Analysis of dye extracted from Phellodendron bark and its identification in archaeological textiles

Abstract

Methanol extractions of 20 archaeological textiles excavated from burial sites in Korea were examined in comparison to standard natural dyes, freshly dyed silk, and naturally-aged dyed silks for the purpose of identifying archaeological textiles dyed with yellow dye from the Phellodendron bark. Phellodendron bark dye was identified by the presence of both berberine and palmatine using the retention times of HPLC-DAD chromatograms and mass spectral data generated from the electrospray ionization (ESI) source. Extracted ion chromatograms (EIC) were used to quantify the concentration of berberine and palmatine in each sample. The results indicated that three of the 20 archaeological textile samples were dyed with Phellodendron bark based on the significant amount of berberine and palmatine detected in these samples, while twelve samples were determined not to have been dyed with natural dyes from Phellodendron bark. Identification of dye in these textiles would allow for appropriate conservation procedures to avoid further dye loss.

Keywords

berberine palmatine HPLC-DAD-MS color change dye identification archaeological textiles

Introduction

A well preserved textile recovered from a burial site can provide valuable information about the culture, the technology, and the everyday lives of people of the past. Due to their organic nature, however, textiles are rarely recovered intact from archaeological sites. Color change is a visually detectable consequence of degradation that a textile undergoes within the burial context. Staining by organic material present in the soil is another factor which can prohibit visual detection of any dye residue in the textile. Such change of color has been one of the major problems in evaluating the composition and condition of the textiles recovered from a number of excavations of the 17^th century burials of Korea. Most of these burials belong to high, if not prestigious, social status individuals.¹ Except for some cases where a blue color remains intact, almost all the other textiles showed a tan to brown hue upon excavation.² Without a trace of the original hue, there is no indication of which dyes were used on these textiles. We may surmise though, that they were dyed with the available natural dyes of the period. Since natural dyes react differently to conservation procedures such as wet washing, identification of the dye in each archaeological textile is necessary for selection of the appropriate conservation procedures that would preserve the textiles without incurring further dye loss. With the lack of original hue and chemical degradation of the dye, it is difficult to minimize the list of possible dyestuffs that could have been used. This paper focuses on the identification of natural plant dye obtained from Phellodendron bark, which produces a bright yellow color, on archaeological textiles. The purpose of this research was to identify the textiles dyed with Phellodendron bark among a number of textiles excavated from the two 17^th century burials of Korea using HPLC-DAD-MS analyses. The target of the analyses was to find the presence of both berberine and palmatine in the archaeological samples as a fingerprint for Phellodendron bark dye.

Phellodendron (cork tree) is a deciduous tree in the family Rutaceae that is native to east and northeast Asia. Aside from its importance in traditional and modern pharmaceuticals,^2–6 Phellodendron bark had been one of the major sources of yellow dye in Asia for fibers and paper materials not only for its color but also for its insect repellency.^2,7 Record of its usage as medicine and dye material are found in historic references such as Ben Cao Gang Mu⁸ that originates from China in the 16^th century. Traditional recipes for dyeing textiles using Phellodendron bark are recorded in Kyuhapchongso⁹ a 19^th century Korean reference. Pharmaceutical effects and the yellow color of Phellodendron bark all come from compounds of a single chemical group, namely the protoberberine alkaloids. A recent analysis of constituents of Phellodendron bark of two different species collected from different regions of China identified 24 different alkaloids using HPLC-DAD coupled with electrospray ionization mass spectrometry.⁴ Berberine and palmatine were the most abundant alkaloids, with the species Phellodendron chinense Schneid having higher berberine content than Phellodendron amurense Rupr.⁴ (Figure 1). Simultaneous detection of berberine and palmatine in Phellodendron bark has been studied extensively in pharmaceuticals by using HPLC.^5,10–12

Berberine has received much attention in the field of dyes and dyeing since it is one of the few cationic colorants of the natural plant dyes, and it has a high affinity to protein fibers with potential anti-bacterial effect.^13,14 To our knowledge palmatine has not been the target of study as a dye although the chemical structure of palmatine is comparable to berberine in dyeing properties and antibacterial characteristics. Using HPLC-MS, the presence of berberine and possible palmatine as well, was found in the analysis of yellow varnish on the metallic foil of a 19–20^th century Tibetan altar.¹⁵ Presence of berberine, palmatine, and jatrorrhizine, another alkaloid compound of Phellodendron bark in the ancient document Dunhuang Diamond Sutra has been investigated using HPLC in comparison with reference standards.⁷ So far, there has been no study that applied an HPLC-DAD-MS analysis on dye extracted from Phellodendron bark for the simultaneous identification of berberine and palmatine in archaeological textiles.

In this research, we incorporated silk samples naturally-aged for 10 years after being dyed with seven different known plants,¹⁶ as intermediaries between the freshly dyed samples and the unknown archaeological samples, so that we could validate the HPLC-DAD-MS methodology for detection of any remains of Phellodendron bark dye in the unknown archaeological samples. The analytical process of this research included four steps; first, analysis of standard dyes to establish the HPLC-DAD-MS protocol; second, analysis of freshly-dyed standard silks to establish the procedure for extraction and analysis of dye in fiber; third, analysis of 10-year-old silk dyed with known natural dyes to verify the methods; and fourth, analysis of archaeological samples to identify Phellodendron bark dyed textiles among the textiles excavated from the two 17^th century burials (Figure 2). The analysis and identification of dye was based on the HPLC-DAD chromatogram and the fingerprint molecular ion mass spectral data generated from the electrospray ionization (ESI) source in the positive scanning modes following the previous research in pharmaceuticals.^4,17

