
==== Front
Environ Sci Technol
Environ Sci Technol
es
esthag
Environmental Science & Technology
0013-936X
1520-5851
American Chemical Society

39163399
10.1021/acs.est.4c05165
Article
Nontargeted Analysis of Reactive Nitrogenous Compounds in Suwannee River Standard Reference Materials and Authentic River Water Samples
Shen Qiming †
Zhao Tingting ‡
Wawryk Nicholas J. P. †
Chau K.N.Minh †
Zhang Di †§
Carroll Kristin †
https://orcid.org/0000-0002-3457-3507
Chu Wenhai §
https://orcid.org/0000-0001-6295-2435
Huan Tao *‡
https://orcid.org/0000-0003-1844-7700
Li Xing-Fang *†
† Division of Analytical and Environmental Toxicology, Department of Laboratory Medicine and Pathology, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta T6G 2G3, Canada
‡ Department of Chemistry, Faculty of Science, University of British Columbia, Vancouver Campus, 2036 Main Mall, Vancouver, British Columbia V6T 1Z1, Canada
§ State Key Laboratory of Pollution Control and Resources Reuse, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China
* Email: thuan@chem.ubc.ca.
* Email: xingfang.li@ualberta.ca. Tel: 1-780-492-5094.
20 08 2024
03 09 2024
58 35 1580715815
29 05 2024
09 08 2024
08 08 2024
© 2024 The Authors. Published by American Chemical Society
2024
The Authors
https://creativecommons.org/licenses/by-nc-nd/4.0/ Permits non-commercial access and re-use, provided that author attribution and integrity are maintained; but does not permit creation of adaptations or other derivative works (https://creativecommons.org/licenses/by-nc-nd/4.0/).

Concerns over toxic nitrogenous disinfection byproducts (N-DBPs) necessitate identifying their precursors in source water. Natural organic amino compounds are known precursors to N-DBPs. Three Suwannee River (SR) standard reference materials (SRMs), humic acids (HA), fulvic acids (FA), and natural organic matter (NOM), are commonly used to study DBP formation, but the chemical makeup of amino compounds in SRSRMs remains largely unknown. To address this, we combined stable hydrogen/deuterium isotope labeling, HDPairFinder bioinformatics, and nontargeted high-performance liquid chromatography–high-resolution mass spectrometry (HPLC-HRMS) to characterize these compounds in SRSRMs. This method classifies reactive amines, provides accurate masses and MS/MS spectra, and quantifies intensities. We identified 2707 high-quality features with primary and/or secondary amines in SRSRMs and 75% of them having an m/z < 300. Across all three SRSRMs, 327 amino features were detected, while 856, 794, and 200 unique features were found in SRNOM, SRHA, and SRFA, respectively. In North Saskatchewan River (NSR) samples, a total of 6449 amino features were detected, 818 of them matched those in SRSRMs, and 87% of them were different between the two rivers. Using chemical standards, we confirmed 10 compounds and tentatively identified 5 more. This study highlights similarities and differences in reactive N-precursors in SRSRMs and local river water, enhancing the understanding of geo-differences in reactive N-precursors in different source waters.

New insights into reactive amino compounds in SRSRMs and rivers enable future discovery of unknown N-DBPs of health importance and water research.

Nontargeted analysis
Suwannee River standard reference materials (SRSRMs)
amino compounds
nitrogenous disinfection byproducts (N-DBPs)
stable isotopic labeling
HDPairFinder
Natural Sciences and Engineering Research Council of Canada 10.13039/501100000038 NA Alberta Innovates 10.13039/501100009192 NA Canada Research Chairs 10.13039/501100001804 NA document-id-old-9es4c05165
document-id-new-14es4c05165
ccc-price
Special Issue

Published as part of Environmental Science & Technologyspecial issue “Non-Targeted Analysis of the Environment.”
==== Body
pmc1 Introduction

Nitrogenous organic compounds in source water can react with disinfectants such as chlorine and chloramine to produce nitrogenous disinfection byproducts (N-DBPs).1,2 The use of source water containing high organic nitrogen content and the switch from chlorination to chloramination both contribute to enhanced formation of N-DBPs.3−5 Recent studies have shown that N-DBPs are the drivers of DBP cytotoxicity in drinking water collected from several cities in the United States,6,7 and N-DBPs are generally more toxic than the regulated DBPs.5,8 Potential health concerns regarding N-DBP exposure have raised the importance of identifying nitrogenous precursors and N-DBPs. Therefore, the characterization of nitrogenous precursors in source water is essential for their removal and understanding and controlling N-DBP formation.

The International Humic Substances Society (IHSS) has provided standard reference materials (SRM) of humic acids (HA), fulvic acids (FA), and NOM extracted from the Suwannee River (SR) since 1981. Over the past decades, the Suwannee River standard reference materials (SRSRMs) have been widely used as a benchmark to study DBP formation because of their well-characterized elemental composition, batch-to-batch consistency, and wide availability.9−11 Despite the extensive use of SRSRMs in DBP research, these standard reference materials have remained incompletely characterized.

