Tri- and Tetrameric Proanthocyanidins with Dentin Bioactivities from Pinus massoniana

Abstract

Guided by dentin biomechanical bioactivity, this phytochemical study led to the elucidation of an extended set of structurally demanding proanthocyanidins (PACs). Unambiguous structure determination involved detailed spectroscopic and chemical characterization of four A-type dimers (2, 4–6), seven trimers (10–16), and six tetramers (17–22). New outcomes confirm the feasibility of determining the absolute configuration of the catechol monomers in oligomeric PACs by 1D and 2D NMR. Electronic circular dichroism (ECD) as well as phloroglucinolysis followed by MS and chiral phase HPLC analysis generated the necessary chiral reference data. In the context of previously reported dentin-bioactive PACs, accurately and precisely assigned ¹³C NMR resonances enabled absolute stereochemical assignments of PAC monomers via (i) inclusion of the ¹³C NMR γ-gauche effect, and (ii) determination of differential ¹³C chemical shift values (Δδ_C) in comparison with those of the terminal monomer (unit II) in the dimers 2 and 4−6. Among the 13 fully elucidated PACs, eight were identified as new, and one structure (11) was revised based on new knowledge gained regarding the subtle, stereospecific spectroscopic properties of PACs.

Graphical Abstract

INTRODUCTION

Proanthocyanidins (PACs) are a structurally diverse and complex group of condensed, oligomeric flavan-3-ols that are widespread in woody and some herbaceous plants.¹ PACs contain several subtypes, such as procyanidins, prodelphinidins, and propelargonidins based on the corresponding basic units catechin and epi-catechin (syn. epicatechin), gallocatechin and epi-gallocatechin (syn. epigallocatechin), as well as afzelechin and epi-afzelechin (syn. epiafzelechin).² B-type PACs have one interflavanyl bond originate from coupling C-4 with one of the nucleophilic C-6 or C-8 of a subsequent monomer. A-type PACs have an additional ether bond between C-2 in the “top” and an A-ring hydroxyl moiety in the “bottom” unit. Since their discovery in 1947, PACs have attracted scientific attention for both their complicated structures and medicinal value.³ While PACs have been associated with a broad spectrum of bioactivities, including antioxidant,⁴ anticancer,⁵ anti-inflammatory,⁶ antidiabetic,⁷ neuroprotective⁸ effects, and play an important role in the fields of nutrition and health. The focus of our research at the chemistry-dentistry interface has been the exploration and understanding of their structure-specific abilities to enhance the biomechanical properties of dentin, the soft collagenous tissue of teeth.^9–11

The structural complexity of PACs increases exponentially with their degree of polymerization (DP), the linkages (A- and B-type) and absolute configuration of their interflavanyl linkages (IFLs), as well as the combination of the possible stereoisomeric monomers. Additional analytical challenges are posed by their formation of atropisomers, which interferes with chromatographic separation and often makes NMR spectra acquired at room temperature uninterpretable due to exceedingly broad lines. Additionally, in some cases, resonances can be “missing” because of H-D exchange in deuterated NMR solvents, even for hydrogens that might initially be considered “not likely to exchange”.^9–11 Collectively, these properties make the structural elucidation of trimeric or higher oligomeric PACs an often challenging undertaking. In general, reliance on literature data is hazardous as structure-related publications seldom report full assignment of the ¹H and ¹³C NMR data. This provides grounds for questioning the ultimate reliability of at least some PAC structures and their spectroscopic assignments and rationalizes the importance of developing highly cohesive data sets.

Pinus massoniana is a pine species within the Pinaceae family that has a PAC-rich inner bark. Our previous studies showed that P. massoniana PACs can specifically enhance dentin biomechanical properties and reduce proteolytic dentin degradation.^9–11 Studies of PACs action on dentin initially established that the affinity of PACs for type-I collagen correlates with the degree of polymerization and certain structural/stereochemical features.^9–11 These previous studies showed that tri- and tetrameric PACs are the most promising dentin strengthening agents. Accordingly, this study was focused on phytochemical isolation and structural analysis of tri- and tetrameric PACs from P. massoniana pine bark. Carried out in a scaled-up operation, the goal was to provide solid structural evidence for and sufficient amounts of the most bioactive PAC oligomers required to explore their mechanism of action on dentin, including a variety of biomechanical properties such as static and dynamic stiffness, biodegradation resistance, and resin bonding properties.

In addition to the four known A-type dimers, epicatechin-(2β→O→7,4β→8)-catechin (2),⁹ epicatechin-(2β→O→7,4β→8)-epicatechin (4),⁹ epicatechin-(2β→O→7,4β→8)-ent-catechin (5),¹⁰ and epicatechin-(2β→O→7,4β→8)-ent-epicatechin (6)¹⁰ reported previously from this plant, seven trimers (10−16) and six tetramers (17−22) were isolated and characterized (Figure 1). Among these 13 PACs, eight are new structures, and one (11)⁹ was revised based on new NMR evidence. In order to achieve rigorous structural evidence for these PACs, the following strategies were applied: (i) all NMR data sets were acquired at low temperatures (278 and/or 255 K) to achieve assignable spectra with high resolution. (ii) 2D HSQC, HMBC, and NOESY spectra were used to establish the planar structures; particularly, 4→6 and 4→8 IFLs can be distinguished from the NOESY data, by the differential chemical shifts of the C-6 and C-8 atoms in the monomers. The absolute configuration of the constituent monomers were determined by a combination of (iii) NMR evidence involving NOESY correlations, the “γ-gauche effect” in the ¹³C domain, differential ¹³C chemical shift (Δδ_C) values relative to the four stereochemically fully characterized dimers (terminal unit II), (iv) ECD data, and (v) phloroglucinolysis combined with LC-MS and chiral phase HPLC analysis (Figures 5 and S104–S106, Supporting Information). Factoring in previously reported NMR assignment, these spectroscopic and chemical methods allowed the unambiguous assignment of the structures of all isolated PACs. The presented detailed analytical profiles of 13 PACs (Tables 1–10, sorted by ¹H/¹³C domains and increasing degree of oligomerization) enlarge significantly the space of available NMR data of oligomeric PACs with solid structural information. This also expands the library of PACs that can be used to explore their dentin biomodification mechanisms of action systematically and supports the ongoing establishment of structure-activity relationships.

Figure 1.

Structures of 1−22. Compounds 1, 3, and 9 represent the phloroglucinol adducts.

Figure 5.

Phloroglucinolysis products of selected 10, 11, 15, 17−19 identified by chiral phase HPLC and MS analysis.

Table 1.

¹H and ¹³C NMR (400/100 MHz, resp.) data of 1, 3, and 9 in CD₃OD.

Unit	Ring	No	1a		3b		9b
			1H (mult., J in Hz)	13C	1H (mult., J in Hz)	13C	1H (mult., J in Hz)	13C
I	C	2	5.03, brs	77.1		100.1		100.3
		3	3.89, dd (2.0, 1.0)	73.2	4.05, d, (3.4)	68.3	4.13, d, (3.5)	67.8
		4	4.49, dd (2.0, 0.9)	37.0	4.45, d, (3.4)	29.5	4.20, d, (3.5)	29.5
	A	5		157.6*		157.1		156.8
		6	AB	AB	6.03, d (2.4)	98.3	5.89, d (2.4)	98.0
		7		157.8*		158.1		158.2
		8	AB	AB	6.08, d (2.4)	96.7	6.04, d (2.4)	96.5
		9		158.1*		154.2		154.0
		10		101.7		104.4		103.9
	B	1′		132.7		132.6		132.4
		2′	6.87, d (1.9)	115.0	7.16, d (2.1)	115.7	7.13, d (2.2)	115.7
		3′		145.8		145.6		145.6
		4′		145.5		146.7		146.8
		5′	6.71, d (8.2)	115.7	6.82, d (8.4)	115.6	6.80, d (8.4)	115.6
		6′	6.71, dd (8.2, 1.9)	119.1	7.03, dd (8.4, 2.1)	119.8	7.01, dd (8.4, 2.2)	119.9
II	F	2			5.31, brs	79.1	4.54, d (9.8)	84.5
		3			3.98, brd (1.9)	72.9	4.57, dd (9.8, 8.4)	73.0
		4			4.56, brs	37.9	4.48, brd (8.4)	38.6
	D	5				157.0		156.1
		6			6.06, s	96.3	5.97, s	97.6
		7				152.4		151.9
		8				106.8		107.0
		9				152.4		152.2
		10				105.6		108.8
	E	1′				131.8		131.3
		2′			7.13, d (2.1)	116.2	7.07, d (2.1)	116.3
		3′				145.9		146.2
		4′				146.2		146.8
		5′			6.80, d (8.3)	116.1	6.84, d (8.2)	116.2
		6′			6.96, dd (8.3, 2.1)	120.6	6.95, dd (8.2, 2.1)	121.1
phloroglucinol		1″		107.4		AB		107.5
		2″&6″		158.7		AB		159.7
		3″&5″	5.75, s	95.1	AB	AB	5.79, s	95.5
		4″		160.0		AB		160.1

^a,bNMR data were acquired at ^a255 K, and ^b298 K

^*interchangeable assignment; AB: absence of signals.

Table 10.

Comparison of the ¹³C NMR resonances of the terminal units in 11, 17, and 19 with those of the corresponding carbons in 2 and 5^a.

No	2	5	11	11vs2	11vs5	17	17vs2	17vs5	19	19vs2	19vs5
	δC	δC	δC	ΔδC	ΔδC	δC	ΔδC	ΔδC	δC	ΔδC	ΔδC
2	84.43	83.80	84.44	0.01	0.64	85.1	0.67	1.3	84.49	0.06	0.69
3	68.09	68.32	68.08	−0.01	−0.24	68.18	0.09	−0.14	68.13	0.04	−0.19
4	29.04	28.84	29.18	0.14	0.34	29.22	0.18	0.38	29.14	0.10	0.30
5	156.15	156.14	156.09	−0.06	−0.05	156.15	0	0.01	155.98	−0.17	−0.16
6	96.39	96.37	96.30	−0.09	−0.07	96.60	0.21	0.23	96.45	0.05	0.07
7	152.19	152.17	152.11	−0.08	−0.06	152.08	−0.11	−0.09	152.52	0.33	0.35
8	106.69	106.4	106.79	0.10	0.39	106.39	−0.3	−0.01	107.12	0.43	0.72
9	151.41	150.80	151.36	−0.05	0.56	151.29	−0.12	0.49	151.00	−0.41	0.20
10	103.01	102.69	103.11	0.10	0.42	103.08	0.07	0.39	102.80	−0.22	0.11
1’	130.47	130.77	130.37	−0.10	−0.40	129.95	−0.52	−0.82	130.43	−0.04	−0.34
2’	115.65	115.30	115.83	0.18	0.53	116.31	0.66	1.01	115.70	0.05	0.40
3’	146.30	146.35	146.21	−0.09	−0.14	146.11	−0.19	−0.24	146.20	−0.10	−0.15
4’	146.74	146.70	146.70	−0.04	0.00	146.92	0.18	0.22	146.70	−0.04	0.00
5’	116.22	116.24	116.28	0.06	0.04	116.78	0.56	0.54	116.00	−0.22	−0.24
6’	120.66	120.30	120.81	0.15	0.51	121.31	0.65	1.01	120.74	0.08	0.44
ΣΔδC				√0.22	2.47		√2.03	4.28	84.49	√−0.06	2.19

^aThe ¹³C NMR data (100 MHz) was acquired in CD₃OD with the temperature set as 278 K. ^√With smaller ΣΔδ_C value.