Experimental

Materials

Phellodendron bark was purchased from the Kyungdong traditional pharmaceutical market of Korea. Berberine chloride, palmatine chloride hydrate, alizarin, and purpurin were purchased from Sigma-Aldrich (USA) and used as reference standards. Mordants, aluminum potassium sulfate [AlK(SO₄)₂·12H₂O)] and iron sulfate (FeSO₄·7H₂O) were purchased from Shinyo Pure Chemicals, Co (Osaka, Japan). Methanol (Acros Organics) and acetonitrile (EMD Chemicals) were HPLC grade. Formic acid and HPLC water were purchased from Mallinckrodt Baker. Hydrochloric acid (36–38 wt %) was purchased from VWR Inc. Silk fabric (KS K0905 Standard Adjacent Fabrics for Colorfastness Test) was purchased from the Korea Apparel Testing & Research Institute. Glass fiber enhanced 0.45 µm syringe filters were purchased from Alltech (Deerfield, IL). Water used for dyeing and sample preparation was deionized using a Corning Mega-pure MP6 with Barnstead Nanopure System.

Preparation of standards

Standard liquid dyes for HPLC analysis

Powdered Phellodendron bark (13 g) and methanol (130 mL) were refluxed for 1 hour and centrifuged for 5 min at 5,000 rpm, 21°C. About 2 mL of the supernatant was filtered into an HPLC vial using a glass fiber enhanced 0.45 µm syringe filter and used as the standard Phellodendron extract. 0.01 g aliquots of each of berberine chloride and of palmatine chloride hydrate were dissolved in 10 mL methanol in separate volumetric flasks and were used as the reference standards. Berberine and palmatine standards with various concentrations (0.0001 to 0.0022 mg/mL) were prepared using volumetric flasks. The standard dyes were filtered into the HPLC vials using the glass fiber enhanced 0.45 µm syringe filters and analyzed in HPLC-DAD-MS to obtain standard calibration curves. Alizarin and purpurin standards were prepared and analyzed with HPLC-DAD-MS to determine the retention times of the two species, alizarin and purpurin. All the samples were analyzed using the same HPLC-DAD-MS method.

Standard silk freshly dyed with berberine, palmatine, or Phellodendron extract

Berberine and palmatine dye liquors were prepared by dissolving 0.3 g of each of the standard compounds in 50 mL water. Phellodendron bark dye liquor was prepared as follows. For the first extraction, 5 g of powdered Phellodendron bark and 50 mL water were heated on a hotplate at 80°C for 1 hour. The second extraction was carried out by adding 50 mL water onto the residue. The two extracts were mixed and used as the Phellodendron dye liquor. Following the procedures described by Nam,¹⁸ silk was dyed using the pre-mordanting procedure and repeating the dyeing process twice to ensure sufficient dye uptake. Approximately 2 g of silk was prepared each for dyeing of the standard silks with Phellodendron bark, berberine, and palmatine. Each silk specimen was mordanted with 0.3 g of aluminum potassium sulfate in 50 mL of water for 30 min at 60°C. Mordanted silk was dyed with dye liquors of Phellodendron bark, berberine, or palmatine for 1 hour at 60°C, and the dye liquor was saved. Dyed silk was mordanted second time using fresh aluminum potassium sulfate solution. Then the silk was dyed a second time using the dyed liquor saved from the first dyeing. After each mordanting and dyeing, the silk was rinsed first with running tap water until the water was clear, then soaked in the beaker with fresh deionized water for final rinsing; excess water was removed by gently squeezing the silk.

Silk samples dyed 10 years ago with known natural dyes

Silk fabrics dyed approximately 10 years ago, by one of the authors, using the dye extracted from Phellodendron bark, pagoda tree, turmeric, cape jasmine, safflower yellow, madder, and sappanwood were used as the known samples (Table 1). An extraction of 40 g of each plant was made using 5 L tap water in a stainless steel bucket of approximately 10 L capacity for 1 hour at approximately 80°C. The extraction medium was brought up to the boil using a gas burner in preparation for dyeing the fabric.¹⁶ Mordanting steps of the 10-year-old specimens were the same as that used for the freshly dyed silks, i.e. mordanting twice with aluminum potassium sulfate before and after the first dyeing step. However, for Phellodendron bark one piece of silk was mordanted using the prior method and another piece of silk was mordanted first with aluminum potassium sulfate and second with iron sulfate following a procedure published in the previous literature.¹⁸ For each mordanting step, 5 L of a fresh 2% o.w.f. (on weight of fabric) aluminum potassium sulfate or iron sulfate solution was prepared in tap water. The dyeing procedure used for 10-year old silk was the same as that used for the fresh dyeing. For dyeing with the extract from Phellodendron bark, two pieces of silk mordanted following the above process were dyed together in a single dyebath. The dyed silk was completely covered with aluminum foil after dyeing and stored in a laboratory cabinet at room temperature and relative humidity with no light exposure for 10 years. This way the known silk samples were naturally-aged for 10 years with low possibility of fading by light.¹⁶

Table 1.