Several studies have assessed the size distribution, elemental composition, and structural details of SRSRMs using techniques such as size-exclusion chromatography, elemental analysis, organic acidity analysis, and various spectroscopic methods, including ultraviolet–visible (UV–vis), NMR, and fluorescence spectroscopy.12−16 The extensive use of SRSRMs in DBP research is often based on the general characteristics of these reference materials. To date, the detailed molecular information on chemicals in SRSRMs, in particular, nitrogenous compounds, remains unclear. High-resolution mass spectrometry (HRMS) emerges as a powerful tool for characterizing SRSRMs at the molecular level, providing elemental compositions and molecular formulas.17,18 Owing to the high resolving power, Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS) is widely employed to determine the molecular formula of the compounds in SRSRMs.19−21 The predicted molecular formula enables classification of thousands of compounds in these reference materials using van Krevelen diagram, double bond equivalent (DBE), and aromaticity index.22,23 However, the assigned molecular formulas alone cannot elucidate the functional groups of the molecules, thus impeding the identification of the detected compounds. In particular, the compositions of nitrogenous compounds containing amine groups in SRSRMs have not been well characterized.

Amino compounds are key precursors to N-DBPs,24−26 but their detection and identification in source water is challenging because of the complex matrices and trace levels.27 To address this challenge, we developed an isotopic labeling-based liquid chromatography–high-resolution mass spectrometry (LC–HRMS) method for the identification of amino compounds in source water.27 In this approach, the reductive amination reaction with formaldehyde or deuterated formaldehyde generates pairs of hydrogen- and deuterium-labeled amino compounds with a certain mass distance (e.g., 2, 4 Da). The characteristic mass pattern facilitates the prioritization analysis of amino features from raw data generated by nontargeted HPLC-HRMS analysis. To streamline the identification of the labeled amino features within complex raw data, we developed the HDPairFinder program and archived 38314 possible reactive amino compounds in the AMINES library.28 This tool automatically corrects the retention time shift, removes the false chemical features, and provides putative compound annotation. These functions significantly enhance the sensitivity, specificity, and efficiency in identifying amino compounds.

In this study, we aimed to characterize reactive amino compounds in these SRSRMs because they are important precursors to N-DBPs. To specifically detect the reactive amino compounds, we integrated three steps: (1) applying the H/D stable isotopic labeling to target amino compounds in SRSRMs, followed by mixed-mode cation exchange (MCX) extraction that selectively extracted basic amino compounds; (2) using HPLC-electrospray ionization-quadrupole-time of flight-mass spectrometry (HPLC-ESI-QTOF-MS) performed in positive mode to enhance the detection of basic amino compounds; (3) using HDPairFinder to extract the labeled amino compounds in the nontargeted data and search databases for identification of potential candidates. Using the same approach, we further investigated amino compounds in the local North Saskatchewan River (NSR) water. This study will provide new knowledge about the similarities and differences in the compositions of amino compounds in SRSRMs and NSR water, which will support future DBP research.

2 Materials and Methods

2.1 Chemicals and Materials

Formaldehyde solution (CH2O, contains 10–15% methanol as a stabilizer, 37 wt % in H2O), formaldehyde-d2 solution (CD2O, ∼20 wt % in D2O, 98 atom % D), sodium cyanoborohydride (NaBH3CN, 95%), sodium bicarbonate (NaHCO3, 99.7%), and formic acid (99%) were purchased from Sigma-Aldrich (St. Louis, MO). Optima water, methanol, acetonitrile (ACN), and ammonium hydroxide (30%) were obtained from Fisher Scientific (Ottawa, ON). Suwannee River NOM (RO isolation) (SRNOM, Catalog 2R101N), Suwannee River Humic Acid Standard III (SRHA, Catalog 3S101H), and Suwannee River Fulvic Acid Standard III (SRFA, Catalog 3S101F) were purchased from the International Humic Substances Society (IHSS) (St. Paul, MN). The amino acids tyrosine (Tyr), leucine (Leu), isoleucine (Ile), and phenylalanine (Phe) were obtained from Sigma-Aldrich (St. Louis, MO). The dipeptides Phe-Pro and Tyr-Gly were purchased from Bachem AG (Bubendorf, Switzerland). The other dipeptides Tyr-Phe, Leu-Leu, Tyr-Ala, and Phe-Gly were obtained from Sigma-Aldrich (St. Louis, MO). Oasis MCX cartridges (6 mL, 150 mg of sorbent) were obtained from Waters (Milford, MA). Syringe filters (0.45 μm, poly(vinylidene difluoride) (PVDF)) were purchased from Dikma (Markham, ON).

2.2 H/D Isotopic Labeling Reaction and Sample Preparation

The Suwannee River samples were prepared in triplicate by dissolving SRNOM, SRHA, or SRFA (2 mg/L as C) and NaHCO3 (80 mg/L as CaCO3) in Optima water. The pH levels of the SRNOM, SRHA, and SRFA samples were 8.0, 7.9, and 8.0, respectively. The H/D isotopic labeling process followed the previously developed method.27,28 In brief, CH2O (1.8 M), CD2O (1.8 M), and NaBH3CN (0.6 M) were freshly prepared in Optima water before use. The Suwannee River sample (2 L) was equally divided into two portions. For hydrogen (H) labeling, CH2O (2 mL, 1.8 M) and NaBH3CN (2 mL, 0.6 M) solutions were added to one portion. For deuterium (D) labeling, CD2O (2 mL, 1.8 M) and NaBH3CN (2 mL, 0.6 M) solutions were added to the other portion. The reactions were conducted under magnetic stirring at room temperature for 4 h. After the reaction, formic acid (2 mL) was added to each solution to protonate the amine groups, enhancing the cationic extraction efficiency in the subsequent step. The labeled Suwannee River samples were preconcentrated using solid-phase extraction (SPE). Oasis MCX cartridges (6 mL, 150 mg of sorbent) were used to extract and concentrate the labeled amino compounds because positively charged amino compounds can be retained on the anionic polymeric sorbent through electrostatic interaction. The MCX cartridges were rinsed with methanol (2 mL) and pure water (4 mL, 0.2% FA). Sequentially, the labeled solutions passed through the MCX cartridges at a rate of 2–3 mL/min. Next, the MCX cartridges were washed with water (2 mL, 0.2% FA) and dried for 15 min under a vacuum. After being washed, analytes were eluted with ammonium hydroxide solution (10 mL, 5 wt % in methanol). Then, a gentle nitrogen stream (20–50 kPa) was applied to concentrate the eluent down to 0.1 mL. Finally, the H-and D-labeled extracts were mixed and analyzed using high-performance liquid chromatography–high-resolution mass spectrometry (HPLC-HRMS) (Sciex QTOF x500R).