RESULTS AND DISCUSSION

The present study focused on those fractions of pine bark extract that were enriched in tri- and tetrameric PACs via a two-steps centrifugal partition chromatography (CPC) based on dentin bioassay guidance.^9,10 In view of practical limitations posed by bioassay throughput and costs, one valuable means of targeting the isolation of these PACs is to distinguish them from other compounds by following their bright pink-red colored bands on Si gel TLC plates after spraying with vanillin sulfuric acid and heating. A multi-step approach for the further purification of enriched bioactive fractions was applied to isolate the 13 PACs. The approach included Sephadex LH-20 and RP-18 column chromatography with eluents of different selectivity, as well as semi-preparative HPLC (see Experimental Section).

The isolation efforts focused on 12.1 g of enriched tri- and tetrameric PACs, obtained upon combining 6.5 g of fraction A and 5.6 g of fraction B that were prepared from 200 g of pine bark extract by centrifugal partition chromatography (CPC) as described previously.¹² Dentin treated with both fractions A and B presented statistically higher apparent moduli of elasticity compared to control (p<0.001), with an approximate 7- and 9-fold increase in the apparent modulus of elasticity values, respectively. As no statistically significant differences occurred among the PAC-treated groups (p<0.05; Figure 2). Considering a certain degree of overlap in constituents as per HPLC and TLC analysis, the materials were combined for the isolation of individual PACs.

Figure 2.

Mean and standard deviation of the apparent modulus of elasticity (MPa) of dentin beams treated with the CPC fractions A and B. Different symbols depict statistical differences among groups (p< 0.05).

Compounds 10 and 17 are the most abundant tri- and tetrameric PACs, respectively, in the investigated pine bark. Their structures are identical with those of the peanut procyanidins D and E,¹³ as established by comparison of their ¹H and ¹³C NMR data (Tables 3, 5–7). Phloroglucinolysis of 10 and 17 (Figure 5) further confirmed their structures. Purification of the phloroglucinolysis products resulted in the isolation of 1 and 2 from 10, as well as 2 and 3 from 17, respectively. In addition to dimer 2, the phloroglucinolysis monomer 1 and dimer 3 were used as stereochemical reference points in the subsequent structure elucidation of other trimers and tetramers.

Table 3.

The ¹H NMR data of 10−14 (CD₃OD, 800 MHz).

Unit	Ring	No	10a		11b	12b	13b	14b
			1H (mult., J in Hz)		1H (mult., J in Hz)	1H (mult., J in Hz)	1H (mult., J in Hz)	1H (mult., J in Hz)
			Rotamer A	Rotamer B	Major	Major	Major	Major A	Minor B
I	C	2	5.48, s	5.05, s	5.10, s	5.01, s
		3	3.80, brs	3.88, brs	3.96, brs	3.89, brs	3.45, d (3.5)	3.46, d (3.5)	4.07, d (3.3)
		4	4.64, brs	4.65, brs	4.82, brd (1.9)	4.77, brs	3.99, d (3.5)	3.99, d (3.5)	4.42, d (3.3)
	A	6	5.86, d (1.7)	5.92, d (1.7)	6.00, brs	5.95, d (1.9)	5.92, d (2.4)	5.92, d (2.3)	6.01, d (2.3)
		8	6.17, d (1.7)	5.98, d (1.7)	6.04, brs	5.98, d (1.9)	5.99, d (2.4)	5.99, d (2.3)	6.06, d (2.3)
	B	2’	6.93, d (2.1)	6.85, d (2.1)	6.87, d (1.8)	6.80, d (2.0)	7.08, d (2.2)	7.08, d (2.2)	7.14, d (2.2)
		5’	6.72, d (8.2)	6.68, d (8.2)	6.72, d (8.3)	6.66, d (8.2)	6.85, d (8.2)	6.85, d (8.2)	6.81, d (8.3)
		6’	6.61, dd (8.3, 2.1)	6.64, dd (8.3, 2.1)	6.68, dd (8.3, 1.8)	6.61, dd (8.2, 2.0)	6.93, dd (8.2, 2.2)	6.93, dd (8.2, 2.2)	7.02, dd (8.3, 2.2)
II	F	2					5.51, s	5.56, s	5.39, s
		3	4.05, d (3.6)	4.09, d (3.5)	4.18, d (3.5)	4.23, d (3.4)	4.04, brd (1.9)	4.06, brd (1.9)	4.00, brs
		4	4.06, d (3.6)	4.19, d (3.5)	4.35, d (3.5)	4.49, d (3.4)	4.53, brs	4.54, brs	4.72, brs
	D	6			5.98, s	5.85, s	5.82, s	5.83, s	6.08, s
		8	6.20, s	6.03, s
	E	2’	7.09, d (2.2)	7.11, d (2.2)	7.35, d (2.0)	7.30, d (2.2)	7.23, d (2.1)	7.61, brd (8.7)	7.48, brd (8.6)
		3’						6.86, brd (8.7)	6.81, brd (8.6)
		5’	6.81, d (8.1)	6.79, d (8.1)	6.85, d (8.1)	6.81, d (8.2)	6.84, d (8.2)	6.86, brd (8.7)	6.81, brd (8.6)
		6’	7.00, dd (8.1, 2.3)	7.01, dd (8.1, 2.3)	7.22, dd (8.1, 2.0)	7.19, dd (8.2, 2.2)	7.09, dd (8.2, 2.1)	7.61, brd (8.7)	7.48, brd (8.6)
III	I	2	4.36, d (9.4)	4.75, d (8.2)	4.78, d (7.9)	5.06, s	3.94, d (9.2)	3.96, d (9.1)	4.94, d (5.5)
		3	4.20, ddd (10.0, 9.4, 6.2)	3.89, ddd (8.8, 8.2, 5.6)	4.20, ddd (8.4, 7.9, 5.0)	4.26, dd (4.3, 2.0)	3.66, ddd (10.3, 9.2, 6.2)	3.66, ddd (10.2, 9.10, 6.2)	4.13, m
		4	2.47, dd (16.2, 10.0)	2.57, dd (16.2, 8.8)	3.02, dd (16.2, 5.0)	2.87, dd (16.9, 2.0)	2.41, d (16.1, 10.3)	2.41, d (16.1, 10.2)	2.63, m
			3.08, dd (16.2, 6.2)	3.00, dd (16.2, 5.6)	2.62, dd (16.2, 8.4)	2.94, dd (16.9, 4.3)	3.04, dd (16.1, 6.2)	3.04, dd (16.1, 6.2)	2.63, m
	G	6	6.11, s	6.11, s	6.09, s	6.03, s	6.09, s	6.10, s	5.86, s
	H	2’	6.74, d (2.2)	6.99, d (2.2)	6.98, d (1.6)	7.15, d (2.1)	6.75, d (2.01)	6.77, d (2.1)	6.87, d (2.2)
		5’	7.03, d (8.1)	6.86, d (8.1)	6.84, d (8.4)	6.84, d (8.2)	6.76, d (8.10)	6.73, d (8.1)	6.71, d (8.2)
		6’	6.79, dd (8.2, 2.2)	6.91, dd (8.2, 2.2)	6.88, dd (8.4, 1.6)	6.95, dd (8.2, 2.1)	6.67, dd (8.18, 2.04)	6.65, dd (8.1, 2.1)	6.88, dd (8.2, 2.1)

^aThe NMR data were acquired at 255 K, and

^bwere acquired at 278 K.

Table 5.

The ¹³C NMR data of 10−16 (CD₃OD, 100 MHz).

Unit	Ring	No	10a		11b	12b	13b	14b		15a		16b
			δC		δC	δC	δC	δC		δC		δC
			Rotamer A	Rotamer B	Major	Major	Major	Major A	Minor B	Major A	Minor B	Rotamer A	Rotamer B
I	C	2	77.7	77.0	77.1	77.1	100.0	100.0	100.0	76.7	78.0	100.2	100.2
		3	73.5	72.8	73.2	73.3	67.1	67.2	68.4	72.9	73.0	67.7	67.7
		4	36.7	37.5	37.1	37.2	28.8	28.8	29.4	37.5	36.6	29.4	29.4
	A	5	158.2	156.8	158.1	158.1	156.7	156.7	157.2	157.7	157.8	156.7	156.7
		6	95.6	95.5	95.9	96.3	98.3	98.2	98.2	95.6	95.9	97.8	97.8
		7	157.8	157.9	158.0	158.1	157.8	157.8	158.1	157.7	157.5	158.1	158.1
		8	95.3	95.7	96.3	96.1	96.4	96.4	96.5	95.6	95.3	96.4	96.4
		9	157.7	157.8	157.9	157.9	154.2	154.2	154.2	157.8	156.9	153.9	153.9
		10	103.4	102.2	101.9	101.9	105.0	104.9	104.2	102.1	103.2	103.7	103.8
	B	1’	132.9	132.6	132.5	132.6	132.4	132.4	132.6	132.0	132.7	132.2	132.2
		2’	115.2	115.1	115.1	115.1	115.7	115.8	115.6	115.0	115.5	115.5	115.5
		3’	145.9	145.8	145.7	145.8	145.5	145.5	145.6	145.5	145.5	145.5	145.5
		4’	145.6	145.5	145.4	145.5	146.6	146.6	146.7	145.8	145.5	146.7	146.7
		5’	115.8	115.6	115.8	115.8	115.7	115.6	115.8	115.6	115.8	115.5	115.5
		6’	119.4	119.1	119.3	119.3	120.0	120.0	119.7	119.1	120.1	119.8	119.8
II	F	2	99.9	100.1	100.5	100.8	78.6	78.5	79.0	100.0	99.7	84.8	84.8
		3	67.6	67.7	67.4	67.5	72.4	72.3	72.7	67.6	68.0	73.7	73.7
		4	29.1	29.6	29.3	29.5	38.3	38.2	37.7	29.5	29.0	38.7	39.5
	D	5	154.9	154.1	155.2	155.3	155.9	155.9	157.0	154.5	155.1	156.1	156.0
		6	111.3	110.6	99.2	99.0	95.9	95.9	96.1	110.4	110.9	97.6	97.3
		7	156.2	157.1	157.0	157.1	151.0	151.0	152.4	157.2	156.5	151.8	152.1
		8	96.4	97.5	108.3	108.4	106.1	106.1	106.6	97.4	96.5	107.0	106.8
		9	152.0	152.1	151.1	151.2	151.7	151.8	152.4	152.1	152.2	152.5	152.5
		10	104.2	103.9	103.8	103.9	106.5	106.4	105.8	104.2	103.8	108.5	108.8
	E	1’	132.1	132.1	132.2	132.3	131.5	130.8	130.3	132.5	132.1	131.1	131.1
		2’	115.5	115.5	115.5	115.6	116.5	130.6	131.2	115.5	115.5	116.0	116.0
		3’	145.6	145.6	145.6	145.7	146.0	116.0	115.8	145.6	145.4	146.1	146.1
		4’	146.7	146.8	146.7	146.8	146.3	158.3	158.2	146.7	146.6	146.7	146.7
		5’	115.4	115.4	115.7	115.7	116.1	116.0	115.8	115.4	115.5	116.0	116.0
		6’	119.7	119.7	120.0	120.1	121.0	130.6	131.2	119.8	119.7	121.0	121.0
III	I	2	85.7	85.2	84.4	80.8	83.2	83.2	81.9	82.8	82.6	82.5	82.5
		3	68.2	68.9	68.1	67.2	70.1	70.0	68.3	67.0	66.9	68.7	68.8
		4	31.2	29.2	29.2	29.6	30.7	30.7	26.6	30.7	30.4	28.7	29.2
	G	5	156.3	156.3	156.1	156.7	155.4	155.4	155.4	157.0	156.8	154.6	155.6
		6	96.7	96.4	96.3	96.3	96.3	96.3	96.9	96.6	96.5	110.7	110.9
		7	152.2	152.2	152.1	152.1	155.6	155.6	156.3	152.3	151.7	156.4	155.6
		8	106.9	106.3	106.8	107.0	108.7	108.7	108.2	106.8	106.5	96.0	97.1
		9	151.8	151.2	151.4	151.3	155.3	155.3	153.9	152.2	152.6	154.7	154.6
		10	103.8	103.1	103.1	101.9	101.5	101.5	100.3	101.9	102.4	101.9	101.5
	H	1’	129.6	129.9	130.4	131.4	132.6	132.7	132.5	130.4	131.0	132.0	132.0
		2’	117.9	115.8	115.8	115.1	115.7	115.6	114.3	116.3	115.8	115.1	115.1
		3’	145.1	146.2	146.2	146.2	145.8	146.0	146.1	145.8	145.5	146.1	146.1
		4’	146.5	146.9	146.7	146.2	145.8	145.8	145.8	146.7	146.2	146.1	146.1
		5’	118.2	116.6	116.3	116.1	116.1	116.0	116.0	115.9	117.2	116.0	116.0
		6’	119.6	121.2	120.8	119.3	119.9	120.0	119.3	121.3	120.9	120.0	120.0

