Tandem mass spectrometry of small-molecule antiviral drugs: 3. antiviral agents against herpes, influenza and other viral infections

Abstract

For the treatment of various viral infections, antiviral drugs may be used. Liquid chromatography–mass spectrometry (LC–MS) with tandem mass spectrometry (MS–MS) operated in selected-reaction monitoring (SRM) mode is the method of choice in quantitative bioanalysis of drugs, e.g., to establish bioavailability, to study pharmacokinetics, and later on possibly for therapeutic drug monitoring. In this study, the fragmentation in MS–MS of small-molecule antiviral drugs against herpes and influenza viruses is reviewed. In this way, insight is gained on the identity of the product ions used in SRM. Fragmentation schemes of antiviral agents are also relevant in the identification of drug metabolites or (forced) degradation products. As information of the fragmentation of antiviral drugs in MS–MS and the identity of the product ions is very much scattered in the scientific literature, it was decided to collect this information and to review it. In this third study, attention is paid to small-molecule antiviral agents used against herpes and influenza virus infections. In addition, some attention is paid to broad-spectrum antiviral agents, that are investigated with respect to their efficacy in challenging virus infections of this century, e.g., involving Ebola, Zika and corona viruses, like SARS-CoV-2, which is causing a world-wide pandemic at this very moment. The review provides fragmentation schemes of ca. 35 antiviral agents. The identity of the product ions used in SRM, i.e., elemental composition and exact-m/z, is tabulated, and more detailed fragmentation schemes are provided.

Graphical abstract

Highlights

• •Detailed study on the fragmentation of 15 Herpes-related and 9 influenza A-related antiviral drugs in MS–MS.
• •MS–MS of nucleoside analogues and neuraminidase inhibitors.
• •Tabulated product ions (with elemental composition) used in SRM.
• •Fragmentation scheme, confirmed using accurate-m/z data.
• •Broad-spectrum antiviral agents.

1. Introduction

In the third part of this series of review papers (see Refs. [1,2] for previous parts), attention is paid to the fragmentation in MS–MS of antiviral agents against herpes and influenza viruses. Thus, structures of relevant compounds were searched on the internet. Subsequently, data on product ions used in selected-reaction monitoring (SRM) for quantitative bioanalysis and MS–MS spectra of these compounds were collected and interpreted. This information is provided in the text below.

Human herpes viruses comprise a large family of viruses, containing double-stranded DNA, some of which are widespread among humans [3]. Clinical manifestation of viral infection vary widely. The herpes family is divided in three subfamilies: α-, β-, and γ-Herpesvirinae. The α-class includes the Herpes simplex viruses type 1 and 2 (HSV-1 and HSV-2) and the varicella zoster virus (VZV; the cause of chickenpox). The β-class includes the human cytomegalovirus (CMV), the human herpes virus 6 (HHV6), and human herpes virus 7 (HHV7). The γ-class includes the Epstein-Barr virus (EBV), the human herpes virus 4 (HHV4), and Kaposi’s sarcoma-associated herpes virus (KSHV; HHV8). A variety of compounds are licensed for use as antiviral agents or are currently under development. Most of these compounds are nucleoside analogues.

There are several types of the influenza virus that affect humans [4,5]. Influenza A and B viruses contain negative-sense single-stranded RNA. Usually, the virus is spread through the air by aerosols from coughs or sneezes. Influenza spreads around the world in yearly outbreaks, which differ in severity, generally with a substantial number of deaths. Larger outbreaks of pandemic nature are less frequent. Well-known influenza pandemics, all related to the Influenza A virus, are the Spanish influenza (H1N1) in 1918, the Asian influenza (H2N2) in 1957, the Hong Kong influenza (H3N2) in 1968, and the Mexican influenza (H7N9) in 2018. Influenza may also affect animals like pigs and birds, e.g., the outbreak of bird flu (H5N1) in 2004. Influenza vaccines are available, but not always successful due to mutations in the virus. The most common treatment of influenza is the use of acetaminophen to relieve fever and muscle pain associated with the disease. There are two classes of antiviral drugs used against influenza, the antagonists of the influenza virus A Matrix-2 proton channel (adamantane derivatives) and the neuraminidase (sialidase) inhibitors (oseltamivir, zanamivir, laninamivir and peramivir).

A multiresidue analysis of 14 anti-influenza and anti-herpes drugs in chicken muscle has been reported [6]. The analytes included anti-influenza drugs (amantadine, rimantadine, oseltamivir, oseltamivir carboxylate, memantine, arbidol, and moroxydine), anti-herpes drugs (acyclovir, ganciclovir, famciclovir, penciclovir, ribavirin and its main metabolite TCONH2), and an immunomodulator (imiquimod). Other multiresidue studies of antiviral agents in chicken meat have been reported as well [7,8].

During the preparation of these review papers, which started in October 2019, the world was struck by a pandemic due to a new corona virus, the so-called severe acute respiratory syndrome (SARS) corona virus 2 (SARS-CoV-2). Therefore, a short section was added at the end of this third part.

2. Methods

The m/z data to interpret the fragmentation in MS–MS, to derive elemental composition and exact-m/z for the product ions used in SRM and to develop the fragmentation schemes, have been collected by searching the literature using the PubMed search website with search terms like “compound name MS” or “compound name metabolites”. The papers found in this way were screened for relevant MS–MS data, i.e., MS–MS spectra, (tabulated) m/z-values for SRM transitions, and, when available, structure proposals for the product ions. Special attention was paid to finding accurate-m/z data, mostly from literature dealing with identification of metabolites or (forced) degradation products. For some compounds, accurate-m/z data were available in an available MS–MS spectral library for toxicological unknown screening [9].

