Is Our Natural Food Our Homeostasis? Array of a Thousand Effect-Directed Profiles of 68 Herbs and Spices

The beneficial effects of plant-rich diets and traditional medicines are increasingly recognized in the treatment of civilization diseases due to the abundance and diversity of bioactive substances therein. However, the important active portion of natural food or plant-based medicine is presently not under control. Hence, a paradigm shift from quality control based on marker compounds to effect-directed profiling is postulated. We investigated 68 powdered plant extracts (botanicals) which are added to food products in food industry. Among them are many plants that are used as traditional medicines, herbs and spices. A generic strategy was developed to evaluate the bioactivity profile of each botanical as completely as possible and to straightforwardly assign the most potent bioactive compounds. It is an 8-dimensional hyphenation of normal-phase high-performance thin-layer chromatography with multi-imaging by ultraviolet, visible and fluorescence light detection as well as effect-directed assay and heart-cut of the bioactive zone to orthogonal reversed-phase high-performance liquid chromato-graphy−photodiode array detection−heated electrospray ionization mass spectrometry. In the non-target, effect-directed screening via 16 different on-surface assays, we tentatively assigned more than 60 important bioactive compounds in the studied botanicals. These were antibacterials, estrogens, antiestrogens, androgens, and antiandrogens, as well as acetylcholinesterase, butyrylcholinesterase, α-amylase, α-glucosidase, β-glucosidase, β-glucuronidase, and tyrosinase inhibitors, which were on-surface heart-cut eluted from the bioautogram or enzyme inhibition autogram to the next dimension for further targeted characterization. This biological-physicochemical hyphenation is able to detect and control active mechanisms of traditional medicines or botanicals as well as the essentials of plant-based food. The array of 1,292 profiles (68 samples × 19 detections) showed the versatile bioactivity potential of natural food. It reveals how efficiently and powerful our natural food contributes to our homeostasis.


INTRODUCTION
Herbs and spices are widely used for nutrition, flavoring, cosmetics, dyeing, or fragrances (Guldiken et al., 2018). They are also applied in medicine due to their known beneficial effects on human health (Yuan et al., 2016;Caesar et al., 2019), inspired by traditional healers who have used botanical extracts since ancient times (Belwal et al., 2018b). The knowledge of biologically active plants, their harvesting, production, preparation, and administration has been passed down through thousands of years of traditional medicine (Yuan et al., 2016). Particularly phenols were reported to have antibacterial, antiviral, and antioxidant effects, as well as the ability to modulate enzyme activity and transduction pathways (Krüger et al., 2017;Tresserra-Rimbau et al., 2018). Some studies have quantified the total amount of healthful constituents in herbal extracts and calculated the recommended intake of antioxidants from culinary herbs (Halvorsen et al., 2002;Wojdylo et al., 2007). However, their multifactorial relevance in homeostasis is underexplored. It is evident that the use of the whole natural plant extract is more powerful for homeostasis due to the versatility of the gentle mechanisms of active compounds than the use of isolated compounds (Morlock and Heil, 2020).
In a typical screening for potential drug candidates, plant extracts are currently freed from assay-interfering tannins by solid-phase extraction, separated with an HPLC gradient (42 min/sample including equilibration), and collected in fractions, which are screened for bioactivity in a microtiter plate assay (Kongstad et al., 2015). Therefore, bioactivity can only be assigned to a fraction containing several analytes via a costly and time-consuming workflow, which subsequently requires analytical separation and testing of each peak to assign the individual bioactive compounds (Caesar et al., 2019). In routine, there has only been a little progress in nontarget screening of food for bioactive compounds at an affordable price. Most methods deal with illicit additions, organic contaminants , adulterated foods (Díaz et al., 2012), and migrants from packaging (Rusko et al., 2020;Su et al., 2020). Also, generic chromatography-based high-resolution mass spectrometric methods were examined to cover as many substances as possible within a single analysis (Díaz et al., 2012). However, one drawback is the high load of interfering matrix caused by the diversity and abundance of substances in such natural products as spices and herbs (Caesar et al., 2019;Morlock and Heil, 2020). Elaborate sample preparation (which is selective and error-prone) would otherwise limit the validity and significance of the results. The state of the art is setting an intensity threshold and focusing on highly abundant signals (Wu et al., 2016). But even the smallest signal can have an important biological effect. Ignoring minor signals from the set instrumental threshold will produce grossly negligent results. Moreover, compounds may not ionize well or at all with standard settings of mass spectrometric recording. That is why routine analysis of natural extracts is still tailored and limited to marker compounds. However, the important active portion of natural food needs to be under (analytical) control, which is presently not the case.
To overcome these limitations and expand the analytical toolbox, a high-throughput eight-dimensional (8D) hyphenation was recently developed, and its proof of principle was shown for cinnamon samples detected with an antibacterial bioassay (Schreiner and Morlock, 2021). It demonstrated the information gained by combining effect-directed assays (EDA) with normalphase high-performance thin-layer chromatography including multi-imaging by ultraviolet, visible, and fluorescence light detection (NP-HPTLC−UV/Vis/FLD) (Morlock, 2021). Heartcut elution and transfer of the bioactive compound zone to an orthogonal reversed-phase high-performance liquid chromatography (RP-HPLC) system was exploited to separate potentially coeluting bioactive substances. The subsequent photodiode array detection (DAD) and heated electrospray ionization mass spectrometry (HESI-MS) were used for additional straightforward characterization of the bioactive substances.
The advantage of NP-HPTLC−UV/Vis/ FLD−EDA−heart-cut RP-HPLC−DAD−HESI-MS is that it prioritizes and reduces the thousands of compounds in such natural samples to the most important bioactive compounds. As the previously developed hyphenation was only shown for cinnamon and one antibacterial bioassay, this study intended to examine the influence of 68 different plant matrices and 16 different assays on the robustness of the new 8D hyphenation. It was of interest to prove its universal validity and significance, to figure out potential limitations, and to verify its suitability as generic activity screening. Such straightforward effect-directed profiling could be applied to reveal, understand, and control the mode of action of traditional medicines, botanicals, and plantbased food.

