Profiling of secondary metabolites in root exudates of Arabidopsis thaliana Author Strehmel, Nadine , & Leibniz Institute of Plant Biochemistry, Department of Stress and Developmental Biology, Weinberg 3, 06120 Halle (Saale), Germany Author Böttcher, Christoph Author Schmidt, Stephan Author Scheel, Dierk text Phytochemistry 2014 2014-12-31 108 35 46 http://dx.doi.org/10.1016/j.phytochem.2014.10.003 journal article 10.1016/j.phytochem.2014.10.003 1873-3700 10490933 2.4. Compounds identified in root exudates of A. thaliana 2.4.1. Nucleosides, amino acids and dipeptides Although the established method was not intended to target polar metabolites, several polar aromatic compounds including three nucleosides ( 1–3 ), two deoxynucleosides ( 4 , 5 ), a nucleoside derivative ( 6 ) and the aromatic amino acids Phe ( 7 ), Tyr ( 8 ) and Trp ( 9 ) were detected in root exudates of A. thaliana . In addition, a simple β- carboline alkaloid ( 10 ) previously reported to accumulate in tomato fruits ( Yahara et al., 2004 ) was identified using a synthetic reference compound prepared from Trp and formaldehyde ( Tilstra et al., 1990 ). While other proteinogenic amino acids were not detected using the established analytical method, the more hydrophobic side chain-elongated Met homologs ( 11 , 12 ) and their S - oxides ( 13 , 14 ) were putatively identified based on the elemental composition, the presence of three exchangeable protons and the occurrence of characteristic neutral losses of 46.01 Da (HCOOH) and 48.00 Da ( CH 3 SH)/64.00 Da ( CH 3 SOH) in the CID mass spectra of the [M+H] + ions. Compounds 11–14 represent biosynthetic intermediates of corresponding ω- methylsulfinylalkyl/ω- methylthioalkyl glucosinolates of chain length 7 and 8 which are known constituents in roots of the accession Col-0 ( Brown et al., 2003 ; Petersen et al., 2002 ). Beside nucleosides and amino acids, an array of 20 dipeptides ( 15–34 ) composed of the amino acids Gly, Val, Leu, Ile, Pro, Phe and Tyr was identified in root exudates. Structural assignments are in agreement with the elemental composition, the number of exchangeable protons and the occurrence of characteristic y-type fragment ions upon CID of the [M+H] + ions. To discriminate Leufrom Ile-containing peptides, a total of 16 dipeptides were synthesized and used for authentication. The spectrum of dipeptides detected in root exudates matches the set of 14 dipeptides reported to accumulate in roots of A. thaliana in a cell-type specific manner ( Moussaieff et al., 2013 ). 2.4.2. Degradation products of methionine-derived glucosinolates Met-derived glucosinolates are the predominating phytoanticipins in most organs of A. thaliana . Despite their significant accumulation level in root tissue ( Brown et al., 2003 ; Petersen et al., 2002 ) glucosinolates were not consistently detected in root exudates using the established experimental setup, which is in contrast to a previous study on root exudates of Brassica rapa ( Schreiner et al., 2011 ) . Instead of glucosinolates, a set of 8 putative glucosinolate degradation products ( 35–42 ) derived from 8-methylthiooctyl, 8-methylsulfinyloctyl and 7-methylsulfinylheptyl gluco sinolate were identified based on the occurrence of characteristic neutral losses of 48.00 Da and 64.00 Da upon CID of the [M+H] + ions which are indicative for methylthio and methylsulfinyl moieties, respectively. Among them, classical glucosinolate breakdown products known to be produced upon tissue damage were found including two nitriles ( 35 , 36 ) and an isothiocyanate ( 37 ). Related to these compounds, a carboxylic acid ( 38 ) and its amide ( 39 ) as well as three amines ( 40–42 ) were identified. It was suggested, that these compounds represent glucosinolate breakdown products, which are formed in intact tissue by either nitrilase-mediated hydrolysis of nitriles or by degradation of isothiocyanate–glutathione conjugates ( Bednarek et al., 2009 ; Wittstock and Burow, 2010 ). 