Chinese Chemical Letters 33 (2022) 4746–4749

Contents lists available at ScienceDirect

Chinese Chemical Letters

journal homepage: www.elsevier.com/locate/cclet

Simultaneous determination of indole metabolites of tryptophan in rat

feces by chemical labeling assisted liquid chromatography-tandem
mass spectrometry
Qin-Feng Zhang a , Hua-Ming Xiao a , Jin-Tao Zhan a , Bi-Feng Yuan a,b,∗ , Yu-Qi Feng a,b,∗
a
Department of Chemistry, Wuhan University, Wuhan 430072, China
b
School of Public Health, Wuhan University, Wuhan 430071, China

a r t i c l e i n f o a b s t r a c t

Article history: As the connecting part of diet and host physiology, intestinal microbes can convert the ingested diet
Received 22 November 2021 into a huge number of physiologically active small molecules. Indole metabolites of tryptophan are pre-
Revised 10 December 2021
cursors or signal molecules for many biologically active substances, which are involved in serotonin
Accepted 4 January 2022
and microbial catabolism pathways. To understand the influence of tryptophan metabolism in the in-
Available online 12 January 2022
testinal environment on the neurological and immune systems at the molecular level, it is important
Keywords: to establish a high-coverage analytical method to comprehensively analyze the metabolites involved
Chemical labeling in tryptophan metabolism. However, due to a small molecular weight and poor response during mass
Liquid chromatography-mass spectrometry spectrometry analysis, as well as weak retention on the reversed-phase chromatography, determina-
Indole metabolites tion of indole metabolites of tryptophan is challenging. Here, we proposed a method for the simulta-
Tryptophan neous determination of 20 indole metabolites of tryptophan in a single run on reversed-phase chro-
Rat feces
matography by chemical labeling coupled to liquid chromatography-tandem mass spectrometry analy-
sis. 4-(Dimethylamino)benzaldehyde (DMAB) was used for the labeling of indole metabolites of trypto-
phan, which could significantly improve the detection sensitivities and retention of these metabolites
on reversed-phase chromatography. With the developed method, we realized the sensitive detection and
comprehensive analysis of 15 endogenous indole metabolites of tryptophan in rat feces samples with
functional dyspepsia intervention by acupuncture. The developed method offers a useful tool for study-
ing tryptophan metabolism-related diseases.
© 2022 Published by Elsevier B.V. on behalf of Chinese Chemical Society and Institute of Materia
Medica, Chinese Academy of Medical Sciences.

Intestinal microorganisms generate a "diet from the microbiome croorganisms [7,15,16]. These metabolites act as ligands and are
to the host axis" that turns the consumed meal into a huge num- essential for intestinal homeostasis [17–19].
ber of physiologically active small molecules [1–3]. A variety of To understand the influence of tryptophan metabolism in the
metabolites produced by intestinal microorganisms play essen- intestinal environment on the neurological system and immune
tial roles in metabolism, immunology and neurohomeostasis [4– system from the molecular level, it is important to establish a
6]. Tryptophan and its indole metabolites are precursors or signal high-coverage analytical method to analyze the indole metabo-
molecules for many biologically active substances [7–10]. The in- lites of tryptophan. At present, the analysis of tryptophan indole
dole metabolites in the serotonin pathway play important roles in metabolites mainly relies on liquid chromatography-mass spec-
the "gut-brain axis", which activates specific 5-hydroxytryptamine trometry (LC-MS) [20–23]. It has been reported that LC-MS-based
receptors in the gastrointestinal tract and is engaged in a wide methods were used to determine tryptophan metabolites in vari-
range of physiological functions [11–14]. Tryptophan also produces ous biological samples [24–26]. For example, Chen et al. reported
a variety of indole metabolites under the action of intestinal mi- an analytical method for the simultaneous analysis of 16 tryp-
tophan indole metabolites using LC-MS with selected reaction
monitoring mode [27]. The fragments of these tryptophan indole
metabolites under collision-induced dissociation (CID) are similar
∗
Corresponding authors at: Department of Chemistry, Wuhan University, Wuhan because they all have modest molecular weights and share the
430072, China. same indole skeleton structure. In general, the fragmentation be-
E-mail addresses: bfyuan@whu.edu.cn (B.-F. Yuan), yqfeng@whu.edu.cn (Y.-Q. havior of parent ions under CID is employed to select distinc-
Feng).