HPTLC-EDA
For bioprofiling, 16 silica gel 60 F 254 MS-grade chromatograms were prepared. The buffer and assay solutions were piezoelectrically sprayed (Derivatizer, CAMAG) if not stated otherwise. To remove acidic traces left on the planar chromatogram (which can interfere with pH-sensitive bioassays), the chromatogram was neutralized with 1.5 ml phosphate buffer (80 mg/ml Na 2 HPO 4 , pH 7.5 adjusted with NaOH; yellow/green nozzle, level 6) for enzymatic and bacterial assays, 1.4 ml citrate buffer (6 mg/ml citric acid monohydrate, 10 mg/ ml Na 2 HPO 4 , adjusted to pH 12 with NaOH, yellow ultra-nozzle, level 2) (Klingelhöfer et al., 2020) for the hormonal-effective bioassays, 1.25 ml sodium bicarbonate buffer (2.5%, yellow nozzle, level 3) for Frontiers in Pharmacology | www.frontiersin.org December 2021 | Volume 12 | Article 755941 α-amylase bioassay or twice with 2.8 ml sodium bicarbonate buffer for SOS-Umu-C bioassay (neutralization procedure for SOS-Umu-C bioassay was investigated during this study, Supplementary Figure S1). The moist chromatogram was dried as mentioned in 2.3. A positive control was applied at three different concentrations at the top plate edge to verify the proper bioassay performance. The assay solutions/suspensions were applied as follows. For incubation, the plates were horizontally placed in a moistened polypropylene KIS box (26.5 cm × 16 cm × 10 cm, ABM, Wolframs-Eschenbach, Germany) pre-saturated with 30 ml water at 37°C (30°C for hormonal-effective bioassays) for 30 min. The procedure was documented at FLD 366 nm and white light illumination in transmission, reflection, and reflection/transmission mode.

Bacillus subtilis Bioassay
For the Gram-positive B. subtilis inhibition bioassay, 80 µl of stock solution was suspended in 20 ml Müller-Hinton Broth and incubated overnight at 37°C. Before usage, the cell number was determined using a spectrophotometer (M501, Camspec, Garforth, United Kingdom) at 600 nm. At an optical density (OD 600 ) between 0.8 and 1.1, the culture was ready to use for EDA. An aliquot of the bacteria suspension (2 ml) was sprayed on the planar chromatogram (red nozzle, level 6) (Morlock et al., 2021b). The plate was incubated at 37°C for 2 h. As substrate solution (2 mg/ml), MTT was freshly prepared in phosphatebuffered saline. After the application of 250 µl substrate solution (blue nozzle, level 6), the plate was incubated again for 30 min at 37°C. Inhibitory zones appeared colorless (white) on a formazanpurple background. The positive control was tetracycline (10 μg/ml in ethanol, 0.4, 0.8, and 1.2 μl/band).

Aliivibrio fischeri Bioassay
The bioluminescent marine Gram-negative bacteria A. fischeri were cultured according to DIN EN ISO 11348-1, Section 5 (European Committee for Standardization, 2009). Therefore, 200 µl of cryostock were suspended in 20 ml medium. The cultivation was performed overnight (18-24 h) in a 100 ml Erlenmeyer flask at room temperature by shaking at 75 rpm.
Once the culture showed brilliant blue fluorescence by shaking in the dark, it was ready for use. An aliquot of the bacteria suspension (3 ml) was sprayed on the plate (blue nozzle, level 6) and directly recorded (BioLuminizer 2, CAMAG) (Morlock et al., 2021a;Morlock et al., 2021b). The native bioluminescence (depicted as a greyscale image) was documented in ten images at time intervals of 3 min. Exposure time was set to 120 s. Antibacterial components were detected as dark zones, whereas metabolism-enhancing substances appeared as bright zones on the bioluminescent background. The positive control was caffeine (1 mg/ml in methanol, 0.5, 1.5, and 3 µl/band).

Tyrosinase Inhibition Assay
The tyrosinase inhibitor potential was investigated according to an improved (Morlock et al., 2021b) workflow (Taibon et al., 2015). To prepare the substrate solution, 45 mg levodopa, 25 mg CHAPS, and 75 mg PEG 8000 were dissolved in 10 ml of phosphate buffer (1.4 mg/ml K 2 HPO 4 , 1.68 mg/ml Na 2 HPO 4 , pH 6.8) and stored at 4°C until use. The levodopa substrate solution was sprayed onto the chromatogram (1 ml, blue nozzle, level 6) and subsequently dried for 2 min in a stream of cold air. Then, 1 ml of enzyme solution (400 U/ml in phosphate buffer) was sprayed onto the plate (blue nozzle, level 6), followed by incubation at room temperature for 20 min. After incubation, the plate was immediately dried and documented. Tyrosinase Frontiers in Pharmacology | www.frontiersin.org December 2021 | Volume 12 | Article 755941 inhibition activity was apparent as colorless (white) zones on a greyish-brown background. The positive control was kojic acid (0.1 mg/ml in ethanol, 1, 3, and 6 μl/band).

SOS-Umu-C Bioassay
The planar SOS-Umu-C bioassay was run on HPTLC silica gel plates without a fluorescence indicator. After development, the plates were additionally scanned at 546/>580 nm using the TLC Scanner 3 (CAMAG). Salmonella typhimurium cells (50 µl cryostock) were suspended in 35 ml Lysogeny broth (20 mg/ml with 1 mg/ml D-(+)-glucose and 106 mg/L ampicillin sodium salt) and incubated overnight at 75 rpm and 37°C for 16 h. Before use, the cells were centrifuged (3,000 × g, 10 min). The pelleted cells were re-suspended in fresh medium to obtain the required OD 660 of 0.2 (Meyer et al., 2020). The chromatogram was sprayed with Salmonella suspension (2.8 ml, yellow nozzle, level 3) and incubated at 37°C for 3 h. Substrate solution [15 µl resorufin-β-D-galactopyranoside solution (20 mg/ml in DMSO) in 2.1 ml phosphate buffer and 0.9 ml glycerol] was sprayed onto the plate (2.5 ml, red nozzle, level 6). Incubation followed at 37°C for 1 h. The plates were documented at white light illumination and 366 nm. The generated resorufin fluorescence was measured at 546/>580 nm. Genotoxic substances were detectable either as pink zones on the colorless background at white light illumination or as pink fluorescent zones on a brown-reddish background at 254 nm or 366 nm. The positive control was 4nitroquinoline-1-oxide (1 μg/ml in methanol, 1 µl/spot).