2.4.3. Degradation products of tryptophan-derived glucosinolates and other indolic compounds In addition to aliphatic glucosinolates indolic glucosinolates accumulate in A. thaliana and, in particular in roots, relatively large amounts of these compounds are found ( Brown et al., 2003 ). As observed for Met-derived glucosinolates, the major Trp-derived glucosinolates of roots (indol-3-yl methyl, 4-methoxy and 1-methoxyindol-3-yl methyl glucosinolate) were absent in root exudates. Again, corresponding degradation products including indol-3- ylmethyl amine ( 43 ), its 4-methoxy ( 44 ) and 1-methoxy derivative ( 45 ) were detected. Here, isomers 44 and 45 can be discriminated based on the number of exchangeable protons determined for the dominating [M+H-NH 3 ] + - type in-source fragment ion. Amine 43 was recently identified as a product of the atypical myrosinase PENETRATION2 ( At 2g 44490 ) which is involved in an inducible pre-invasion resistance mechanism conferring broad-spectrum resistance to fungal pathogens ( Bednarek et al., 2009 ; Lipka et al., 2005 ). In addition, further indolic compounds were identified in root exudates including indole-3-carbaldehyde ( 46 ), 4-hydroxyindole-3-carbaldehyde ( 47 ), indole-3-carboxylic acid ( 48 ), a malonyl-glucoside of 6-hydroxyindole-3-carboxylic acid ( 49 ) and an unknown indole derivative ( 50 ). Compounds 46–49 represent known metabolites which constitutively accumulate (in case of 47 and 48 as glucose conjugates) in leaf and root tissue of A. thaliana and exhibit a pathogen-inducible accumulation pattern ( Bednarek et al., 2005 ; Bottcher et al., 2014 ; Hagemeier et al., 2001 ). It should be noted that camalexin, the major indolic phytoalexin of A. thaliana , was not detected as root exudate in sterile nutrient solutions. In accordance to literature data describing the exudation of this phytoalexin following treatment with a conserved 22- amino-acid peptide from bacterial flagellin (flg22) ( Millet et al., 2010 ), camalexin became clearly detectable when analyzing microbially contaminated nutrient solutions. This suggests that the camalexin level could be used as an additional marker to judge the sterility of the hydroponic culture system. 2.4.4. Salicylic and jasmonic acid degradation products Based on their elemental composition and the occurrence of consecutive neutral losses of 132.04 Da (C 5 H 8 O 4 ) and 43.99 Da (CO 2 ) upon CID of the [M – H] – ions, two exuded compounds ( 51 , 52 ) were annotated as dihydroxybenzoic acid (DHBA) pentosides. Pentoside 51 was identified as 2,3-DHBA 3- O -β- D- xyloside using an authenticated standard ( Bartsch et al., 2010 ) whereas the other was annotated as 2,5-DHBA conjugate based on acid-catalyzed hydrolysis experiments. As recently shown, both DHBA pentosides represent the major salicylic acid catabolites in young and senescent A. thaliana rosette leaves ( Zhang et al., 2013 ). In addition, another plant hormone catabolite ( 53 ) was detected in root exudates (Supplementary Fig. 5A). The CID mass spectrum obtained from the [M – H] – ion of 53 exhibited a prominent neutral loss of 79.96 Da (SO 3 ) and a corresponding fragment ion at m / z 96.96 (HSO 4 – ) indicating the presence of a sulfate moiety (Supplementary Fig. 5B). The elemental composition of 53 was determined to be C 12 H 20 O 7 S suggesting that this compound might represent a sulfated dihydro-hydroxyjasmonic acid derivative. To prove this hypothesis, A. thaliana seedlings were grown in liquid culture, treated for 24 h with [2-D 2 ]-jasmonic acid, extracted and analyzed by UPLC/ESI-QTOFMS. Following this treatment, 53 and its deuterated isotopolgue became detectable (Supplementary Fig. 5C). On-line H/ D exchange experiments revealed the presence of two exchangeable protons (Supplementary Fig. 5D). Thus, compound 53 possesses a saturated side chain and a ring keto group rather than Table 1 Annotated compounds reproducibly detected by UPLC/ESI-QTOFMS in solid-phase extracted nutrient solution of hydroponically grown Arabidopsis thaliana Col-0. Nutrient solution was collected overaone-week period from six-week-old plants.