Generally, the interpretation of the MS–MS spectra is initially performed using nominal m/z values of the product ions, based on logical neutral losses from the structure of the compound [10]. No software tools were applied other than a structure drawing software (ACD/ChemSketch software; version 2018.1.1; www.acdlabs.com), which was set to automatically calculate exact-m/z values. The elemental compositions and exact-m/z values of the proposed structures are then verified against and/or confirmed by the accurate-m/z data. If in this way no structure proposal is reached for a particular product ion, the process is reversed: a plausible structure proposal is derived from elemental compositions calculated from the accurate-m/z data. Accurate-m/z data of product ions can be used to calculate their possible elemental composition. Such calculations were performed using a ±3 mDa window, which generally (but not always) is appropriate to the available data. With respect to the product ions in an MS–MS spectrum, the elemental composition of a product ion is restricted by the elemental composition of the precursor ion (except in some rare, specific cases, when ion-trap instruments are used [11]). This greatly reduces the number of possible hits. Thus, the proposed structures of the product ions have been evaluated, at least in terms of elemental composition and logical neutral losses [10]. Whenever needed, names of product ions or neutral losses were generated with the “Generate Name for Structure” option available in the ACD/ChemSketch software. If the elemental composition derived from available accurate-m/z data is not in agreement with the elemental composition of the literature-proposed structure of a product ion, alternative structures are proposed for such product ions; this discrepancies have been discussed in the text.

Throughout the text, the m/z values, when expressed as accurate-m/z (measured) or exact-m/z (calculated), are given with three significant digits. With the current mass accuracy of typically 1–3 ppm of high-resolution MS instruments and ions with m/z below 1000, the error is in the third decimal place. As SRM transitions in the Tables have been acquired on different instruments from various manufacturers, it was decided not to add information on the collision energy applied, as a particular set collision energy value may yield different levels of fragmentation between instruments from different manufacturers. It must be added, that information on the applied collision energy in SRM and/or MS–MS spectrum acquisition is often missing in the literature studied for this review.

3. Antivirals against Herpes viruses

In this section, the MS–MS fragmentation of the antiviral agents against Herpes viruses is reviewed. As ribavirin is also used in HCV treatment, it was already discussed in Part 2 [2]. Characteristic SRM transitions used in their analysis in biological matrices are collected in Table 1a, Table 1b; its fragmentation scheme can be found elsewhere [2]. Some compounds are analyzed in the positive-ion mode (Table 1 a), some others in negative-ion mode (Table 1 b), whereas a few compounds can be analyzed in either positive-ion or negative-ion mode. No LC–MS data were found for foscarnet.

Table 1a.

SRM transitions for antiviral drugs against Herpes viruses in positive-ion mode.

Compound	m/z of [M+H]+	Formula	m/z of SRM product-ion	Formula	Literature
aciclovir/acyclovir	226.093	[C8H12N5O3]+	152.0577	[C5H6N5]+	[[6], [7], [8],[12], [13], [14]]
			135.030	[C5H3N4O]+	[6,8,13]
cidofovir	280.069	[C8H14N3O6P]+	282.059	[C8H13N3O5P]+	[15]
famciclovir	322.151	[C14H20N5O4]+	280.140	[C12H18N5O3]+	[6]
			136.062	[C5H6N5]+	[6]
ganciclovir	256.104	[C9H14N5O4]+	152.0577	[C5H6N5]+	[6,[12], [13], [14],16]
			135.030	[C5H3N4O]+	[6,13,17]
idoxuridine	354.979	[C9H12IN2O5]+	117.055	[C5H9O3]+	[18]
letermovir	573.212	[C29H29F4N4O4]+	383.101	[C18H15F4N2O3]+	[18]
penciclovir	254.125	[C10H16N5O3]+	152.0577	[C5H6N5]+	[6,20,21]
			135.030	[C5H3N4O]+	[6]
pritelivir	403.089	[C18H19N4O3S2]+	196.076	[C13H10NO]+	[22,23]
ribavirin	245.088	[C8H12N4O5]+	113.046	[C3H5N4O]+	[6,8,24,25]
			96.019	[C3H2N3O]+	[6,8]
ribavirin – Met: TCONH2, triazole carboxamide	113.046	[C3H5N4O]+	96.019	[C3H2N3O]+	[6]
			69.008	[C2HN2O]+	[6]
trifluridine	297.069	[C10H12F3N2O5]+	117.055	[C5H9O3]+	[18]
valaciclovir	325.162	[C13H21N6O4]+	152.0577	[C5H6N5]+	[14,26]
valganciclovir	355.172	[C12H23N6O5]+	152.0577	[C5H6N5]+	[14,17]
vidarabine	268.104	[C10H14N5O4]+	136.062	[C5H6N5]+	[27]
viramidine/taribavirin	244.104	[C8H14N5O4]+	112.062	[C3H6N5]+	[28]

Table 1b.

SRM transitions for antiviral drugs against Herpes viruses in negative-ion mode.