https://doi.org/10.1016/j.cclet.2022.01.004
1001-8417/© 2022 Published by Elsevier B.V. on behalf of Chinese Chemical Society and Institute of Materia Medica, Chinese Academy of Medical Sciences.
Q.-F. Zhang, H.-M. Xiao, J.-T. Zhan et al. Chinese Chemical Letters 33 (2022) 4746–4749

converting indole butyric acid (IBA) into a cationic derivative via

the Ehrlich reaction, which could then be combined with nega-
tively charged gold nanoparticles to form a SERS hot spot and fi-
nally realize the dual signal enhancement of IBA [43].
In the current study, we successfully established a chemical
labeling strategy for indole using 4-(dimethylamino)benzaldehyde
Fig. 1. Chemical labeling of indole metabolites by DMAB. (DMAB) reagent to specifically label indole skeleton in indole
metabolites of tryptophan (Fig. 1). The reaction is a specific elec-
trophilic substitution reaction of DMAB with position 2 on the in-
tive product ions for multiple reaction monitoring (MRM) transi- dole ring under the catalysis of hydrochloric acid. The reaction
tions [28]. As a result, establishing the MRM transitions for indole mechanism is shown in Fig. S1 in Supporting information. The
metabolites is difficult. oxygen atom on the carbonyl group of DMAB is first protonated
Since metabolites with carboxylic groups have a poor response by the hydrogen ion in the strong acid medium. Due to the elec-
in the positive ion mode of mass spectrometry [29–32], some trophilicity of the carbocation intermediate, it can undergo an elec-
tryptophan indole metabolites such as indole-3-lactic acid (ILA), trophilic substitution reaction with the nucleophilic electron-rich
indole-3-carboxylic acid (ICA), and 3-indoxyl sulfate, are difficult to indole ring. DMAB can selectively react with the indole skele-
be analyzed in the positive ion mode. Therefore, researchers used ton, and the N,N-dimethyl on DMAB can enhance the signal re-
polarity switching to classify tryptophan indole metabolites, and sponse during mass spectrometry analysis [44]. In addition, the
analyzed the corresponding metabolites in positive and negative benzene ring on DMAB can increase the hydrophobicity of the an-
ion modes [27]. However, switching between positive and negative alyte, thereby enhancing the retention of the labeled product on
ion modes consumes acquisition time, and frequently reciprocat- the reversed-phase chromatography [45]. The DMAB labeling cou-
ing ion modes switching would affect the MS response to a certain pled with LC-MS/MS analysis enabled the sensitive and simultane-
extent. In addition, tryptophan indole metabolites have nitrogen- ous analysis of 20 indole metabolites of tryptophan (Fig. 2).
containing heterocyclic indole, and the modified groups are usu- We first selected indole metabolites with various substituents,
ally hydrophilic groups such as amino/carboxyl groups, resulting in including indole (IND), indole-3-acetonitrile (IAcn), indole-3-
limited reversed-phase chromatography retention. acetic acid (IAA) and indole-3-acetamide (IAM), as representative
Chemical labeling assisted LC-MS strategies have been shown metabolites to investigate the fragmentation behavior of DMAB-
to improve the detection sensitivities of analytes as well as in- labeled products under CID mode. The results showed that all
crease the retention on reversed-phase chromatography [33–41]. the four representative indole metabolites could be successfully la-
For example, Guo et al. reported a method that used two labeling beled by DMAB (Fig. 3). The DMAB-labeled IND (m/z 249.1) pro-
reagents, N-dimethyl-/N–diethyl-aminonaphthalene-sulfonyl chlo- duced m/z 233.1, m/z 219.1, and m/z 204.1 product ions (Fig. 3a).
ride (Dns/Dens-Cl) to label amino and phenolic hydroxyl groups in The DMAB-labeled IAcn produced a product ion of m/z 247.1
indole metabolites, thereby realizing the analysis of 9 tryptophan (Fig. 3b), and the DMAB-labeled IAA produced product ion of m/z
metabolites [42]. However, this method requires the simultaneous 261.1 (Fig. 3c). The DMAB-labeled IAM produced product ions of
usage of two labeling reagents. Wang et al. reported a strategy for m/z 246.1 and m/z 218.1 (Fig. 3d). All these product ions were pro-

Fig. 2. The chemical structures of indole metabolites of tryptophan involved in serotonin pathway and microbial catabolites.