^aThe NMR data were acquired at 255 K, and

^awere acquired at 278 K.

Table 7.

The ¹³C NMR data of 17−22 (CD₃OD, 100 MHz, 278 K).

Unit	Ring	No	17	18	19		20		21		22
			δC	δC	δC	δC	δC	δC	δC	δC	δC
			Major	Major	Major A	Minor B	Major A	Minor B	Rotamer A	Rotamer B	Major
I	C	2	100.0	100.1	100.0	99.9	100.1	99.9	100.0	100.0	100.2
		3	68.2	68.3	68.3	67.0	68.3	67.0	68.3	67.8	67.4
		4	29.4	29.4	29.4	28.8	29.4	28.8	29.2	29.1	29.2
	A	5	157.1	157.1	157.1	156.7	157.1	156.8	157.1	156.9	156.5
		6	98.2	98.3	98.1	98.2	98.2	98.2	98.3	98.2	97.8
		7	158.0	158.1	158.0	157.7	158.1	157.8	158.1	158.0	158.1
		8	96.6	96.6	96.5	96.5	96.6	96.5	96.5	96.5	96.5
		9	154.1	154.1	154.1	154.1	154.1	154.1	154.1	154.3	154.0
		10	104.3	104.3	104.2	105.0	104.2	105.1	104.3	104.8	104.1
	B	1’	132.5	132.5	132.5	132.4	132.6	132.5	132.6	132.5	132.2
		2’	115.6	115.6	115.7	115.7	115.5	115.8	115.6	115.6	115.7
		3’	145.6	145.6	145.6	145.4	145.6	145.4	145.8	145.6	145.5
		4’	146.7	146.7	146.7	146.5	146.7	146.6	146.7	146.7	146.7
		5’	115.6	115.5	115.7	115.7	115.8	116.0	115.5–116.0	115.5–116.0	115.7
		6’	119.7	119.7	119.7	119.8	119.7	119.8	119.7	119.9	119.9
II	F	2	78.9	78.7	79.0	78.6	79.0	78.7	78.8	79.4	84.6
		3	72.4	72.4	73.0	72.1	73.1	72.1	72.3	73.1	73.3
		4	38.3	38.5	38.1	38.8	38.2	38.9	38.2	37.3	38.8
	D	5	156.5	156.3	156.5	156.5	156.6	156.7	156.6	156.1	155.5
		6	96.3	96.1	96.4	96.0	96.4	96.0	95.9	96.0	97.5
		7	152.1	152.2	152.4	151.2	152.4	151.2	152.2	151.7	151.3
		8	106.7	106.9	106.9	106.0	107.0	105.9	106.7	106.3	106.7
		9	152.3	152.4	152.4	151.4	152.2	151.4	152.4	152.3	153.0
		10	106.0	105.5	105.5	106.0	105.5	106.1	105.9	106.3	109.3
	E	1’	131.6	131.5	131.5	131.6	131.5	131.7	131.6	131.9	131.1
		2’	116.2	116.2	116.1	116.5	116.1	116.6	116.2	116.0	116.3
		3’	145.8	145.9	145.7	145.9	146.0	145.9	146.0	145.7	146.1
		4’	146.1	146.2	146.1	146.3	146.3	146.3	146.2	146.0	146.2
		5’	115.9	115.9	116.0	116.0	116.0	116.0	115.5–116.0	115.5–116.0	116.0
		6’	120.6	120.7	120.6	121.2	120.6	121.2	120.6	120.3	121.2
III	I	2	100.2	100.1	100.6	99.4	100.6	99.3	100.2	100.3	100.3
		3	67.7	67.7	67.5	67.6	67.8	68.0	67.8	67.8	67.7
		4	29.7	29.6	29.4	29.5	29.5	29.6	29.7	29.3	29.2
	G	5	154.1	154.5	155.2	155.0	155.5	155.1	154.4	154.3	154.0
		6	111.2	110.6	99.1	97.9	99.4	98.2	110.8	111.7	111.6
		7	156.7	156.9	156.8	156.1	156.8	156.2	156.8	156.2	156.8
		8	97.5	97.5	108.6	109.5	108.7	109.8	97.6	96.6	96.0
		9	152.1	152.2	151.2	152.5	152.4	152.7	152.0	152.3	152.0
		10	104.2	104.3	103.7	103.3	104.1	103.8	103.9	104.6	104.4
	H	1’	132.0	132.1	132.2	133.1	132.3	133.3	132.0	132.2	132.5
		2’	115.6	115.6	115.5	115.6	115.5	116.0	115.6	115.6	115.6
		3’	145.6	145.6	145.6	145.7	145.4	145.7	145.7	145.6	145.8
		4’	146.7	146.8	146.7	146.1	146.8	146.3	146.8	146.8	146.8
		5’	115.5	115.6	115.5	115.2	115.6	115.2	115.5–116.0	115.5–116.0	115.5
		6’	119.8	119.8	120.0	120.5	120.0	120.5	119.8	119.9	119.8
IV	L	2	85.1	82.7	84.5	84.5	81.8	81.7	80.9	81.1	80.9
		3	68.2	66.9	68.1	68.2	66.8	67.0	67.2	67.3	67.5
		4	29.2	30.5	29.1	29.2	30.0	30.0	29.5	29.4	29.5
	J	5	156.2	157.0	156.1	155.8	156.7	156.4	156.7	156.7	156.8
		6	96.6	96.7	96.4	96.5	96.5	96.5	96.5	96.8	96.5
		7	152.1	152.2	152.2	152.9	152.2	153.0	152.2	151.9	152.0
		8	106.4	106.8	106.8	107.4	107.2	107.8	106.7	106.9	106.7
		9	151.3	152.2	151.3	150.7	152.4	151.5	151.0	151.1	151.0
		10	103.1	102.2	103.1	102.5	102.3	101.7	101.7	101.7	101.9
	K	1’	130.0	130.5	130.4	130.4	131.3	131.1	131.3	131.5	131.8
		2’	116.3	116.6	115.7	115.7	116.0	115.8	115.0	114.9	114.8
		3’	146.1	145.9	146.2	146.2	146.1	146.0	146.2	146.0	146.2
		4’	146.9	146.7	146.7	146.7	146.0	146.3	146.3	145.4	145.3
		5’	116.8	116.1	116.0	116.0	115.8	115.8	115.5–116.0	116.6	116.8
		6’	121.3	121.4	120.8	120.7	120.4	120.2	119.4	119.3	119.5

Compound 1 was assigned as 4-phloroglucinol substituted (ent)-epicatechin. Its molecular formula C₂₁H₁₈O₉ was confirmed based on the (+)-HRESIMS [M + H]⁺ ion at 415.1022 and the ¹³C NMR carbon counts (Table 1). The relative configuration of 1 was assigned as 2,3-cis-3,4-trans, by comparing its ¹³C NMR resonances of C-2 (δ_C 77.1), C-3 (δ_C 73.2), and C-4 (δ_C 37.0) with reported values.¹⁴ The observed high-amplitude positive Cotton effect (CE) at 220−240 nm in the electronic circular dichroism (ECD) spectrum (Figure 3A) of 1 showed that the 4-phloroglucinol group must be β-configured.^15–17 The absolute configuration of 1 was, thus, assigned as (2R, 3R, 4S)-epicatechin-(4β→2)-phloroglucinol.

Figure 3.

ECD spectra of 1 (A) as well as 3, 4 and 6 (B and C showing expansions of the 260–300 nm region).

Compound 2 was identified by comparing its ¹H and ¹³C NMR data (Table 2, Figure S6) and ECD (Figure S7) with reported data.⁹ However, based on the analysis of its 2D NMR data, the assignments of several carbon atoms in 2⁹ should be revised as indicated.

Table 2.

¹³C NMR data (CD₃OD, 100 MHz) of dimers 2 and 4–6 acquired at three different temperatures: 255, 278, and 298 K