Twenty-five source water samples from the North Saskatchewan River were collected at the water treatment plant located in Edmonton, Alberta, Canada, on February 15, February 23, February 28, March 7, March 14, March 16, March 21, March 24, March 28, March 31, April 11, and May 5 in 2022, as well as February 3, February 23, March 2, March 9, March 16, March 20, March 23, March 30, April 2, April 5, April 8, April 11, and April 13 in 2023. The sampling dates were before, during, and after spring runoffs based on water color measurements. All source water samples were filtered by using 1.5 μm glass microfiber filters, followed by 0.45 μm nylon membrane filters. The filtered samples were stored at 4 °C prior to the isotopic labeling experiments. The same H/D labeling process was applied to the filtered source water samples.

For standard labeling, a solution of a mixture of standards (250 μg/L in 10 mL of pure water) is added into a 15 mL amber glass vial. Solutions of CH2O (20 μL, 1.8 M) and NaBH3CN (20 μL, 0.6 M) were added into the 15 mL glass vial. The mixture was stirred at room temperature for 4 h. The resulting solution was then analyzed using HPLC-HRMS. To correct the retention time due to the matrix effect, we also spiked the standards labeled with only formaldehyde (H-labeled standards) into SRSRMs with a concentration of 10–50 ppb.

2.3 HPLC-HRMS Analysis

An Agilent 1260 Series HPLC system (Agilent, Santa Clara, CA), with a Luna C18 column (100 μm × 2 μm × 3 μm pore size, Phenomenex, Torrance, CA), was used for sample separation. Mobile phase A was H2O/ACN (95/5, v/v, 0.1% FA,) and mobile phase B was ACN (0.1% FA). The flow rate was set to 80 μL/min, and a gradient elution program was set as follows: 0–10 min, 0% B; 10–30 min, a linear increase from 0% to 30% B; 30–45 min, a linear increase from 30% to 90% B; 45–55 min, 90% B; 55–55.01 min, a linear decrease from 90% to 0% B; 55.01–60 min, 0% B. The column temperature was maintained at 30 °C. In the multiple injection analysis, a series of injection volumes (2, 4, 5, 8, 10, and 15 μL) were used. Three replicates were measured, and the average intensities of features detected in these measurements were calculated for analysis. The injection volumes for the source water and standard samples were 20 μL.

Nontargeted analysis was performed using a quadrupole time-of-flight mass spectrometer (Sciex QTOF x500R) operated in an information-dependent analysis (IDA) mode with a positive ion spray voltage of 5500 V. Other parameters were as follows: source gas 1(N2, 35 arbitrary units), source gas 2 (N2, 40 arbitrary units), curtain gas (N2, 30 arbitrary units), temperature (500 °C), and declustering potential (DP, 100 V). For the full scan, the collision energy (CE) was set to 10 V, and the MS1 scan range was m/z 50–1000 with an accumulation time of 0.25 s. For MS/MS collection, CE was set to 35 V with a CE spread of 15 V. The intensity threshold was 1000 cps, and the maximum number of candidate ions was 10. The MS2 scan range was m/z 20–1000, with an accumulation time of 0.1 s.

2.4 Data Process

The HRLC-HRMS data were analyzed using the HDPairFinder program (https://github.com/HuanLab/HDPairFinder), which is detailed in the previous paper.28 Briefly, the raw HPLC-HRMS data files (Sciex wiff2 files) were first converted to mzML using MSConvert (version 3.0). The MSConvert parameters are shown in Figure S1. Then, the mzML files were analyzed using HDPairFinder on a desktop computer (Intel i9–10900 CPU@2.8 GHz with 10 cores and 32 GB memory, Windows 10; 64-bit operating system). The parameters of HDPairFinder were as follows: intensity threshold of H- and D-labeled features was 1000 cps, m/z tolerance of D-labeled features was 20 ppm, the retention time difference between H- and D-labeled features was −0.2 to 0.1 min, the intensity ratio between H- and D-labeled features was 0.4 to 1.4, the threshold for cross-correlation was 0.7, m/z tolerance for feature alignment was 50 ppm, retention time tolerance for feature alignment was 0.5 min, m/z tolerance for gap filling was 20 ppm, retention time tolerance for gap filling was 0.5 min, the intensity threshold for gap filling was 1000 cps, and the relative m/z tolerance for annotation was 30 ppm.

Statistical analyses such as heatmap and principal component analysis (PCA) were conducted using MetaboAnalyst 6.0 (https://www.metaboanalyst.ca/). The aligned feature table underwent the following preprocessing steps: replacement of missing values with 0, log transformation, and intensity autoscaling.