4747
Q.-F. Zhang, H.-M. Xiao, J.-T. Zhan et al. Chinese Chemical Letters 33 (2022) 4746–4749

Fig. 3. MS/MS spectrum and fragmentation of DMAB-labeled tryptophan indole metabolite. (a) DMAB-labeled IND. (b) DMAB-labeled IAcn. (c) DMAB-labeled IAA. (d) DMAB-
labeled IAM.

meet the requirements of subsequent analysis. Collectively, the op-

timized conditions for DMAB labeling were as follows: acid cata-
lyst, 2 mol/L of HCl; concentration of DMAB, 2.5 mmol/L; reaction
time, 9 h.
The calibration curves were prepared to quantify the amounts
of tryptophan indole metabolites using the mixtures of 20 tryp-
tophan indole metabolite standards (0.5, 1, 2, 5, 10, 20, 50, 100,
200, 500, 1000, 2000, 5000 ng/mL), and the fixed concentrations
of internal standards (MEL-d4 , TRM-d4 , NAS-d3 and 5MeOT-d4 ,
10 ng/mL; IAA-d2 , 20 ng/mL; Trp-d3 , 100 ng/mL). The calibration
curves were constructed by plotting the peak area ratio (ana-
lytes/internal standards) to the concentrations of tryptophan indole
metabolites. The coefficient of determination (R2 ) ranged from 0.97
to 0.99, indicating the good linearity (Table S2 in Supporting infor-
mation). The limits of detection (LODs) and limits of quantifica-
tion (LOQs) are defined as the concentration of analyte when the
signal-to-noise ratio (S/N) is 3 and 10, respectively [46]. The results
Fig. 4. The extracted-ion chromatograms of 20 indole metabolites standards of showed that the LODs and LOQs of tryptophan indole metabo-
tryptophan analyzed by DMAB labeling assisted LC-MS/MS analysis. 1. serotonin lites were 0.01–0.30 ng/mL and 0.03–1.00 ng/mL, respectively (Table
(SER), 2. N-acetylserotonin (NAS), 3. 5-hydroxyindoleacetic acid (5HIAA), 4. 5–
S2 in Supporting information). The reproducibility and accuracy of
hydroxy-l-tryptophan (5HLT), 5. melatonin (MLT), 6. indole (IND), 7. 3-indolelactic
acid (ILA), 8. 5-methoxyindoleacetic acid (MeOIAA), 9. tryptamine (TRM), 10. 3- the method were evaluated by analysis of samples with different
indoleacetic acid (IAA), 11. 3-methylindole (3MI), 12. indole-3-acetamide (IAM), spiked concentrations (5, 50 and 100 ng/mL). The intra-day preci-
13. l-tryptophan (TRP), 14. N-methyltryptamine (NMT), 15. indole-3-propionic acid sion was evaluated by repeating 5 times a day, and the inter-day
(IPA), 16. 5-methoxytryptamine (MeOT), 17. indole-3-carboxylic acid (ICA), 18. 3- precision was evaluated for 3 consecutive days. The result showed
indoleacetonitrile (IAcn), 19. tryptophanol (TOL), 20. indole-3-pyruvic acid (IPyA).
that the method had acceptable intra-day and inter-day precision,
and the RSD value was less than 13.7%. In addition, the recovery
duced by the fragmentation of the substituent at the indole ring rate was 78.5%−113.5% (Table S3 in Supporting information).
C3. In summary, the results showed that DMAB could successfully We then evaluated the matrix effect, the influence of the ma-
label indole metabolites with different substituents. The optimized trix on the labeling reaction and the extraction efficiency of the
MRM parameters of the DMAB-labeled products are shown in Ta- proposed method. The matrix effect was defined as the influ-
