to explain the remaining unaccounted atoms of C₆H₁₀O₂N₄. The ¹³C NMR signals to be rationalized consisted of four quaternary carbons (δ_C 159.5, 150.8, 146.3, and 113.7). Also needing a specific assignment were the two N-methyls (δ_C/_H 28.8/3.39 and 28.8/3.24), noted above. A 5,6-dinitrogen-substituted uracil with N-methylation at both ring nitrogens, as shown in substructure C of Figure 1, was envisioned and was consistent with these data. Validation of this proposal was partially provided by gHMBC correlations, outlined in Figure 1, from H_CH3-N-1′ (δ_H 3.24) to C-2′ (δ_C 150.8) and to C-6′ (δ_C 159.5) and from H_CH3-N-3′ (φ_H 3.39) to C-2′ (δ_C 150.8) and to C-4′ (δ_C 146.3).

The process of merging structures B and C was not straightforward because, as noted above, all attempts to visualize the four NH ¹H NMR resonances were unsuccessful. To summarize the constraints established above, there were three bridging atom sites in B and two in C, resulting in a total of six possible working structures, shown in Figure 2. These structures could be divided into two sets on the basis of the presence of an NH₂ group at C-5′ or at C-4′, consisting of I-III and IV-VI, respectively. Two structures, I and IV, of Figure 2, follow the substitution pattern seen in the naamidines¹⁶ and pyronaamidines,¹⁷ as they retain the commonly observed 2-aminoimidazole ring. However one significant change includes a pyrimidinedione moiety attached to the exocyclic nitrogen and the N-substitution at C-4 is without biosynthetic precedent. The same unprecedented situation, N-substitution at C-4, holds for all other structures (II, III, V, and VI). Attempts to distinguish among these six possibilities aided by NOE data were unsuccessful, as no correlations were observed between the N-Me groups of substructure C and the sp³ and sp² protons of substructure B. The only NOE correlations observed were from H-6 (′_H δ_H 4.05) to H-8 (δ_H 6.97) and H-12 (δ_H 6.79) and vice versa.

Figure 2.

Evaluation of the six possible working structures of leucosolenamine (5) via ¹³C chemical shifts. The carbon numbers are listed above each bar. R = benzo-1,3-dioxole moiety.

We next turned to analyzing the ¹³C chemical shifts expected for these possibilities and focused on three observed signals, at δ 113, 146, and 148, that appeared to be distinctive. The calculated ¹³C chemical shifts (ACD software) versus experimental data are shown in Figure 2 (and Figure S9), and agreement of these data within (5 ppm was considered to be diagnostic. The three key observations were as follows: (a) only structures I, II, and III showed good agreement for the data at C-5[.minute], (b) the data set at C-4/C-4′ showed agreement for structures I and II, and not III or V, and (c) the data set at C-5 and C-2 showed the best agreement for structure I. Thus I emerged as the most consistent candidate, as shown by the excellent agreement of the δ_C-calc to δ_C-obsv and was confidently concluded to be the final structure of 5 (Figure 2). As seen here, employing database software to calculate ¹³C shifts is becoming an invaluable tool to confirm structural assignments.¹⁸

To further substantiate the above conclusion, 5 was subjected to MS-MS in both positive and negative ion modes. The negative ion experiment gave an m/z 382.17 [M - H]^-, and further MS-MS on this ion produced a single fragment at m/z 340.01 corresponding to a loss of CN₂H₂, as shown in Figure 3. MS-MS on m/z of 340.01 resulted in a fragment ion of m/z 312.17 from a loss of CO. The direct loss of CN₂H₂ was not compatible with structure III or VI due to differences in regiochemistry, leaving two remaining possible mechanisms to explain these results. We preferred fragmentation pathway A beginning with I or IV involving two C-N bond cleavages at N-1-C-2 and N-3-C-4. As explained below, pathway B was less favored and resulted from a C-N bond cleavage occurring from II or V at the C-4-C-5 and C-2-N-3 bonds. Additional observations supporting pathway A were obtained by MS-MS on N,N-dimethylnaamidine D (8, Figure S15) and leucettamine B (2, Figure S16). The MS-MS on the [M + H]⁺ at m/z 352.33 for 8 resulted in a loss of an anisole moiety to give fragment m/z 244.14, and further MS-MS on this fragment afforded the peak at m/z 188.04 rationalized by a loss of C₂N₂H₄ and is consistent with a N-1-C-2 and N-3-C-4 cleavage, shown in Figure 3A. A similar cleavage process was observed when a leucettamine B (2) [M - H]^- peak at m/z 244.09 was subjected to MS-MS, eventually affording the peak at m/z 159.00 from loss of C₂N₂H₃ also via N-1-C-2 and N-3-C-4 cleavage, shown in Figure 3A. The key observation here is that the MS-MS data of 2 and 8 failed to show the cleavage at the C-4-C-5 bond expected via pathway B (as illustrated in Figure S16B). Some additional observations from the MS-MS data were not diagnostic as follows. Additional MS-MS on m/z 340.01 for 5 gave a new ion at m/z 312.17 from loss of CO to structure 5ii. Positive ion mode MS-MS analysis of 5 (Figure S17) also gave interesting results but similar to the preceding data did not differentiate between structures I, II, IV, and V.

