Four Chemical Markers Reveal How Desert Medicinal Plant Adapts to Saline-Alkali Soil

A parasitic medicinal herb long harvested from the deserts of northwestern China is now being grown successfully on salty coastal land, and scientists have worked out exactly which molecules set the salty-soil harvest apart. In a study published in the journal Chinese Medicine, researchers report that Cistanches Herba, the fleshy stem of the parasitic plant Cistanche tubulosa, accumulates markedly higher levels of four characteristic compounds when cultivated on saline-alkali land in the Yellow River delta compared with plants grown in traditional desert sandy land.

The work matters because the traditional habitat of Cistanches Herba is the desert of northwestern China, where the plant parasitizes the roots of the sand-fixing shrub Tamarix chinensis. To improve saline-alkali soil in eastern coastal regions such as the Yellow River delta, growers have begun cultivating the herb on salt-tolerant Tamarix chinensis there, creating a model that links ecological restoration with the production of a valued medicinal resource. What remained unclear was whether this new environment changes the plant’s secondary metabolism—the suite of compounds that largely defines its quality as a medicinal material. Saline-alkali stress is known to reshape secondary metabolite profiles in other medicinal plants, including Glycyrrhiza and Salvia miltiorrhiza, but the systematic chemical response of Cistanches Herba had not been mapped before.

The research team, led by Mengjiao Li, Yaqi Wang, and Chenghao Bi as co-first authors with Liwen Han among the senior authors, collected twenty batches of plant material from two standardized artificial cultivation systems. Ten batches came from saline-alkali planting sites in Lijin County, Shandong Province, where the sandy loam soil has a pH of 8.32 to 8.68, a salt content of 1.75 to 2.07 parts per thousand, and mild saline-alkali character. Ten batches came from long-term desert cultivation bases in Xinjiang, spanning Aral Township in Hotan, Fuhai County in Altay, and Yutian and Minfeng counties, where the aeolian sandy soil has a pH of 7.4 to 7.83 and total salt content below 0.7 percent. Both systems followed identical protocols: seeds were inoculated onto one-year-old Tamarix chinensis roots every April from 2021 to 2024, and fleshy stems were harvested in November using random sampling points spaced at least three to five meters apart.

Every sample was authenticated before analysis. Dr Fengdan Guo of the Shandong Provincial Academy of Traditional Chinese Medicine verified the specimens morphologically as the fleshy stems with scale leaves of Cistanches Herba, and DNA barcoding of the internal transcribed spacer 2 region of nuclear ribosomal DNA, followed by NCBI BLAST comparison, confirmed that all sequenced samples belonged to Cistanche tubulosa. This dual verification is important because the study’s conclusions rest on comparing chemically similar material grown under two genuinely different soil regimes, not on comparing different species or unverified market samples.

The first screening layer used untargeted metabolomics on an ultra performance liquid chromatography-quadrupole time-of-flight mass spectrometry platform operating in negative electrospray ionization mode. The team identified 74 chemical constituents in total: 40 phenylethanoid glycosides, 13 iridoid glycosides, 9 lignans, and 12 constituents of other classes. Phenylethanoid glycosides, the dominant class, share a β-glucopyranose core attached to a hydroxyphenylethyl aglycone, and their fragmentation in negative mode follows recognizable routes—loss of a caffeoyl moiety or its dehydrated form, loss of an acetyl group, and cleavage of rhamnosyl and glucosyl residues. One representative constituent, showing a quasi-molecular ion at m/z 639.1940, was tentatively identified as campneoside II through this stepwise fragmentation pattern. Only negative-mode data entered the formal analysis because positive-mode data showed insufficient quality and incomplete metabolic coverage.

Statistical modeling then separated the two habitats cleanly. A principal component analysis showed a clear separation trend between the saline-alkali and sandy-land samples, and an orthogonal projections to latent structures-discriminant analysis model achieved R2Y and Q2(cum) values of 0.835 and 0.787, respectively. A permutation test with 200 permutations confirmed the model was not overfitted, with all permuted R2 and Q2 values falling below the original values and a negative intercept on the Q2 regression line. Applying thresholds of VIP greater than 1, P less than 0.05, and fold change of at least 1.5 or at most 0.5, the researchers screened out 30 potential differential metabolites. Volcano plot analysis indicated that 42 metabolite features changed significantly, with 24 constituents increased and 18 decreased in the saline-alkali samples relative to the desert samples.