Compound	m/z of [M-H]–	Formula	m/z of SRM product-ion	Formula	Literature
brivudine	330.994	[C11H12B5N2O5]–	78.919	[Br]–	[29]
cidofovir	278.055	[C8H12N3O6P]–	237.065	[C7H14N2O5P]–	[30]
idoxuridine	352.964	[C9H10IN2O5]–	126.905	[I]–	[18]
trifluridine	295.055	[C10H10F3N2O5]–	179.007	[CH2F3N2O2]–

3.1. Nucleoside analogues

Most acyclic nucleoside analogues used against Herpes viruses are guanine-related compounds. Aciclovir ([M+H]⁺ with m/z 226), ganciclovir ([M+H]⁺ with m/z 256), penciclovir ([M+H]⁺ with m/z 254), valaciclovir ([M+H]⁺ with m/z 325), and valganciclovir ([M+H]⁺ with m/z 355) all show protonated guanine as a common product ion with m/z 152 ([C₅H₆N₅O]⁺), due to the loss of guanine N ⁹-substituent (Fig. 1 ). For valaciclovir and valganciclovir, the complementary ions due to the loss of the guanine base are observed as well, i.e., the ions with m/z 174 ([C₈H₁₆NO₃]⁺) and with m/z 204 ([C₉H₁₈NO₄]⁺), respectively. Cleavage of the carbonyl C–C bond to the aminoalkyl group results in protonated 2-methylpropan-1-imine with m/z 72 ([C₄H₁₀N]⁺) for both valaciclovir and valganciclovir (Fig. 1). Valaciclovir and valganciclovir are examples of amino-acid prodrugs of nucleoside antivirals. In this context, other analogues are under investigation, like valomaciclovir, valcyclopropavir, valtorcitabine, valopicitabine [31].

Fig. 1.

Structures and fragmentation schemes, indicating the major product ions in the positive-ion mode, of the acyclic nucleoside analogues used against Herpes viruses.

Several carbocyclic nucleoside analogues are used in the treatment of Herpes infections. They can often be analyzed in both positive-ion and negative-ion mode.

Famciclovir ([M+H]⁺ with m/z 322) shows product ions with m/z 280 due to the loss of ethenone (C₂H₂O), m/z 262 due to the loss of acetic acid (CH₃COOH), which in turn leads to ions with m/z 220 and 202 due to subsequent loss of C₂H₂O and CH₃COOH, respectively. The ion with m/z 136 is the protonated 2-aminopurine ([C₅H₆N₅]⁺), which is due to the loss of the N ⁹-substituent (Fig. 2 ). Penciclovir, discussed earlier, is the active metabolite of famciclovir.

Fig. 2.

Structures and fragmentation schemes, indicating the major product ions in positive-ion and/or negative-ion mode, of the carbocyclic nucleoside analogues used against Herpes viruses.

Cidofovir can be analyzed in both positive-ion and negative-ion mode. In positive-ion mode for cidofovir ([M+H]⁺ with m/z 280), the product ion with m/z 262 due to the loss of water is used in SRM. Based on similar structures, e.g., lamivudine and zalcitabine, a product ion with m/z 112 ([C₄H₆N₃O]⁺) is most likely observed for cidofovir as well. In negative-ion mode, cidofovir ([M−H]^– with m/z 278) shows a product ion with m/z 235, which was suggested to be due to the loss of methanediimine (CH₂N₂) [30], but the loss of hydrogen isocyanate (HNCO) is more likely. Cleavage of N¹–C bond to the cytosine substituent with charge retention of the phosphate-containing moiety results in the ion with m/z 167, and subsequent loss of propenal (C₃H₄O) leads to the ion with m/z 110 (Fig. 2). Brindicofovir ([C₂₇H₅₃N₃O₇P]⁺; [M+H]⁺ with m/z 562) is under investigation as a prodrug for cidofovir.

For vidarabine, trifluiridine, brivudine, and idoxuridine, which are nucleoside reverse transcriptase inhibitors (NRTIs), the loss of the N-substituent is observed in positive-ion MS–MS. This leads to the protonated 6-aminopurine or adenine ([C₅H₆N₅]⁺) with m/z 136 for vidarabine ([M+H]⁺ with m/z 268), to the protonated 5-trifluoromethyl uracil with m/z 181 for trifluridine ([M+H]⁺ with m/z 297) and to the protonated 5-iodouracil ([C₄H₄IN₂O₂]⁺) with m/z 239 for idoxuridine ([M+H]⁺ with m/z 355) (Fig. 2). However, for both trifluridine and idoxuridine, the 4-hydroxy-5-(hydroxymethyl)oxolan-2-ylium ion ([C₅H₉O₃]⁺) with m/z 117 is also observed, next to two consecutive water losses [18].

In negative-ion MS–MS, trifluridine ([M−H]^– with m/z 295) shows the loss of the sugar unit to the deprotonated 5-(trifluoromethyl)uracil ([C₅H₂F₃N₂O₂]^–) with m/z 179 and an ion with m/z 42 ([NCO]^–), whereas for brivudine ([M−H]^– with m/z 331) and idoxuridine ([M−H]^– with m/z 353), a minor product ion due to the loss of the dideoxyribose unit (C₅H₈O₃) is observed to the deprotonated 5-(2-bromovinyl)uracil ([C₆H₄BrN₂O₂]^–) and 5-iodouracil ([C₄H₂IN₂O₂]^–) with m/z 215 and 237, respectively. However, the major product ions are bromide (Br⁻) with m/z 79 and iodide (I⁻) with m/z 127 (Fig. 2). Under ion-trap MS² conditions, idoxuridine also shows the loss of hydrogen isocyanate (HNCO) to an ion with m/z 310 due to a retro-Diels-Alder fragmentation and an ion with m/z 42 ([NCO]^–) [18].

3.2. Other antiviral agents against Herpes viruses

Two other compounds under development as Herpes antivirals are letermovir, targeted at CMV infections, and pritelivir, an antiviral helicase-primase inhibitor. Limited information of MS–MS behavior is available.