α-Amylase Inhibition Assay
The latest α-amylase inhibition method, which used immersion of the plate into enzyme and substrate solutions (Agatonovic-Kustrin and Morton, 2017; Agatonovic-Kustrin et al., 2019), was adjusted and transferred to a piezoelectric spraying procedure, in which the enzyme solution (62.5 U/mL in sodium acetate buffer) was sprayed onto the chromatogram (1 ml, red nozzle, level 5), followed by 30 min incubation at 37°C. As substrate 2%-soluble starch solution was sprayed onto the wet plate (0.5 ml, red nozzle, level 5). After 20 min incubation at 37°C, Gram's iodine solution was sprayed (250 μl, yellow nozzle, level 5) for visualization. The α-amylase inhibition activity was observed as violet zones on a colorless background. The positive control was acarbose (0.1 mg/ml in methanol, 0.3, 0.6, and 0.9 µl/band).

Instrumental Setup of the 8D-Hyphenation
The multipotent bioactive zones were further characterized with RP-HPLC-DAD-HESI-MS directly after EDA. The UPLC system (Acquity H Class, Waters, Eschborn, Germany) was equipped with the quaternary solvent manager, solvent degasser, sample manager, column oven, photodiode array detector (DAD), and HESI-MS (single quadrupole QDa, Waters). The bioactive target zone was heart-cut eluted with an oval elution head (4 mm × 2 mm) of the TLC-MS Interface 2 (CAMAG) with 90% aqueous methanol. A standalone pump supplied the solvent (515 HPLC pump, Waters). Analytes were transferred through a biocompatible inline filter (IDEX Health and Science, Oak Harbor, WA, United States) containing a PEEK frit (0.5 µm, Techlab, Brunswick, Germany) to an online desalting RP precolumn/defender guard (Accucore RP-MS, 10 mm × 2.1 mm, 2.6 µm, Thermo Scientific, Bellefonte, PA, United States). The online desalting device was installed onto a two-position switching valve (MXT-Series PD715-000, Rheodyne IDEX Health and Science) and served as an analyte trap while discarding the bioassay salts as waste. By switching, controlled via remote control and Rheodyne TitanMX software, the analytes were transferred to the main RP column (Accucore RP-MS 100 mm × 2.1 mm, 2.6 µm, Thermo Scientific) and separated orthogonally. The column was thermostated at 40°C. The 13 min HPLC gradient consisted of (A) 2.5 mM ammonium acetate (pH 4.5 adjusted with acetic acid) and (B) methanol. Starting Frontiers in Pharmacology | www.frontiersin.org December 2021 | Volume 12 | Article 755941 conditions were 98% A at a flow rate of 0.6 ml/min for the first 2 min. The methanolic portion increased linearly to 20% within the following 2.5 min. At 8 min, a ratio of 10/90% A/B was reached and held for the next 2 min; then it fell to 98% A within 0.1 min, followed by 3 min equilibration time. Detection parameters were set to a wavelength scan from 190 to 400 nm for DAD. The MS was operated in polarity-switching mode, while the ESI probe was heated to 600°C and ESI source to 120°C. The sampling frequency was set to 5 Hz and cone voltage to ±10 V in both ionization modes (Schreiner and Morlock, 2021). The MassLynx V4.2 software (Waters) was used to evaluate and process the data.

Outline of the Study
A total of 68 very different powdered plant extracts (botanicals added to food products in food industry) and 16 different effectdirected assays were selected to investigate and prove the suitability of the biological-physicochemical 8D hyphenation for generic screening ( Figure 1). Among the plants (Table 1) were such ones that are commonly used as culinary spices and herbs or in traditional medicine. Their diverse and varying compositions represent different matrix loads for the analytical system. Moreover, the nine biological and seven biochemical assay media differed over a wide range in salt and nutrient composition. This represents the diversity of possible compositions of a bioactive zone (to be heart-cut and transferred to the next dimension) and was therefore considered a good worst-case scenario to test whether the developed generic hyphenation method is suitable for routine analyses. First, the bioactivity screening was evaluated per assay ( Figures 2-6), whereby some botanicals were mentioned repeatedly, i.e. galangal (no. 13) yerba mate (no. 37), orange peel (no. 41), licorice (no. 55), and Siberian ginseng (no. 56). Then, these botanicals were subjected to heart-cut RP-HPLC−DAD−HESI-MS analysis (Figures 7-11). All botanicals were extracted and applied analogously. Thus, the effect profiles of each assay could directly be compared by their response pattern. The most effective and important botanicals were highlighted at a glance in side-by-side comparison. The band broadening (diffusion) depended on the assay incubation time.
The most important bioactive compounds discovered were tentatively assigned based on information obtained about spectral (UV/Vis/FLD), polarity (hR F values with a deviation of ±1), and molecular (mass signal) properties. Since there was no access to a high-resolution mass spectrometry system, the assignments were verified by comparing with reference standards. Using the Gram-negative Aliivibrio fischeri and Gram-positive Bacillus subtilis bioassays, natural antibacterial compounds were detected that can subtly fight infections and contribute to longer shelf life and better preservation of products. Natural AChE and BChE inhibitors, which can provide symptomatic benefits for the cognitive decline of Alzheimer's patients (Tundis et al., 2016), were revealed by the respective planar enzyme inhibition assays. The tyrosinase inhibition assay was used to screen for plant-based skin whiteners or inhibitors of the enzymatic food browning. The β-glucuronidase inhibition assay was used to detect compounds that prevent the gut-bacterial reversion of detoxification via glucuronidation. Additional α-amylase, αand β-glucosidase inhibition assays were employed to determine natural compounds with benefits for diabetes patients. Phytoandrogens and phytoestrogens were detected via the human estrogen receptor α (hERα) or hERβ or in combination with the S9 enzyme mixture simulating liver metabolization, and respective antagonists were investigated using recombinant yeast cells equipped with the human estrogen/androgen receptor. Genotoxins were detected with a recombinant Salmonella typhimurium strain equipped with the SOS-Umu-C repair mechanism. Hence, the spectrum of effects in the investigated botanicals may be linked to the mitigation of bacterial infects, the improvement of cholinergic transmission for Alzheimer's patients, the decrease of blood glucose levels in diabetics, the reduction of skin abnormalities, and to the balance of the steroid hormonal system, among other disorders.