Figure 3.

Proposed MS-MS fragmentation pattern of 5. Pathway A rationalizes the fragmentation patterns for structures I and IV, while pathway B details the fragmentation patterns for II and V. See Figure 2 for candidate structures I, II, IV, and V. CE = collision energies set at 30 eV.

An especially critical step in evaluating the structure of 5 pertained to focusing on the major tautomer between the two possibilities shown in Figure 4. The NMR data of 5 clearly showed that only one major tautomer (structure I of Figure 2) was present, and this was concluded to be 5a on the basis of the results of relative energies from MOPAC calculations and the best fit between observed and experimental ¹³C shifts (Figure S18). Interestingly, the fully minimized structure of both tautomers (Figure S19) predicted that the phenyl ring and the pyrimidine ring should be positioned away from each other as shown and is consistent with the lack of NOE correlations between the proton-containing residues of these two rings.

Figure 4.

Relative energies of the two tautomeric forms of 5

We sought to extend the results obtained above by examining two additional PNG collections of this species. The two collections (nos. 05413 and 06123) were selected, and their crude extracts were combined and processed as described for collection no. 03528 (Figure S2). From the crude methanol extract six preparative HPLC fractions were obtained. Fraction 4 contained 5 (5.6 mg) and fraction 3 was further purified to give leucosolenamine B (6) (1.3 mg).

The structure elucidation of 6 began with obtaining a molecular formula from the HRESITOFMS [M + H]⁺ m/z 397.1658, indicating C₁₈H₂₁N₈O₃ (Δ 7.0 mmu vs calc) and required 13 degrees of unsaturation. Comparing this formula to that of 5 showed a difference of +C₁H₃N and -O, making it evident that one of the CdO in 5 was replaced by a CdNCH₃ group in 6. Using the new NMR data in Table 2 for a side-by-side comparison with that of Table 1 indicated many structural similarities between the two compounds. The only differences in their ¹³C NMR are an upfield shift of δ_C-6[.minute] in 6 at 145.3 versus δ_C-6′ 159.5 in 5 plus there was an extra methyl signal at ′ 38.1. From these NMR comparisons it was clear that 6 also contains a dioxybenzo-2-aminoimidazole and a 5,6-dinitrogen-substituted uracil moiety. These were confirmed by gHMBC correlations as illustrated in Figure 1, substructure B, and in Figure 5, substructure D. The important gHMBC correlations to fix the position of the CdNCH₃ are shown and were as follows. One set involved the correlation from methyl protons δ 3.55 and 3.22 to δ_C-6[.minute] 145.3, and the other set involved correlations from methyl protons δ 3.17 and 3.22 to δ_C-2[.minute] 150.8. The chemical shift of C-6[.minute] is consistent with that observed in synthetic purinone derivatives with their N-methylimino carbons resonating at ~δ_C 144-147 (Figure S26).¹⁹

Table 2.