Letermovir ([C₂₉H₂₉F₄N₄O₄]⁺; [M+H]⁺ with m/z 573) shows two complementary product ions with m/z 383 ([C₁₈H₁₅F₄N₂O₃]⁺) and m/z 191 ([C₁₁H₁₅N₂O]⁺) due to the cleavage of the C–N bond between the 1,6-dihydropyrimidine and the piperazine moieties (Fig. 3 ) [19]. Pritelivir ([C₁₈H₁₉N₄O₃S₂]⁺; [M+H]⁺ with m/z 403) shows an acylium ion with m/z 196 ([C₁₃H₁₀NO]⁺) due to the cleavage of the amide bond (Fig. 3) [22,23].

Fig. 3.

Structures and the major product ions in positive-ion mode of letermovir and pritelivir used against Herpes viruses.

4. Antiviral agents against influenza viruses

In this section, the MS–MS fragmentation of the antiviral agents against influenza viruses is reviewed. There are two classes of antiviral drugs used against influenza, the antagonists of the influenza virus A Matrix-2 proton channel (adamantane derivatives) and the neuraminidase (sialidase) inhibitors (oseltamivir, zanamivir, laninamivir and peramivir). All compounds are analyzed in positive-ion mode. Characteristic SRM transitions used in their analysis in biological matrices are collected in Table 2 .

Table 2.

SRM transitions for antiviral drugs against influenza in positive-ion mode.

Compound	m/z of [M+H]+	Formula	m/z of SRM product-ion	Formula	Literature
amantadine	152.143	[C10H18N]+	135.117	[C10H15]+	[[6], [7], [8],[32], [33], [34]]
			93.070	[C7H9]+	[[8], [8],34]
			79.054	[C6H7]+	[32,33]
laninamivir	347.156	[C13H22N4O7]+	60.056	[CH6N3]+	[35]
moroxydine	172.119	[C6H14N5O]+	130.097	[C5H12N3O]+	[6,7]
			113.071	[C5H9N2O]+	[[4], [5], [6], [7], [8]]
			69.045	[C3H5N2]+	[8]
oseltamivir	313.212	[C16H29N2O4]+	225.123	[C11H17N2O3]+	[6,7,36]
			166.086	[C9H12NO2]+	[6,7]
oseltamivir carboxylate	285.181	[C14H25N2O4]+	138.055	[C7H8NO2]+	[6]
			197.092	[C8H13N2O3]+	[6,36]
peramivir	329.218	[C15H29N4O4]+	270.170	[C14H24NO4]+	[[37], [38], [39]]
			100.112	[C6H14N]+	[40]
rimantadine	180.175	[C12H22N]+	163.148	[C12H19N]+	[[6], [7], [8],33]
			135.117	[C10H15]+	[34]
			121.101	[C9H13]+	[7]
			93.070	[C7H9]+	[33,34]
			81.070	[C6H9]+	[6,8]
umifenovir (arbidol)	477.084	[C22H26BrN2O3S]+	432.026	[C20H19BrNO3S]+	[4]
			278.988	[C12H10BrNO2]+•	[4]
	479.082	[C22H2681BrN2O3S]+	434.024	[C20H1981BrNO3S]+	[41]
			280.986	[C12H1081BrNO2]+•	[42]
zanamivir	333.140	[C12H21N4O7]+	60.056	[CH6N3]+	[[43], [44], [45]]
			121.028	[C7H5O2]+	[45]

4.1. M2 protein inhibitors against influenza A

Amantadine and rimantadine are inhibitors of the Matrix-2 protein, a proton-selective viroprotein, integral in the viral envelope of the influenza A virus. They have been used against influenza A, but because of their antagonistic properties to the N-methyl-d-aspartate receptor (NMDAR), they are also used in the treatment of parkinsonism and multiple sclerosis (like the structural analogue memantine). In order to prevent potential resistance issues in human, the use of amantadine and rimantadine as antiviral agents during poultry farming has been banned in many countries. Therefore, sensitive methods for the analysis of these antiviral agents in chicken muscle tissues are needed [33,34].

Amantadine ([M+H]⁺ with m/z 152) shows the loss of ammonia (NH₃) to an ion with m/z 135, followed by cleavages of the adamantyl skeleton, e.g., to ions with m/z 107 ([C₈H₁₁]⁺), m/z 93 ([C₇H₉]⁺), m/z 79 ([C₆H₇]⁺), m/z 67 ([C₅H₇]⁺), and m/z 55 ([C₄H₇]⁺) (Fig. 4 ) [46]. The identity of these product ions has been confirmed by accurate-m/z data [9].

Fig. 4.

Structures and the major product ions in positive-ion mode of amantadine and rimantadine used against Influenza A viruses.

Rimantadine ([M+H]⁺ with m/z 180) shows the loss of ammonia (NH₃) to an ion with m/z 165, followed by cleavages of the adamantyl skeleton, e.g., to ions with m/z 135 ([C₁₀H₁₅]⁺), m/z 121 ([C₉H₁₃]⁺), m/z 107 ([C₈H₁₁]⁺), m/z 95 ([C₇H₁₁]⁺), m/z 93 ([C₇H₉]⁺), and m/z 81 ([C₆H₉]⁺) (Fig. 4) [33].