Compounds Inhibiting Bacteria
In traditional medicines, plant-based extracts are used to assist in the treatment of bacterial infections (Brantner and Grein, 1994;Palombo and Semple, 2001). Antibacterial activities can be so effective that plant extracts are also used as preservatives in food products. For example, the ingredients carnosol and carnosic acid of rosemary extract are marketed as preservative E 392. Rosemary (no. 46) is also screened here, using non-pathogenic bacterial representatives which are easier to handle in the laboratory. The Gram-positive B. subtilis bioassay is based on an oxidoreductase enzyme reaction. Intact enzymes of viable B. subtilis cells reduce the tetrazolium salt MTT to the insoluble purple formazan (Marston, 2011). Cell death is visualized as colorless zones indicating antibacterial compounds. Most antibacterials detected were located at hR F values ≥ 90 (Figures 2-4C). In eucalyptus (no. 9), marjoram (no. 35), yerba mate green (no. 37), Siberian ginseng (no. 56), thyme (no. 57), hawthorn leaves (nos. 63 and 64), and cinnamon bark (no. 66) additional antibacterials were detected in the lower hR F range. Essential oils from herbs such as oregano and thyme are known for their antimicrobial activity, especially against Gram-positive bacteria (Soković et al., 2010).
The Gram-negative A. fischeri bacteria are able to convert metabolic energy into turquoise bioluminescence via luciferase. A change in bioluminescence intensity is correlated with substances enhancing or reducing the cell metabolism. Such effects are visualized as lightened or dark zones on the bioluminescent background of the bioautogram (depicted as greyscale image). Almost all botanical extracts showed antimicrobial activity against A. fischeri (Figures 2-4D). Intense antimicrobial zones were detected in yerba mate green (no. 37), passionflower (no. 43), peppermint (no. 44), rooibos (no. 45), licorice (no. 55), Siberian ginseng (no. 56), and cinnamon bark (no. 66). Most samples showed at least one dark antimicrobial zone, while more than half of all samples had two or more. The more universally and sensitively detecting A. fischeri bioassay proved to be a good starting assay to investigate complex mixtures.

Compounds Inhibiting AChE and BChE
In traditional or ayurvedic medicine, plants and their phytoconstituents are used to assist in the treatment of Alzheimer´s disease (Azadniya et al., 2021). The pathophysiology of Alzheimer's disease is often associated with cholinergic system dysfunction. Therefore, many synthetic drugs target the inhibition of cholinesterases. Both AChE and BChE catalyze the hydrolysis of the neurotransmitter acetylcholine into acetic acid and choline. While AChE is highly selective for acetylcholine, BChE can also convert other substrates, e.g., butyrylcholine, succinylcholine, or organophosphates. Besides synthetic drugs, also natural compounds are able to inhibit this enzyme mechanism. Particularly polyphenols interact with amino acid residues of the active side of the enzymes terminating the splitting from acetylcholine into acetic acid and choline and thus maintaining colinergic neurotransmission and improve cognition of Alzheimer's patients (Jabir et al., 2018). The enzyme-inhibiting potential is revealed as colorless zones on a purple background. Acerola fruit (no. 1) showed a remarkably strong inhibition zone at hR F 41 ( Figures 2E,F). While roots and branches from the acerola tree are known to have anti-cholinesterase activity through the norfriedelins A-C (Liu et al., 2013), only cytotoxic, anti-HIV, antioxidant, antihyperglycemic, skin whitening, and antimicrobial activities are described for extracts from the fruits (Motohashi et al., 2004;Belwal et al., 2018a). This zone showed not only a strong response in the AChE and BChE assays, but also in most other assays ( Figure 2

Compounds Inhibiting β-Glucuronidase
The detoxification mechanism via glucuronidation can be reversed with β-glucuronidase from opportunistic Enterobacteriaceae such as Escherichia coli, resulting in gastrointestinal malfunction. This can be prevented by inhibiting the microbial β-glucuronidase, where such inhibitors address the extra loop in the bacterial enzyme (in contrast to the mammalian one) (Mahran et al., 2020). The enzyme inhibitors do not cleave the chromogenic substrate and are therefore detected   (Figures 2-4G). If the application volume is reduced by a factor of 4 for these botanicals, the individual inhibitors become evident (Supplementary Figure S2). All botanical extracts showed β-glucuronidase inhibitory potential at least in the solvent front (hR F 99). For such screening results, repetition using a mobile phase of reduced solvent strength is recommended in order to better differentiate the individual inhibitors (Supplementary Figure S3C,D). Some isolated flavonoid standards, i.e., isorhamnetin, kaempferol, liquiritigenin, daidzein, etc., already proved to be active against β-glucuronidase (Sun et al., 2020). Their activity and also that of additional flavonoids have been confirmed by our study directly in herbs and spices.

Compounds Inhibiting Tyrosinase
Polyphenoloxidases are responsible for the browning of cut or injured fruits or plant tissues. In mammalian cells, the corresponding tyrosinase controls melanogenesis by catalyzing the hydroxylation of phenols with subsequent oxidation to quinones. An overproduction of melanin induces pigmentary abnormality, freckles, or age spots. Preferably, the cosmetics industry is interested in naturally derived tyrosinase inhibitors (Taibon et al., 2015). In this context, ethnobotanicals are brought into focus. In South Africa, herbal extracts are traditionally used as skin care products to treat burns, abcesses, wounds and acne (Lall and Kishore, 2014). Chinese herbal medicines with antityrosinase activity are traditionally used as folk skin whiteners. Among the studied botanicals, Ginkgo biloba (no. 14), Panax ginseng (no. 15), and Zingiber officinale (no. 25) were reported to inhibit mushroom tyrosinase (Ye et al., 2010;Hu et al., 2020). Our screening results showed that ginseng (no. 15) and ginger (no. 25) played a minor role in tyrosinase inhibition compared to other botanicals. However, differences in effects can be caused for example by plant subspecies, climate, soil, environmental and agricultural conditions as well as extraction mode. In the planar assay, tyrosinase inhibitors are shown as colorless zones on a greyish-brown background. Many botanical extracts showed multiple tyrosinase inhibitors (Figures 2-4H). While acerola (no. 1) (Belwal et al., 2018a), ginkgo (no. 14) (Shu et al., 2020), licorice (no. 55) , and hawthorn (nos. 62-64) (Rocchetti et al., 2020) are known for their anti-tyrosinase activity, the screening results proved similarly potent tyrosinase inhibitors in artichoke (no. 4), plantain (no. 35), yerba mate green (no. 37), rosemary (no. 46), and yarrow (no. 50).