NMR Data of Leucosolenamine B (6) in DMSO-d₆ at 500 MHz (¹H) and 125 MHz (¹³C)

C#	δ C	δH mult. J (Hz), int.	gHMBC
2	159.7
4	144.3
5	155.5
6	38.2	4.01, s, 2H	4, 5, 7, 8, 12
7	133.6
8	110.0	6.93, d, 2.5, 1H	9, 10, 12
9	146.8
10	145.3
11	107.9	6.75, d, 8.0, 1H	7, 9, 10, 12
12	122.1	6.80, dd, 8.0, 2.5, 1H	6, 7, 8, 11
13	100.6	5.91, s, 2H	9, 10
2′	150.8
4′	143.8
5′	115.4
6′	145.3
CH3-N1′	29.9	3.22, s, 3H	2′, 6′
CH3-N3′	28.7	3.17, s, 3H	2′, 4′
CH3-N6′	38.1	3.55, s, 3H	6′

Figure 5.

Substructure D showing ¹³C chemical shifts and gHMBC correlations.

Substructures B and D were then combined on the basis of parallel ¹³C chemical shifts and biogenetic comparison to 5 to give the final structure of 6. The analysis according to the strategy of Figure 4 (Figure S18) supported the tautomer shown. Last, a ROESY experiment on 6 showed only correlations from H-6 (′_H 4.01) to H-8 (′_H 6.39) and H-12 (δ_H 6.80), while no correlations were observed between the three N-Me protons of substructure D or from these protons to those of substructure B. From these observations, it was concluded that CH₃-N6[.minute] is trans to CH₃-N1′, as shown in the final structure of 6, and that the phenyl ring and the pyrimidine ring should be positioned away from each other as discussed above for 5.

Our investigation of the biological properties was restricted to an evaluation of 5 due to sample limitation and stability. The data set summarized in Table 3 employed our standard disk diffusion assay¹⁴ and revealed that 5, at 180 μg/disk, inhibited murine C-38 cells with a 10.5 mm zone as compared to 0 mm against CFU-GM cells. This is the profile for a mildly potent selective cytotoxin when compared to the in vivo active agent bengamide E,²⁰displaying at 7.8 μg/disk inhibition zones of 21 mm against C-38 and 16.5 mm against CFU-GM cells. A similar pattern was observed for a semipure fraction (coded XFMP1) that yielded thymidine (7), and it showed a zone of inhibition of 21 mm against C-38 cells versus 4.5 mm against CFU-GM at 180 μg/disk. However, pure 7 was inactive when reassayed against C-38 and CFU-GM cells. Unfortunately, it has not been possible to date to isolate the active component of this semipure fraction.

Table 3.

Zone Unit Differentials in the Disk Diffusion Soft Agar Colony-Forming Assay^a

compound	conc, μg/disk	ZC–38 – ZCFU –gm/mm
bengamide E (standard)	7.8	4.5
crude MeOH extract	180	9
5	180	10.5
XFMP1	180	16.5
7	180	0

^aMurine cell lines: C-38, colon adenocarcinoma; CFU-GM, colony-forming unit-granulocyte macrophage; normal hematopoietic. Significant selectivity is defined by a difference of ≥7.5 mm zone unit differentials. XFMP1 is semipure and contains thymidine.

The overall structures of leucosolenamines A (5) and B (6) while not overly complex are unprecedented and presented an extreme challenge to characterize. Their core structure retains the 2-amino imidazole array observed in other calcareous sponges. The substitution of N,N-dimethyl-5,6-diaminopyrimidine-2,4-dione and 5,6-di-amino-1,3-dimethyl-4-(methylimino)-3,4-dihydropyrimidin-2(1H)-one at C-4 is unique and represents a new structural theme in calcareous imidazole alkaloid chemistry. Substructure C has been observed as a natural product in 1,3-dimethyl-4,5-diamino-2,6-dihydroxypyrimidine from an Aspergillus nidulans²¹ and in hymeniacidin from a marine Hymeniacidon sponge.²² Of additional note, this substructure is embedded in the pteridine ring of leucettidine (9), isolated from a Bermuda collection of Leucetta microraphis.²³ The discovery of 5 and 6 from Leucosolenia represents the first natural products from this genus and provides a stimulus to explore the other unstudied orders of the class Calcarea.