4.2. Neuraminidase inhibitors against influenza A

The initial fragmentation of the neuraminidase inhibitors laninamivir, peramivir, and zanamivir is the loss of 59 Da, which can either involve the loss of guanidine (CH₅N₃) or of acetamide (C₂H₅NO), and the formation of the protonated guanidine with m/z 60 ([CH₆N₃]⁺) or protonated acetamide with m/z 60 ([C₂H₆NO]⁺). Fragmentation of [¹³C₁ ¹⁵N₂]-zanamivir (labeling at the guanidine moiety, as indicated with ∗ in Fig. 6) indicates the loss of guanidine [43], as also confirmed by accurate-m/z data, available for zanamivir [47]. Either the ion with m/z 60 or the product ion resulting from the loss of guanidine is generally used in SRM. Thus, laninamivir ([M+H]⁺ with m/z 347) shows a product ion with m/z 60, whereas peramivir ([M+H]⁺ with m/z 329) shows a product ion with m/z 270 due to the loss of guanidine.

Fig. 6.

Structures and the major product ions in positive-ion mode of peramivir and zanamivir used against Influenza A viruses. The labeling position in [¹³C₁¹⁵N₂]-zanamivir, used as internal standard, is indicated with ∗.

Oseltamivir (Tamiflu; [M+H]⁺ with m/z 313) shows product ions with m/z 296 due the loss of ammonia (NH₃), with m/z 243 due to the loss pentene (C₅H₁₀) to an ion, with m/z 225 due to the loss of 3-pentanol (C₅H₁₂O), with m/z 208 and m/z 166 due to the subsequent loss of ammonia or acetamide (C₂H₅NO) (Fig. 5 ) [36,48]. Its carboxylate (desethyl) metabolite shows similar product ions with a −28-Da shift (Fig. 5).

Fig. 5.

Structures and the major product ions in positive-ion mode of oseltamivir and its carboxylate (desethyl) metabolite used against Influenza A viruses.

Peramivir ([M+H]⁺ with m/z 329) shows a product ion with m/z 270 due to the loss of guanidine (CH₅N₃); the protonated guanidine with m/z 60 is also observed (Fig. 6 ). Further fragmentation of the ion with m/z 270 involves the subsequent losses of water to m/z 252, of ethenone (C₂H₂O) to m/z 210, of ammonia (NH₃) to m/z 183, and of either carbon dioxide (CO₂) or formic acid (CH₂O₂) to the ions with m/z 149 and m/z 147, respectively. Another fragmentation route involves the cleavage of a C–C bond between the cyclopentane moiety and the tertiary C-atom, resulting in an ion with m/z 142 ([C₈H₁₆NO]⁺), which shows secondary fragmentation to the ion with m/z 100 due to the loss of ethenone and with m/z 83 due to the subsequent loss of ammonia (Fig. 6).

The fragmentation of zanamivir is studied in detail in the course of an environmental photodegradation study [47]. Zanamivir ([M+H]⁺ with m/z 333) shows a product ion with m/z 274 due to the loss of guanidine (CH₅N₃); the protonated guanidine with m/z 60 is also observed. The ion with m/z 274 shows subsequent fragmentation to m/z 215 due to the loss of acetamide (C₂H₅NO), which in turn shows three consecutive water losses to the ions with m/z 197, 179, and 151. Alternatively, the ion with m/z 274 shows consecutive losses of water, ethenone (C₂H₂O), ammonia (NH₃), formaldehyde (CH₂O), and formic acid (CH₂O₂) to the ions with m/z 256, 214, 197, 167, and 121, respectively (Fig. 6). The elemental compositions of the product ions in confirmed by accurate-m/z data [47].

4.3. Other antiviral agents against influenza A viruses

Umifenovir (arbidol) is only approved for use against influenza in Russia and China. It inhibits membrane fusion, and thereby prevents contact between the virus and target host cells. Umifenovir ([M+H]⁺ with m/z 477) shows product ions with m/z 432 due to the loss of dimethylamine ((H₃C)₂NH), with m/z 387 due to the subsequent loss of ethanal (C₂H₄O), with m/z 323 due to the loss of dimethylamine and a phenylsulfanyl radical (C₆H₅S^•), with m/z 294 and with m/z 279 due to the subsequent loss of an ethyl radical (^•C₂H₅) or ethanal, respectively (Fig. 7 ). The fragmentation is confirmed by accurate-m/z data, acquired in the course of a metabolite identification study [49].

Fig. 7.

Structures and the major product ions in positive-ion mode of umifenovir and moroxydine used against Influenza A viruses.

Moroxydine ([M+H]⁺ with m/z 172) shows extensive fragmentation in the guanidine substituent of the molecule, that is product ions with m/z 155 due to the loss of NH₃, with m/z 130 due to the loss of methanediimine (CH₂N₂), with m/z 113, which is protonated N-cyanomorpholine ([C₅H₉N₂O]⁺), and its complementary ion with m/z 60, protonated guanidine ([CH₆N₃]⁺), and ions with m/z 85, which is protonated N-cyanoguanidine ([C₂H₅N₄]⁺), and its complementary ion with m/z 88, which is protonated morpholine ([C₄H₁₀NO]⁺) (see Fig. 7). The ion with m/z 113 shows secondary fragmentation involving the loss of ethanal (C₂H₄O) to the 2-(cyanoamino)eth-1-ylium ion with m/z 69 ([C₃H₅N₂]⁺). The elemental compositions of these ions is confirmed by accurate-m/z data [9].

5. Antiviral agents against SARS and MERS

5.1. Introduction

During the first 20 years of the 21st century, several outbreaks of virus infections took place, causing serious health problems, e.g., the severe acute respiratory syndrome (SARS), caused by SARS-CoV in 2002 and 2003, mainly in China, the Middle East respiratory syndrome (MERS, also known as camel flu, caused by MERS-CoV) from 2012 onwards, the Ebola virus disease between 2013 and 2016 in West Africa, and the Zika virus epidemic in 2015–2016, starting from Brazil. Most of these outbreaks result from viruses of zoonotic nature, transferred to humans.