Compounds Inhibiting αand β-Glucosidase
The enzymes αand β-glucosidase hydrolyze the saccharide dimers and oligomers, as well as glucosides dependent on the anomeric glycosidic bond, into resorbable monomers such as glucose and into aglycones. In the treatment of hyperglycemic blood levels of type 2 diabetes patients, enzyme inhibitors are of therapeutic interest by reducing postprandial glucose uptake (Simões-Pires et al., 2009;Turkiewicz et al., 2019). Screening results showed several αand β-glucosidase inhibitors as a specific pattern (at hR F 42, 48, 62, and 80) for artichoke (no. 4) in both assays ( Figures 2I,J).  All of them showed both αand β-glucosidase inhibitory activities. Worldwide more than 1,000 herbal remedies were traditionally deployed for the maintenance and treatment of high blood glucose levels and thus diabetes. Both, the ethnobotanicals used and their mode of application (as tincture or extract, orally or as infusion) differ between local communities (Cock et al., 2021). Since western medical treatment methods for type 2 diabetes focus on hypoglycemic drugs, such as insulin, ethnopharmacological remedies are considered to be safe and to have less toxic side effects. According to the theory of traditional Chinese medicine, flavonoids (Bai et al., 2019) and polyphenols (Umeno et al., 2016) are attributed to have antidiabetic effects via several mechanisms. Glycyrrhizic acid from Glycyrrhiza glabra (no. 55), apigenin and its derivates as well as quercetin, found in many botanicals ( Table 2), are known to target α-glucosidase (Bai et al., 2019).

Compounds With Agonistic/Antagonistic Hormonal Effects
Disrupting endocrine signaling pathways can have a severe impact on hormonal balance and cause feminization/ masculinization, infertility, acne, and menstrual cycle disorders. Also phytochemicals present in food (Morlock and Klingelhöfer, 2014) or commodities in daily use (Klingelhöfer et al., 2020) can affect the human hormone system. HPTLC hyphenated to the planar yeast androgen/estrogen screens (pYAS/pYES) and their antagonistic versions (pYAAS/pYAES) showed positive reactions to a limited extent ( Figures  2-4K-N). In the pYAS bioassay, a few 4-methylumbelliferoneblue and thus androgenic responses were detected in acerola (no.  (Mbachu et al., 2020), the pYES bioassay was also performed via the hERβ for the first time. The assay was analogously run for both receptors, but no remarkable difference was observed in the results obtained by both ( Figure 5A versus Figure 5B). The hERα seemed to be less selective. However, a significant difference in the estrogenic pattern was observed after metabolization with the S9 mixture ( Figure 5C). The use of genetically modified yeast cells containing the hERβ for the pYES bioassay and the simulated metabolization via the S9 liver enzyme system were reported here for the first time.
In the respective antagonistic assays, only a few zones were detected. For acerola (no. 1, hR F 99), fenugreek (no. 6, hR F 99), FIGURE 6 | First report of synergy profiles, exemplarily shown for the NP-HPTLC-pYAES bioassay. Separation of the applied botanicals (nos. 7-19, assigned in Table 1; 4 µl/band each; solvent blank B for comparison) on HPTLC plate silica gel 60 F 254 MS-grade with ethyl acetate-toluene-formic acid-water (16:4:3:2, V/V/V/V) up to 70 mm; each separated track was oversprayed by an 1 mm × 70 mm area of 17-β-estradiol (10 ng/band) and after bioassay application detected at FLD 366 nm. Fortified estrogenic fluorescence was interpreted as synergistic effects, which were not observed in solvent blank B and no. 7 (only a slight anti-estrogenic effect in the lower hR F range), partially observed as cotton-swab-shaped 17-β-estradiol band (nos. 11, 15 and 18) Figure S4) to exclude any false-positive response as observed for the physicochemical fluorescence reduction by pigments or dyes (Klingelhöfer et al., 2020). Seven botanical samples (nos. 6, 15, 16, 31, 41, 55, and 68) showed truly antiandrogenic activities. The same verification test was    Figure S5). In the antiestrogenic assay, synergistic effects were evident. The overlapping 17β-estradiol area (10 pg/70 mm area) is partially enhanced on sample tracks, apparent as a cotton swab shape if compared to the solvent blank (Figures 2-4N). This observation revealed synergistic effects between distinct botanical ingredients in most samples (except for nos. 7, 29, and 31) and 17-β-estradiol, resulting in fortified estrogenic activities ( Figure 6). This overlapped experimental setup can be transferred to all other assays to identify synergy. Such a  Figure 3 (A), followed by heart-cut elution of the bioactive zone at hR F 40 (marked*) directly out of the BChE autogram and online desalting to obtain the RP-HPLC−DAD−HESI-MS chromatograms (B) and respective UV and mass spectra for the peaks at RT 2.78, 3.70, and 4.44 min, which were assigned to the isomeric structures of caffeoylquinic acid with the most predominant depicted, (C) as all three peaks provided the same signals (D).

Compounds With Genotoxic Effects
In recent decades, there has been a steady trend away from industrial medical care towards phytotherapy based on medicinal herbs, which have been used in traditional medicine for years. The common belief that phytochemicals are gentler than synthetic medicines may prove to be a fallacy, as most toxic compounds originate in nature (Efferth and Kaina, 2011). The SOS-Umu-C assay is a reporter gene assay indicating genotoxicity based on a genetically modified test organism Salmonella typhimurium TA1535 [pSK1002]. Under genotoxic stress, the SOS-DNA repair mechanism is induced. The SOS-Umu response activates the lacZ gene, encoding for β-galactosidase, which enables substrate cleavage into detectable products (Meyer et al., 2020). As the analyzed botanicals are commonly used as spices, in herbal medicine, or as tea, no genotoxic substances were expected. Nevertheless, pyrrolizidine alkaloids are known for genotoxic properties and their occurrence in herbal formulations (Habs et al., 2017), e.g., only a few micrograms per liter of tea (6 μg/L) (Mulder et al., 2015). The European Food and Safety Authority calculated a margin of exposure to 237 g/kg body weight per day for pyrrolizidine alkaloids and their N-oxides (Knutsen et al., 2017;Kaltner et al., 2020). No genotoxicity was detected as pink fluorescent zone for the 68 botanical extracts applied at 400 µg/band (Figures 2-4O). This bioassay was repeated to prove for the absence of genotoxic effects at a 2.5-fold (Supplementary Figure S6) and 12.5-fold higher amount applied, which latter at 5 mg/band was closest to overloading the chromatographic system (Supplementary Figure S7).