No.	Compound	vl1	tr [s]	Elemental composition	Quantifier ion Type	m / z	# exchang. protons2	Observed fragment ions upon CID3 m / z , precursor ion marked in bold
1	Adenosine	S	44	C10H13N5O4	[M+H] +	268.10	5	268 , 136
2	Guanosine	S	46	C10H13N5O5	– – [M H]	282.08	6	282 , 150, 133
3	Uridine	S	41	C9H12N2O6	[M – H] –	243.06	4	243 , 200, 153, 152, 140, 122, 110
4	2 0 -Deoxyadenosine	S	46	C10H13N5O3	[M+H] +	252.11	4	252 , 136, 117
5	Thymidine	S	77	C10H14N2O5	[M – H] –	241.08	3	241 , 198, 151, 125
6	2 0-O -Methyladenosine	S	64	C11H15N5O4	[M+H] +	282.12	4	282 , 136
7	Phe	S	75	C9H11NO2	[M+H] +	166.09	3	166 , 149, 131, 120, 107, 103
8	Tyr	S	42	C9H11NO3	[M+H] +	182.08	4	182 , 165, 147, 136, 123, 119, 91
9	Trp	S	144	C11H12N2O2	[M+H] +	205.10	4	205 , 188, 159, 132
10	3-Carboxy-1,2,3,4-tetrahydro-β- carboline	S	187	C12H12N2O2	[M+H] +	217.10	3	217 , 188, 173, 171, 144, 132
11	Pentahomo-Met	I	276	C10H21NO2S	[M+H] +	220.14	– a	220 , 174, 126, 109
12	Hexahomo-Met	I	327	C11H23NO2S	[M+H] +	234.15	3	234 , 188, 140, 123
13	Pentahomo-Met S -Oxide	I	86	C10H21NO3S	[M+H] +	236.13	3	236 , 190, 172, 126
14	Hexahomo-Met S -Oxide	I	164	C11H23NO3S	[M+H] +	250.15	3	250 , 204, 186, 140
15	H-Phe-Gly-OH	I	99	C11H14N2O3	[M+H] +	223.11	4	223 , 120, 103
16	H-Gly-Phe-OH	I	138	C11H14N2O3	[M+H] +	223.11	– a	223 , 177, 166, 120
17	H-Ile-Val-OH	S,L	114	C11H22N2O3	[M+H] +	231.17	4	231 , 213, 185, 118, 86
18	H-Val-Ile-OH	S,L	135	C11H22N2O3	[M+H] +	231.17	4	231 , 213, 185, 132, 86, 72
19	H-Leu-Val-OH	S,L	139	C11H22N2O3	[M+H] +	231.17	4	231 , 213, 185, 118, 86
20	H-Val-Leu-OH	S,L	159	C11H22N2O3	[M+H] +	231.17	4	231 , 213, 185, 132, 86, 72
21	H-Leu-Pro-OH	S	154	C11H20N2O3	[M+H] +	229.15	3	229 , 116, 86
22	H-Leu-Tyr-OH	S,L	166	C15H22N2O4	[M+H] +	295.17	5	295 , 277, 249, 182, 165, 136, 86
23	H-Tyr-Ile-OH	S	174	C15H22N2O4	[M+H] +	295.17	5	295 , 278, 277, 250, 249, 136, 132, 130, 121, 119
24	H-Tyr-Leu-OH	S	192	C15H22N2O4	[M+H] +	295.17	5	295 , 278, 277, 250, 249, 136, 132, 130, 121, 119
25	H-Phe-Val-OH	I	179	C14H 20N2O3	[M+H] +	265.15	4	265 , 219, 120, 118
26	H-Val-Phe-OH	I	196	C14H20N2O3	[M+H] +	265.15	4	265 , 219, 166, 120, 72
27	H-Ile-Ile-OH	S,L	185	C12H24N2O3	[M+H] +	245.19	4	245 , 199, 132, 86
28	H-Leu-Ile-OH	S,L	199	C12H24N2O3	+ [M+H]	245.19	4	245 , 199, 132, 86
29	H-Ile-Leu-OH	S,L	204	C12H24N2O3	[M+H] +	245.19	4	245 , 199, 132, 86
30	H-Leu-Leu-OH	S,L	217	C12H24N2O3	[M+H] +	245.19	4	245 , 199, 132, 86
31	H-Phe-Ile-OH	S,L	226	C15H22N2O3	[M+H] +	279.17	4	279 , 262, 261, 233, 132, 120, 86