ble S1 in Supporting information. With the established MRM mass ence of the matrix on the MS response of the analytes and was
transitions, we obtained the optimal chromatographic separation calculated by the formula (Cstd+endo − Cendo )/Cstd . Cstd+endo repre-
of these by optimizing mobile phase compositions, gradients and sents the peak area of tryptophan indole metabolites in rat fe-
flow rates (Fig. 4). ces samples after spiked with standards, and Cendo represents
We next optimized the DMAB labeling reaction conditions, in- the peak area of tryptophan indole metabolites in the blank rat
cluding the types of acid catalyst, the concentrations of acid cat- feces sample. Cstd represents the peak area of tryptophan in-
alysts, the amount of DMAB, and the reaction time. The results dole metabolites in the standard solution. The influence of the
showed that, when HCl was utilized as the acid catalyst in the re- matrix on the labeling reaction was calculated by the formula
action, the best MS response of the DMAB-labeled product could (Cpost-spiked − Cendo )/(Cstd+endo − Cendo ), where Cpost-spiked represents
be obtained (Fig. S2a in Supporting information). We further op- the peak area of tryptophan indole metabolites in rat feces samples
timized the concentration of HCl. The result showed that 2 mol/L spiked with standard solution mixtures after extraction. The matrix
of HCl offered the best reaction efficiency (Fig. S2b in Supporting effect and the effect of matrix on the labeling reaction were be-
information). In addition, the signal of the DMAB-labeled product tween 80.8%−111.1% and 81.4%−113.2% in rat feces samples, respec-
increased with increasing concentration of the DMAB, and reached tively (Table S4 in Supporting information), indicating that rat feces
the plateau when the concentration of DMAB was 2.5 mmol/L (Fig. samples have minor effects on the ionization and labeling reaction
S2c in Supporting information). However, due to the ion suppres- of analytes. The extraction efficiency was calculated by the formula
sion, further increasing of DMAB will cause a certain degree of (Cpre-spiked − Cendo )/(Cpost-spiked − Cendo ), where Cpre-spiked represents
signal reduction. Moreover, the signal of the DMAB-labeled prod- the peak area of tryptophan indole metabolites in rat feces samples
uct was gradually increased with the increased reaction time, and spiked with standard solution mixtures before extraction. The ex-
reached a plateau when the reaction time was 9 h (Fig. S2d in traction efficiency was between 78.5%−92.1% in rat feces samples
Supporting information). We continued to monitor the signal of (Table S4 in Supporting information), indicating that the extraction
the DMAB-labeled tryptophan indole metabolites at 4 °C for 24 h method can effectively extract the analytes in rat feces samples.
to investigate the stability of the DMAB-labeled product. The re- We evaluated the slopes of the quantitative regression curves of
sults showed that the DMAB-labeled product signal within 24 h the standards solution mixtures of tryptophan indole metabolites
was within ± 20% (Fig. S3 in Supporting information), which could and the spiked rat feces samples. The results showed that after