Unit	Ring	No	2			4			5			6
			δC			δC			δC			δC
			298 K	278 K	255 K	298 K	278 K	255 K	298 K	278 K	255 K	298 K	278 K	255 K
I	C	2	100.28	100.22	100.13	100.16	100.05	99.94	100.40	100.32	100.24	100.42	100.37	100.30
		3	67.74	67.75	67.71	68.06	67.99	67.98	67.64	67.56	67.50	67.73	67.70	67.65
		4	29.16	29.13	29.12	29.23	29.13	29.05	29.22	29.12	29.03	29.23	29.17	29.07
	A	5	156.71	156.78	156.84	156.99	156.98	157.00	156.66	156.63	156.64	156.70	156.73	156.72
		6	98.14	98.01	97.83	98.29	98.11	97.92	98.14	98.02	97.89	97.95	97.84	97.71
		7	158.08	158.11	158.06	158.11	158.03	158.00	158.12	158.10	158.05	158.10	158.10	158.05
		8	96.52	96.39	96.26	96.61	96.49	96.36	96.58	96.47	96.35	96.54	96.43	96.31
		9	154.19	154.19	154.16	154.24	154.18	154.14	154.08	154.02	153.98	154.14	154.12	154.06
		10	104.00	103.89	103.76	104.25	104.12	103.98	104.03	103.91	103.80	104.06	103.97	103.86
	B	1’	132.24	132.22	132.15	132.43	132.34	132.28	132.21	132.10	132.01	132.30	132.24	132.16
		2’	115.66	115.59	115.48	115.65	115.54	115.43	115.71	115.60	115.50	115.63	115.53	115.43
		3’	145.59	145.61	146.70	145.64	145.57	145.52	145.65	145.59	145.55	145.65	145.64	145.60
		4’	146.73	146.74	145.56	146.75	146.67	146.62	146.80	146.74	146.69	146.78	146.76	146.72
		5’	115.62	115.51	115.41	115.54	115.53	115.43	115.63	115.54	115.43	115.70	115.67	115.52
		6’	119.81	119.77	119.71	119.78	119.72	119.67	119.85	119.79	119.74	119.85	119.8	119.75
II	F	2	84.41	84.43	84.36	81.73	81.60	81.48	83.87	83.80	83.75	80.82	80.74	80.61
		3	68.06	68.09	68.10	66.96	66.93	66.97	68.38	68.32	68.29	67.13	67.10	67.05
		4	28.90	29.04	29.04	29.88	29.87	29.90	28.83	28.84	28.90	29.46	29.51	29.55
	D	5	156.09	156.15	156.14	156.59	156.55	156.56	156.15	156.14	156.16	156.62	156.65	156.66
		6	96.52	96.39	96.20	96.48	96.31	96.12	96.52	96.37	96.22	96.46	96.30	96.14
		7	152.14	152.19	152.20	152.28	152.24	152.24	152.21	152.17	152.16	152.06	152.05	152.02
		8	106.75	106.69	106.60	107.20	107.10	107.03	106.54	106.40	106.29	106.92	106.84	106.75
		9	151.36	151.41	151.42	152.13	152.13	152.18	150.83	150.80	150.79	151.28	151.28	151.27
		10	103.09	103.01	102.89	102.41	102.28	102.21	102.81	102.69	102.58	101.89	101.78	101.66
	E	1’	130.51	130.47	130.39	131.19	131.18	131.22	130.92	130.77	130.63	131.40	131.37	131.33
		2’	115.71	115.65	115.56	115.92	115.79	115.64	115.42	115.30	115.18	115.20	115.07	114.91
		3’	146.29	146.30	146.25	145.97	145.88	145.81	146.41	146.35	146.33	146.20	146.2	146.16
		4’	146.73	146.74	146.70	146.28	146.17	146.08	146.75	146.70	146.69	146.20	146.17	146.11
		5’	116.31	116.22	116.10	116.03	115.92	115.81	116.34	116.24	116.14	116.19	116.09	115.98
		6’	120.68	120.66	120.64	120.37	120.29	120.18	120.33	120.30	120.31	119.42	119.30	119.15

Compound 3 was recognized as a phloroglucinol substituted A-type dimer, and its molecular formula was determined as C₃₆H₂₈O₁₅ by the (+)-HRESIMS [M + H]⁺ ion at 701.1504 and the ¹³C NMR data (Table 1). The ECD spectrum of 3 exhibited a high-amplitude positive CE in the region 220−240 nm (Figure 3B), which assigned the double linkages (2→O→7,4→8) and the 4-phloroglucinol moiety as being β-oriented. Comparison of the ECD spectra of 3 with those of 4 and 6 revealed that compound 3 exhibited a negative CE in the region 280−290 nm analogous to that of 4 (Figure 3C). The absolute configuration of (2R, 3R) in unit II of 3 was thus determined.^15–17 Considering the diagnostic power of the γ-gauche effect, the upfield shifted C-2 (δ_C79.1, ring F) compared to the corresponding carbon in 4 (δ_C 81.5) suggested H-2 and H-4 to be trans-positioned.^18,19 This further corroborated the absolute configuration in unit II as being 2R, 3R. Accordingly, the structure of 3 had to be epicatechin-(2β→O→7,4β→8)-epicatechin-(4β→2)-phloroglucinol or [PAC 4]-(4β→2)-phloroglucinol.

Collectively, this evidence also assigned the structures of the major constituents, 10 and 17, unambiguously as being assembled by two corresponding moieties, 2 and 4, assigning them as the trimer, epicatechin-(4β→6)-[PAC 2], and the tetramer, [PAC 4]-(4β→6)-[PAC 2].

Compound 9 was purified from the phloroglucinolysis products of 16, and was identified as a phloroglucinol trapped A-type dimer, an assignment supported by its molecular formula C₃₆H₂₈O₁₅ as confirmed by the (+)-HRESIMS [M + H]⁺ peak at 701.1546 and the ¹³C NMR data (Table 1). Analyzing the ¹H NMR spectra (Figure S16) of the phloroglucinolysis product of 17 revealed that its products were the same, but present in different molar ratios compared to those of 16. The products of 17 consisted of 3 as major and 9 as minor in terms of their molar ratio, while 9 was found as major and 3 as minor phloroglucinolysis products of 16. Compounds 3 and 9 were, thus, assigned as C-2 epimerization products during phloroglucinolysis.²⁰ Accordingly, 9, was recognized as a C-2 epimer of 3 and elucidated as [PAC 5]-(4β→2)-phloroglucinol.

Compound 11 was isolated and identified in a previous study,⁹ during which the relative configuration of H-2 and H-4 in the C-ring was determined as being cis-configured based on the NOE between H-2 and H-4 observed in the NOESY spectrum. However, it is known that NOESY data can be ambiguous or inconclusive in case of spin diffusion. As the previously reported 2,3-cis-3,4-cis relative configuration of natural PACs is rarely found in the literature, we challenged the prior evidence and pursued additional evidence to establish both the relative and absolute configuration. The upfield shifted C-2 (δ_C 77.1, C-ring; Table 5) compared to the signal from the corresponding carbon in epicatechin (δ_C 79.1) suggested that H-2 and H-4 are in fact trans-positioned when considering earlier knowledge about the γ-gauche effect.^18,19 Difference in the J_3,4 values in 2,3-cis-3,4-cis-(3.8~4.5 Hz) vs. 2,3-cis-3,4-trans-(1~2 Hz) configured moieties further indicated the need to revise the prior assignment of 11, because its J_3,4 (ca 1.9 Hz) was congruent with 2,3-cis-3,4-trans relative configuration.^14,21,22 Reductive cleavage of 11 by phloroglucinol formed compounds 1 and 2, which were verified by analysis of their MS data and chiral chromatograms (Figure 5). In summary of the new evidence, the structure of 11 should be revised to that of epicatechin-(4β→8)-[PAC 2].

Compound 12 shared the molecular formula C₄₅H₃₆O₁₈ with 11 based on the (+)-HRESIMS [M + H]⁺ ion at 865.1972 and the ¹³C NMR carbon counts. Comparison of the NMR data (Tables 3 and 5) revealed that 12 is structurally closely related with 11, with the major differences were observed in resonances within unit III. Additionally, close similarities in the ECD curves of 12, especially the positive CE in diagnostic 220−240 nm region (Figure 4A), as compared to the ECD of 11, suggested that the only differences between 12 and 11 were the configurations of C-2 and C-3 in unit III. A cis-configuration of III-H-2 and III-H-3 could be gleaned from the unresolved thus small J_2,3 value. Accordingly, units II and III had to be linked like in PACs 4 or 6. Owing to the spatial distance between the upper unit I and the terminal unit III, the chemical shift impact of unit I on unit III had to be small, especially in the ¹³C domain due to the shielding of the molecular hydrogen “envelope”. Differential carbon chemical shift (Δδ_C) could, thus, be used to determine the absolute configuration of unit III: units II and III were determined to be doubly connected as in 6 based on the close similarity of the ¹³C NMR data of the terminal unit III to that of the corresponding carbons in 6 (Table 8). Thereby, the structure of 12 including its absolute configuration was elucidated as epicatechin-(4β→8)-[PAC 6]. Although this structure had been reported earlier,²³ the insufficiency of available NMR data and prior assignment puts its validity into question. Chiral phase HPLC and MS analysis of the phloroglucinolysis products of 12 further confirmed the compound to be assembled from dimer 6 and epicatechin (Figure S104).

Figure 4.

ECD spectra of the trimeric PACs 10−16 (A) and the tetrameric PACs 17−22 (B).

Table 8.

Comparison of the ¹³C NMR resonances of the terminal catechol units in 12, 18, and 20−22 with those of the corresponding carbons in 4 and 6^a.

NO	6	4	12	12vs6	12vs4	18	18vs6	18vs4	20	20vs6	20vs4	21	21vs6	21vs4	22	22vs6	22vs4
	δC	δC	δC	ΔδC	ΔδC	δC	ΔδC	ΔδC	δC	ΔδC	ΔδC	δC	ΔδC	ΔδC	δC	ΔδC	ΔδC
2	80.74	81.60	80.79	0.05	−0.81	82.72	1.98	1.12	81.71	0.97	0.11	81.00	0.26	−0.60	80.91	0.17	−0.69
3	67.10	66.93	67.16	0.06	0.23	66.90	−0.20	−0.03	66.94	−0.16	0.01	67.25	0.15	0.31	67.54	0.44	0.61
4	29.51	29.87	29.58	0.07	−0.29	30.50	0.99	0.63	30.00	0.49	0.13	29.43	−0.09	−0.45	29.48	−0.03	−0.39
5	156.65	156.55	156.73	0.08	0.18	156.96	0.31	0.41	156.56	−0.09	0.00	156.73	0.08	0.18	156.80	0.15	0.25
6	96.30	96.31	96.31	0.01	0	96.70	0.40	0.39	96.47	0.17	0.16	96.63	0.33	0.31	96.53	0.23	0.22
7	152.05	152.24	152.05	0	−0.19	152.24	0.19	0.00	152.61	0.55	0.36	152.06	0.00	−0.19	151.97	−0.08	−0.27
8	106.84	107.10	106.96	0.12	−0.14	106.75	−0.09	−0.35	107.51	0.67	0.41	106.80	−0.05	−0.31	106.70	−0.14	−0.40
9	151.28	152.13	151.34	0.06	−0.79	152.40	1.12	0.27	151.94	0.66	−0.19	151.06	−0.22	−1.07	150.95	−0.33	−1.18
10	101.78	102.28	101.88	0.1	−0.4	102.19	0.41	−0.09	102.01	0.22	−0.28	101.71	−0.07	−0.57	101.87	0.09	−0.41
1’	131.37	131.18	131.37	0	0.19	130.47	−0.90	−0.71	131.21	−0.16	0.03	131.39	0.02	0.21	131.76	0.39	0.58
2’	115.07	115.92	115.08	0.01	−0.84	116.60	1.53	0.68	115.90	0.83	−0.02	114.97	−0.10	−0.95	114.80	−0.27	−1.12
3’	146.20	145.88	146.24	0.04	0.36	145.86	−0.34	−0.02	146.05	−0.15	0.17	146.08	−0.12	0.20	146.21	0.01	0.33
4’	146.17	146.17	146.21	0.04	0.04	146.70	0.53	0.53	146.15	−0.02	−0.02	145.84	−0.33	−0.33	145.26	−0.91	−0.91
5’	116.09	115.54	116.11	0.02	0.57	116.07	−0.02	0.53	115.80	−0.29	0.26	NA	NA	NA	116.81	0.72	1.27
6’	119.30	120.29	119.28	−0.02	−1.01	121.41	2.11	1.12	120.33	1.03	0.04	119.36	0.06	−0.93	119.47	0.17	−0.82
ΣΔδC				√0.64	−2.9		8.02	√4.48		4.72	√1.18		√−0.08	−4.17		√0.61	−2.93

^aThe ¹³C NMR data (100 MHz) was acquired in CD₃OD with the temperature be set as 278 K. ^√With smaller ΣΔδ_C value.