Because differential abundance alone cannot distinguish metabolites genuinely coupled to saline-alkali stress from those fluctuating randomly, the team added a second filtering layer based on weighted gene co-expression network analysis. Re-analyzing the 74 constituents, the method produced four co-expression modules: a blue module of 23 constituents, a turquoise module of 26, and a brown module of 18, with unassigned metabolites in a grey module that was excluded. The blue module showed a significant positive correlation with the saline-alkali group, with a Pearson correlation coefficient of 0.56 and P less than 0.05. Twelve constituents fell both inside the blue module and among the putative differential metabolites—nine phenylethanoid glycosides, one iridoid glycoside, one lignan, and one compound from another class—forming the candidate pool for the next stage.

The third layer brought absolute quantification. Using an AB SCIEX Triple Quad 5500 triple quadrupole system in multiple reaction monitoring mode, the researchers measured the twelve candidates across all twenty batches. The method performed well: calibration correlation coefficients ranged from 0.9989 to 0.9997, sample solutions stored at room temperature for 24 hours showed relative standard deviations below 2 percent, precision RSDs ranged from 0.47 to 2.21 percent with relative errors between 98.08 and 103.78 percent, repeatability RSDs spanned 0.79 to 2.25 percent, and extraction recoveries fell between 97.39 and 104.40 percent. Seven of the twelve constituents were elevated in the saline-alkali samples, and isoacteoside showed the most pronounced increase, averaging 2313.60 plus or minus 745.76 micrograms per gram in desert-origin samples versus 6425.60 plus or minus 1289.63 micrograms per gram in saline-alkali samples.

The final selection combined principal component analysis with the quantitative results. The dataset passed suitability checks, with a Kaiser–Meyer–Olkin value of 0.726 and a Bartlett’s sphericity statistic of 290.418 with P equal to 0. Two principal components were extracted under the eigenvalue-greater-than-1 criterion, together explaining 79.211 percent of the variance. Isoacteoside, verbascoside, pinoresinol, and cistanoside F loaded highly on the first component, while tubuloside A and echinacoside loaded highly on the second. Of these six, four—verbascoside, echinacoside, isoacteoside, and pinoresinol—were significantly more abundant in the saline-alkali samples, and they were confirmed as the characteristic chemical markers of the saline-alkali habitat. Cluster analysis of the twelve measured constituents grouped most saline-alkali samples together and most sandy-land samples together, revealing geographically dependent differences in compound content without divergence at the species level.

The study also looked at where these compounds travel in the body, using desorption electrospray ionization mass spectrometry imaging on mouse tissues after the animals were given an intragastric gavage of the saline-alkali-grown herb at 5 grams per kilogram. Because verbascoside and isoacteoside could not be differentiated under imaging conditions and pinoresinol gave insufficient signal, the imaging work focused on echinacoside, verbascoside, cistanoside F, and tubuloside A, using an optimized 98 percent aqueous methanol spray solvent after formic acid was found to suppress target ion signals. Echinacoside was detectable in all organs examined—brain, heart, spleen, lung, and kidney—but was most prominently enriched in brain tissue, with higher relative contents in the cerebrum, pons, and olfactory bulb than in the cerebellum. Verbascoside and tubuloside A showed markedly higher levels in lung tissue than in other organs, and verbascoside was significantly more abundant in the renal medulla than in the renal cortex, while echinacoside and tubuloside A showed the opposite pattern. Cistanoside F was enriched specifically in heart tissue and was not detected in the renal medulla. These are observations of spatial distribution in mice after administration, not demonstrations of therapeutic effect, but the authors argue that such in situ histological patterns offer clues for interpreting the pharmacological potential of the screened metabolites. Overall, the team says the hierarchical strategy—broad differential screening, stress-correlated module selection, absolute quantification, and dimensionality reduction—provides a scientific basis for quality control and resource development of this medicinal material grown on reclaimed coastal land.

Subject of Research: Identification of habitat-specific chemical markers in the medicinal plant Cistanches Herba cultivated on saline-alkali land

Article Title: Systematic identification of characteristic chemical constituents in Cistanches Herba from saline-alkali land based on hierarchical filtering of mass spectrometry data and spatial metabolomics technology

Article References: Li, M., Wang, Y., Bi, C., Wang, R., Guo, F., Shen, G., Hu, Z., Liu, K., Li, X., Guan, R., Wang, S., Zhong, H., Wang, X., & Han, L. (2026). Systematic identification of characteristic chemical constituents in Cistanches Herba from saline-alkali land based on hierarchical filtering of mass spectrometry data and spatial metabolomics technology. Chinese Medicine, 21(1), Article 229. https://doi.org/10.1186/s13020-026-01526-6

Image Credits: AI Generated

DOI: 10.1186/s13020-026-01526-6

Keywords: Cistanches Herba, Cistanche tubulosa, saline-alkali land, metabolomics, UPLC-QTOF-MS, WGCNA, phenylethanoid glycosides, echinacoside, verbascoside, DESI-MSI, Yellow River delta, mass spectrometry imaging