3 Results and Discussion

3.1 Reactive Amino Features in SRSRMs

The amino compounds in SRNOM, SRHA, and SRFA were labeled with H/D methyl groups using the stable H/D isotope labeling method we established.27 The labeled samples were analyzed using HPLC-HRMS, yielding a total of 65,610 features across all three SRSRMs (Figure 1a). Using HDPairFinder, 7695 labeled amino features were extracted from this pool of 65,610 features (Figure 1b). To remove the unreliable features, we applied the established multiple injection method to evaluate the quality of the extracted features.28 The qualified features met the three criteria: (1) detection in at least 4 out of 6 injection volumes, (2) intensity exceeding twice the method blank, and (3) Pearson correlation coefficient ≥0.9 between intensity and injection volume, as depicted in Figure S2. Based on multiple injection analysis, as many as 2707 high-quality features were found to meet the criteria above, as shown in Figure 1c.

Figure 1 Scatter plots of total HRMS features of SRSRMs (a), HDPairFinder extracted amino features of SRSRMs (b), and high-quality amino features of SRSRMs after multiple injection analysis (c).

Figure 2a presents the number of high-quality features, 1596, 1544, and 751 in SRNOM, SRHA, and SRFA, respectively. The total MS intensities of these features in SRNOM, SRHA, and SRFA were 1.26 × 107, 1.16 × 107, and 6.27 × 106, respectively. In general, the total MS intensity was proportional to the number of features, indicating that the abundance of features is roughly proportional to the total MS intensity across the SRNOM, SRHA, and SRFA samples. Compared with SRNOM and SRHA, SRFA exhibited fewer than half of the amino features, while SRNOM and SRHA demonstrated a comparable number of qualified features. Based on the elemental composition analysis shown in Table S1 obtained from IHSS,29 SRFA has a significantly lower nitrogen content, which explains the fewer amino compounds observed in SRFA. Figure S3 exhibits heatmaps of the relative intensity of the features in the method blank, SRNOM, SRHA, and SRFA, allowing the visualization of the difference across the three SRSRMs. The three SRSRMs are clearly distinct from the method blanks. The results clearly show the similarities and differences between the amino features in SRNOM, SRHA, and SRFA. Similarities include common features shared among all three SRSRMs. However, differences are evident, as a significant number of features appear exclusively in one or two of the SRSRMs, indicating unique compositions within each reference material. The differences are explained by the extraction conditions of SRNOM, SRHA, and SRFA. SRHA and SRFA are extracted from the Suwannee River water sample using XAD-8 resin and separated based on their solubilities under acidic conditions,30 resulting in compositional differences between SRHA and SRFA. On the contrary, SRNOM is isolated using the reverse osmosis method, theoretically containing most compounds in SRHA and SRFA.12,31,32 However, the use of the reverse osmosis method to isolate SRNOM may have inevitably resulted in the loss of low-molecular-weight compounds,31,32 leading to the observed differences between SRNOM and the other two reference materials.

Figure 2 (a) Number of high-quality amino features and their total MS intensities in each SRSRM with a 15 μL injection volume. (b) Venn diagram of qualified features in SRNOM, SRHA, and SRFA.

A Venn diagram was then used to illustrate the distribution of the total 2707 qualified features in SRSRMs as well as the intersections and differences among SRNOM, SRHA, and SRFA (Figure 2b). A total of 327 features were present in all three SRSRMs, accounting for 20.5% of the features in SRNOM, 21.2% in SRHA, and 43.5% in SRFA. The intensities of these shared features contributed to 31.3, 34.3, and 47.5% of the total amino feature intensities in SRNOM, SRHA, and SRFA, respectively. This large intensity contribution from a relatively small number of features indicates that these 327 common features are dominant in the SRSRMs. In addition, SRNOM, SRHA, and SRFA individually exhibited 856, 794, and 200 unique features, respectively. Moreover, 444 features are common to both SRFA and SRHA, while 434 features are shared between SRFA and SRNOM, and 633 features are common to both SRNOM and SRHA. We further investigated the intensity ratio of the shared features between any two of the three SRSRMs (Figure S4). Overall, around 90% of the shared features between any two SRSRMs did not exhibit significant differences in intensities, although some shared features were more prevalent in one reference material than the other. For example, a feature with m/z = 183.1717 is detected in both SRHA and SRFA. However, its intensity in SRFA exceeds that in SRHA by over 1000 times. Therefore, both the distinct amino compounds in each SRSRM and certain shared amino compounds with varying prevalence among different SRSRMs may contribute to diverse N-DBP profiles formed from each reference material.

3.2 Labeling Tags in SRSRMs

Another unique advantage of the H/D labeling method is that it enables the classification of primary, secondary, and polyamines. Depending on the type and number of amine groups on the molecules, different numbers of labeling methyl tags can be attached. For example, primary amines can be labeled with two tags, while secondary amines can be labeled with only one tag. To characterize the SRSRMs, we further investigated the number of labeling tags in the high-quality features. In SRNOM and SRHA, around 47.5 and 49.5% of features contained one tag, while approximately 30% had two tags (Figure 3a,b). SRFA showed a similar distribution but with a slight increase in one-tag features (54.1%) and a decrease in two-tag features (26.5%), as shown in Figure 3c. Generally, one and two tags were predominant across all three SRSRMs, indicating the prevalence of primary or secondary amine groups.

Figure 3 Number of labeling tags attached to the high-quality features in SRNOM (a), SRHA (b), and SRFA (c). The m/z distribution of the features with different labeling tags in SRNOM (d), SRHA (e), and SRFA (f). The m/z values represent the original m/z values of the amino compounds and exclude the labeled methyl tags. The curves in panels (d–f) represent the Gaussian distribution of m/z for the labeled amino features.