Compounds Inhibiting α-Amylase
In the context of hypoglycemic drugs from nature, not only the mentioned αand β-glucosidase inhibitors, but also α-amylase inhibitors play a role in ethnopharmacological remedies with less toxic side effects. Hyperglycemic blood levels could be reduced by inhibiting α-amylases of the saliva and pancreas. The inhibition reduces the cleavage of starch into oligo-and disaccharides, and so the release of glucose molecules absorbed postprandially into the blood. Plant-derived flavones as luteolin present in celery, parsley, broccoli, carrot, peppers, cabbage, and apple peel can inhibit α-amylase (Bai et al., 2019). HPTLC coupled to α-amylase assay (Figures 2-4P) showed only a few inhibiting signals in white horehound (no. 2, hR F 43, 83, 91, and 93), artichoke (no. 4, hR F 94), fenugreek (no. 6, hR F 81), ginkgo (no. 14, hR F 46), licorice (no. 55, hR F 59 and 68), and lemon peel (no. 67, hR F 93 and 95). An α-amylase inhibitory activity has already been described for artichoke extracts from bracts (Turkiewicz et al., 2019), but it was not assigned to any single compound. In Ginkgo biloba extracts the compound sciadopitysin (C 33 H 24 O 10 ) was found to potentially inhibit α-amylase (Petersen et al., 2019).

Assigning Bioactivity to Single Compounds
In previous work, the samples were applied twice (as two sets). After plate cut, the assay was performed on one plate part, and the positions of bioactive zones were analogously marked on the other plate part for zone elution and transfer into the MS (Krüger et al., 2017). The challenge of this parallel handling was the accurate positioning of the elution head on the zone. Instead, the elution directly from the bioassay plate would avoid any possible mismatch. In addition, the orthogonal HPLC separation would solve coelution. Hence, zones of interest were heart-cut eluted directly out of the bioassay plate and transferred through an online desalting device to RP-HPLC−DAD−HESI-MS. This reduced the interfering salts and nutrients from the bioassay, separated possibly coeluting substances via the orthogonal chromatographic system, and added value via spectral and mass spectrometric data (Schreiner and Morlock, 2021). Through this straightforward comprehensive workflow, the 60 most bioactive compounds in the botanicals were assigned ( Table 2). Several plant extracts out of the 68 botanicals contained multipotent compounds which were active in different assays. Among striking botanicals, such as galangal (no. 13), ginkgo (no. 14), yerba mate green (no. 37), orange peel (no. 41), licorice (no. 55), or Siberian ginseng (no. 56), four botanicals were exemplarily highlighted to demonstrate the targeted assignment of bioactive compounds (orange peel, Figure 7; licorice, Figure 8; galangal, Figure 9; yerba mate green, Figure 10). By transferring residual bioassay salts, different ion species were formed for a molecule, which was found to be helpful as it confirmed the assignment made several times. The adduct [M+62] − has not been described in HPTLC-MS literature so far, but appeared frequently during this study (Figures 7C,D,Figures 8C,D,Figures 11A,C). It was proposed to be the nitrate adduct. Nitrate is taken up by the root and transported via the xylem to leaves, shoots, and grains. If too much nitrate is available in the short term, which is the case as it is discussed as a global environmental challenge (Zhang et al., 2021), it passes through the cytoplasm into the vacuoles to be stored and thus can be found in botanicals (Dechorgnat et al., 2011). Consequently, it may also appear as a pronounced adduct in the mass spectrum.

Hormonal Antagonists in Orange Peel (No. 41)
For orange peel, manifold positive responses across the different assays were observed at hR F 34 ( Figure 7A, marked*). Using the pYEAS bioautogram, this zone was heart-cut eluted directly to RP-HPLC−DAD−HESI-MS. In the DAD chromatogram, three compound peaks were evident in the range of RT 6.87-7.07 min, apart from the 4-methylumbelliferone background signal at RT 5.71 min ( Figure 7B). The corresponding mass spectral data led to the tentative identification of naringin ( Figure 7C), hesperidin ( Figure 7D), and meranzin ( Figure 7E, coeluting with an unknown marked blue) and is discussed as follows. The total ion current (TIC) peak at RT 6 Figure 7D). The absorption spectrum revealed a maximum wavelength at 284 nm. Based on this data, hesperidin was assumed. The third peak at RT 7.06 min had a major response in TIC-HESI + and a minor response in TIC-HESI − and DAD chromatograms. The UV absorbance spectrum and respective mass spectra suggested two different constituents.  (Fan et al., 2012). Other UV maxima at 223 and 288 nm were possibly induced by the unknown.
Citrus sinensis is rich in beneficial secondary metabolites and therefore traditionally used in the treatment of gastrointestinal malfunction, diseases of the upper respiratory tracts, or menstrual disorders (Favela-Hernández et al., 2016). Both naringin and hesperidin have been shown to bind to the antagonist pocket (3ERT) of hERα, and thereby cause an antiestrogenic effect (Puranik et al., 2019). This effect was verified for both via cochromatography of standards and samples (Supplementary Figure S8E, Supplementary Table S1), whereby an estrogenic activity was not observed for the applied amount (4 µg/band). Furthermore, naringin was reported to slightly bind to the androgen receptor (Fang et al., 2003). The pronounced antiandrogenic effects observed in the pYAAS bioautogram confirmed this ( Figure 3M, no. 41). Tyrosinase activity attributed to naringin (Itoh et al., 2009) and hesperidin (Zhang et al., 2007) has already been demonstrated. The only bioactivity reported for meranzin was no effect (Smyth et al., 2009) or a minor (Rosselli et al., 2007) antibacterial effect against B. subtilis. After NP-HPTLC-FLD comparison to a standard, meranzin was located at hR F 99 (Supplementary Figure S8G), where antibacterial activity against B. subtilis and A. fischeri was detected in orange peel (no. 41). Hence, the assumption that meranzin was co-eluting with hesperidin and naringin at hR F 34 was discarded. Licorice (No. 55) and Galangal (No. 13) In licorice (no. 55) a few bioactive analytes were found in the zone at hR F 31 with antibacterial, tyrosinase, and β-glucuronidase inhibitory, antidiabetic, and endocrine-antagonistic properties ( Figure 8A, marked*). These results illustrate the diverse pharmacological activities of this root, which has long been used in traditional medicine as remedy to treat gastrointestinal problems (β-glucuronidase inhibition) and respiratory infections (antibacterial activity). Moreover, Glycyrrhiza glabra extracts were subject of many pharmacological studies showing neuroprotective, antimicrobial, estrogenic and skin-whitening activity (Pastorino et al., 2018). The second chromatography produced five individual signals with pure mass spectra ( Figures 8B-E). The mass spectral data extracted from the peaks at RT 6.52/6.63 min (colored blue) were identical, indicating a single analyte in different configurations ( Figure 8C) +H] + could be assigned to fragments with a loss of carbohydrates. Additionally, the experimental UV absorbance spectra were consistent with the ones of liquiritin apioside found in literature (Wong et al., 2018). The isomeric isoliquiritin apioside probably caused the second peak at RT 6.63 min. ESI + mass signals at RT 7.10 min were m/z 563 [M + H] + , 585 [M + Na] +, and 601 [M + K] + ( Figure 8D). (Iso)liquiritin apioside was found to inhibit capsaicin-induced cough, confirming its traditional use (Pastorino et al., 2018). Lacking a π-electron system and conjugated double bonds, the UV absorbance spectrum showed background absorbance maxima at 251 nm. Glycyrrhizic acid was reported to have anti-inflammatory effects (Yu et al., 2015) similar to those of glucocorticoids (Pastorino et al., 2018), antitussic activity through increased tracheal mucus secretion (Sharma et al., 2018), and neuroprotective (Kao et al., 2009) properties.