The SARS and MERS viruses are corona viruses. Since the mid 1960’s, the human coronaviruses (HCoV), represented by HCoV-OC43 and HCoV-229E, were generally considered relatively harmless viruses. This status changed dramatically with the emergence of SARS-CoV in 2002/2003. The SARS-CoV epidemic took 774 lives around the globe and infected more than 8000 people in 29 countries. The epidemic was halted in 2003 by a highly effective global public health response, and SARS-CoV is currently not circulating in humans [50].

During the preparation of these review papers, a new SARS-CoV-2 was discovered in the city of Wuhan in China in December 2019. This virus, causing the so-called corona virus disease 2019 (Covid-19), was found to be extremely contagious. It resulted in a world-wide pandemic. When finishing these reviews (mid April 2020), Covid-19 have caused ca. 2 million confirmed disease cases and more than 120,000 deaths world-wide, whereas when finishing the revision (early June 2020) there were over 6 million confirmed disease cases and more than 375,000 deaths world-wide (https://coronavirus.jhu.edu/map.html). Obviously, this pandemic lead to a boost of research activities towards ways to control the disease and the stop spreading of the virus (more than 18,000 results upon searching NCBI SARS-CoV-2 literature, sequence, and clinical content: https://www.ncbi.nlm.nih.gov/sars-cov-2/; search early June 2020). These research activities explore multiple directions, including (1) the search for immunotherapy against the virus, (2) the development of vaccines against the virus, and (3) the search for applicable antiviral agents, e.g., Ref. [51].

In this context, it is interesting to review studies performed earlier in relation to the SARS-CoV outbreak in 2002–2003. Several molecular targets could be considered for discovery and development of antiviral agents [52], but apparently with little results in terms of new antiviral agents. Therapies for coronaviruses were reviewed, with focus on viral entry inhibitors [53] and on inhibitors of the intracellular life cycle [54]. An extensive overview of drug discovery and therapeutic options against coronaviruses have been complied as well [55]. Candidate therapies against MERS-CoV have also been reviewed [56], including monoclonal antibodies, antiviral peptides, vaccines, interferons, teicoplanin-derived antibiotics, nucleoside analogues such as 6-mercaptopurine (6 MP) and 6-thioguanine (6 TG), and protease inhibitors, such as the small-molecule inhibitors camostat, nafamostat, as well as lopinavir and ritonavir, known from HIV treatment. When SARS-CoV and MERS-CoV outbreaks were under control, research funds probably dried out, ignoring the possible occurrence of an outbreak of a similar virus in the future. Hardly any developments on CoV-related antiviral agents has been described, until recently.

5.2. Broad-spectrum antiviral agents

In the past few months, numerous studies have been reported on possible pharmacotherapeutic treatment for Covid-19 (e.g., Refs. [51,[57], [58], [59], [60]]) and many more will follow in the months to come.

In one of these studies [60], 120 agents were listed and evaluated as possible broad-spectrum antiviral agents (BSAAs). The rationale behind BSAAs is that different viruses utilize similar pathways and host factors to replicate in the cell. Therefore, compounds that inhibit one of these processes for one virus may do the same for other viruses. This is the drug repurposing approach, which is extremely attractive, because the development of specific drugs for each individual virus will take far too much time. In addition, the safety-in-man assessment of such drugs was already done for their initial therapeutic spectrum. This will (hopefully) greatly facilitate and speed up the approval for their use as antiviral agent.

Among the 120 agents selected as potential BSAAs [60], drugs from many therapeutic classes are listed, such as anticancer, antibacterial, antiprotozoal drugs, but also immunosuppressants, antidepressants, and anti-arrhythmic and lipid-lowering agents. Whether the compounds listed are actually effective may be a matter of fierce discussion, as for instance demonstrated in the discussion on the effectiveness of the antimalaria drugs chloroquine and hydroxychloroquine in the treatment of Covid-19 [61]. Adequate clinical studies will be necessary.

Among the 120 compound listed as potential BSAAs [60] are actually 27 compounds which are approved or investigated as antiviral agents. In the context of the present review, the focus is on collecting and interpreting available MS–MS data of these 27 antiviral agents. The status of these compounds in this respect is summarized in Table 3 . Characteristic SRM transitions for compounds not previously discussed, i.e., camostat, favipiravir, lobucavir, nafamostat, remdesivir, and tilorone, as they are used in their analysis in biological matrices, are collected in Table 4a, Table 4b a and b for positive-ion and negative-ion mode, respectively The available MS–MS data for these compounds is reviewed here.

Table 3.

MS–MS data and interpretation of the 27 antiviral agents, considered as broad-spectrum antiviral agents [60].

Compound class	Compounds	Status
HIV-related agents	lamuvidine, lopinavir, nelfinavir, ritonavir, tenofovir	Discussed in part 1 [1].
HBV- or HCV-related agents	ribavirin, sofusbuvir	Discussed in part 2 [2].
Herpes-related agents	aciclovir, brindicofovir, cidofovir, ganciclovir, letermovir, valaciclovir	Discussed in this paper.
Influenza-related agents	umifenovir, zanamivir	Discussed in this paper.
Protein or peptide antiviral agents.	recombinant interleukin-7 (CYT-107; ∼17.5 kDa), the cyclic peptide alisporivir (1215.9 Da), and the peptide thymalfasin (3106.5 Da)	Outside scope: no small-molecule antivirals.
No MS–MS data found in public domain.	filociclovir, foscarnet, galidesivir, N-methanocarbathymidine
MS–MS data available.	camostat, favipiravir, lobucavir, nafamostat, remdesivir, tilorone	Fragmentation discussed below.