32	H-Ile-Phe-OH	S,L	235	C15H22N2O3	[M+H] +	279.17	4	279 , 261, 233, 166, 120, 86
33	H-Phe-Leu-OH	S,L	247	C15H22N2O3	[M+H] +	279.17	4	279 , 262, 261, 233, 132, 120, 86
34	H-Leu-Phe -OH	S,L	249	C15H22N2O3	[M+H] +	279.17	4	279 , 261, 233, 166, 120, 86
35	7-MeSO-Heptyl-CN	I,L	266	C9H17NOS	[M+H] +	188.11	– a	188 , 124, 107, 105, 96, 91, 82, 81, 79, 77, 65
36	8-MeSO-Octyl-CN	I,L	323	C10H19NOS	[M+H] +	202.13	0	202 , 138, 121, 110, 105, 96, 95, 93, 91, 82, 81, 79, 77, 65
37	8-MeSO-Octyl-NCS	S,L	483	C10H19NOS2	[M+H] +	234.10	0	234 , 175, 170, 161, 137, 136, 128, 122, 114,108, 69, 65, 55
38	8-MeSO-Octyl-CO2H	S,L	296	C10H20O3S	[M+H] +	221.12	1	221 , 203, 175, 139, 137, 121, 119, 111, 109, 95, 93
39	8-MeSO-Octyl-CONH2	I	233	C10H21NO2S	[M+H] +	220.14	2	220 , 203, 175, 139, 137, 121
40	7-MeSO-Heptyl-NH2	S	61	C8H19NOS	[M+H] +	178.13	2	178 , 161, 160, 114, 112, 97, 95
41	8-MeSO-Octyl-NH2	S,L	133	C9H21NOS	[M+H] +	192.14	2	192 , 175, 174, 128, 111, 109, 97
42	8-MeS-Octyl-NH2	I	311	C9H21NS	[M+H] +	176.15	2	176 , 159, 128, 111, 103
43	Indol-3-ylmethylamine (I3CH2NH2)	S,L	111	C9H10N2	[M+H-NH] + 3	130.07	1b	130 , 128, 103, 77
44	4-MeO-I3CH2NH2	S	192	C10H12N2O	[M+H-NH] + 3	160.08	1b	160 , 148, 145, 132, 130, 117
45	1-MeO-I3CH2NH2	I,L	223	C10H12N2O	[M+H-NH] + 3	160.08	0b	160 , 145, 132, 128, 117
46	Indole-3-carbaldehyde (I3CHO)	S	289	C9H7NO	[M – H] –	144.04	1	144 , 142, 126, 116, 115
47	4-HO-I3CHO	S	307	C9H7NO2	– – [M H]	160.04	2	160 , 158, 132, 131, 130, 114, 104, 102
48	Indole-3-carboxylic acid (I3CO2H)	S,L	284	C9H7NO2	[M+H] +	162.06	2	162 , 144, 118
49	6-(Malonyl-GlcO)-I3CO2H	I,L	194	C18H19NO11	[M-H-CO] – 2	380.10	– a	424 , 380, 338, 320, 218, 176, 175, 132
50	Unknown indole derivative	I	391	C10H9NO3	[M+H] +	192.07	– a	192 , 177, 174, 161, 159, 148, 133, 132, 117,116, 105, 104
51	2,3-DHBA 3- O -Xyl	S,L	176	C12H14O8	[M – H] –	285.06	5	285 , 153, 109
52	2,5-DHBA Pent	I,L	153	C12H14O8	[M – H] –	285.06	5	285 , 153, 152, 108

( continued on next page ) Table 1 ( continued )

No.	Compound	vl1	tr	Elemental	Quantifier ion	# exchang. protons2	Observed fragment ions upon CID3
[s]	composition	Type	m / z	m / z , precursor ion marked in bold
53	9,10-Dihydrohydroxy-JA Sulfate	I	222	C12H20O7S	– – [M H]	307.09	2	307 , 227, 97
54	Coniferin	I,L	179	C16H22O8	[M+Na] +	365.12	5	387 ([M-H+FA] – ), 341, 207, 179
55	Syringin	I,L	193	C17H24O9	[M+Na] +	395.13	5	417 ([M-H+FA] – ), 209, 207,194
56	G(8- O -4)G	I	275/	C20H24O7	[M – H] –	375.15	4	375 , 327, 195, 179, 165
281
57	G(8- O -4)S Glycerol	I	188/	C21H28O10	[M – H] –	439.16	6	439 , 391, 243, 195