Enzyme Inhibition and Endocrine Activity in
Galangal extract (no. 13) responded in almost all assays at hR F 99 ( Figure 9A, marked*). After separating this zone with RP-HPLC-DAD-HESI-MS, multiple signals were observed ( Figure 9B). The spectral details and tentative identifications are listed in the table below ( Figure 9C). Traditionally, Alpinia officinarum is used against cold, which antibacterial effects were confirmed by the Gram-negative A. fischeri and Gram-positive B. subtilis bioassays. Other traditional applications were described for gynecological disorders, diabetes treatments, and skin washing (Abubakar et al., 2018). All these bioactivities were verified by the respective assays. To assign the bioactivity to one single component out of the coeluting substances from HPTLC, fractionation after column separation is necessary. The fractions can be applied again on a new plate followed by EDA and MS characterization. Alternatively, the mobile phase for planar chromatography has to be optimized in order to separate the previously coeluting substances during HPTLC analysis.

Separating Multipotent Isomers in Yerba Mate Green (No. 37) via 8D-Hyphenation
Biologically active isomers were also separated and detected, shown for example in yerba mate green (no. 37). A multipotent bioactive compound zone (hR F 40) was observed in AChE, tyrosinase, and Frontiers in Pharmacology | www.frontiersin.org December 2021 | Volume 12 | Article 755941 α-/β-glucosidase inhibition autograms ( Figure 10A, marked*). This zone was separated into three distinct signals via RP-HPLC ( Figure 10B), all providing the same absorbance and mass spectra ( Figure 10D) and assigned to chlorogenic acid isomers, which differed only in quinic acid positioning ( Figure 10C).  (Huang et al., 2017), anti-inflammatory (Willems et al., 2016;Huang et al., 2017), and anti-HIV (Tamayose et al., 2019) properties, this screening revealed even more bioactive potential for these phenolics. For instance, the assigned bioactive compound zone also showed anti-Alzheimer, antidiabetic, and skin-whitening effects in this study. Co-chromatography against standards (Supplementary Figure  S8A,B and D, Supplementary Table S1) confirmed that these beneficial health effects of yerba mate green (no. 37, Figure 10) come from the mono-caffeoylquinic acids. The various effects of Ilex paraguariensis traditionally consumed as herbal beverage qualify this botancial for its new role as functional food (Cardozo Junior and Morand, 2016).