Dyes of the 10-year old silk samples dyed with known dye source

Sample number of knowns	Name of plant			Part used for dyeing	Color of dyed fabric	Major pigments
	Botanical		Common			Name	Chemical formula	Molecular weight	UV λ_max (190–400) nm	Solubility in cold water
	Family	Species	Common			Name	Chemical formula	Molecular weight	UV λ_max (190–400) nm	Solubility in cold water
1	Rutaceae	Phellodendron	amur cork tree	bark	yellow	Berberine^4,20	[C₂₀H₁₈NO₄]⁺	336.37	228 265 348⁴	Poor ²⁰
						Palmatine^4,20	[C₂₁H₂₂NO₄]⁺	352.42	226 265 347⁴	Poor ²⁰
2	Fabaceae	Sophora japonica L.	pagoda tree	flower buds	yellow	Rutin^21,22	C₂₇H₃₀O₁₆	610.51	210 260 356²¹	Poor ²²
						Quercetin^21,23	C₁₅H₁₀O₇	302.23	210 254 356²¹	Poor ²³
3	Zingiberaceae	Curcuma longa L.	turmeric	rhizome	yellow	Curcumin^24,25	C₂₁H₂₀O₆	368.37	263²⁴	Insoluble²⁵
4	Rubiaceae	Gardenia jasminoides	cape jasmine	fruit	yellow	Crocin²⁶	C₄₄H₆₄O₂₄	977.00	-	Soluble ²⁶
5	Asteraceae	Carthamus tinctorius L.	safflower yellow	flower	yellow	safflomin A²⁷	C₂₇H₃₂O₁₆	612.53	-
6	Rubiaceae	Rubia tinctorum L.	madder	roots	red	Alizarin^28,29	C₁₄H₈O₄	240.20	248²⁸	Slightly soluble²⁹
						Purpurin²⁸	C₁₄H₈O₅	256.20	256 294²⁸
7	Fabaceae	Caesalpinia Sappan L	sappanwood	roots	red	Brazilin²⁹	C₁₆H₁₄O₅	286.27	-	Soluble²⁹

Excavated archaeological textiles dyed with unknown natural dyes

Twenty two fragments which are part of twenty archaeological textiles from the late 17^th century burials of Korea were kindly provided by the Curators of the Seok Juseon Memorial Museum of Dankook University and the Damin Clothing Art Hall of Ewha Womans University Museum (Table 2). Burials of Kupori and Youngam belong to Choi Sook, and Admiral Choi Won Lip, respectively, who were high standing nobles of the Joseon dynasty.^1,19 During the excavations a number of textile relics used as shrouds or fillers for coffins were uncovered from each burial (Table 3). Shortly after the excavation, curators of the two museums that were designated to manage the textiles provided the fragment samples to the author either before or after the wet washing treatment of the textiles.

Table 2.

Description of the two burials: the source of the excavated archaeological textile samples

Burial label	Designated name of burial	Burial site of	Life of deceased	Year excavated	Excavated location in Korea	Museum that provided the samples
Kupori¹	Kupori Soosung Choisi burial	Choi Sook	1636–1698	1995	Hwasung-gun Kupori, Kyunggido	Seok Juseon Memorial Museum, Dankook University
Youngam¹⁹	Youngam Choi Won Lip Jangun burial	Admiral Choi Won Lip	1618–1690	2001	Youngam-goon, Cheollanamdo	Damin Clothing Art Hall, Ewha Womans University Museum

Table 3.

Archaeological textile samples with unknown dye type from the two burial sites^1,19

Sample number	Burial	Museum treatment	Characteristics of mother piece	Visual color
1	Kupori	wet washed	cloth band (denim or goreum)	tan
		unwashed		tan
2	Kupori	wet washed	small blanket (soryungeum), shell	tan
3	Kupori	unwashed	large blanket (deryungeum), shell	tan
4	Kupori	wet washed	shroud danryung, shell	tan
5	Kupori	wet washed	shroud danryung, liner	tan
6	Kupori	wet washed	small blanket (soryungeum), liner	tan
7	Kupori	wet washed	fabric of half boots (mokwha)	dark brown
		unwashed		dark brown
8	Kupori	unwashed	large blanket (deryungeum), liner	tan
9	Kupori	wet washed	unknown	tan
10	Kupori	unwashed	pants	tan
11	Youngam	unwashed	coffin inner cover	tan
12	Youngam	unwashed	filler clothes, gyupchangui	tan
13	Youngam	unwashed	inner shirt (jeogori) collar, shell	tan
14	Youngam	unwashed	inner shirt (jeogori) collar, shell	tan
15	Youngam	unwashed	inner sleeve	tan
16	Youngam	unwashed	unknown piece, shell	tan
17	Youngam	unwashed	unknown piece, liner	tan
18	Youngam	unwashed	pants	tan
19	Youngam	unwashed	left sleeve	tan
20	Youngam	unwashed	unknown piece, shell	dark brown

The excavated archaeological textile samples with unknown dye source were examined to identify those textiles dyed with Phellodendron bark under the condition of complete lack of information on the original hue (Table 3). The archaeological samples were all made of silk, and the visual colors were mostly tan, some showing a dark brown tone.^1,19 Among 20 excavated silk samples, sample #1 and sample #7 each had two specimens, one that had been wet washed and the other that was not washed.¹ Due to severe staining within the coffin environment, washing was done by museum staff by laying the specimen flat in the tank with tap water at room temperature without the addition of soap or other cleaning agent.¹ Five archaeological samples (#2, 4, 5, 6, and 9) were provided only after the washing treatment, the rest of the archaeological samples (13 samples) were all unwashed.^1,19

Extraction of dye from the textiles

Dye was extracted from the freshly dyed silk, 10-year old silk samples dyed with known natural dyes, and excavated archaeological samples with unknown dye source using the conventional solvent system of HCl/methanol/water (2:1:1 v/v/v) with a slight modification in the procedure.^30–33 An aliquot of 400 µL of HCl/methanol/water (2:1:1 v/v/v) was placed in a 20 mL beaker with approximately 0.5–0.8 mg (about 5×5 mm) of textile specimen. The beaker was placed in an oven at 105°C for 15 min followed by rapid cooling in cold water. Then, the beaker was placed in a vacuum desiccator above NaOH pellets until the liquid was completely removed. After this stage, the beaker contained only completely dried silk specimen, and the specimen had turned tan in color. An aliquot of 1.1 mL methanol was added to the beaker, slightly rotating the beaker so that the silk specimen came in complete contact with methanol. With this action the dye was extracted from the silk making the methanol yellow and the silk specimen white. The methanol extract was filtered into the HPLC vial using the glass fiber enhanced 0.45 µm syringe filters and analyzed with HPLC-DAD-MS.