Then, we investigated the mass (m/z) distribution of the features with different numbers of tags in SRNOM, SRHA, and SRFA. In SRNOM, features with one, two, three, and four tags had median m/z values of 218.13, 255.63, 222.13, and 192.12 (Figure 3d). SRHA displayed similar one- and two-tag medians (223.67 and 259.16) but lower values for three- and four-tag features (182.62 and 179.11), as shown in Figure 3e. In contrast, SRFA (Figure 3f) exhibited lower median m/z values across all tag numbers (199.67, 209.15, 165.62, and 149.59), suggesting smaller molecular weights compared to SRNOM and SRHA. This finding aligns with the reports that SRFA has better solubility and lower average molecular weight.33,34 Overall, the majority of amino compounds in SRSRMs had a molar mass less than 300, suggesting that small amino compounds made up a significant portion of the amino compounds in the SRSRMs. In addition, the features labeled with one and two tags had higher m/z values than those features labeled with three and four tags among all three SRSRMs. This may be partially attributed to the degradation of compounds through amide bonds, which break them down into smaller compounds with additional amine groups. Another possible reason is that we set the mass scan range between 50 and 1000 Da to investigate small precursors that are not easily removed by water treatment processes, which may have excluded larger amino compounds (m/z > 1000 Da) with three or four tags.

3.3 Identification of Amino Compounds in SRSRMs

Among the high-quality amino features, 61.5% (1664/2707) of these features were putatively annotated as potential amino compounds by matching their accurate mass against the AMINES library of HDPairFinder (Figure S2). Based on the putative annotations, we further investigated the experimental MS/MS spectra to confirm the structures of the detected amino compounds. However, due to the nontargeted high-resolution MS/MS analysis being performed on the tandem mass spectrometer in IDA mode, numerous amino features extracted by HDPairFinder lacked MS/MS spectra, posing a challenge to the identification process. Through the analysis of MS/MS spectra using SCIEX OS 2.0, we tentatively identified 15 amino compounds, and 10 of these identified compounds were confirmed using chemical standards, as summarized in Table S2.

Here, we demonstrate the identification of compound Tyr-Phe as an example in Figure 4. The measured mass for the H-labeled feature of m/z 357.1808 (compound No. 1) matched that of methyl-labeled Tyr-Phe ([C20H24N2O4 + H]+, m/z 357.1809), with an error of 0.2 ppm. In Figure 4a, the H-labeled SRSRMs exhibit an XIC peak (black) for the feature of m/z = 357.1808 at 26.7 min. Spiking the H-labeled Tyr-Phe standard into the H-labeled SRSRMs increased this XIC peak, as shown by the red peak in Figure 4a. The MS spectrum of compound No. 1 in H/D-labeled SRSRMs matched that of the H/D-labeled Tyr-Phe standard (Figure 4b). An ion pair (m/z = 357.1808 and 361.2064) with a mass difference of 4.0256 was detected in both the measured MS spectrum of H/D-labeled SRSRMs and the H/D-labeled Tyr-Phe standard. The mass difference of 4.0256 indicates that two methyl tags were added to the amino groups of compound No. 1. The isotope peaks were observed in H/D-labeled Tyr-Phe standard but not in H/D-labeled compound No. 1 because of the low MS intensity of compound No. 1. In addition to MS spectra, MS/MS spectra were used to identify and confirm compound No. 1. MS/MS spectra of H- and D-labeled compound No.1 in SRSRMs are shown in Figure 4c. Comparing the peaks in the H- and D-labeled MS/MS spectra helps to infer the presence of labeling tags in fragment ions. For example, the H/D fragment ion pair (m/z = 164.1078 and 168.1327, with Δm/z = 4.0249) indicated the presence of an amine group with two methyl labeling tags, while the H/D fragment ion pair (m/z = 149.0841 and m/z = 151.0969, with Δm/z = 2.0128) suggested the presence of one methyl-labeled amine group. Finally, we compared the MS/MS spectrum of H-labeled compound No. 1 with that of the H-labeled Tyr-Phe standard (Figure 4d). All three fragment ions of H-labeled compound No. 1 matched those of H-labeled Tyr-Phe. Additionally, all three fragments were explained by the putative structures of H-labeled Tyr-Phe, as shown in Figure 4d. Tag information from Figure 4c was crucial in proposing these structures. The fragment with m/z = 164.1078 includes two methyl tags on the nitrogen atom, aligning with previous H/D-labeled MS/MS analysis. The fragment with m/z = 149.0841 results from removing one labeled methyl group from the m/z = 164.1078 fragment, consistent with the tag information. These findings support the identification of the unlabeled amino compound as Tyr-Phe. It is worth noting that the nitrogen in amide groups cannot be labeled because the lone-pair electrons on the amide nitrogen delocalize onto the carbonyl oxygen, forming a partial double bond,35 rendering the lone-pair electrons on amide nitrogen unavailable for the labeling reaction. Following the same processes, the other 9 compounds were confirmed and 5 tentatively identified based on their XIC, MS, and MS/MS spectra, as shown in Figures S5–S18.

Figure 4 (a) XIC of H-labeled Tyr-Phe identified in SRSRMs (black) and a H-labeled Tyr-Phe chemical standard spiked in SRSRMs (red). The pink bonds within the structure denote the labeling tags. (b) MS spectra of H/D-labeled Tyr-Phe identified in SRSRMs (black) and H/D-labeled Tyr-Phe chemical standard (red). (c) MS/MS spectra of H-labeled (black) and D-labeled (blue) Tyr-Phe in SRSRMs. (d) MS/MS spectra of H-labeled Tyr-Phe (black) identified in SRSRMs and H-labeled Tyr-Phe chemical standard (red).