Universal Potential of 8D-Hyphenation
The characterization of additional multipotent bioactive zones, partially plant-specific, is shown in Figure 11. Artichoke (no. 4) showed bioactivity at hR F 42 in the tyrosinase and α-/β-glucosidase inhibition assays. Transferring this zone to RP-HPLC-DAD-HESI-MS provided five signals ( Table 2), one of which is specific for artichoke. At RT 6.30 min, cynarascoloside C was assumed ( Figure 11A). Spectral data showed no UV absorbance, referring to structural properties of cynarascoloside C which possess neither a π-electron system nor conjugated double bonds.  Figure 10). Nevertheless, Cynara scolymus has been used since the 4th century B.C. as medicinal product due to its health benefits and bioactive constituents, responsible for the hypoglycemic, anti-inflammatory, antimicrobial, and antioxidant properties (Turkiewicz et al., 2019). Eucalyptus (no. 9) is known for its beneficial health properties against infections of the upper respiratory tracts or as antiseptic and is therefore widely used in the pharmaceutical industry (Hasegawa et al., 2008;Ács et al., 2018). The species Eucalyptus is closely associated with herbal medicine and traditional health care in various human cultures. Its leaf extracts are administered to fight against cold and cough, bacterial infections, high blood glucose levels, and to boost the immune system and skin health (Salehi et al., 2019). Planar bioanalytical screening confirmed most of those traditional uses, showing many positive responses through the assays in a wide hR F range ( Figure 11B). Focusing on hR F 57, highly plant-specific eucaglobulin was identified with UV (λ max 222 and 261 nm) and mass spectral data (Boulekbache-Makhlouf et al., 2010). Both the positive ion species at m/z 521 [M + Na] + and 537 [M + K] + , and the deprotonated molecule [M−H] − confirmed this suspicion. Anti-melanogenesis activity was attributed to eucaglobulin (Hasegawa et al., 2008) and proved with a planar tyrosinase bioassay. Extracts of eucalyptus fruits were demonstrated to have antibacterial effects against B. subtilis (Boulekbache- Makhlouf et al., 2013), which so far have not been directly correlated to the monoterpene conjugate eucaglobulin. The anti-cholinesterase activity was only described for the whole methanolic extract of Eucalyptus globulus (Amat-ur-Rasool et al., 2020) but not directly correlated to eucaglobulin.
Extracts of Ginkgo biloba seeds and leaves were applied as herbal remedies all over the globe. Originating from traditional Chinese medicine, it is used to treat bacterial skin diseases, cognitive decline (Chassagne et al., 2019), and cardiovascular diseases (Shu et al., 2018). Affirming the ethnopharmacological usage, EDA of ginkgo leaf extract (no.14) demonstrated a variety of bioactivity at hR F 93, i.e., antibacterial and antidiabetic effects, as well as β-glucuronidase and tyrosinase inhibition, and antiandrogenic activity ( Figure 11C). RP-HPLC-DAD-HESI-MS analysis revealed the three ginkgolides A-C incorporated in this bioactive zone. Ginkgolide C eluted earlier from RP column (RT 6.42 min). Despite the second chromatography, ginkgolides A and B were not separated (both RT 6.86 min) and were thus detected in the same UV and mass spectra. Structurally they only differ in one additional hydroxy group of ginkgolide B. The high abundance of different ion species describing the two compounds is listed in Table 2 and displayed in Figure 11C. The wide bioactive spectrum of these two diterpenes was confirmed against standards (Supplementary Figure S8B, Supplementary Table S1).
As an alternative to coffee, the leaves of Ilex paraguariensis are widely consumed as beverage in Latin America. Due to its chemical composition, mainly alkaloids and polyphenols, yerba mate exhibits many bioactive effects, e.g., antibacterial, cardiovascular-protective, neuroprotective, and antidiabetic activities (Gan et al., 2018). In yerba mate green (no. 37), a multipotent bioactive analyte zone was detected at hR F 20 ( Figure 11D in ESI + at an RT of 6.84 min, the flavonoid rutin was assumed. Its antibacterial (Orhan et al., 2010) potential against B. subtilis, antityrosinase activity (Kishore et al., 2018), and glucosidase inhibition (Li et al., 2009) have already been demonstrated.
The assumption was verified against a standard (Supplementary Figure S8A and D, Supplementary Table S1).
Salvia rosmarinus, predominantly growing in Mediterranean regions, is independently used in Mexican and Spanish ethnopharmacology. As medicinal plant, it was used to fight bacterial skin diseases, colds, intestinal parasites, and headaches (Heinrich et al., 2006). A very broad spectrum of bioactivities was confirmed by the presented study. The comprehensive NP-HPTLC-EDA-heart-cut RP-HPLC-DAD-HESI-MS analysis of rosemary (no. 46) showed antibacterial activity in the A. fischeri bioassay and antidiabetic properties in the β-glucosidase assay at hR F 25. Several ions at m/z 387 [M−H] − , 406 [M + NH 4 ] + , 411 [M + Na] + , and 427 [M + K] + in positive and negative ion mode were correlated to the neutral molecular weight of 388 Da ( Figure 11E). The mass spectrometric data substantiate the suspicion that this bioactivity is related to medioresinol. In rosemary extracts, various bioactive compounds have been described. Carnosic acid and carnosol are known to inhibit the growth of human cancer cell lines and operate as anti-inflammatory agents (Bai et al., 2010;Wang et al., 2018) or antioxidants (Loussouarn et al., 2017). Rosmarinic acid and rosmanol were found to have antioxidant potential (Vallverdú-Queralt et al., 2014), but to our knowledge, there is no bioactivity reported for medioresinol.
Many more examples could be explained in detail. All have in common that the effect profiles we obtained explain why consuming green tea and using fresh herbs and spices as seasoning can reduce the risk of diabetes by inhibiting αand β-glucosidases, why healthy nutrition can protect the intestinal flora from severe impairment caused by β-glucuronidases from Enterobacteriaceae, why plant-derived cosmetics can reduce skin abnormalities via its tyrosinase-inhibiting potential, and why Alzheimer's disease can be prevented by daily intake of chlorogenic acid or rosmarinic acid from herbs such as artichoke, lemon balm, peppermint, thyme and rosemary, and so on. The wealth of effect information obtained inspires the mind and could fuel further studies. An enormous diversity in bioactivity was revealed in the effect-profiles of the 68 botanicals, contributing to human health by drug-like properties. It clearly shows the potential and spectrum of nature as basis for alternative medicines.

CONCLUSION
Since the major part, and especially, the important active part of natural food and traditional medicines is presently not under analytical control, a paradigm shift from quality control based on marker compounds to effect profiles is postulated for plant-based samples. Considering the global production chain, whose influences on the product cannot be controlled, at least entries or changes regarding the effect should be kept under control. The Chemical Abstracts database (www.cas.org) contains over 190 million chemicals, and thousands of compounds are added daily. There is of little help, if we can measure some thousands of them with ever lower limits of determination. By doing so, this does not come close to doing justice to the complexity of plant extracts, nor to the metabolic networking of natural processes, nor to the contamination-prone global production chain. The implementation of effect-directed profiles would subtantially improve quality control, ensure the expected activites and detect unexpected activities. Even small amounts of compounds can be highly active. Disruptive thinking is essential to better control our natural food and traditional medicines. Sophisticated instrumentation does not solve pressing challenges, but combining orthogonal areas does. The complexity of plant extracts requires modern non-target methods that combine chromatography with effect-directed assays to prioritize active compounds that show an effect and thus require utmost attention. This combination is indispensable for routine quality control to prioritize substances among the thousands of individual compounds in a plant extract. The modular NP-HPTLC−UV/Vis/FLD−EDA-heart-cut RP-HPLC−DAD−HESI-MS coupling can be used in routine and makes it easy to recognize the essence. It is said that a picture is worth a thousand words, but the effect image is worth even more. It visualizes the impressive power of nature to supply the body with important building blocks (multipotent chemicals). Several innovations were demonstrated. The pYES equipped with the hERβ or in combination with the simulated S9 metabolization were applied for the first time. In the antagonistic hormonal assays, the proof for false-positive results was newly included. Synergistic effects were revealed in the hormonal effect-profiles for the first time. The developed workflows can be transferred to any other assays or samples. The array of effect-directed profiles clearly showed that natural food has the power to contribute to our homeostasis in various effective ways. The 1,292 profiles (68 samples x 19 detections) obtained within a few weeks showed the versatility of the activity potential of natural food. Artificial intelligence could help evaluate the wealth of information obtained. Exemplarily, the 60 most bioactive components were identified as proof of principle. The developed non-targeted effect-directed hyphenation highlights the advantages of analytical speed, efficiency, and economy. First, the samples were freed from the interfering matrix via planar chromatographic separation. Secondly, the focus was exclusively laid on bioactive compounds, providing targeted characterization. Calculated per sample, the robust profiling takes 3-15 min and costs 0.5-1 Euro, depending on incubation time and material consumption, respectively. One current limitation is the low resolution of the single quadrupole MS instrument. Potential drug candidates were only tentatively assigned and confirmed in an additional run against standards. At the same time, this limitation can be an opportunity for further research. Upgrading the MS instrument to a highresolution MS with fractionation possibility enables unambiguous assignment of molecular formulas and structure elucidation through fragmentation. This makes effect profiling even more attractive for routine analysis.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

FUNDING
This work was supported by the Instrumentation was partially funded by the Deutsche Forschungsgemeinschaft (DFG; INST 162/536-1 FUGG).