The structure of 13 was established as that of cinnamtannin D1 by comparison of its ¹H and ¹³C NMR data (Tables 3 and 5), as well as it ECD spectrum (Figure 4A) with reported data.^24–26 Phloroglucinolysis dissociated 13 into two major compounds, 3 and catechin, the structures of which were verified by chiral phase HPLC and MS (Figure S104), collectively confirming 13 as [PAC 4]-(4β→8)-catechin, previously named cinnamtannin D1.

Comparison of their NMR (Tables 3 and 5) and ECD data (Figure 4A) indicated that the structure of 14 was closely related with that of 13. The major difference was the presence of para-hydroxyphenyl signals in 14 vs. the AMX ¹H NMR resonances of 13. One pair of highly coupled signals at (δ_H 7.61 and 6.86, each 2H, br d, J = 8.6) was characteristic for a para-substituted phenyl, which was identified as a E-ring via HMBC correlations (Figure S41) connecting II-H-2′&6′ (δ_H 7.61) to II-C-2 (δ_C 78.5). Thus, 14 was shown to contain epiafzelechin as unit II, which was also consistent with the (+)-HRESIMS [M + H]⁺ peak at 849.2045, 16 mass units less than that of 13. Collectively, the structure of 14 was assigned as II-3′-OH-nor-cinnamtannin D1. Phloroglucinolysis with chiral HPLC and MS data (Figure S105) affirmed the terminal unit III as catechin.

Compound 15 exhibited a (+)-HRESIMS [M + H]⁺ peak at 865.1960, corresponding to a molecular formula of C₄₅H₃₆O₁₈, which was consistent with its ¹³C NMR data. Compound 15 was thus assigned as a trimeric PAC with one A- and one B-type IFL. Two major rotamers in a ~3:1 ratio were observed in the ¹H and ¹³C NMR spectra (Tables 4 and 5) acquired at low temperature (255 K) for restricted rotation around the B-type linkage.^9–11,13 The doubly linked interflavanyl bonds (2→O→7,4→8) between units II and III were verified by the HMBC correlations (Figure S54) from both II-H-4 (δ_H 4.33) and III-H-2 (δ_H 4.95) to III-C-9 (δ_C 152.2). Unit I was attached to unit II through a 4→6 bond as derived from the NOESY correlations from II-H-8/II-H-2′ and II-H-6′ (Figure S55), as well as the downfield shifted II-C-6 carbons (avg. δ_C 110.6).¹³ The ECD data (Figure 4A) showed a high-amplitude positive CE at 220−240 nm, which assigned a β-configuration to the interflavanyl bonds. Both units I and III were confirmed as (ent)-epicatechin based on the singlet signals I-H-2 (δ_H 5.04/5.63, s) and III-H-2 (δ_H 4.95/4.76, s). The upfield shifted chemical shift of the C-ring C-2 carbons (avg δ_C 77.4) suggested H-2 and H-4 to be trans-configured based on the γ-gauche effect.^18,19 Unit I was, thus, identified as epicatechin with a (2R, 3R)-absolute configuration. The 3,4-trans configuration in the C-ring was assigned based on the NOESY correlation between II-H-3 and III-H-6.¹⁶ Finally, the diagnostic Δδ_C patterns were applied to determine the absolute configuration of the terminal unit III: it was identified as epicatechin as its ¹³C resonances were significantly closer to those of the corresponding carbons in 4 than to those in 6 (Table 9). To further confirm this conclusion, phloroglucinolysis of 15 was performed. The reaction products were identified as 1 and 4 by MS and chiral phase HPLC (Figure 5). Thus, compound 15 was unambiguously elucidated as epicatechin-(4β→6)-[PAC 4]. While this structure has been reported previously,²⁷ the prior NMR data did not match the present data set. This mismatch can be explained by the incorrect assignment of II-C-10 (δ_C 104.9) in the previous report,²⁷ which led to establishment of an incorrect 4→6 linkage through interpretation of HMBC correlations. The chemical shift previously reported for II-C-6 (δ_C 108.9) better matches the corresponding carbons in 11 and 12. This indicated that the compound reported earlier²⁷ in fact had a 4β→8 IFL. Thus, its structure should be revised to epicatechin-(4β→8)-epicatechin-(4β→8,2β→7)-epicatechin.²⁸ However, considering the subtle substituent chemical shift (s.c.s.) effects uncovered in the present study, it is not surprising that the ¹³C NMR data corresponding to the structure of epicatechin-(4β→8)-epicatechin-(4β→8,2β→7)-epicatechin in refs ²⁷ and ²⁸ are inconsistent. Recognition of such inconsistencies and detection of potential misassignment in PAC structures remains challenging for authors and reviewers alike, and this emphasizes the urgent need for consolidated collections of raw NMR spectroscopic data.²⁹

Table 4.

The ¹H NMR data of 15, 16, and 22 (CD₃OD, 800 MHz).

Unit	Ring	No	15a		16b		22b
			1H (mult., J in Hz)		1H (mult., J in Hz)		1H (mult., J in Hz)
			Major A	Minor B	Rotamer A	Rotamer B	Major
I	C	2	5.04, s	5.63, s
		3	3.92, d (1.8)	3.96, d (1.8)	4.12, d (3.4)	4.09, d (3.4)	4.25, d (3.5)
		4	4.65, d (1.8)	4.66, d (1.8)	4.21, d (3.4)	4.19, d (3.4)	3.91, d (3.5)
	A	6	5.92, d (1.7)	5.86, brs	5.89, d (2.4)	5.89, d (2.4)	5.84, d (2.4)
		8	5.98, d (1.7)	6.17, brs	6.04, d (2.4)	6.04, d (2.4)	6.03, d (2.4)
	B	2’	6.86, d (2.1)	7.00, d (2.1)	7.10, d (2.2)	7.11, d (2.2)	7.16, d (2.1)
		5’	6.70, d (8.1)	6.72, d (8.2)	6.79, d (8.3)	6.79, d (8.3)	6.80, d (8.2)
		6’	6.64, dd (8.1, 2.1)	6.88, dd (8.3, 2.1)	6.98, dd (8.3, 2.2)	6.99, dd (8.3, 2.2)	7.03, dd (8.3, 2.1)
II	F	2			4.57a	4.54a	4.34, d (9.8)
		3	4.14, d (3.5)	4.00, d (3.6)	4.53a	4.51a	4.37, dd (9.8, 7.8)
		4	4.33, d (3.5)	4.19, d (3.6)	4.55a	4.43, d (7.6)	4.44, d (7.8)
	D	6			6.00, s	5.98, s	5.86, s
		8	6.04, s	6.21, s
	E	2’	7.13, d (2.2)	7.10, d (2.2)	7.07, d (2.2)	7.07, d (2.2)	6.99, d (2.2)
		5’	6.82, dd (8.1)	6.79, dd (8.1)	6.84, d (8.2)	6.84, d (8.2)	6.89, d (8.1)
		6’	7.03, dd (8.1, 2.2)	6.99, dd (8.1, 2.3)	6.96, dd (8.2, 2.2)	6.96, dd (8.2, 2.2)	6.80, dd (8.1, 2.2)
III	I	2	4.95, s	4.76, s	4.50*	4.57*
		3	4.17, m	3.98, m	3.97*	3.98*	4.14, d (3.6)
		4	2.83, brd (16.9)	2.79, brd (16.5)	2.46, dd (15.8, 8.4)	2.60, dd (16.1, 8.2)	4.34, d (3.6)
			3.00, dd (16.9, 4.5)	2.90, dd (16.5, 4.2)	2.78, dd (15.8, 5.4)	2.91, dd (16.1, 5.3)
	G	6	6.13, s	6.12, s
		8			6.00, s	5.90, s	6.18, s
	H	2’	7.29, d (2.2)	6.84, d (2.2)	6.84, d (2.0)	6.82, d (2.0)	7.13, d (2.2)
		5’	6.82, d (8.1)	6.97, d (8.1)	6.74, d (8.1)	6.75, d (8.1)	6.81, dd (8.3)
		6’	6.99, dd (8.2, 2.2)	7.02, dd (8.2, 2.2)	6.72, dd (8.1, 2.0)	6.71, dd (8.1, 2.0)	7.02, dd (8.1, 2.3)
IV	L	2					5.21, brs
		3					4.32, m
		4					2.86, brd (17.0)
							2.96, dd (17.0, 4.4)
	J	6					6.07, s
	K	2’					7.28, d (2.1)
		5’					6.94, d (8.1)
		6’					7.12, dd (8.2, 2.1)

^*Signals are overlapped

^aThe NMR data were acquired at 255 K, and

^bwere acquired at 278 K.

Table 9.

Comparison of the ¹³C NMR resonances of the terminal units in 10 and 15 with those of the corresponding carbons in 2/5 and 4/6^a.

No	2	5	10	10vs2	10vs5	4	6	15	15vs6	15vs4
	δC	δC	avg δC	ΔδC	ΔδC	δC	δC	avg δC	ΔδC	ΔδC
2	84.36	83.75	85.45	1.09	1.70	81.48	80.61	82.66	2.05	1.18
3	68.10	68.29	68.51	0.41	0.22	66.97	67.05	66.92	−0.13	−0.05
4	29.04	28.90	30.17	1.13	1.27	29.90	29.55	30.55	1.00	0.65
5	156.14	156.16	156.30	0.16	0.14	156.56	156.66	156.91	0.25	0.34
6	96.20	96.22	96.56	0.36	0.34	96.12	96.14	96.54	0.40	0.42
7	152.20	152.16	152.2	0.00	0.04	152.24	152.02	152.01	−0.01	−0.23
8	106.60	106.29	106.595	0.00	0.30	107.03	106.75	106.60	−0.15	−0.43
9	151.42	150.79	151.495	0.08	0.71	152.18	151.27	152.40	1.12	0.21
10	102.89	102.58	103.41	0.52	0.83	102.21	101.66	102.19	0.53	−0.02
1’	130.39	130.63	129.74	−0.65	−0.89	131.22	131.33	130.69	−0.64	−0.53
2’	115.56	115.18	116.85	1.29	1.67	115.64	114.91	116.06	1.15	0.42
3’	146.25	146.33	145.67	−0.58	−0.66	145.81	146.16	145.67	−0.49	−0.14
4’	146.70	146.69	146.685	−0.01	0.00	146.08	146.11	146.43	0.32	0.35
5’	116.10	116.14	117.38	1.28	1.24	115.81	115.98	116.56	0.58	0.75
6’	120.64	120.31	120.375	−0.27	0.06	120.18	119.15	121.09	1.94	0.91
ΣΔδC				√4.80	6.97				8.89	√4.46

^aThe ¹³C NMR data (100 MHz) was acquired in CD₃OD with the temperature set as 255 K. ^√With smaller ΣΔδ_C value.