4748
Q.-F. Zhang, H.-M. Xiao, J.-T. Zhan et al. Chinese Chemical Letters 33 (2022) 4746–4749

sample pretreatment including extraction and DMAB labeling, the References

regression curves of tryptophan indole metabolites in the stan-
dard solution mixtures and spiked rat feces samples had compa- [1] G. Agirman, E.Y. Hsiao, Cell 184 (2021) 2524.
[2] E.M. Teichman, K.J. O’Riordan, C.G.M. Gahan, T.G. Dinan, J.F. Cryan, Cell Metab.
rable slopes (Table S5 in Supporting information), indicating that 31 (2020) 448–471.
the quantitative approach can meet the analytical requirements. [3] H. Luan, X. Wang, Z. Cai, Mass Spectrom. Rev. 38 (2019) 22–33.
We also compared the LOQs of this method with other analyti- [4] Z. Kuang, Y. Wang, Y. Li, et al., Science 365 (2019) 1428–1434.
[5] F. Bishehsari, R.M. Voigt, A. Keshavarzian, Nat. Rev. Endocrinol. 16 (2020)
cal methods for tryptophan indole metabolites (Table S6 in Sup- 731–739.
porting information). Previous studies mainly focused on the seven [6] A. Agus, J. Planchais, H. Sokol, Cell Host Microbe 23 (2018) 716–724.
metabolites of the serotonin pathway, including 5HLT, SER, MeOT, [7] S. Krishnan, Y. Ding, N. Saedi, et al., Cell Rep. 23 (2018) 1099–1111.
[8] C. Michaudel, H. Sokol, Cell Metab. 32 (2020) 514–523.
5HIAA, MeOIAA, NAS and MEL [26,47–49]. In the current study, we
[9] V. Rothhammer, D.M. Borucki, E.C. Tjon, et al., Nature 557 (2018) 724–728.
could simultaneously detect 20 tryptophan indole metabolites due [10] H.M. Roager, T.R. Licht, Nat. Commun. 9 (2018) 3294.
to the improved detection sensitivity. Therefore, our method could [11] F. De Vadder, E. Grasset, L. Manneras Holm, et al., Proc. Natl. Acad. Sci. U. S. A.
115 (2018) 6458–6463.
cover the serotonin and microbial catabolism pathways of tryp-
[12] A. Lavelle, H. Sokol, Nat. Rev. Gastroenterol. Hepatol. 17 (2020) 223–237.
tophan, which is helpful to understand the effect of tryptophan [13] Y. Peng, J.D. McCorvy, K. Harpsoe, et al., Cell 172 (2018) 719–730.
metabolism on the “brain-gut-microbe axis”. [14] S.J. Allison, Nat. Rev. Nephrol. 14 (2018) 71.
We further applied the established DMAB labeling-assisted LC- [15] L. Laurans, N. Venteclef, Y. Haddad, et al., Nat. Med. 24 (2018) 1113–1120.
[16] L.H. Morais, H.L.T. Schreiber, S.K. Mazmanian, Nat. Rev. Microbiol. 19 (2021)
MS/MS method for simultaneous analysis of a series of metabolites 241–255.
involved in the serotonin metabolic pathway of tryptophan and [17] E.C. Rosser, C.J.M. Piper, D.E. Matei, et al., Cell Metab. 31 (2020) 837–851.
the microbial catabolism pathway. With the developed method, [18] A. Sadik, L.F.S. Patterson, S. Ozturk, et al., Cell 182 (2020) 1252–1270.
[19] G.Z. Wei, K.A. Martin, P.Y. Xing, et al., Proc. Natl. Acad. Sci. U. S. A. 118 (2021)
we successfully detected metabolites in the serotonin pathway (Ta- e2021091118.
ble S7 in Supporting information), including 5HLT, SER, 5HIAA, [20] A.A. Kolodziejczyk, D. Zheng, E. Elinav, Nat. Rev. Microbiol. 17 (2019) 742–
MeOIAA, NAS and MEL in rat feces samples. In addition, we de- 753.
[21] P.D. Cani, M. Van Hul, C. Lefort, et al., Nat. Metab. 1 (2019) 34–46.
tected the existence of a variety of indole metabolites in the micro- [22] J. Zhang, S.D. Petersen, T. Radivojevic, et al., Nat. Commun. 11 (2020) 4880.
bial degradation pathway, including IND, 3MI, ICA, ILA, TRM, NMT, [23] J. Lloyd-Price, C. Arze, A.N. Ananthakrishnan, et al., Nature 569 (2019)