HPLC-DAD-MS

An Agilent 1200 series binary HPLC-DAD-MS system (Foster City, CA) with diode-array detector (DAD), and mass selective detector (MSD) consisting of a single quadrupole mass analyzer with the multi-mode source electrospray ionization (ESI) interface in the positive mode was used to detect the target compounds. LC separation was achieved by 150 mm× 4.6 mm i.d. stainless steel C18 column, 5 µm particle size (Restek Corporation, Bellefonte, PA). A linear gradient elution of solvent A (acetonitrile) and solvent B (0.1–0.5% formic acid in water) was tested in various programs to determine optimum conditions. The gradient elution applied in the analysis with solvent A (acetonitrile) and solvent B (0.5% formic acid in water) was: 0–5.7 min, 90–20% B; 5.7–10 min, 20–61% B; 10–15 min, 61% B. The flow rate was 1.0 mL/min, and the injection volume was 20 µL. The detection wavelength for DAD was set at 275 nm.³⁴ The column temperature of the MSD was 25°C and the electrospray ionization source operated with drying gas (N₂) temperature 350°C, vaporizer 230°C, capillary voltage of 3 kV in positive ion mode, fragmentor voltage 160 V.¹⁷ Extracted ion chromatograms (EIC) were used to quantify berberine and palmatine using the standard calibration curves.

Results and discussion

HPLC results of the reference standard dyes

Reference standards of berberine chloride and palmatine chloride hydrate were analyzed by HPLC-DAD-MS, and the results were used to identify berberine and palmatine in other samples. Berberine eluted at 5.21 min on the LC chromatogram, and the MS spectra showed the molecular cation m/z = 336^17,,35,36 (Figure 3). The LC chromatogram of the palmatine standard showed the retention time of 5.12 min on the reversed-phase LC column, and the MS spectra showed the molecular cation m/z = 352^35,36 (Figure 4). Berberine and palmatine exhibited maximum absorption at three wavelengths, 232 nm, 264 nm, and 344–348 nm, within the UV range that is consistent with the UV spectra of quaternary alkaloids reported in the literature.^36,37 HPLC-DAD chromatograms of both berberine and palmatine in positive ESI mode were used to record and identify these two species in plant extract and textile samples with known and unknown dye source.¹⁷ Extracted ion chromatograms (EIC) were used to quantify berberine and palmatine, using standard calibration curves shown in Figure 5.

HPLC-DAD-MS results for dye extracted from Phellodendron bark

HPLC-DAD-MS results revealed that dye extracted from Phellodendron bark contained both berberine and palmatine. Peaks in the HPLC-DAD chromatogram showed the same retention times as the pure standards berberine and palmatine, 5.21 min and 5.12 min respectively. As reported in the previous literature,¹⁵ the peaks of these compounds overlap due to their close retention time (Figure 6a). The MS spectra (Figure 6b) confirmed the presence of these two compounds in the dye of Phellodendron bark. The extracted ion chromatograms (EIC) for berberine (m/z = 336) and palmatine (m/z = 352) of Phellodendron bark show good separation of these compounds (Figure 7). Two other peaks with relatively high intensity were observed in the chromatogram at retention times of 1.48 min and 1.90 min (Figure 6a). Extracted ion chromatograms (EIC) revealed that their molecular ions are m/z = 180 and m/z = 342 respectively. The peak at 1.90 min (m/z = 342) that had the highest intensity among the three peaks could be Phellodendrine which has been reported as another minor type of alkaloid found in Phellodendron bark.⁴ The peak at 1.48 min could be that of phenol.⁴ However, the identification of these species is a subject for further research.

The concentration of berberine (3230 µg/L) and palmatine (1670 µg/L) in the dye extracted from Phellodendron bark was calculated using the standard calibration curves. Berberine is present at a concentration of approximately twice that of palmatine (1.93:1). The chemical constituents of Phellodendron bark are known to differ due to a variety of factors such as species, growth condition, or part in plant.^{4, 38,39} The present results including the ratio of berberine and palmatine were in agreement with some of the extractions of Phellodendron amurense Rupr. previously reported.⁴

HPLC-DAD-MS analysis of silk freshly dyed

Extractions of dye from silk samples freshly dyed with Phellodendron bark, berberine, and palmatine were analyzed using the HPLC-DAD-MS. Peaks in the HPLC-DAD chromatogram of the silk freshly dyed with Phellodendron bark showed a slight shift in the retention times compared to the corresponding standards and plant extract, 5.3 min for berberine and 5.2 min for palmatine, but the MS spectra confirmed the identity of these peaks. The concentration of berberine (269 µg/L) and palmatine (258 µg/L) in the Phellodendron bark dyed silk was similar with only 4% difference, while the standard Phellodendron extract showed berberine concentration twice that of palmatine, as reported in a previous section. This interesting difference is believed to be due to differences in solubility of the two compounds in water and methanol. As reported previously, Phellodendron bark was extracted using hot water for dyeing silk, while it was extracted using methanol at room temperature for preparing the standard Phellodendron extract. Berberine and palmatine are said to be poorly soluble in cold water but highly soluble in boiling water while their solubility in cold alcohol is said to be poor.^20,40,41 Solubility of berberine and palmatine in water and methanol in the temperature range used in this study was not available in the literature. According to our experiment, however, concentration of berberine and palmatine in each of their standard solutions of methanol was 4.54 µg/L and 2.26 µg/L respectively, which is almost identical to the ratio between berberine and palmatine in the Phellodendron extract. The present result suggests that palmatine has higher solubility in hot water than in methanol at ambient temperature, and such solubility seems to have affected the difference in concentration of berberine and palmatine between the dye extracted from silk and the standard Phellodendron extract.