3.4 Comparison of SRSRMs and NSR Samples

To compare the amino compounds in SRSRMs with those in river samples, we collected 25 authentic river water samples from the North Saskatchewan River (NSR) during the periods of February to May in both 2022 and 2023. The water samples were labeled with H/D methyl groups and analyzed using HPLC-HRMS. Using HDPairFinder, a total of 6449 H/D-labeled amino features were recognized from the water samples, and 73.0% (4705/6449) of the features were putatively annotated at the MS1 level by searching against the AMINES library. To compare the amino features between SRSRMs and NSR, we performed principal component analysis (PCA) on both the total features and high-quality features in SRSRMs, along with the features recognized in 25 NSR samples. As shown in Figure 5a, the amino features in NSR samples exhibited a small overlap with the total amino features recognized in SRSRMs and the high-quality amino features of SRSRMs. A detailed comparison between the amino features recognized in river samples and the high-quality amino features of SRSRMs revealed that 818 features matched the high-quality features identified in SRSRMs, while 5631 features were unique to the river samples (Figure 5b). Because of day-to-day variation, many low-abundance amino features in authentic river samples were not frequently detected. Therefore, we examined the features frequently detected across the 25 water samples. As depicted in Figure 5, there were 118 features present in >50% of the samples. Among the 118 features, 48 features matched the high-quality amino features present in SRSRMs, while 70 features were unique to the river samples. Upon raising the detection frequency threshold to 80% of the water samples, only 6 features were retained, and only one matched a high-quality amino feature in SRSRMs. The seven identified amino features from SRSRMs, namely, Leu-Leu, Leu, Ile, Phe, Tyr, N,4-dimethylaniline, and N-benzylglycine, were also detected in NSR and verified by aligning their XIC, MS, and MS/MS spectra with those in SRSRMs, as shown in Figures S19–S25. The retention time shifts observed between features in SRSRMs and NSR samples were attributed to the matrix effect. The results suggest the presence of some common amino compounds shared between SRSRMs and the NSR samples. However, a significant portion of amino compounds in NSR was not found in SRSRMs.

Figure 5 (a) PCA plot of the total amino features in NSR, high-quality amino features in SRSRMs, and total amino features in SRSRMs. (b) Number of amino features from NSR samples that are present or absent in the high-quality features of SRSRMs. Detection frequency refers to the percentage of samples in which a particular feature was identified out of the 25 river samples. The term “total” denotes features that were identified in at least one sample out of the 25 collected. The term “>50%” and “>80%” refers to features that were recognized in at least 50 or 80% of the 25 samples.

We further investigated the number of labeling tags in the NSR samples. As shown in Figure S26a, 55.9, 27.2, 10, and 6.9% of the features in NSR samples contained one, two, three, and four labeling tags, respectively. When compared with SRSRMs, the prevalence of one- and two-tag features in NSR samples exceeded that in SRNOM and SRHA, but it was comparable to that in SRFA. We then examined the mass distribution of the NSR features with different numbers of tags, as shown in Figure S26b. The median m/z values for the one-, two-, three-, and four-tag features in NSR were 316.15, 320.20, 322.20, and 311.11, respectively. In contrast to SRSRMs, the features with one and two tags in NSR did not demonstrate a larger molecular weight than those with three and four tags. Most amino features in NSR fell within the mass range of 300–400, which is greater than the majority of amino features in SRSRMs, where the mass range was between 200 and 300. Based on the above comparison, NSR samples exhibited a distinct composition of amino compounds compared to those found in SRSRMs, attributed to the intrinsic differences in organic amines between the two river systems. Variations in amino precursor composition between SRSRMs and authentic river water samples can result in different N-DBP formations with varying concentrations during the water treatment process. Therefore, distinct treatment strategies may be needed for different river systems.

4 Implication

Small water-soluble amino compounds present in source water pose challenges in removal during the water treatment process.36 More significantly, these amino compounds can serve as potential precursors of toxic N-DBPs. SRSRMs are frequently used as a standard humic substance for the study of the water treatment process and DBP formation.9−11 While many studies have characterized SRSRMs as bulk analysis, few of them have investigated the amino compounds in these reference materials at the molecular level. This study demonstrated stable HD isotopic labeling that can specifically methylate the amine groups using both formaldehyde and deuterated formaldehyde in SRSRMs and in source water. The HPLC-HRMS combined with HDPairFinder enabled the identification of labeled amino compounds. This unique tool reveals the differences and similarities of labeled amino compounds in SRSRMs and source water samples, representing different N-precursors existing in different water systems. This study demonstrated the capabilities of this unique tool for characterizing amino organics and monitoring them in various source waters. The knowledge of amino chemicals in SRSRMs and source waters will enable the future development of technologies for the removal of small reactive organic amine precursors and identification of currently unknown N-DBPs. The comprehensive analysis will provide insights into understanding and managing N-DBP formation in water treatment processes.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.4c05165.Additional information on data processing and data cleaning; comparison between SRSRMs; identification of amino compounds in SRSRMs; amino compounds identified in both SRSRMs and NSR, and labeling tag and mass distribution of amino features in NSR (PDF)

Annotation of amino compounds in SRSRMs (XLSX)

Annotation of amino compounds in NSR (XLSX)

Supplementary Material

es4c05165_si_001.pdf

es4c05165_si_002.xlsx

es4c05165_si_003.xlsx

The authors declare no competing financial interest.