IAA and IPA (Table S7 in Supporting information). 655–662.
[24] L. Whiley, L.C. Nye, I. Grant, et al., Anal. Chem. 91 (2019) 5207–5216.
In summary, we developed a DMAB labeling strategy in com-
[25] V.Y. Rubio, J.G. Cagmat, G.P. Wang, R.A. Yost, T.J. Garrett, Anal. Chem. 92 (2020)
bination with LC-MS/MS analysis for the simultaneous and sen- 2550–2557.
sitive detection of 20 tryptophan metabolites with indole struc- [26] F. Tomosi, G. Kecskemeti, E.K. Cseh, et al., J. Pharm. Biomed. Anal. 185 (2020)
113246.
ture, allowing for a high-coverage analysis of tryptophan metabo-
[27] G.Y. Chen, W. Zhong, Z. Zhou, Q. Zhang, Anal. Chim. Acta 1037 (2018) 200–
lites. This method showed good accuracy and precision. Using this 210.
method, we successfully detected 15 endogenous indole metabo- [28] X. Wang, G. Yan, X. Zhou, P. Yang, Chem. J. Chin. Univ. 41 (2020) 1505–1512.
lites of tryptophan in rat feces samples with functional dyspepsia [29] Y. Feng, N.B. Xie, W.B. Tao, et al., CCS Chem. 2 (2020) 994–1008.
[30] Y. Feng, J.J. Chen, N.B. Xie, et al., Chem. Sci. 12 (2021) 11322–11329.
intervention by acupuncture. This method could be applied to re- [31] Y. Tang, J. Xiong, H.P. Jiang, et al., Anal. Chem. 86 (2014) 7764–7772.
search on tryptophan metabolism-related diseases in future study. [32] Y. Tang, S.J. Zheng, C.B. Qi, Y.Q. Feng, B.F. Yuan, Anal. Chem. 87 (2015)
3445–3452.
[33] Y. Dai, C.B. Qi, Y. Feng, et al., Anal. Chem. 93 (2021) 6938–6946.
Declaration of competing interest [34] F.L. Liu, T.T. Ye, J.H. Ding, et al., Anal. Chem. 93 (2021) 6848–6856.
[35] Q.Y. Cheng, J. Xiong, F. Wang, B.F. Yuan, Y.Q. Feng, Chin. Chem. Lett. 29 (2018)
The authors declare that they have no known competing finan- 115–118.
[36] M.D. Lan, B.F. Yuan, Y.Q. Feng, Chin. Chem. Lett. 30 (2019) 1–6.
cial interests or personal relationships that could have appeared to [37] C.B. Qi, J.H. Ding, B.F. Yuan, Y.Q. Feng, Chin. Chem. Lett. 30 (2019) 1618–1626.
influence the work reported in this paper. [38] C.B. Qi, H.P. Jiang, J. Xiong, B.F. Yuan, Y.Q. Feng, Chin. Chem. Lett. 30 (2019)
553–557.
[39] Y. Dai, B.F. Yuan, Y.Q. Feng, RSC Chem. Biol. 2 (2021) 1096–1114.
Acknowledgments [40] M.Y. Chen, Z. Gui, K.K. Chen, et al., Chin. Chem. Lett. 33 (2022) 2086–2090.
[41] M.Y. Chen, C.B. Qi, X.M. Tang, et al., Chin. Chem. Lett. 33 (2022) 3772–3776.
This work is supported by the National Key Research and [42] H. Guo, Y. Jiao, X. Wang, et al., J. Chromatogr. A 1504 (2017) 83–90.
[43] F. Wang, X. Gu, C. Zheng, et al., Anal. Chem. 89 (2017) 8836–8843.
Development Program of China (No. 2018YFA0900400), the Na-
[44] B.L. Qi, P. Liu, Q.Y. Wang, et al., TrAC Trends Anal. Chem. 59 (2014) 121–132.
tional Natural Science Foundation of China (Nos. 21635006 and [45] S. Zhao, L. Li, TrAC Trends Anal. Chem. 131 (2020) 115988.
21721005). [46] Q.W. Yu, S.J. Liu, F. Zheng, et al., Chin. Chem. Lett. 31 (2020) 482–486.
[47] K. Wicha-Komsta, R. Skibinski, T. Kocki, W.A. Turski, L. Komsta, J. Chromatogr.
Sci. 59 (2021) 40–46.
Supplementary materials [48] L.S. Wang, M.D. Zhang, X. Tao, et al., J. Chromatogr. B Anal. Technol. Biomed.
Life Sci. 1112 (2019) 24–32.
Supplementary material associated with this article can be [49] E.K. Cseh, G. Veres, M. Szentirmai, et al., Anal. Biochem. 574 (2019) 7–14.

found, in the online version, at doi:10.1016/j.cclet.2022.01.004.

4749