The HPLC-DAD chromatograms of both berberine-dyed and palmatine-dyed textiles showed the co-occurrence of berberine and palmatine in a single extraction. The concentration of palmatine (39.7 µg/L) in berberine-dyed textile sample was 4.5% of the concentration of berberine (875 µg/L), and the concentration of berberine (85.6 µg/L) in palmatine-dyed textile sample was about 6.7% of that of palmatine (1270 µg/L). Although unexpected, the result was in agreement with the previous literature³⁵ that reported a minor peak of palmatine in a berberine standard using the capillary electrophoresis ESI mass spectrometry.

Analysis of 10-year old silk dyed with known natural plant dyes

The 10-year-old silk, sample dyed with Phellodendron bark contained both berberine and palmatine. However, small amounts of these two species were found in the extractions of the samples dyed with other plants; the madder dyed sample showed both berberine and palmatine, and the turmeric, cape jasmine, and sappanwood dyed samples showed palmatine only (Table 4). Their retention times were consistent with the textile sample freshly dyed with Phellodendron bark. The HPLC-DAD chromotograms were consistent with those of quaternary alkaloid,^36,37 and the MS spectra confirmed that they were berberine and palmatine, respectively. The concentration of berberine and palmatine in the madder-dyed sample was 3% and 15% respectively compared to the concentrations of berberine and palmatine in the 10-year old sample dyed with Phellodendron bark. The turmeric-dyed sample contained a similar concentration of palmatine to the madder-dyed sample. Detection of berberine and palmatine in these samples raises questions in the identification of dye in unknown archaeological samples using HPLC-DAD-MS. Whether cross contamination occurred during extraction and dyeing processes or whether berberine or palmatine are natural constituents in these plants remain unresolved issues.

Table 4.

HPLC-DAD-MS results of the 10-year old fabric samples

Sample number of known	Dye plant	Berberine (µg/L)*	Palmatine (µg/L)	Weight ratio of berberine to palmatine
1	Phellodendron bark	328	107	3.1:1
2	pagoda tree	ND	ND	-
3	turmeric	ND	12.8	-
4	cape jasmine	ND	4.64	-
5	safflower yellow	ND	ND	-
6	madder	10.9	16.5	0.7:1
7	sappanwood	ND	5.77	-

ND: No detection.

Berberine and palmatine are known to be the constituents of an extensive number of plants in different plant families such as Berberidaceae, Fumariaceae, or Papaveraceae.⁴² Among the plant families related to the plant dyes of the 10-year old samples (Table 1), berberine was known to be isolated from the genus Hymenodictyon in the Rubiaceae family,^43–45 the family of the madder plant. According to our search of the literature, there was no record of the isolation of berberine from madder specifically, and there was no record of palmatine in the plant families such as those of madder, turmeric, cape jasmine, or sappanwood. The results of the present investigation suggest a whole different realm of study on the possible isolation of berberine and palmatine in new plant families. However, for our purpose it can be concluded that given the same length of time for natural ageing, the textile dyed with Phellodendron bark contains both berberine and palmatine and their amount is distinctly greater than textiles dyed with any other plant that may contain either or both berberine and palmatine.

Large amounts of both berberine and palmatine were detected in the 10-year-old known samples dyed with Phellodendron bark (Table 4). Two obvious peaks appeared at retention times of 5.3 min and 5.2 min on the HPLC-DAD chromatogram (Figure 8A); this is consistent with that reported for quaternary alkaloids,^36,37 and the MS spectra confirmed that they were berberine and palmatine, respectively. In this naturally-aged fabric dyed with Phellodendron bark, berberine was three times more abundant than palmatine (Table 4), and this was similar to the results of standard Phellodendron extract but differs from that observed for freshly-dyed silk. The large difference in the concentration of the two compounds between the naturally-aged sample and the freshly-dyed silk may be due to the nature of plant species used for dyeing or to the differential rate of degradation between berberine and palmatine. Another possibility is that berberine and palmatine may exhaust differently from the dyebath. Such dyeing characteristics of berberine and palmatine need to be investigated in future research. The possibility of the different plant species can be elucidated from research published in the literature that examined constituents of Phellodendron bark of different species. Phellodendron chinense Schneid, known to be the major Phellodendron species of China, was found to have higher berberine content than Phellodendron amurense Rupr. that is known to be the major Phellodendron species of Korea.^4,39,46 Since it is widely accepted in Korea that the majority of the dry herbs sold in the traditional pharmaceutical market are imported from China and that the country of origin is often not specified, it would not be surprising if the Phellodendron bark used for fresh dyeing of silk in this research and that used in the 10-year old silk were from different species. We were unable to find previous work on the differential degradation rate of berberine and palmatine.

Additional analysis was conducted on a silk dyed 10 years ago with extract from Phellodendron bark comparing two different mordanting procedures. One specimen used two mordanting steps with aluminum potassium sulfate, and the other used aluminum potassium sulfate in the first mordanting and iron sulfate in the second mordanting step. Mordanting with alum/iron sulfate is the method widely used in the past to fix natural dyes on fabric.^18,47 The HPLC-DAD chromatogram of silk mordanted with iron sulfate showed no difference from silk mordanted with aluminum potassium sulfate alone in the retention times of berberine and palmatine, 5.2 min and 5.1 min respectively (Figure 8). However, the concentrations of the two compounds in the silk mordanted with alum/iron sulfate were less than those in the silk mordanted with aluminum potassium sulfate alone. Berberine and palmatine concentrations in the silk mordanted with alum/iron sulfate (berberine 206 µg/L, palmatine 80.1 µg/L) were only 63% and 75% of the silk mordanted with aluminum potassium sulfate alone (berberine 328 µg/L, palmatine 107 µg/L) (Figure 9).