Experimental Section

General Experimental Procedures

NMR spectra were obtained using a Varian Unity 500+ at 500 MHz for ¹H and 125 MHz for ¹³C employing internal standards: DMSO-d₆ at δ_H 2.50 and δ_C 39.57. Multiplicities of ¹³C NMR peaks were determined using DEPT and gHMQC data. ¹³C NMR data of 6 were measured using a direct detection ¹³C nano probe. Low- and high-resolution mass spectrometry was performed on a benchtop Mariner electrospray ionization time-of-flight instrument (ESITOF). Ultrahigh-resolution mass measurements were obtained via positive mode on a Finnigan FTICR hybrid mass spectrometer, which combines ion trap and Fourier transform ion cyclotron resonance.¹⁵ MS-MS measurements were done using a LTQ mass spectrometer. Preparative HPLC was carried out using a single-wavelength (λ = 254 nm) UV detector and evaporative light-scattering detector (ELSD) in series with a Waters RP C18, 5 μm particle column. Computer modeling and calculations were done using ACD/¹³C NMR DB version 8.00 (May 4th, 2005).

Biological Material

The sponge samples (UCSC coll. nos. 03528, 05413, 06123) were collected by scuba from a variety of sites off Milne Bay in Papua New Guinea at depths of 30—70 ft. The sponge was identified by Prof. Rob van Soest as a Leucosolenia sp., and a voucher sample has been deposited at the Zoological Museum of Amsterdam (ZMAPOR 17556). Voucher specimens and underwater photos are available from the author (P.C.).

Isolation

The sponge samples were preserved in the field according to our standard procedure and transported back to the laboratory at ambient temperature.²⁴ The preserved sponge (03528-1.06 kg, wet weight) was thoroughly dried, ground, and extracted with the accelerated solvent extractor (ASE) first with hexanes, next with CH₂Cl₂, and last with MeOH under high pressure (1700 psi) at 110 °C (Figure S2).²⁵ The MeOH extract was evaporated in vacuo to produce a brown, viscous oil. This oil was then subjected to preparative reversed-phase (RP) HPLC using a Waters C18, 5 μm particle column to afford five fractions. Fraction 3, which was the major metabolite, was eluted at 60% CH₃CN/H₂O to yield 5 (9.0 mg). Fraction 1 was also further purified by HPLC to yield 7 (12.2 mg).

Dried sponge samples (collections numbers 05413 and 06123, 1.31 kg wet weight) were extracted using the ASE employing the procedure outlined above (Figure S2). The MeOH-soluble fraction labeled XFM was subjected to preparative RP HPLC using 10-100% CH₃CN/H₂O to afford six fractions. Fraction 4 contained 5 (5.6 mg) while fraction 3 was further purified by semipreparative RP HPLC at 40% CH₃CN/ H₂O to yield 6 (1.3 mg).

Leucosolenamine A (5)

9.2 mg, brown solid; ¹H NMR (DMSO-d₆, 500 MHz) and ¹³C NMR (DMSO-d₆, 125 MHz), see Table 1 and Supporting Information; ¹H NMR of 5 was also taken in MeOH-d₄, acetone-d₆, dioxane-d₈, and THF-d₆ with and without TFA but no NH’s were observed. NOESY correlations were observed from H-6 (δ_H 4.05) to H-8 (δ_H 6.97) and H-12 (δ_H 6.79) and vice versa; HRFTICRMS [M + H]⁺ obsd 384.14161 (Δ0.04 mmu vs calc), calcd for C₁₇H₁₈N₇O₄ 384.14165. Hydrogenolysis of 5 was unsuccessful.

Leucosolenamine B (6)

1.3 mg, brown solid; ¹H NMR (DMSO-d₆, 500 MHz) and ¹³C NMR (DMSO-d₆, 125 MHz), see Table 2 and Supporting Information; ROESY correlations observed were from H-6 (δ_H 4.01) to H-8 (δ_H 6.39) and H-12 (δ_H 6.80) and vice versa; HRESITOFMS [M + H]⁺ obsd 397.1628 (Δ7.0 mmu vs calc), calcd for C₁₈H₂₁N₈O₃ 397.17311.

Thymidine (7)

12.2 mg, white needles; [α]_D +78.0 (c 0.61, H₂O); ¹H NMR (MeOH-d₄, 500 MHz) and ¹³C NMR (MeOH-d₄, 125 MHz), see Supporting Information; HRESITOFMS [M + H]⁺ obsd 243.10319 (Δ5.6 mmu vs calc), calcd for C₁₀H₁₅N₂O₅ 243.09755.