The A + B-type trimeric PAC nature of 16 was supported by its molecular formula C₄₅H₃₆O₁₈, which was established by the (+)-HRESIMS [M + H]⁺ peak at 865.2007 and its ¹³C NMR data. Two sets of ¹H NMR data (Table 4) representing two rotamers (~1:1) were observed in the low temperature experiment at 278 K as indicative of the presence of atropisomers in PACs involving B-type IFLs. Both units II and III were identified as (ent)-catechin based on the J_3,4 = 7.6 Hz for unit II and J_3,4 = (8.4 and 5.4 Hz) for unit III. Units I and II were doubly linked as an A-type dimer through 2→O→7/4→8 IFLs, as shown by the NOESY cross-peaks from II-H-2′ and II-H-6′ to I-H-4. The terminal unit III was shown to be linked to unit II by a 4→6 bond based on the shift of II-C-6 (δ_C 110.8) to lower field compared to the corresponding carbons II-C-8 (δ_C 108.7) of cinnamtannin D1 and II-C-8 (δ_C 108.8) of aesculitannin B.²⁵ The β-configuration of both the 2→O→7/4→8 and the 4→6 IFLs were assigned from the positive CE in the diagnostic region 220−240 nm (Figure 4A). Hydrogens II-H-2 and II-H-4 were determined to be cis-positioned based on the similar chemical shifts of II-C-2 (δ_C 84.7) and its corresponding carbon in 5 (δ_C 83.9). This also established the absolute configuration for unit II as 2S, 3R, 4S. The absolute configuration of I-H-3 was assigned as R, based on the trans-orientation between I-H-3 and I-H-4. This assignment was verified by the NOESY correlation between I-H-3 and II-H-6. Finally, phloroglucinolysis was performed to determine the stereochemistry of the terminal unit III. Analysis of the reaction products revealed the presence of two major compounds: catechin (8) and compound 9 (Figure S105), which was found to be [PAC 5]-(4β→2)-phloroglucinol. Thus, the structure of 16 was established as [PAC 5]-(4β→6)-catechin.

In addition to 17, five more 2A+B-type tetramers were isolated, as was gleaned from their identical molecular formulas C₆₀H₄₆O₂₄ established via HRESIMS (Figures S70, S76, S82, S89, S96, and S103) and associated ¹³C NMR carbon counts. Comparison of their NMR (Tables 6 and 7) and ECD data (Figure 4B) revealed that 18 had the same planar structure as peanut procyanidin F.¹³ Phloroglucinolysis of 18 led to the isolation of two major compounds 3 and 4, the structures of which were confirmed by MS and chiral phase HPLC (Figure 5). Thus, the structure of 18 was unambiguously determined as [PAC 4]-(4β→6)-[PAC 4].

Table 6.

The ¹H NMR data of 17−21 (CD₃OD, 800 MHz, 278 K).

Unit	Ring	No	17	18	19		20		21
			1H (mult., J in Hz)	1H (mult., J in Hz)	1H (mult., J in Hz)	1H (mult., J in Hz)	1H (mult., J in Hz)	1H (mult., J in Hz)	1H (mult., J in Hz)	1H (mult., J in Hz)
			Major	Major	Major A	Minor B	Major A	Minor B	Rotamer A	Rotamer B
I	C	3	4.10, d (3.3)	4.09, d (3.3)	4.09, d (3.3)	3.33, d (3.5)	4.08, d (3.3)	3.28, d (3.5)	4.05, d (3.27)	4.08, d (3.5)
		4	4.45, d (3.3)	4.44, d (3.3)	4.44, d (3.3)	4.22, d (3.5)	4.42, d (3.3)	4.18, d (3.5)	4.39, d (3.31)	4.29, d (3.5)
	A	6	6.04, d (2.4)	6.03, d (2.4)	6.04, d (2.3)	5.98, d (2.3)	6.02, d (2.3)	5.97, d (2.3)	6.00, d (2.34)	6.03, d (2.3)
		8	6.09, d (2.4)	6.08, d (2.4)	6.08, d (2.3)	6.03, d (2.3)	6.07, d (2.3)	6.02, d (2.3)	6.05, d (2.34)	6.07, d (2.3)
	B	2’	7.15, d (2.1)	7.15, d (2.1)	7.15, d (2.1)	7.11, d (2.1)	7.15, d (2.1)	7.11, d (2.1)	7.12, d (2.1)	7.17, d (2.1)
		5’	6.82, d (8.3)	6.81, d (8.3)	6.82, d (8.3)	6.86, d (8.2)	6.82, d (8.2)	6.86, d (8.2)	6.79, d (8.2)	6.84, d (8.2)
		6’	7.03, dd (8.3, 2.1)	7.02, dd (8.3, 2.1)	7.04, dd (8.3, 2.1)	6.90, dd (8.3, 2.1)	7.03, dd (8.2, 2.1)	6.90, dd (8.2, 2.1)	7.00, dd (8.2, 2.1)	7.04, dd (8.2, 2.1)
II	F	2	5.35, s	5.23, s	5.26, s	5.52, s	5.20, s	5.55, s	5.29, s	5.44, s
		3	3.96, brd (2.3)	4.00, brd (2.3)	3.97, brs	4.17, brs	3.92, brd (2.4)	4.18, d (2.3)	3.89, brd (2.3)	3.84, brd (2.3)
		4	4.74, brs	4.69, brs	4.82, brs	4.53, brs	4.80, brs	4.52, brs	4.55, brs	4.60, brs
	D	6	6.07, s	6.07, s	6.11, s	5.86, s	6.10, s	5.84, s	6.00, s	5.90, s
	E	2’	7.12, d (2.1)	7.12, d (2.1)	7.06, d (2.1)	7.31, d (2.1)	7.05, d (2.1)	7.32, d (2.1)	7.07, d (2.1)	7.03, d (2.1)
		5’	6.78, d (8.2)	6.78, d (8.2)	6.76, d (8.2)	6.83, d (8.2)	6.73, d (8.2)	6.83, d (8.2)	6.75, d (8.2)	6.82, d (8.2)
		6’	6.94, dd (8.2, 2.1)	6.94, dd (8.2, 2.1)	6.91, dd (8.2, 2.1)	7.18, dd (8.2, 2.1)	6.84, dd (8.2, 2.1)	7.19, dd (8.2, 2.1)	6.89, dd (8.2, 2.1)	6.85, dd (8.2, 2.1)
III	I	3	4.12, d (3.6)	4.13, d (3.6)	4.13, d (3.4)	3.68, d (3.2)	4.13, d (3.4)	3.68, d (3.2)	4.19, d (3.4)	4.20, d (3.4)
		4	4.21, d (3.6)	4.33, d (3.6)	4.48, d (3.4)	4.17, d (3.2)	4.48, d (3.4)	4.37, d (3.2)	4.42, d (3.4)	4.31, d (3.4)
	G	6			5.99, s	6.13, s	5.99, s	6.16, s
		8	6.05, s	6.05, s					6.03, s	6.21, s
	H	2’	7.13, d (2.3)	7.13, d (2.3)	7.30, d (2.2)	7.03, d (2.3)	7.29, d (2.2)	7.02, d (2.2)	7.13, d (2.2)	7.132, d (2.2)
		5’	6.82, d (8.1)	6.82, d (8.1)	6.80, d (8.1)	6.79, d (8.3)	6.79, d (8.1)	6.78, d (8.1)	6.83, d (8.1)	6.81, d (8.1)
		6’	7.02, dd (8.1, 2.3)	7.02, dd (8.1, 2.3)	7.17 dd (8.1, 2.2)	7.01, dd (8.2, 2.3)	7.16, dd (8.1, 2.2)	7.00, dd (8.1, 2.2)	7.03, dd (8.1, 2.2)	7.02, dd (8.1, 2.2)
IV	L	2	4.75, d (8.4)	4.96, brs	4.74, d (8.0)	4.69, d (8.1)	4.93, brs	4.90, brs	5.20, brs	4.94, brs
		3	4.17, m	4.19, m	4.17, m	4.13, m	4.24, m	4.22, m	4.22, m	4.05, m
		4	2.57, dd (16.2, 9.1)	2.84, dd (17.0)	2.58, dd (16.2, 8.5)	2.56, dd (15.8, 8.8)	2.78a	2.78a	2.88a	2.85a
			3.03, dd (16.2, 5.4)	2.99, dd (17.0, 4.6)	2.98, dd (116.2, 5.4)	2.97, dd (15.8, 5.5)	2.95a	2.95a	2.91a	2.90a
	J	6	6.12, s	6.14, s	6.04, s	6.16, s	6.04, s	6.15, s	6.10, s	6.11, s
	K	2’	6.99, d (2.1)	7.27, d (2.1)	6.94, d (2.1)	6.93, d (2.1)	7.19, d (2.1)	7.16, d (2.1)	7.09, d (2.1)	7.03, d (2.1)
		5’	6.87, d (8.1)	6.82, d (8.1)	6.84, d (8.1)	6.84, d (8.1)	6.81, d (8.1)	6.82, d (8.1)	6.80, d (8.1)	6.76, d (8.1)
		6’	6.94, dd (8.2, 2.1)	7.00, dd (8.2, 2.1)	6.86, dd (8.2, 2.1)	6.84, dd (8.2, 2.1)	7.02, dd (8.2, 2.1)	6.99, dd (8.2, 2.1)	6.93, dd (8.2, 2.1)	6.59, dd (8.2, 2.1)

^aSignals overlapped.

Analysis of the ¹H and ¹³C NMR spectra of 19, recorded at 278 K, showed two major rotamers (1.3:1 ratio). Four pairs of resonances for AX spin systems were observed at δ_H 3.33 (d, J = 3.5 Hz), 4.22 (d, J = 3.5 Hz), 3.68 (d, J = 3.2 Hz), and 4.17 (d, J = 3.2 Hz) for rotamer A; and at δ_H 4.09 (d, J = 3.3 Hz), 4.44 (d, J = 3.3 Hz), 4.13 (d, J = 3.2 Hz), and 4.48 (d, J = 3.2 Hz) for rotamer B. This revealed the presence of two doubly linked A-type linkages. Three coupled methines (δ_H 5.52, 4.17, and 4.53 for rotamer A; and δ_H 5.26, 3.92, and 4.82 for rotamer B) indicated the presence of one B-type linkage, which must be located between units II and III as evident from the HMBC correlations (Figure S80) from II-H-4 to III-C-7 and III-C-9. Units II and IV were assigned as (ent)-epicatechin and (ent)-catechin based on the near 0 Hz value of J_2,3 in the F-ring and the 8.1 Hz J_2,3 value in unit IV, respectively. The A-type linkage between units III and IV was confirmed as 2β→O→7/4β→8 via the HMBC correlation from IV-H-2 and III-H-4 to IV-C-9, as well as the positive CE at 220−240 nm in the ECD spectrum (Figure 3B). Similarly, the 2β→O→7/4β→8 linkage for units I and II was assigned via the NOESY correlations shown in Figure S81 between I-H-4 and II-H-2, supported by the ECD data. A 4β→8 B-type linkage between units II and III was assigned by the NOESY correlations from II-H-4 to III-H-2′, −5′, and −6′, as well as the chemical shifts of III-C-8 in the rotamers (avg. δ_C 109.2 vs. 111.2 for 17). Trans-configuration for H-3 and H-4 was verified by the NOESY correlations from H-3 (C- and I- rings) to H-6 (D- and J- rings, respectively). Trans-positions of H-2 and H-4 in the F-ring was confirmed via the upfield shifted II-C-2 resonances (avg. δ_C 78.8) as compared to the corresponding carbon in 4 (δ_C 81.6). Comparison of the ¹³C NMR data of unit IV in 19 with the corresponding resonances in 2 and 5 showed that unit IV more closely resembles those in 2 (Table 10). Thus, 19 was confirmed as [PAC 4]-(4β→8)-[PAC 2]. Assignment of MS data and chiral HPLC (Figure 5) was consistent with the structures of 3 and 2 as two major phloroglucinolysis products of 19 further confirmed this conclusion.