These results present a significant implication on the subject of natural mordant dyeing. It is possible that silk mordanted with alum/iron sulfate had less dye uptake of Phellodendron dye than silk mordanted with aluminum potassium sulfate alone or that both berberine and palmatine in dyed silk are more liable to loss when the silk is mordanted with alum/iron sulfate. It is also possible that mordant type may affect the completeness of extraction. This research result indicates that caution is needed in the identification of dye in unknown archaeological textiles. Whether there is a difference either in the physicochemical attraction of berberine and palmatine toward different mordants or in the degree of dye loss needs further investigation.

HPLC-DAD-MS analysis of excavated archaeological textile samples

From 20 excavated archaeological textiles, the results of HPLC-DAD-MS revealed that three samples contained both berberine and palmatine and five samples contained palmatine only (Table 5). Neither berberine nor palmatine was observed in twelve samples, #8–19. These archaeological specimens are believed not to have been dyed using Phellodendron bark and they propose strong evidence that for the other samples which were confirmed to contain berberine and/or palmatine, Phellodendron bark may be the source of their dye. For eight samples which had berberine and/or palmatine, the retention times were 5.3 min for berberine and 5.2 min for palmatine, consistent with the freshly dyed silk and the 10-year old silks dyed with Phellodendron bark (Figure 10). The three samples that contained both berberine and palmatine were all from the Kupori burial site (#1, 3, 7). Sample #1 and 7 for which we had both washed and unwashed pieces revealed that the unwashed piece had significantly higher concentrations of berberine and palmatine than the washed specimens. In sample #1, the berberine content of the washed piece was less than 10%, and palmatine was less than 25% of that of the unwashed specimen. In sample #7, neither berberine nor palmatine was detected in the washed specimen. The results strongly suggest that these samples were dyed with Phellodendron bark and that wet washing had removed a large amount of dye. This result has a significant implication on the museum conservation treatments suggesting that wet washing can be detrimental to the preservation of dyes in archaeological textiles and that it should be prevented or minimized. And there is also a need for the study evaluating the loss of dye as a result of wet cleaning of aged textiles. As discussed previously, both berberine and palmatine were detected in the textiles dyed with Phellodendron bark and madder (Table 4). Therefore, we cannot rule out the possibility that madder could also be the source of dye for the excavated archaeological samples #1, 3, and 7. In order to determine if madder was used to dye these samples, alizarin standard and purpurin standard, which are the major constituents of plant madder dye²⁸ (Table 1), were analyzed and their retention times were determined in the HPLC-DAD-MS using the same instrumental method. As a result, we observed that alizarin eluted at 7.1 min from the alizarin (m/z = 241) standard and purpurin (m/z = 257) eluted at 7.6 min from the purpurin standard. The retention times of alizarin and purpurin were crosschecked in the chromatograms of the three samples, and neither alizarin nor purpurin was found in the chromatograms for these archaeological samples. Thus, it is strongly suggested that Phellodendron bark was the source of dye for the archaeological samples #1, 3, and 7.

Table 5.

HPLC-DAD-MS results of archaeological samples

Sample number	Burial	Museum treatment	Berberine (µg/L)	Palmatine (µg/L)
1	Kupori	wet washed	1.33	14.6
		unwashed	13.9	59.4
2	Kupori	wet washed	ND	9.57
3	Kupori	unwashed	1.39	9.62
4	Kupori	wet washed	ND	3.07
5	Kupori	wet washed	ND	1.77
6	Kupori	wet washed	ND	2.73
7	Kupori	wet washed	ND	ND
		unwashed	1.23	0.98
8	Kupori	unwashed	ND	ND
9	Kupori	wet washed	ND	ND
10	Kupori	unwashed	ND	ND
11	Youngam	unwashed	ND	ND
12	Youngam	unwashed	ND	ND
13	Youngam	unwashed	ND	ND
14	Youngam	unwashed	ND	ND
15	Youngam	unwashed	ND	ND
16	Youngam	unwashed	ND	ND
17	Youngam	unwashed	ND	ND
18	Youngam	unwashed	ND	ND
19	Youngam	unwashed	ND	ND
20	Youngam	unwashed	ND	0.56

ND: No detection.

In five excavated archaeological samples only palmatine was detected; samples #2, 4, 5, 6, and 20. It is difficult to determine whether the samples were dyed with Phellodendron bark or not. From the results of the naturally-aged textiles, we found that palmatine was detected in the textiles dyed with other plant sources as well, such as turmeric, cape jasmine, or sappanwood. Thus, it is possible that these five archaeological samples were dyed with plants other than Phellodendron bark. However, since they were all wet washed except for sample #20, it is also possible that the textiles were dyed with Phellodendron bark but a large portion of dye was lost during the washing procedure. Such questions raise the need for future research in the HPLC-DAD-MS analyses of berberine and palmatine after repeated washing. More information is needed on the effect of specific mordants; an elemental analysis for identifying mordants in archaeological samples would be useful.⁴⁸ For the current investigation, archaeological samples #2, 4, 5, and 6 may have been dyed using a number of different natural dyes with Phellodendron bark being a highly likely candidate. As for archaeological sample #20 which was not washed but showed the least concentration of palmatine (less than 1% of that in unwashed sample #1), it is doubtful that such small amount of palmatine alone came from dyeing with Phellodendron bark. Thus it is highly likely that the textile was not dyed with Phellodendron bark. Palmatine might have come from the contamination of the textile piece within the burial environment.