192
58	G(8-5)G	I	344	C20H22O6	O] – [M-H-H2	339.12	3	357 ([M – – ), 339, 327, 324 H]
59	S(8-5)G	I	340	C21H24O7	[M-H-H O] - 2	369.13	3	387 ([M – H] – ), 369, 357, 354
60	S(8-8)S	I	377	C22H26O8	[M – H] –	417.16	2	417 , 402, 387, 181, 166
61	Lariciresinol	S,L	337	C20H24O6	[M-H-	329.14	3	359 ([M – H] – ), 344, 329, 192, 178, 175, 164, 160, 159
CH O] – 2
62	Lariciresinol Hex	I,L	272	C26H34O11	[M – H] –	521.20	6	521 , 359, 329, 192, 178, 175, 160
63	G(8- O -4)S(8-5)G	I	366/	C31H36O11	[M – H] –	583.22	5	583 , 535, 505, 369, 357, 195, 165
379
64	G(8- O -4)S(8-8)S	I	401/	C32H38O12	[M – H] –	613.23	– a	613 , 565, 535, 417, 402, 387, 195, 181, 165, 150
415
65	G(8- O -4)G(8-8)S/G(8- O -4)S(8-8)G	I	407/	C31H36O11	[M – H] –	583.22	– a	583 , 535, 505, 387, 357, 195, 181, 180, 165, 150
421
66	Esculetin	S,L	201	C9H6O4	[M – H] –	177.02	2	177 , 149, 133, 121, 105, 93, 89, 81, 77
67	Scopoletin	S,L	268	C10H8O4	[M+H] +	193.05	1	191 ([M – H] – ), 176, 148, 120, 104
68	Esculin	S,L	166	C15H16O9	[M – H] –	339.07	5	339 , 177, 133
69	Scopolin	I,L	196	C16H18O9	[M-H+FA] –	399.09	4	399 , 353, 207, 191, 176
70	Scopoletin Hex-Pent	I	197	C21H26O13	[M – H] –	485.13	6	485 , 191, 176, 148
71	Scopoletin Benzoyl-Hex-Pent	I	329	C28H30O14	[M – H] –	589.16	5	589 , 467, 191, 148, 121
72	G(8- O -4)Scopoletin	I	280	C20H20O8	[M – H] –	387.11	3	387 , 339, 195, 191, 176, 165, 150
73	G(8- O -4)Esculetin	I	377	C19H16O7	[M – H] –	355.08	2	355 , 337, 325, 179, 177, 176
74	S(8- O -4)Esculetin	I	370	C20H18O8	– – [M H]	385.09	2	385 , 367, 355, 209, 177, 176
75	Scopoletin dehydrodimer	I	315	C20H14O8	[M – H] –	381.06	2	381 , 366, 337, 325, 322, 321, 307, 305
76	Unknown oligolignol	I	300	C41H46O20	[M – H] –	857.25	8	857 , 661, 485, 467, 389, 371, 341, 195, 193, 191, 176, 175, 165
77	G(8- O -4)FA Sulfate	I	237/	C20H22O11S	[M – H] –	469.08	4	469 , 389, 341, 275, 245, 203, 195, 193
256
78	G(8-5)FA	S	358	C20H20O7	[M-H-H O] – 2	353.10	3	371 , 353, 341, 338, 326, 322, 309, 297, 294, 282, 235, 191
79	G(8-5)FA Sulfate	I	323	C20H20O10S	[M – H] –	451.07	3	451 , 371, 353, 341, 338, 326, 309, 297, 294, 282, 267, 245, 97
80	FA dimer	I	242/	C20H20O8	– – [M H]	387.11	4	387 , 339, 324, 309, 235, 220, 217, 203, 202, 176, 151, 136
249
81	Didehydro-di(coumaroyl)spermidine Sulfate	I	219	C25H29N3O7S	[M – H] –	514.17	4	516 ([M+H] +), 436, 419, 362, 348, 347, 291, 290, 265, 203, 202, 145
82	Kaempferol 3- O -Rha(1+2)Glc 7- O -Rha	I,L	228	C33H40O19	[M – H] –	739.21	11	739 , 593, 430, 284