Table 4a.

SRM transitions for broad-spectrum antiviral drugs (BSAAs), not previously discussed, in positive-ion mode. A question mark (?) in the column “formula” indicates that the proposed elemental composition is uncertain, as it is not confirmed by accurate-m/z data.

Compound	m/z of [M+H]+	Formula	m/z of SRM product-ion	Formula	Literature
camostat	399.166	[C20H23N4O5]+
lobucavir (BMS-180194)	266.125	[C11H16N5O3]+	152.057	[C5H6N5O]+	[62]
nafamostat	348.146	[C19H18N5O2]+	187.087	[C11H11N2O]+	[63]
			162.066	[C8H8N3O]+	[63]
pleconaril	382.137	[C18H19F3N3O3]+	298.092	[C14H13F3N2O2]+• ?	[64]
remdesivir	603.233	[C27H36N6O8P]+	200.057	[C9H6N5O]+ ?	[65]
GS-441524 (remdesivir metabolite)	292.104	[C12H14N5O4]+	163.061	[C7H7N4O]+	[65]
tilorone (amixin)	411.264	[C25H35N2O3]+	100.112	[C6H14N]+	[66,67]

Table 4b.

SRM transitions for broad-spectrum antiviral drugs (BSAAs), not previously discussed, in negative-ion mode. A question mark (?) in the column “formula” indicates that the proposed elemental composition is uncertain, as it is not confirmed by accurate-m/z data.

Compound	m/z of [M−H]–	Formula	m/z of SRM product-ion	Formula	Literature
favipiravir (T-705)	156.021	[C5H3FN3O2]–	113.016	[C4H2FN2O]–	[68]
			66	?	[68]

Favipravir is selectively inhibiting viral RNA-dependent RNA polymerase and is applied as antiviral agent against new influenza strains in Japan. Favipiravir ([M−H]^– with m/z 156) is analyzed in negative-ion mode. The major product ion with m/z 133 results from the loss of hydrogen isocyanate (HNCO) (Fig. 8 ) The product ion with m/z 66 is not understood.

Fig. 8.

Structures and the major product ions of the broad-spectrum antiviral agents (BSAAs) favipiravir in negative-ion mode and lobucavir in positive-ion mode.

Lobucavir ([M+H]⁺ with m/z 266) is a nucleoside analogue with broad-spectrum activity. It is analyzed in SRM using the product ion with m/z 152, which is the protonated guanine and due to the loss of guanine N ⁹-substituent (Fig. 8).

Tilorone ([M+H]⁺ with m/z 411), an interferon inducer, shows extensive fragmentation. The product ion with m/z 100, the 2-(diethylamino)ethylium ion ([C₆H₁₄N]⁺) in used in SRM. The complementary ion with m/z 312 is also observed, as well as an ion with m/z 239 due to the subsequent loss of diethylamine (C₄H₁₁N) from the ion with m/z 312. Cleavage of a C–N bond leads to the complementary ions with m/z 74 and 338, whereas the cleavage of the α-β-C–N bond results in the N,N-diethylmethaniminium ion ([C₅H₁₂N]⁺) with m/z 86 (Fig. 9 ).

Fig. 9.

Structures and the major product ions in positive-ion mode of the broad-spectrum antiviral agent tilorone and the serine protease inhibitors camostat and nafamostat, investigated as possible antiviral agents against MERS-CoV.

Camostat and nafamostat were not listed as BSAAs, but as these serine protease inhibitors have been suggested as possible antiviral agents against MERS-CoV [56], their fragmentation is also discussed here.

The fragmentation of camostat and two of its metabolites (GBPA and GBA) has been reported [69]. Camostat ([C₂₀H_{2 3}N₄O₅]⁺; [M+H]⁺ with m/z 399.166) shows the loss of ammonia (NH₃) to an ion with m/z 382. The loss of dimethylamine ((H₃C)₂NH) leads to a product ion with m/z 354. Cleavage of the ester bond leads to two complementary ions, that is the ion with m/z 104 (protonated 2-hydroxy-N,N-dimethylacetamide; [C₄H₁₀NO₂]⁺) and the ion with m/z 294 ([C₁₇H₁₆N₃O₂]⁺). The loss of carbon dioxide (CO₂) from the ion with m/z 296 results in the ion with m/z 254. Cleavage of the other ester bond leads to an acylium ion with m/z 162 ([C₈H₉N₃O]⁺), with shows secondary fragmentation involving the loss of either ammonia (NH₃) to an ion with m/z 145 or of methanediimine (CH₂N₂) to an ion with m/z 120 (Fig. 9).

The metabolite GBPA ([C₁₆H_{1 6}N₃O₄]⁺; [M+H]⁺ with m/z 314) shows the cleavage of the ester bond to an acylium ion with m/z 162 ([C₈H₉N₃O]⁺), which also shows secondary fragmentation involving the loss of either ammonia (NH₃) to an ion with m/z 145 or of methanediimine (CH₂N₂) to an ion with m/z 120. The 4-(carboxymethyl)phen-1-ylium ion with m/z 135 ([C₈H₇O₂]⁺) is also observed (Fig. 9).