Conclusion

Extractions from the excavated textiles of the 17^th century burials of Korea were examined using HPLC-DAD-MS in comparison with the dye extracted from Phellodendron bark and the naturally-aged silk dyed with seven different known plants. The results of this study provide strong evidence that the three archaeological samples (#1, 3, 7) from the Kupori burial site were dyed with Phellodendron bark. According to a 19^th century reference, Phellodendron bark was not only used independently but it was also used in conjunction with indigo dye to produce light green colored fabric.⁹ Therefore, while the present results suggest that the three textiles with unknown dye composition were dyed with Phellodendron bark there is still question on the possibility of using two or more dye types in one fabric. Considering the visual color of the three textiles (Table 3), it is also clear that the source of dye cannot be determined by the color of archaeological textiles when they have undergone severe color change. Subsequent research is needed on the effect of different mordants and possible mixing of dyestuffs on the color change of archaeological textiles.

Footnotes

Acknowledgement

The authors greatly acknowledge Professor and Chair Ann T. Lemley Ph.D. of the Department of Fiber Science & Apparel Design of Cornell University for her kind support for this research and providing open access to the HPLC-DAD-MS instrument.

Funding

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology [grant no. 2010-0017075] and by the College of Human Ecology at Cornell University.

References

Ahn

Cho

Kim

. Analysis of non-fibrous matters in the exhumed textiles from Kupori Hwasung. Hankookboksik 1996; 14: 27–48.

Choo

CKK

Lee

. Analysis of dyeings produced by traditional Korean methods using colorants from plant extracts. Color Technol 2002; 118: 35–45.

Chan

Chu

Mok

DKW

Chau

. Analysis of berberine and total alkaloid content in Cortex Phellodendri by near infrared spectroscopy (NIRS) compared with high-performance liquid chromatography coupled with ultra-visible spectrometric detection. Anal Chim Acta 2007; 592: 121–131.

Sze

SCW

Tong

. Quality assessment of Cortex Phellodendri by high-performance liquid chromatography coupled with electrospray ionization mass spectrometry. Biomed Chromatogr 2010; 24: 438–453.

Lee

Eom

. Narrowbore high performance liquid chromatography of berberine and palmatine in crude drugs and pharmaceuticals with ion-pair extraction using cobalt thiocyanate reagent. J Pharm Biomed Anal 1999; 21: 59–63.

Drasar

Moravcova

. Recent advances in analysis of Chinese medical plants and traditional medicines. J Chromatogr B 2004; 812: 3–21.

Gibbs

Seddon

Brovenko

Petrosyan

Barnard

. Analysis of ancient dyed Chinese papers by high-performance liquid chromatography. Anal Chem 1997; 69(10): 1965–1969.

Kim

. Wooriui saek jaihyunul wiehan jeontong yeomsaek. In: Korea Cultural Heritage Foundation (ed.) The Korean Traditional Natural Dyeing and Crafts. Seoul: Korea Cultural Heritage Foundation, 2005, pp. 85–120.

Pinghogak Yi Ssi. Kyuhapchongso. Trans Jung YW. Seoul: Po Chin Chai Printing Company Ltd., 1997, p.151.

10.

Suto

Kakinuma

Ito

Sagara

Iwasaki

Itokawa

. Determination of berberine and palmatine in Phellodendri Cortex using ion-pair supercritical fluid chromatography on-line coupled with ion-pair supercritical fluid extraction by on-column trapping. J Chromatogr A 1997; 786: 371–376.

11.

Pang

Deng

Liu

Liang

Liu

. A sensitive and specific liquid chromatography mass spectrometry method for simultaneous determination of berberine, palmatine, coptisine, epiberberine and jatrorrhizine from coptidis rhizoma in rat plasma. Int J Mass Spectrom 2007; 268: 30–37.

12.

Yin

Han

. Simultaneous determination of 11 active components in two well-known traditional Chinese medicines by HPLC coupled with diode array detection for quality control. J Pharm Biomed Anal 2009; 49: 1101–1108.

13.

Lee

Hwang

Jung

Kim

. Dyeing and deodorizing properties of cotton, silk, wool fabrics dyed with amur corktree, Dryopteris crassirhizoma, Chrysanthemum boreale, Artemisia extracts. J Appl Polym Sci 2010; 115: 2246–2253.

14.

Kim

Park

. Effects of introduced chemical groups on the dyeability of cotton fabrics with Phellodendron amurense Rupr. Dyes Pigm 2007; 75: 351–355.

15.

Zhang

Laursen

. Application of LC-MS to the analysis of dyes in objects of historical interest. Int J Mass Spectrom 2009; 284: 108–114.

16.

Ahn

. Professor. University of Incheon. Personal Communication, 20 July, 2011.

17.

Hua

Ding

Chen

Gong

. Determination of berberine in human plasma by liquid chromatography-electrospray ionization-mass spectrometry. J Pharm Biomed Anal 2007; 44: 931–937.

18.

Nam

. Dyeing with natural dyes. Fiber Technol Industry 1998; 2(2): 238–257.

19.

Ahn

. Youngamgun Keumjoungmyeon Tamjin Choissi chooltoboksik boonsuk. In: Ewha Womans University Museum (ed.) The 17^th-Century Joseon military attire excavated from the tomb of General Choi Won-Lip. Seoul: Publishing Company Minsokwon, 2006, pp. 205–222.

20.

Liu

Zhao

Wang

Zheng

. Simultaneous determination of berberine and palmatine in rabbit plasma by LC-MS-MS and its application in pharmacokinetic study after oral administration of coptidis and coptidis-gardeniae couple extract. Chromatographia 2009; 70(7/8): 1113–1119.

21.

Liu

Wang

Wei

. Quality evaluation of platycladus orientalis (L.) Franco through simultaneous determination of four bioactive flavonoids by high-performance liquid chromatography. J Pharm Biomed Anal 2006; 41: 1186–1190.

22.

Pedriali

Fernandes

Bernusso

LdC

Polakiewicz

. The synthesis of water-soluble derivative of rutin as an antiradical agent. Quim Nova 2008; 31(8): 2147–2151.

23.

Alvarez-Diduk

Ramirez-Silva

Alarcon-Angeles

Galano

Rojas-Hernandez

Romero-Romo

Palomar-Pardave

. Electrochemical characterization of quercetin in aqueous solution. ECS Trans 2009; 20(1): 115–122.

24.

Valianou

Stathopoulou

Karapanagiotis

Magiatis

Pavilidou

Skaltsounis

Chryssoulakis

. Phytochemical analysis of young fustic (cotinus coggygria heartwood) and identification of isolated colourants in historical textiles. Anal Bioanal Chem 2009; 394: 871–882.

25.

Kurien

Singh

Matsumoto

Sofield

. Improving the solubility and pharmacological efficacy of curcumin by heat treatment. Assay Drug Dev Technol 2007; 5(4): 567–576.

26.

Chen

Zhang

Tian

Zhao

Cai

Liu

. Antioxidant potential of crocins and ethanol extracts of gardenia jasminoides ELLIS and crocus sativus L.: A relationship investigation between antioxidant activity and crocin contents. Food Chem 2008; 109(3): 484–492.

27.

Sato

Nojiri

Onodera

. Studies on the synthesis of safflomin-A, a yellow pigment in safflower petals: Oxidation of 3-c-b-d-glucopyranosyl-5-methylphloroacetophenone. Carbohydr Res 2005; 340(3): 389–393.

28.

Szostek

Orska-Gawrys

Surowiec

Trojanowicz

. Investigation of natural dyes occurring in historical Coptic textiles by high-performance liquid chromatography with uv–vis and mass spectrometric detection. J Chromatogr A 2003; 1012: 179–192.

29.

Budavari S, O'Neil MJ, Smith A and Heckelman PE. The Merck Index: An Encyclopedia of Chemicals, Drugs, and Biologicals. 11th ed. N.J.: Merck & Co., Inc., 1989, p. 180, 6951.

30.

Wouters

. High performance liquid chromatography of anthraquinones: Analysis of plant and insect extracts and dyed textiles. Stud Conserv 1985; 30(3): 119–128.

31.

Surowiec

Nowik

Trojanowicz

. Identification of ‘insoluble’ red dyewoods by high performance liquid chromatography–photodiode array detection (HPLC-PDA) fingerprinting. J Sep Sci 2004; 27: 209–216.

32.

Berghe

Gleba

Mannering

. Towards the identification of dyestuffs in early Iron Age Scandinavian peat bog textiles. J Archaeol Sci 2009; 36: 1910–1921.

33.

Wouters

Maes

Germer

. The identification of haematite as a red colorant on an Egyptian textile from the second millenium B.C. Stud Conserv 1990; 35(2): 89–92.

34.

Petroviciu

Albu

Medvedovici

. LC/MS and LC/MS/MS based protocol for identification of dyes in historic textiles. Microchem J 2010; 95: 247–254.

35.

Chen

Wen

Her

. Analysis of coptisine, berberine and palmatine in adulterated Chinese medicine by capillary electrophoresis-electrospray ion trap mass spectrometry. J Chromatogr A 2000; 866: 273–280.

36.

Deevanhxay

Suzuki

Maeshibu

Tanaka

Hirose

. Simultaneous characterization of quaternary alkaloids, 8-oxoprotoberberine alkaloids, and a steroid compound in Coscinium fenestratum by liquid chromatography hybrid ion trap time-of-flight mass spectrometry. J Pharm Biomed Anal 2009; 50: 413–425.

37.

Zhang

Jin

Dong

Xia

Feng

Xue

. Systematic screening and characterization of tertiary and quaternary alkaloids from corydalis yanhusuo W.T. Wang using ultra-performance liquid chromatography-quadrupole-time-of-flight mass spectrometry. Talanta 2009; 78: 513–522.

38.

Ohga

Kira

Koga

. Comparison of berberine content in Phellodendron amurense Rupr. inner bark in relation to its variety, stem position, season and tree age. Bull Kyushu Univ For 1992; 66: 45–53.

39.

Wang

Yan

. Effects of water stress on berberine, jatrorrhizine and palmatine contents in amur corktree seedlings. Acta Ecologica Sinica 2007; 27(1): 58–64.

40.

Chem

Berberine.

http://chemyq.com/En/xz/xz1/5664hecoy.htm (2005, accessed August 2011).

41.

LKT Laboratories, Inc. Palmatine chloride hydrate http://www.lktlabs.com/products/Palmatine_Chloride_Hydrate-345-23.html (2010, accessed August 2011).

42.

Marek

Seckarova

Hulova

Marek

Dostal

Sklenar

. Palmatine and berberine isolation artifacts. J Nat Prod 2003; 66: 481–486.

43.

Broughton

. On a false cinchona bark of India. Pharm J Trans Second series 1868; 9: 418–421.

44.

Suffness

Cordell

. Antitumor Alkaloids. In: Brossi

(ed.) The Alkaloids. Orlando: Academic Press, 1985, pp. 190–190.

45.

Kar

. Pharmacognosy and Pharmacobiotechnology. New Delhi: New Age International Publishers, 2003, pp. 463–463.

46.

Read

Zasada

. Phellodendron amurense Rupr.: Amur corktree. In: Bonner

Karrfalt

(eds) The Woody Plant Seed Manual. Agric Handbook No. 727. Washington DC: US Department of Agriculture, Forest Service, 2008, pp. 783–785.

47.

. Color evaluation of wool fabric dyed with Rhizoma coptidis extract. J Appl Polym Sci 2006; 101: 3376–3380.

48.

Ahn

. Application of spectrochemical analysis in the study of archaeological textiles. J Korean Soc Costume 1999; 49: 49–63.