The ¹H and ¹³C NMR spectra of 20 highly resembled those of 19, including the presence of two rotamers in a ~1.5:1 ratio, and pointed to the terminal unit IV as the only major difference. The two A- and one B-type linkages in 20 were confirmed in analogy to those in 19 using HMBC (Figure S87), NOESY (Figure S88), ECD (Figure 4B), and ¹³C NMR data (Table 7). The terminal unit IV was identified as 2,3-cis-oriented (ent)-epicatechin based on the singlet like δ_H of IV-H-2 in the ¹H NMR spectrum. Analysis of the ¹³C NMR data of unit IV showed that they much more closely correlated with the corresponding carbons of 4 rather than those of 6 (Table 8). Thus, 19 was established as [PAC 4]-(4β→8)-[PAC 4], which was further confirmed by phloroglucinolysis (Figure S106). This structure was described previously and the compound was named as aesculitannin G,³⁰ although no NMR data were reported.

Based on the similarity of its ¹H and ¹³C NMR data with those of 18, 21 was recognized as a PAC tetramer in which two A-type dimers were also connected through a B-type IFL. It also showed two major rotamers (~1:1) in low temperature NMR at 278 K. Both A-type linkages between units I/II and III/IV were confirmed as being 2β→O→7/4β→8 via the NOESY correlations (Figure S95) from I-H-4 to II-H-2, and from III-H-4 to IV-H-2′/6′, respectively, as well as through observation of a strong positive CE at 220−240 nm in the ECD spectrum (Figure 4B). A 4β→6 B-type linkage between units II and III was established by the ECD data and the NOESY correlations from III-H-8 to III-H-2′/6′. The deduction was corroborated by the similar chemical shift of the III-C-6 carbons (avg. δ_C 111.2) relative to those of the corresponding carbons in 17 (avg. δ_C 111.2) and 18 (avg. δ_C 110.9), which were significantly different from those in 19 (avg. δ_C 109.0) and 20 (avg. δ_C 109.2). Unit IV was identified as ent-epicatechin based on the closer resemblance of its ¹³C NMR data to that of 6 relative to those of 4 (Table 8). The structure of 21 was, thus, established as [PAC 4]-(4β→6)-[PAC 6], and this assignment was further confirmed by the identification of 3 and 6 as the phloroglucinolysis products of 21 (Figure S106).

Analysis of the ¹H and ¹³C NMR spectra of 22 revealed it was also an A+B+A-type tetrameric PAC with two major rotamers (~2:1 ratio). Units II and IV were assigned as (ent)-catechin and (ent)-epicatechin, respectively, based on the corresponding J_2,3 values of 9.6 and ~0 Hz. A 2β→O→7/4β→8 linkage between units I and II was evident from the NOESY correlations (Figure S102) of I-H-4/II-H-2′ and 6′, as well as the high amplitude CE at 220−240 nm in the ECD spectrum (Figure 4B). Similarly, the 2β→O→7/4β→8 A-type linkage between units III and IV was verified by the NOESY correlations of III-H-4/IV-H-2′ and 6′, as well as through ECD evidence. Units II and IV were connected by a 4β→6 IFL based on the ECD data and the NOESY cross-peaks from III-H-8 to III-H-2′/6′, as well as through the specific chemical shift of III-C-8 (δ_C 111.6). Trans-configuration for H-3 and H-4 in both the C- and I-rings was verified by the NOESY correlations from H-3 (C- and I-ring) to H-6 (D- and J-ring). The cis-configuration between H-2 and H-4 in the F-ring was determined via the chemical shift of II-C-2 (δ_C 84.6). Thus, units I and II had to be doubly linked as in 5. The Δδ_C method was used to determine the absolute configuration of unit IV. Unit IV was identified as ent-epicatechin based on the closer resemblance of its ¹³C NMR resonance patterns with those of corresponding carbons in 6 rather than with those in 4 (Table 8). Collectively, this verified the structure of 22 to be [PAC 5]-(4β→6)-[PAC 6].

CONCLUSIONS

The presented combination of chiral spectroscopic (electronic circular dichroism [ECD]) and chemical (phloroglucinolysis plus MS and chiral HPLC) analysis generated the necessary framework of chiral reference data that enables the determination of absolute configuration of catechol monomers in oligomeric PACs by 1D and 2D NMR. Access to the full 3D structure of PAC trimers, tetramers, and even higher oligomers in their native, underivatized form is critical to establish structure activity relationships from the subsequent biological evaluation, such as for the dentin biomodification potential of the compounds presented here, which were isolated from fractions with known bioactivity. The presented structural data expands both our prior reports^9–11 as well as other PAC literature by providing accurately and precisely assigned ¹³C NMR resonances. This forms the knowledge base for making absolute stereochemical assignments of PAC monomers via inclusion of the γ-gauche effect observed in the ¹³C NMR resonances of PACs, as well as the determination of differential ¹³C chemical shift values (s.c.s; Δδ_C) relative to those in the terminal monomers and analogous dimers (here: 2 and 4−6).

Forming part of a closely-knit interdisciplinary approach to PACs as dentin biomodification agents, the present but also past and ongoing studies have frequently pointed out that the structural complexity of PACs is paired with substantial heterogeneity of the PAC literature in terms of the interplay between reported biological/bioactivity and chemistry/structural information. In this context, robust NMR data play a key role as being most significant for rigor and reproducibility: accurate and precise ¹H and ¹³C assignments and shared spectra are key to achieving consistency among the chemical space of PACs and their biological profiles. While the relatively close resemblance of the spectra of analogous compounds adds to the challenge, the presented outcomes show how unambiguous assignments can transform potentially confusing spectral similarity into definitive structural information. It should be noted that the presence of atropisomeric forms not only dictates the use of low-temperature NMR, but also poses a challenge in terms of PAC separation, which can be addressed by applying multiple steps and utilizing a variety of chromatographic techniques with as much as possible orthogonal character (here: countercurrent separation, silica gel based and Sephadex LH-20-based size-exclusion/adsorption column chromatography, as well as semi-preparative HPLC).

The current study demonstrated the unambiguous elucidation of 13 trimeric and tetrameric PACs by combination of analytical approaches, including HRESIMS, low temperature NMR, ECD, and phloroglucinolysis with MS and chiral phase HPLC. To ensure the correct structure elucidated that use all techniques for each PAC oligomer.

The detailed analysis of ¹³C NMR data proved to be a key element for achieving absolute stereochemical determination in the monomeric moieties in a series of diastereomeric PACs. Especially the γ-gauche effect and consideration of s.c.s. effects via comparison of Δδ_C values can corroborate assignments and help detect inconsistencies in reported data. The presented structures can not only serve as well-characterized components of a small PAC library for the further exploration of their dentin bioactivity, but also contribute reliable, fully interpreted and raw spectroscopic data to the available knowledge base. As high-quality structural data of trimeric and tetrameric PACs is in high demand, the presented outcomes also help avoid future ambiguous or erroneous assignments and increase structural reproducibility and integrity.²⁹

EXPERIMENTAL SECTION

General Experimental Procedures

High resolution electrospray ionization mass spectrometric measurements were carried out by using a Bruker Impact II, quadrupole time of flight (q-TOF). ECD spectra were acquired on a JASCO-715 spectrometer with a 0.2 cm quartz cuvette, sample concentration was less than ≤ 0.1 mg/mL in MeOH. All 1D/2D NMR spectra were acquired at 255 K, 278 K and/or 298 K on an 800 MHz Bruker Avance spectrometer equipped with a 5 mm triple resonance inverse TCI RT probe. The ¹³C NMR spectra of all compounds were acquired on a JEOL (Jeol USA, Peabody, MA, USA). ECZ 400 MHz spectrometer with an FTS cooling system that consists of XR AirJet and TC-84 temperature controller (SP Industries, Warminster, PA, USA). C₁₈ reversed-phase (RP-18) silica gel (Macherey-Nagel, Bethlehem, PA, USA) and Sephadex LH-20 gel (Pharmacia, Uppsala, Sweden) were used for column chromatography (CC). TLC was performed on SIL G/UV254 (Macherey-Nagel, Inc, Bethlehem, PA, USA) with visualization under UV light (254 and 365 nm) and spraying with vanillin-sulfuric acid reagent followed by heating. Semi-preparative HPLC was performed on a Shimadzu HPLC (Kyoto, Japan) connected to a PDA detector (Shimadzu, model SPD-20A) and equipped with a YMC-Pack ODS-AQ (250 × 10 mm, S-5, 12 nm) or CHIRAPAK IA (250 × 10.0 mm, S−5 μm, Chiral Tech., West Chester, PA, USA). All solvents used were obtained from Fisher Scientific (Fair Lawn, NJ, USA) or Sigma-Aldrich (St. Louis, MO, USA). Phloroglucinol, ascorbic acid, hydrogen chloride, and sodium acetate were used in ACS grade and purchased from Sigma-Aldrich (St. Louis, MO, USA). All NMR solvents were purchased from Cambridge Isotope Laboratories (Tewksbury, MA, USA).

Plant Material

Extract powder of the inner bark of Pinus massoniana was purchased from Xi’an Chukang Biotechnology in China in 2012 (No. PB120212).

Extraction and Isolation

12 g of enriched tri- and tetrameric proanthocyanidins, being separated as 6.5 g of fraction A and 5.6 g of fraction B, were prepared from 200 g pine bark extract by centrifugal partition chromatography (CPC) described previously.¹² Both CPC fractions A and B were chromatographed on Sephadex LH-20 column (EtOH), affording corresponding 6 subfractions (A1−A6) and 7 subfractions (B1−B7), respectively. Fraction A1 (800 mg) contained mainly (±)-catechin and (±)-epicatechin as confirmed by chiral phase HPLC. Fraction A2 (1.0 g) was fractionated over a RP-18 silica gel column (MeOH/H₂O, 20−80%), and the major two subfractions A2b and A2c were then purified to afford compounds 2 (30 mg), 4 (20 mg), 5 (10 mg), and 6 (10 mg) by the semi-preparative HPLC (23% MeCN in 0.1% formic acid H₂O, 2.5 mL/min, isocratic). Similarly, after being fractionated by RP-18 silica gel column, compounds 15 (20 mg); 20 (2.8 mg) and 22 (0.9 mg); as well as 16 (12 mg) and 18 (7.1 mg) were purified from their corresponding fractions A3, A4, and A5, respectively, by semi-preparative HPLC (23% MeCN in 0.1% formic acid H₂O, 2.5 mL/min, isocratic). Purification of A6 (1.5 g) via an RP-18 silica gel column (MeOH/H₂O, 25−30%) led to the isolation of the major tetramer, 17 (900 mg). Fraction B5 (800 mg) was fractionated over a RP-18 silica gel column (MeOH/H₂O, 20−80%), and the major three subfractions B5a−B5c were then purified via semi-preparative HPLC (18% MeCN in 0.1% formic acid H₂O, 2.5 mL/min) to afford compounds 11 (30 mg), 12 (1.5 mg), 13 (1.9 mg), and 14 (2.1 mg). Fraction B6 (1.6 g) mainly contained trimer 10 (700 mg), which was purified by a RP-18 silica gel column. Compounds 19 (25 mg) and 21 (2.0 mg) were purified from fraction B7 by semi-preparative HPLC (20% MeCN in 0.1% formic acid H₂O, 2.5 mL/min) after pre-fractionation with an RP-18 silica gel column.

