1H nuclear magnetic resonance-based metabolite profiling of guava leaf extract: an attempt to develop a prototype for standardization of plant extracts

Abstract

Background: Herbal medicines are fast gaining popularity. However, their acceptability by modern practitioners is low which is often due to lack of standardization. Several approaches towards standardization of herbals have been employed. The current study attempted to recognize key peaks from ¹H NMR spectra which together would comprise of a spectral fingerprint relating to efficacy of Psidium guajava (guava) leaf extract as an antidiarrhoeal when a number of unidentified active principles are involved.

Methods: Ninety samples of guava leaves were collected from three locations over three seasons. Hydroalcoholic (water and ethanol, 50:50) extracts of these samples were prepared and their ¹H NMR spectra were acquired. Spectra were also obtained for quercetin, ferulic acid and gallic acid as standards. Eight bioassays reflecting different stages of diarrhoeal pathogenesis were undertaken and based on pre-decided cut-offs, the extracts were classified as 'good' or 'poor' extracts. The bioactivity data was then correlated with the ¹H NMR profiles using Regression or Orthogonal Partial Least Square-Discriminant Analysis (OPLS-DA).

Results: OPLS-DA showed seasonal and regional segregation of extracts. Significant models were established for seven bioassays, namely those for anti-bacterial activity against Shigella flexneri and Vibrio cholerae, adherence of E. coli, invasion of E. coli and S. flexneri and production and binding of toxin produced by V. cholerae. It was observed that none of the extracts were good or bad across all the bioassays. The spectral analysis showed multiple peaks correlating with a particular activity. Based on NMR and LC-MS/MS, it was noted that the extracts contained quercetin, ferulic acid and gallic acid. However, they did not correlate with the peaks that segregated extracts with good and poor activity.

Conclusions: The current study identified key peaks in ¹H NMR spectra contributing to the anti-diarrhoeal activity of guava leaf extracts. The approach of using spectral fingerprinting employed in the present study can thus be used as a prototype towards standardization of plant extracts with respect to efficacy.

Keywords: 1H NMR; India; Metabolomics; Psidium guajava.

Fig. 1

A representative ¹H NMR spectrum of the guava leaf extract

Fig. 2

PCA plots for seasonal differentiation. a) PCA plot for seasonal data; b) OPLS-DA plot for season B and C; c) OPLS-DA plot for season C and D; d) OPLS-DA plot for season B and D. B: collection season May 2013; C: collection season October 2013; D: collection season March 2014

Fig. 3

PCA plots for regional differentiation. a) PCA plot for regional data; b) OPLS-DA plot for W and R; c) OPLS-DA plot for R and Da; d) OPLS-DA plot for W and Da. W: Leaves collected from region Shirwal; R: Leaves collected from region Rahata; Da: Leaves collected from region Dapoli

Fig. 4

OPLS-RA model developed for antibacterial activity against Shigella flexneri. B: Poor activity; G: Good activity

Fig. 5

OPLS-RA model developed for antibacterial activity against Vibrio cholerae. B: Poor activity; G: Good activity; I: Intermediate activity

Fig. 6

OPLS-DA model developed for effect on adherence of EPEC to HEp-2 cells. B: Poor activity; G: Good activity

Fig. 7

OPLS-DA model developed for effect on invasion of EIEC into HEp-2 cells. B: Poor activity; G: Good activity

Fig. 8

OPLS-DA model developed for effect on invasion of Shigella flexneri into HEp-2 cells. B: Poor activity; G: Good activity

Fig. 9

Regression model developed for production of cholera toxin by Vibrio cholerae. B: Poor activity; G: Good activity; I: Intermediate activity

Fig. 10

OPLS-DA model developed for binding of cholera toxin to GM1 receptor. B: Poor activity; G: Good activity