Acknowledgments

The authors acknowledge the funding support by Alberta Innovates, the Canada Research Chair Program (Li), and the Natural Sciences and Engineering Research Council of Canada.
==== Refs
References

Zhang D. ; Chu W. ; Yu Y. ; Krasner S. W. ; Pan Y. ; Shi J. ; Yin D. ; Gao N. Occurrence and stability of chlorophenylacetonitriles: a new class of nitrogenous aromatic DBPs in chlorinated and chloraminated drinking waters. Environ. Sci. Technol. Lett. 2018, 5 (6 ), 394–399. 10.1021/acs.estlett.8b00220.
Shah A. D. ; Krasner S. W. ; Lee C. F. T. ; von Gunten U. ; Mitch W. A. Trade-offs in disinfection byproduct formation associated with precursor preoxidation for control of N-nitrosodimethylamine formation. Environ. Sci. Technol. 2012, 46 (9 ), 4809–4818. 10.1021/es204717j.22463122
Ding S. ; Chu W. Recent advances in the analysis of nitrogenous disinfection by-products. Trends Environ. Anal. Chem. 2017, 14 , 19–27. 10.1016/j.teac.2017.04.001.
Li X.-F. ; Mitch W. A. Drinking water disinfection byproducts (DBPs) and human health effects: multidisciplinary challenges and opportunities. Environ. Sci. Technol. 2018, 52 (4 ), 1681–1689. 10.1021/acs.est.7b05440.29283253
Shah A. D. ; Mitch W. A. Halonitroalkanes, halonitriles, haloamides, and N-nitrosamines: a critical review of nitrogenous disinfection byproduct formation pathways. Environ. Sci. Technol. 2012, 46 (1 ), 119–131. 10.1021/es203312s.22112205
Allen J. M. ; Plewa M. J. ; Wagner E. D. ; Wei X. ; Bokenkamp K. ; Hur K. ; Jia A. ; Liberatore H. K. ; Lee C.-F. T. ; Shirkhani R. ; et al. Drivers of disinfection byproduct cytotoxicity in US drinking water: should other DBPs be considered for regulation?. Environ. Sci. Technol. 2022, 56 (1 ), 392–402. 10.1021/acs.est.1c07998.34910457
Le Roux J. ; Plewa M. J. ; Wagner E. D. ; Nihemaiti M. ; Dad A. ; Croué J.-P. Chloramination of wastewater effluent: Toxicity and formation of disinfection byproducts. J. Environ. Sci. 2017, 58 , 135–145. 10.1016/j.jes.2017.04.022.
Plewa M. J. ; Wagner E. D. ; Muellner M. G. ; Hsu K.-M. ; Richardson S. D. Comparative Mammalian Cell Toxicity of N-DBPs and C-DBPs; ACS Publications, 2008.
Han J. ; Zhang X. ; Jiang J. ; Li W. How much of the total organic halogen and developmental toxicity of chlorinated drinking water might be attributed to aromatic halogenated DBPs?. Environ. Sci. Technol. 2021, 55 (9 ), 5906–5916. 10.1021/acs.est.0c08565.33830743
Watson K. ; Farré M. J. ; Knight N. Comparing three Australian natural organic matter isolates to the Suwannee River standard: Reactivity, disinfection by-product yield, and removal by drinking water treatments. Sci. Total Environ. 2019, 685 , 380–391. 10.1016/j.scitotenv.2019.05.416.31176223
Zhang X. ; Minear R. A. Formation, adsorption and separation of high molecular weight disinfection byproducts resulting from chlorination of aquatic humic substances. Water Res. 2006, 40 (2 ), 221–230. 10.1016/j.watres.2005.10.024.16343584
McDonald S. ; Pringle J. M. ; Bishop A. G. ; Prenzler P. D. ; Robards K. Isolation and seasonal effects on characteristics of fulvic acid isolated from an Australian floodplain river and billabong. J. Chromatogr. A 2007, 1153 (1–2 ), 203–213. 10.1016/j.chroma.2006.08.086.17010354
Driver S. J. ; Perdue E. M. Acidic Functional Groups of Suwannee River Natural Organic Matter, Humic Acids, and Fulvic Acids. In Advances in the Physicochemical Characterization of Dissolved Organic Matter: Impact on Natural and Engineered Systems; ACS Publications, 2014; pp 75–86.
Senesi N. ; Miano T. ; Provenzano M. R. ; Brunetti G. Spectroscopic and compositional comparative characterization of IHSS reference and standard fulvic and humic acids of various origin. Sci. Total Environ. 1989, 81-82 , 143–156. 10.1016/0048-9697(89)90120-4.
Thorn K. A. ; Folan D. W. ; MacCarthy P. Characterization of the International Humic Substances Society Standard and Reference Fulvic and Humic Acids by Solution State Carbon-13 (13C) and Hydrogen-1 (1H) Nuclear Magnetic Resonance Spectrometry; U.S. Dept. of the Interior, U.S. Geological Survey: Books and Open-File Reports Section: Denver, CO, 1989.
Langlais B. ; Reckhow D. A. ; Brink D. R. Ozone in Water Treatment. In Application and Engineering; Taylor & Francis, 1991; Vol. 558 .
Smith D. F. ; Podgorski D. C. ; Rodgers R. P. ; Blakney G. T. ; Hendrickson C. L. 21 T FT-ICR mass spectrometer for ultrahigh-resolution analysis of complex organic mixtures. Anal. Chem. 2018, 90 (3 ), 2041–2047. 10.1021/acs.analchem.7b04159.29303558