Nafamostat ([M+H]⁺ with m/z 348) shows two complementary product ions with m/z 162 and with m/z 187 due to cleavage in the ester bond and charge retention on either side. The loss of ammonia (NH₃) from the ion with m/z 187 yields an ion with m/z 170, whereas the loss of methanediimine (CH₂N₂) from the ion with m/z 162 leads to the ion with m/z 120 (Fig. 9) [63,70].

6. General aspects on the fragmentation observed

The primary fragmentation of protonated molecules [M+H]⁺ involves backbone cleavages of C–heteroatom (N, O, S) bonds via either an inductive cleavage or a four-center H-rearrangement. Such bonds can often we considered to be the “weak links” in the protonated molecule. However, there are a lot of subtle issues involved in this, which unfortunately may be sometimes obscured a little by the level of detail in the available data.

A nice example of this is the fragmentation of compounds with a guanidine moiety. There are a number of such compounds in the current data set, that is peramivir, zanamivir, moroxydine, camostat, its metabolite GBPA, and nafamostat. In principle, one may expect fragmentation involving the loss of ammonia (NH₃, [M+H−17]⁺), methanediimine (CH₂N₂, [M+H−42]⁺) and/or guanidine (CH₅N₃, [M+H−59]^–), as for instance shown for moroxydine (Fig. 7). According to the rule of Field [71], complementary ions may be observed, that is protonated ammonia ([NH₄]⁺ with m/z 18), methanediimine ([CH₃N₂]⁺ with m/z 43) and guanidine ([CH₆N₃]⁺ with m/z 60); the first two ions are generally outside the m/z range the spectrum is acquired with. The formation of the protonated guanidine requires a H-rearrangement to the guanidine moiety. In most cases, a β-H-rearrangement, i.e., a 1,2-elimination, is considered [72], but the situation is probably more complicated [10]. This implies that the ion with m/z 60 and/or the loss of guanidine is readily observed, if such a H-rearrangement is possible, that is when the guanidine is attached to an aliphatic system, such as with peramivir, zanamivir and moroxydine (Fig. 6, Fig. 7). On the other hand, it will not (or with much lower ion abundance, as it should be formally explained) occur, if such a H-rearrangement is less likely, that is when the group is attached to an aromatic system, such as with camostat, GBPA, and nafamostat (Fig. 9). In that case, the loss of methanediimine in observed instead.

The same issues play a role in the fragmentation of amines and ethers. In the di-aliphatic ether, cleavage of either of the two C–O bonds may be observed, as for instance with oseltamivir, whereas in the aliphatic–aromatic ether, there is a preference for the cleavage of the aliphatic-C–O bond rather than the aromatic-C–O bond, as for instance observed in tilorone (Fig. 9). In this context, it must be pointed out that the ion with m/z 135 ([C₈H₇O₂]⁺) for GBPA (Fig. 9) should actually be considered as the 2-(4-hydroxyphenyl)-1-oxoeth-1-ylium ion ([HO–C₆H₄–CH₂–C O]⁺) rather than as the 4-(carboxymethyl)phen-1-ylium ion ([C₆H₄–CH₂–COOH]⁺); both alternatives are shown in Fig. 9.

In some cases, the fragmentation can be considered as a series of small neutral losses. An illustrative example is the fragmentation of zanamivir (Fig. 6).

In yet other cases, the identity of the product ions in terms of elemental composition is clear, but it is more difficult to indicate how such product ions are formed and/or which C-atoms are involved in their generation. This is for instance the case for amantadine and rimantadine (Fig. 4). The choices made in proposing structures for their product ions is primary based on the fact that cleavages at tertiary (and quaternary) C-atoms are more likely to occur, as after fragmentation the positive charge is more stabilized on the resulting secondary (and tertiary) C-atoms.

In this way, a wealth of more generally applicable fragmentation rules can be derived from this type of data sets.

The fragmentation of even-electron ions, like protonated molecules [M+H]⁺, should comply with the even-electron rule [10,73,74]. Therefore, the observation of odd-electron product ions immediately attracts attention. In the current set of compounds, only one shows clearly detectable odd-electron ions. It should be noted, that quite often low-abundant odd-electron “satellite” ions are observed next to the (far) more abundant even-electron ions, that are mentioned in the text. In a number of cases, the occurrence of odd-electron product ions can be predicted from the presence of specific functional groups [10]. After the loss of dimethylamine to the ion with m/z 432, umifenovir (Fig. 7) shows an odd-electron ion with m/z 323 ([C₁₄H₁₄BrNO₃]^{+ •}) due to the loss of a phenylsulfanyl radical (C₆H₅S^•). Sulfur-containing compounds are prone to facilitate radical losses involving the sulfur-containing group. The resulting odd-electron ion may show either the loss of a radical to “return” to an even-electron ion, that is an ethyl radical (^•C₂H₅) to an ion with m/z 294 ([C₁₂H₉BrNO₃]⁺), or the loss of a neutral molecule to “continue” as an odd-electron ion, that is ethanal (C₂H₄O) to an ion with m/z 279 ([C₁₂H₁₀BrNO₂]^+•).

7. Conclusions

The fragmentation in MS–MS of some 30 antiviral agents involved in the treatment of Herpes and influenza infections has been reviewed. MS–MS spectra and SRM transitions were collected and interpreted. For some compounds, confirmation of the interpretation by accurate-m/z data is required. In addition, available MS information on some BSAAs, which were not yet discussed in these review papers, is reviewed as well.

In order to facilitate the access to the information collected in these review papers, a Supplementary File was complied that alphabetically combines all information on SRM transitions for all 112 compounds discussed in the three parts of this review series (antiviral agents as well as some metabolites and anabolites) in two Tables (one for positive-ion data and one for negative ion data).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: