Determining the purity of samples from natural products by coupling HPLC and CCD spectrometry

J. Sep. Sci. 2005, 28, 286–290 www.jss-journal.de i 2005WILEY-VCH Verlag GmbH&Co. KGaA,Weinheim

Bin Chen1

Xianrong Shen2

Jilie Kong1

1Department of Chemistry, FudanUniversity, Shanghai 200433,P.R. China

2Marine Biology Research Center,Naval Medical ResearchInstitute, Shanghai 200433,P.R. China

Determining the purity of samples from naturalproducts by coupling HPLC and CCD spectrometry

Isolation and identification of natural products is a very important and active researchfield. However, establishing the purity of the samples during the isolation process isquite difficult, especially when the retention times are similar for two desired compo-nents in HPLC. Although some technologies, e.g. MS and NMR, offer effective waysof obtaining purity information about the samples, the expensive instrumentationrequired or the off-line nature of coupling (generally speaking) make purity analysissomewhat inconvenient. In this paper, an on-line analytical system coupling HPLCand a CCD spectrometer for determination of purity for each eluate was developed ina thin layer spectrometric cell. The effectiveness of the system was demonstrated bydifferentiating Tanshinone I, Tanshinone IIA, and their mixture. The time-resolvedUV-Vis spectra promptly revealed significant differences between the three sampleswhile conventional single wavelength detection (CSWLD) could not. The system wasthen used to distinguish two steroid compounds which behaved as a single compo-nent in CSWLD. The compounds were isolated from a Chinese marine invertebrateanimal, a marine annelid, Arenicola cristata, referred to here as Stimpson. Themethod reported here provided an efficient, convenient, fast, and inexpensiveapproach holding promise for on-line determination of the purity of samples isolatedfrom natural products.

KeyWords:Natural products analysis; Stimpson; HPLC; CCD spectrometer; On-line Purity identi-fication;

Received: April 5, 2004; revised: May 24, 2004; accepted: December 15, 2004

DOI 10.1002/jssc.200401815

1 Introduction

The isolation and identification of natural products com-mands great interest. Numerous active compounds,including potential novel medicines and materials, havebeen found and utilized in research work [1–3]. The nat-ural products reported are mainly isolated from the sec-ondary metabolites of terricolous organisms [4, 5], marineorganisms [6, 7], andmicroorganisms [8, 9]. The chemicalconstituents of these secondary metabolites are rathercomplicated and in many cases probably belong to thesame class of compounds with similar molecular struc-tures [10]. The purity of the isolated samples thus stronglyaffects the subsequent structural characterization, on theone hand by frequently increasing the analytical difficultyof MS or NMR, and on the other hand by also prolongingthe analysis time by necessitating additional purificationsteps in cases where the MS or NMR signals of the analo-gues greatly interfere with those of the desired compound.The strategies commonly used for identifying the purity ofthe prepared samples are thin layer chromatography(TLC), HPLC-UV-Vis, MS, or LC-MS, and LC-NMR [11–

13]. Although somewhat old, TLC is nevertheless a usefuland simple method for primary isolation and analysis.However, its determination sensitivity is rather limited;thus, it is often very difficult to judge whether a sample ispure or not by using TLC alone. HPLC is a well-knownhighly efficient isolation method. However, only singlewavelength UV-VIS detection is used in most HPLC sep-arations, which sometimes fails to provide enough infor-mation to analyze the purity of the desired samples, espe-cially when two analytes show the same or similar reten-tion time. In this case, what appears as a single perfectpeak in HPLC is usually shown to be mixture by subse-quent MS or NMR measurements. MS/LC-MS or LC-NMR does work as a universal method to authenticate thepurity of the interested samples; however, the expensiveinstrumentation and the time-consuming manipulationlimit their widespread application in the routine on-linescreening of natural products.

For HPLC coupled with some newly developed detectiontechnologies, such as chemiluminescence (CL) and eva-porative light scattering (ELC) detectors [14, 15], thedetection sensitivity was obviously enhanced. In addition,efforts focusing on the improvement of separation effi-ciency for chromatographic columns have also

Correspondence: Jilie Kong, Department of Chemistry, FudanUniversity, Shanghai 200433, P.R. China.E-mail: [email protected].

286 Chen, Shen, Kong

Determining the purity of samples from natural products 287

emerged [16, 17]. However, these methods have rarelybeen applied in routine assessment of the purity of naturalproducts. Utilizing a CCD spectrophotometer coupled toHPLC to verify the purity of samples in the isolation of nat-ural products has not previously been reported, and istherefore the subject of this work.

In our previous work, a CCD spectrophotometer has beensuccessfully used in studying the in-situ electrochromicproperties of MoO3 thin films [18] and the dynamic spec-troelectrochemical behavior of myoglobin in bio-mimicmembranes [19]. Compared to the conventional singlewavelength detector (CSWLD), the CCDspectrophot-ometer has shown great capability in both high-speeddata acquisition and full-wavelength spectra monitoring.In this work, an on-line analytical system coupling HPLCand CCD spectrometer for identifying the purity of the elu-ate of interest was developed on the basis of a thin layerspectrometric cell. The advantage of the system in verify-ing the purity of samples with similar structure wasdemonstrated by using it to differentiate Tanshinone I,Tanshinone IIA, and their mixture, which belong to thesame class of natural products isolated from a well-knownChinese herb, red sage root [20]. Also, the purity of a sam-ple containing unknown analogous compounds isolatedfrom a Chinese marine invertebrate, called Stimpson, wasanalyzed by this method. The results revealed that themethod is simple, fast, efficient, and inexpensive, andrepresents a promising on-line approach for identifyingthe purity of the samples isolated from natural products.Of course, conventional diode-array spectrophotometricdetection can also give very similar results but requireslonger response times and is of greater complexity thanthe CCD system presented here.

2 Experimental

2.1 General experimental procedures

A 10 lL sample was injected into the analytical system viaa six-way valve. The absorbance of each eluate wasdetected by CSWLD, and the time-resolved full-wave-length spectra were consecutively acquired by the CCDspectrophotometer. All data were processed to 3D spec-tra by Origin 5.0 software.

2.2 Standards

Tanshinone I and Tanshinone IIA were purchased fromNational Institute for the Control of Pharmaceutical andBiological Products (Beijing, China). Their molecularstructures (shown in Figure 1 [20]) show them to be diter-pene compounds. The two standards (2.0 mg) were dis-solved in 10 mL HPLC grade methanol for further experi-ments.

2.3 Collection, extraction, and purification

Stimpson were collected in winter of 2002 from the Beachof Zhou Shan Island, located on the coast of the EastChina Sea in Zhejiang province, and frozen immediatelyafter collection. Frozen Stimpson (5000 g dry wt) wereminced and extracted with 85% aqueous ethanol to give500 g of organic extract that was dispersed in 1 L of water,and sequentially partitioned between water and ethylacetate or n-butanol. The n-BuOH layer was repeatedlyfractionated on four kinds of SiO2 column (8 cm ID, 5 cmID, 2.5 cm ID, and 1.5 cm ID), one fraction of which wassubjected to HPLC chromatography on a SiO2 column, togive a roughly pure sample HQY-A (3.3 mg). Then theobtained HQY-A was dissolved in 10 mL HPLC gradeCHCl3-MeOH (9:1) and subjected to further experiments.1H-NMR indicated that HQY-A contains two steroid com-pounds.

2.4 Design of the thin layer flow-throughspectrophotometer cell

The thin layer flow-through spectrometer cell was con-structed from two slices of quartz and one piece of siliconerubber, all of 15-mm diameter. A hole of 5 mm diameterwas drilled at the center of the silicon rubber, and one slotwas made through the whole slice. Then the two stainlesssteel tubes (ID = 0.5 mm) were embedded in the slot andtheir tips pointed to the center of the hole and fixed on theedge of hole. Subsequently, the piece of silicone rubberwas fixed between the two pieces of quartz by epoxy resinand the edge of the cell was airproofed by epoxy. The cellwas left to solidify for 24 h.

2.5 HPLC andCCD spectrophotometer conditions

Liquid flow was powered by a Waters 590 binary pump(Waters Corp., Milford, MA). HPLC separation wasachieved on SiO2 column (4.66250 mm, 5 lm) obtainedfrom Dalian Elite Scientific Instruments Co., Ltd. 98 :2(v/v) n-hexane/isopropyl alcohol was used as mobile


Figure 1. Molecular structure of the standards A) Tanshi-none I; B) Tanshinone IIA.

ShortCommunication


phase and a flow rate of 0.8 mL/min was used in the isola-tion process. The wavelength of the single-wavelengthdetector was set at 254 nm. All full wavelength UV-VISabsorbance data were acquired on a SM-240 CCD spec-trophotometer (CVI spectral instruments, Putnam, CT,USA) whose calibration set was 240-BU60142 and baseaddress was 300h. Before any experiment was started, adark scan and a reference scan were performed.

3 Results and discussion

3.1 Design of the thin layer flow-throughspectrophotometer cell

The design of the flow-through spectrophotometer cell iscrucial to any UV-VIS detector. The efficiency of the chro-matographic column and the sensitivity of the detector aredirectly affected by the volume and the shape of the cell.Turbulent flow, light scattering, flow rate, and temperatureaffect the stability of the detector. In order to enhance thesensitivity of detection, the volume of the cell should beconstructed as small as possible. However, decreasingthe volume will extend the chromatographic peaks. Thedesign of the flow-through spectrophotometer cell used inour research is illustrated schematically. The thickness ofthe silicone rubber layer is 0.4 mm and that of the quartzslice (d = 15 mm) is 1.02 mm. Thus the volume of theformed flow cell (d =mm) is 7.85 lL (cal.), which is consis-tent with the mentioned optimum volume of 5 lL–8 lL fora CSWLD [21]. To avoid light scattering and backgroundabsorption, quartz was used as light-transmitting material.Silicone rubber was used because of its ease of mechan-ical processing and resistance to the solvent used in the

experiment. Epoxy resin was used to ensure good air-proofing.

3.2 Reliability of the analytical system

The retention times measured by CSWLD of HPLC forpure Tanshinone I and Tanshinone IIA were 9.6 min and9.4 min, respectively. The single eluate peak itself con-tained little purity information about the samples injected.In chromatogramsmeasured by the coupled CCD, we cansee that Tanshinone I has UV-VIS absorbance bandsfrom 220 nm to 600 nm with km centered at 245 nm,270 nm, 325 nm, and 420 nm, while Tanshinone IIA hasabsorbance bands from 210 nm to 540 nm with km at224 nm, 250 nm, 268 nm, 273 nm, 352 nm, and 455 nm.These absorbance peaks represent the characteristicabsorption for the two pure compounds [20]. The fullwavelength data were processed to a time-resolved 3Dspectra, from which we can see that km and peak shape inthe 3D full wavelength spectra of Tanshinone I and Tan-shinone IIA are basically the same as what was measuredin a single shot with the coupled CCD. So we can con-clude that the HPLC-CCD analytical system does workwell.

3.3 Detecting Tanshinone IIA containing a smallquantity of Tanshinone I

When 10 lL of a methanol solution of Tanshinone IIAmixed with a small quantity of Tanshinone I, e.g. 1%–3%,was injected into the analytical system via a six-way valve,the eluate peak showed little difference from that of pureTanshinone IIA, as seen in Figure 2(B) under the opti-mum condition of our experiment. In other words, the sin-


Figure 2. UV-VIS absorbance of 97:3 Tanshinone IIA/Tanshinone I. A) The full wavelength UV-VIS absorbance of 97 :3 Tanshi-none IIA/Tanshinone I acquired by coupling HPLC-CCD analytical system. B) Chromatographic signal for 97 :3 Tanshinone IIA/Tanshinone I acquired by CSWLD at 254 nm (zero on z-axis refers to 9th min of retention time in picture A).

Determining the purity of samples from natural products 289

gle wavelength detector gave no purity information aboutthe sample and could not be used as reference to collecteluate for further use. In contrast, the coupled CCD spec-trophotometer gave two different signals corresponding tothe two fractions as shown in Figure 2(A): from the zero’thsecond (referred to the retention time, here and below) tothe 38th second, the UV-VIS absorbance of the eluentshowed a characteristic absorbance spectrum of Tanshi-none IIA; but from the 45th second to the 50th second, theUV-VIS absorbance of the eluent showed a characteristicabsorbance spectrum of Tanshinone I. Between the twotime intervals, i.e. from the 38th to the 45th second, theabsorbance of the eluent consisted of the superposedspectra of the two compounds. From these time-resolvedspectra, one could easily conclude that the sample is not apure compound, in contrast to the situation pertaining onuse of CSWLD. In a series of experiments, different ratiosof Tanshinone IIA and Tanshinone I were examined; theresults indicated that when the content of Tanshinone IIAwas less than 97% of the total, the chromatographic peakfor the mixture acquired by CSWLD tended to broadenwith increasing content of Tanshinone IIA, suggesting theimpurity of the eluate. However, on continuously increas-ing the content of Tanshinone IIA to above 97%, the elu-ate peak appeared as a perfect single peak and could givefalse hint of the absence of Tanshinone I.

3.4 Detecting HQY-A isolated fromStimpson

Figure 3(A) shows the full wavelength absorptionrecorded by the coupled CCD for a 10 lL CHCl3-MeOH(9:1) solution of HQY-A isolated from Stimpson uponinjection into HPLC. The corresponding single peak

measured by CSWLD is shown in Figure 3(B). Asoccurred in case of 97:3 Tanshinone IIA/Tanshinone I,Figure 3(B) also shows a single chromatographic peakwith retention time of about 5.1 min. By way of compari-son, in the full wavelength spectra as shown in Fig-ure 3(A), the dynamic 3D absorption spectra (in the regionof 220 nm to 380 nm) were found basically the same fromthe zero’th second to the 11th second and from the 14thsecond to the 18th second; however, from the 12th sec-ond to the 13th second, different VIS absorbance bandswere observed in the region of 450 nm to 800 nm in the3D spectra. So we can readily conclude that the sample ofHQY-A isolated from Stimpson was not a pure compound.Hence the purity of the sample could be promptly identi-fied on-line by the HPLC-coupled CCD detector.

4 Concluding remarksAn analytical system was developed for determining thetime-resolved spectra obtained by a thin layer spectro-photometer cell with CCD coupled to an HPLC unit. Thesystem was then used to analyze the purity of Tanshino-ne I, Tanshinone IIA, and Tanshinone IIA containing asmall quantity of Tanshinone I. A sample containing theunknown steroid compounds isolated from the inverte-brate animal Stimpson was examined using this method.The result showed that this sample contained one maincompound whose UV-VIS absorbance was from 220 nmto 380 nm and another minor compound whose UV-VISabsorbance was from 400 nm to 800 nm. Themethod pre-sented here is simple, convenient, fast, efficient, and inex-pensive, ad appears promising as an on-line routine tech-


Figure 3. UV-VIS absorbance of HQY-A isolated from Stimpson. A) The full wavelength UV-VIS absorbance of HQY-A isolatedfrom Stimpson acquired by coupling HPLC-CCD analytical system. B) The single wavelength UV-VIS absorbance of HQY-Aacquired by coupling HPLC (zero on z-axis refers to 5th min of retention time in picture A).


nique for verifying the purity of samples in natural productsisolation.

Acknowledgements

This work was supported by NSFC (20335040,20475012), 863 (2002AA639180), Shanghai-SK research& development foundation and Nano-project (0244 nm029, 0452 nm 003) & SKLEAC.

References[1] J. Mann,Nature 2002, 2, 143.

[2] B.R. Constantz, I.C. Ison, M.T. Fulmer, R.D. Poser, R.T.Smith, M. Vanwagoner, J. Ross, S.A Goldstein, Science1995, 267, 1796.

[3] D. Pauly, V. Christensen,Nature 1995, 374, 255.

[4] H. Takayama, T. Ichikawa, T. Kuwajima, M. Kitajima, H.Seki, N. Aimi, M.G. Nonato, J. Am. Chem. Soc. 2000, 36,8635.

[5] D.B. Stierle, A.A. Stierle, T. Bugni, J. Org. Chem. 2003, 68,4966.

[6] A. Zampella, M.V. D’Auria, L.G. Paloma, A. Casapullo, L.Minale, C. Debitus, Y. Henin, J. Am. Chem. Soc. 1996, 118,6202.

[7] A. Zampella, M.V. D’Auria, L. Minale, C. Debitus, C. Roussa-kis, J. Am. Chem. Soc. 1996, 118, 11085.

[8] H. Kakeya, R. Onose, H. Koshino, A. Yoshida, K. Kobaya-shi, S. Kageyama, H. Osada, J. Am. Chem. Soc. 2002, 124,3496.

[9] N. Gonzalez, J. Rodriguez, R.G. Kerr, C. Jimenez, J. Org.Chem 2002, 67, 5117.

[10] Y.J. Lim, C.O. Lee, J. Hong, D.K. Kim, K.S. Im, J.H. Jung, J.Nat. Prod. 2001, 64, 1565.

[11] S.J. Cho, S.K. Lee, B.J. Cha, Y.H. Kim, K.S. Shin, FemsMicrobiol. Lett. 2003, 223, 47.

[12] A. Randazzo, G. Bifulco, C. Giannini, M. Bucci, C. Debitus,G. Cirino, L. Gomez-Paloma, J. Am. Chem. Soc. 2001, 123,10870.

[13] O. Corcoran, M. Spraul,Drug Discov. Today 2003, 8, 6241.

[14] M. M�kinen, V. Piironen, A. Hopia, J. Chromatogr. A 1996,734, 221.

[15] A. Dapkevicius, T.A. van Beek, H.A.G. Niederlander, Anal.Chem. 1999, 71, 736.

[16] S. Strohschein, M. Pursch, K. Albert, J. Pharm. BiomedAnalysis 1999, 21, 669.

[17] J.C. Xie, L.L. Zhu, X.J. Xu,Anal. Chem. 2002, 74, 2352.

[18] J.L. Kong, X.L. Wu, J.X. Zhou, J.F. Rusling, Biophys. Chem.1999, 79, 219.

[19] J. Tang. Y.D. Lu, B.H. Liu, P.Y. Yang, Y.P. Huang, J.L.Kong, J. Solid State Electrochem. 2003, 7, 244.

[20] M.F. Qian, B.J. Yang, W.H. Gu, Z.X. Chen, X.D. Chen, X.Q.Ye,Acta Chim. Sinica 1978, 36, 199.

[21] P. Hewlett, High Sensitivity Detection Cell for HP3D CapillaryElectrophoresis System,12-5965-5984E, Boeblingen, Ger-many, 1997.


Documents

Determining the purity of samples from natural products by coupling HPLC and CCD spectrometry