Lu K. ; Gardner W. S. ; Liu Z. Molecular structure characterization of riverine and coastal dissolved organic matter with ion mobility quadrupole time-of-flight LCMS (IM Q-TOF LCMS). Environ. Sci. Technol. 2018, 52 (13 ), 7182–7191. 10.1021/acs.est.8b00999.29870664
Fievre A. ; Solouki T. ; Marshall A. G. ; Cooper W. T. High-resolution Fourier transform ion cyclotron resonance mass spectrometry of humic and fulvic acids by laser desorption/ionization and electrospray ionization. Energy Fuels 1997, 11 (3 ), 554–560. 10.1021/ef970005q.
Reemtsma T. ; These A. ; Springer A. ; Linscheid M. Differences in the molecular composition of fulvic acid size fractions detected by size-exclusion chromatography–on line Fourier transform ion cyclotron resonance (FTICR−) mass spectrometry. Water Res. 2008, 42 (1–2 ), 63–72. 10.1016/j.watres.2007.06.063.17640699
Kim D. ; Kim S. ; Son S. ; Jung M.-J. ; Kim S. Application of online liquid chromatography 7 T FT-ICR mass spectrometer equipped with quadrupolar detection for analysis of natural organic matter. Anal. Chem. 2019, 91 (12 ), 7690–7697. 10.1021/acs.analchem.9b00689.31117404
Zhang X. ; Han J. ; Zhang X. ; Shen J. ; Chen Z. ; Chu W. ; Kang J. ; Zhao S. ; Zhou Y. Application of Fourier transform ion cyclotron resonance mass spectrometry to characterize natural organic matter. Chemosphere 2020, 260 , 127458 10.1016/j.chemosphere.2020.127458.32693253
Reemtsma T. Determination of molecular formulas of natural organic matter molecules by (ultra-) high-resolution mass spectrometry: status and needs. J. Chromatogr. A 2009, 1216 (18 ), 3687–3701. 10.1016/j.chroma.2009.02.033.19264312
Hua Z. ; Li J. ; Zhou Z. ; Zheng S. ; Zhang Y. ; Fang J. Exploring pathways and mechanisms for dichloroacetonitrile formation from typical amino compounds during UV/chlorine treatment. Environ. Sci. Technol. 2022, 56 (13 ), 9712–9721. 10.1021/acs.est.2c01495.35703371
Sheng D. ; Bu L. ; Zhu S. ; Deng L. ; Shi Z. ; Zhou S. Transfer organic chloramines to monochloramine using two-step chlorination: A method to inhibit N-DBPs formation in algae-containing water treatment. J. Hazard. Mater. 2023, 443 , 130343 10.1016/j.jhazmat.2022.130343.36444058
Rao N. R. H. ; Linge K. ; Li X. ; Joll C. ; Khan S. ; Henderson R. Relating algal-derived extracellular and intracellular dissolved organic nitrogen with nitrogenous disinfection by-product formation. Water Res. 2023, 233 , 119695 10.1016/j.watres.2023.119695.36827767
Liu Z. ; Craven C. B. ; Huang G. ; Jiang P. ; Wu D. ; Li X.-F. Stable isotopic labeling and nontarget identification of nanogram/liter amino contaminants in water. Anal. Chem. 2019, 91 (20 ), 13213–13221. 10.1021/acs.analchem.9b03642.31498582
Zhao T. ; Carroll K. ; Craven C. B. ; Wawryk N. J. ; Xing S. ; Guo J. ; Li X.-F. ; Huan T. HDPairFinder: A data processing platform for hydrogen/deuterium isotopic labeling-based nontargeted analysis of trace-level amino-containing chemicals in environmental water. J. Environ. Sci. 2024, 136 , 583–593. 10.1016/j.jes.2023.02.033.
International Humic Substances Society Elemental Compositions and Stable Isotopic Ratios of IHSS Samples. https://humic-substances.org/elemental-compositions-and-stable-isotopic-ratios-of-ihss-samples/. (Feb 10).
Thurman E. M. ; Malcolm R. L. Preparative isolation of aquatic humic substances. Environ. Sci. Technol. 1981, 15 (4 ), 463–466. 10.1021/es00086a012.22248415
Serkiz S. M. ; Perdue E. M. Isolation of dissolved organic matter from the Suwannee River using reverse osmosis. Water Res. 1990, 24 (7 ), 911–916. 10.1016/0043-1354(90)90142-S.
Sun L. ; Perdue E. ; McCarthy J. Using reverse osmosis to obtain organic matter from surface and ground waters. Water Res. 1995, 29 (6 ), 1471–1477. 10.1016/0043-1354(94)00295-I.
Chin Y.-P. ; Aiken G. R. ; Danielsen K. M. Binding of pyrene to aquatic and commercial humic substances: the role of molecular weight and aromaticity. Environ. Sci. Technol. 1997, 31 (6 ), 1630–1635. 10.1021/es960404k.
Schellekens J. ; Buurman P. ; Kalbitz K. ; Zomeren Av. ; Vidal-Torrado P. ; Cerli C. ; Comans R. N. Molecular features of humic acids and fulvic acids from contrasting environments. Environ. Sci. Technol. 2017, 51 (3 ), 1330–1339. 10.1021/acs.est.6b03925.28102075
Pace V. ; Holzer W. ; Olofsson B. Increasing the reactivity of amides towards organometallic reagents: an overview. Adv. Synth. Catal. 2014, 356 (18 ), 3697–3736. 10.1002/adsc.201400630.
Craven C. B. ; Wawryk N. J. ; Carroll K. ; James W. ; Shu Z. ; Charrois J. W. ; Hrudey S. E. ; Li X.-F. Amino Acids as Potential Precursors to Odorous Compounds in Tap Water during Spring Runoff Events. Environ. Sci. Technol. 2023, 57 (47 ), 18765–18774. 10.1021/acs.est.3c00719.37549310