Epicatechin-(4β→2)-phloroglucinol (1). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 203 (−9.6), 216 (+15.4), 237 (+9.3), 275 (−0.7) nm; ¹H and ¹³C NMR (CD₃OD, 255K), see Table 1, HRMS (ESI) m/z [M + H]⁺ calcd for C₂₁H₁₉O₉, 415.1024; found 415.1022.

Epicatechin-(2β→O→7,4β→8)-catechin (2). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 208 (−6.8), 222 (+4.2), 272 (−1.4); ¹³C NMR (CD₃OD, 298, 277, and 255 K), see Table 2; HRMS (ESI) m/z [M + H]⁺ calcd for C₃₀H₂₅O₁₂, 577.1341; found 577.1360.

Epicatechin-(2β→O→7,4β→8)-epicatechin-(4β→2)-phloroglucinol (3). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 207 (−23.3), 227 (+20.6), 273 (−1.9); ¹H and ¹³C NMR (CD₃OD, 298 K), see Table 1; HRMS (ESI) m/z [M + H]⁺ calcd for C₃₆H₂₉O₁₅, 701.1501; found 701.1504.

Epicatechin-(2β→O→7,4β→8)-epicatechin (4). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 207 (−16.9), 222 (+10.4), 272 (−1.5); ¹³C NMR (CD₃OD, 298, 277, and 255 K), see Table 2; HRMS (ESI) m/z [M + H]⁺ calcd for C₃₀H₂₅O₁₂, 577.1341; found 577.1359.

Epicatechin-(2β→O→7,4β→8)-ent-catechin (5). Light brown, amorphous solid; ECD (MeOH) λ_max (Δε) 207 (−20.7), 224 (+12.7), 270 (−1.4); ¹³C NMR (CD₃OD, 298, 277, and 255K), see Table 2; HRMS (ESI) m/z [M + H]⁺ calcd for C₃₀H₂₅O₁₂, 577.1341; found 577.1357.

Epicatechin-(2β→O→7,4β→8)-ent-epicatechin (6). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 207 (−15.4), 222 (+15.0), 271 (−1.3); ¹³C NMR (CD₃OD, 298, 277, and 255 K), see Table 2; HRMS (ESI) m/z [M + H]⁺ calcd for C₃₀H₂₄O₁₂, 577.1341; found 577.1348.

Epicatechin-(2β→O→7,4β→8)-ent-epicatechin-(4β→2)-phloroglucinol (9). Lightly brown, amorphous solid; ¹H and ¹³C NMR (CD₃OD, 298 K), see Table 1; HRMS (ESI) m/z [M + H]⁺ calcd for C₃₆H₂₉O₁₅, 701.1501; found 701.1540.

Epicatechin-(4β→6)-epicatechin-(2β→O→7,4β→8)-catechin (10). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 212 (−6.6), 229 (+11); 272 (−1.1); ¹H NMR (CD₃OD, 255 K), see Table 3 and ¹³C NMR (CD₃OD, 255 K), see Table 5, HRMS (ESI) m/z [M + H]⁺ calcd for C₄₅H₃₇O₁₈, 865.1974; found 865.1996.

Epicatechin-(4β→8)-epicatechin-(2β→O→7,4β→8)-catechin (11). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 204 (−4.0), 224 (+6.3), 277 (−0.8); ¹H NMR (CD₃OD, 278 K), see Table 3 and ¹³C NMR (CD₃OD, 278 K), see Table 5. HRMS (ESI) m/z [M + H]⁺ calcd for C₄₅H₃₇O₁₈, 865.1974; cound 865.1983.

Epicatechin-(4β→8)-epicatechin-(2β→O→7,4β→8)-ent-epicatechin (12). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 205 (−8.5), 226 (+19.2), 274 (−2.2), 288 (+1.0); ¹H NMR (CD₃OD, 278 K), see Table 3 and ¹³C NMR (CD₃OD, 278 K), see Table 5, HRMS (ESI) m/z [M + H]⁺ calcd for C₄₅H₃₇O₁₈, 865.1974; found 865.1972.

Epicatechin-(2β→O→7,4β→8)-epicatechin-(4β→8)-catechin (13). Light brown, amorphous solid; ECD (MeOH) λ_max (Δε) 208 (−13.7), 229 (+9.8), 272 (−1.1); ¹H NMR (CD₃OD, 278 K), see Table 3 and ¹³C NMR (CD₃OD, 278 K), see Table 5, HRMS (ESI) m/z [M + H]⁺ calcd for C₄₅H₃₇O₁₈, 865.1974; found 865.1966.

Epicatechin-(2β→O→7,4β→8)-epiafzelechin-(4β→8)-catechin (14). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 209 (−35.4), 230 (+29.3), 274 (−3.2); ¹H NMR (CD₃OD, 278 K), see Table 3 and ¹³C NMR (CD₃OD, 278 K), see Table 5, HRMS (ESI) m/z [M + H]⁺ calcd for C₄₅H₃₇O₁₇, 849.2025; found 849.2045.

Epicatechin-(4β→6)-epicatechin-(2β→O→7,4β→8)-epicatechin (15). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 212 (−6.6), 229 (+12.0), 272 (−1.1); ¹H NMR (CD₃OD, 255 K), see Table 4 and ¹³C NMR (CD₃OD, 255 K), see Table 5, HRMS (ESI) m/z [M + H]⁺ calcd for C₄₅H₃₇O₁₈, 865.1974; found 865.1960.

Epicatechin-(2β→O→7,4β→8)-ent-catechin-(4β→6)-catechin (16). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 204 (−12.8), 229 (+17.2), 274 (−1.3); ¹H NMR (CD₃OD, 278 K), see Table 4 and ¹³C NMR (CD₃OD, 278 K), see Table 5, HRMS (ESI) m/z [M + H]⁺ calcd for C₄₅H₃₇O₁₈, 865.1974; found 865.1978.

Epicatechin-(2β→O→7,4β→8)-epicatechin-(4β→6)-epicatechin-(2β→O→7,4β→8)-catechin (17). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 210 (−14.9), 225 (+14.1), 274 (−1.2); ¹H NMR (CD₃OD, 278 K), see Table 6 and ¹³C NMR (CD₃OD, 278 K), see Table 7, HRMS (ESI) m/z [M + H]⁺ calcd for C₆₀H₄₇O₂₄, 1151.2452; found 1151.2415.

Epicatechin-(2β→O→7,4β→8)-epicatechin-(4β→6)-epicatechin-(2β→O→7,4β→8)-epicatechin (18). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 209 (−18.2), 229 (+14.8), 273 (−1.1); ¹H NMR (CD₃OD, 278 K), see Table 6 and ¹³C NMR (CD₃OD, 278 K), see Table 7, HRMS (ESI) m/z [M + H]⁺ calcd for C₆₀H₄₇O₂₄, 1151.2452; found 1151.2468.

Epicatechin-(2β→O→7,4β→8)-epicatechin-(4β→8)-epicatechin-(2β→O→7,4β→8)-catechin (19). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 208 (−22.8), 229 (+20.4), 275 (−1.9); ¹H NMR (CD₃OD, 278 K), see Table 6 and ¹³C NMR (CD₃OD, 278 K), see Table 7, HRMS (ESI) m/z [M + H]⁺ calcd for C₆₀H₄₇O₂₄, 1151.2452; found 1151.2426.

Epicatechin-(2β→O→7,4β→8)-epicatechin-(4β→8)-epicatechin-(2β→O→7,4β→8)-epicatechin (20). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 208 (−30.1), 229 (+25.2), 274 (−2.3); ¹H NMR (CD₃OD, 278 K), see Table 6 and ¹³C NMR (CD₃OD, 278 K), see Table 7, HRMS (ESI) m/z [M + H]⁺ calcd for C₆₀H₄₇O₂₄, 1151.2452; found 1151.2484.

Epicatechin-(2β→O→7,4β→8)-epicatechin-(4β→6)-epicatechin-(2β→O→7,4β→8)-ent-epicatechin (21). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 206 (−8.2), 232 (+13.1), 273 (−0.8); ¹H NMR (CD₃OD, 278 K), see Table 6 and ¹³C NMR (CD₃OD, 278 K), see Table 7, HRMS (ESI) m/z [M + H]⁺ calcd for C₆₀H₄₇O₂₄, 1151.2452; found 1151.2484.

Epicatechin-(2β→O→7,4β→8)-ent-catechin-(4β→6)-epicatechin-(2β→O→7,4β→8)-ent-epicatechin (22). Lightly brown, amorphous solid; ECD (MeOH) λ_max (Δε) 214 (+11.8), 225 (+9.7), 236 (+12.8), 273 (−1.5); ¹H NMR (CD₃OD, 278 K), see Table 4 and ¹³C NMR (CD₃OD, 278 K), see Table 7, HRMS (ESI) m/z [M + H]⁺ calcd for C₆₀H₄₇O₂₄, 1151.2452; found 1151.2427.

Phloroglucinolysis

To further confirm the configurations of all the isolated compounds, phloroglucinolysis was performed.²⁰ Compounds 10 (10 mg) and 17 (10 mg) were individually cleaved using a reaction mixture consisting of 400 mg of phloroglucinol and 100 mg ascorbic acid in 10 mL of 0.1 N HCl in MeOH at 50 °C for 30 mins. Adding of 50 mL of 40 mM sodium acetate solution to the mixture stopped the reaction. The mixture was then extracted twice with ethyl acetate, and the concentrated ethyl acetate layer was loaded on Sephadex LH-20 for purification and resulted in the isolation of 1 and 2 from 10, as well as 2 and 3 from 17. Similarly, phloroglucinolysis of 11 (5.5 mg), 15 (3.1 mg), 16 (2.8 mg), 18 (3.0 mg), and 19 (3.1 mg) by 100 mg of phloroglucinol and 50 mg ascorbic acid in 5 mL of 0.1 N HCl in MeOH at 50 °C for 30 mins were performed. Each reaction mixture was extracted by ethyl acetate, and then purified as corresponding two products as being confirmed by chiral HPLC and MS data. About 0.3~0.5 mg of 12−14, 20, and 21 were cleaved using a reaction mixture consisting of 20 mg of phloroglucinol and 10 mg ascorbic acid in 2 mL of 0.1 N HCl in MeOH at 50 °C for 30 mins. The identity of all reaction products was confirmed by chiral HPLC and MS analysis.

Dental Bioassay

Dentin fragments were sectioned into 0.5 × 1.7 × 7.0 mm (H × W × L) pieces and demineralized using 10% phosphoric acid for 5 h. Specimens were treated with the compounds at 0.65% w/v in 20 mM HEPES buffer (pH 7.2) for 1 h, and a control group was kept in HEPES buffer only (n=5). The apparent modulus of elasticity (E) was measured using a 3-point bending method with universal testing machine as previously described.³¹ Data were statistically evaluated by one-way ANOVA and Tukey’s post hoc tests (α = 0.05).