83	9,12,13-Trihydroxy-10( E ),15( Z )- octadecadienoic	S,L	420	C18H32O5	[M – H] –	327.22	4	327 , 309, 291, 239, 229, 221, 211, 185, 183, 177, 171
acid
84	9,12,13-Trihydroxyoctadec-10-enoic acid	I,L	445	C18H34O5	[M – H] –	329.23	4	329 , 311, 293, 229, 211, 197, 193, 183, 171, 139
85	Fatty acid derivative	I	634	C18H30O4	[M – H] –	309.21	2	309 , 291, 273, 265, 247
86	Azelaoyl lysoPC	I	232/	C17H34NO9P	[M – H] –	426.19	2	428 ([M+H] +), 410, 258, 245, 184, 104
244
87	C4H10O Hex	I	162	C10H20O6	[M+Na] +	259.12	4	237 ([M+H] +), 163, 145, 127, 85
88	C4H10O Malonyl-Hex	I	232	C13H22O9	[M+Na] +	345.11	4	345 , 301, 259
89	C4H10O Hex-DeoxyHex	I	200	C16H30O10	[M – H] –	381.18	6	383 ([M+H]+), 309, 291, 273, 255, 237, 221, 147, 129
90	C4H10O Malonyl-Hex-DeoxyHex	I	255	C19H32O13	[M-H-CO] – 2	423.19	6	467 ([M – H] – ), 423, 381, 277, 235, 217, 161, 159, 143, 113, 101
91	C7H9N5O Hex	I	63	C13H19N5O6	[M – H] –	340.13	5	342 ([M+H] +), 180, 135, 110
92	C9H10O3 Hex	I	198	C15H20O8	[M – H] –	327.11	5	327 , 165, 147, 101
93	C12H16O5 Hex	I	202	C18H26O10	[M – H] –	401.14	– a	447 ([M-H+FA] – ), 401, 255, 205, 161, 113
94	C7H14O4 Malonyl-Hex	I	209	C16H26O12	[M-H-CO] – 2	365.15	– a	409 ([M – H] – ), 365, 323, 203, 161, 149, 143, 101
95	C12H16O6 Hex	I	269/	C18H26O11	[M+Na] +	441.14	4	463 ([M-H+FA] – ), 385, 241, 225, 179, 161, 153, 149, 143, 131, 119
279

an unsaturated side chain and a ring hydroxy group. Beside 12- hydroxyjasmonic acid O -sulfate ( Gidda et al., 2003 ), compound 53 represents a novel sulfated jasmonic acid catabolite produced in A. thaliana . 2.4.5. Phenylpropanoids 2.4.5.1. Lignols. Coniferin ( 54 ) and syringin ( 55 ) represent the predominant phenylpropanoids in dark-grown roots of A. thaliana ( Bednarek et al., 2005 ) and were also detected in root exudates. Both monolignol glucosides were accompanied by related di- and trilignols and derivatives thereof. The identity and the linkage type of these compounds was derived from the analysis of CID mass spectra obtained from corresponding [M – H] – ions ( Morreel et al., 2010a ,b). The CID mass spectra of the β- aryl ether-type dimers 56 and 57 carrying an 8- O -4 linkage exhibited a characteristic neutral loss of 48.02 Da according to a combined loss of water and formaldehyde from the [M – H] – ion. In addition, formation of A – and B – fragment ions by cleavage of the aryl ether bond facilitated the identification of the aromatic units of these dimers ( Morreel et al., 2010b ). For the phenylcoumaran-type dimers 58 and 59 with 8–5 linkage, fragmentation of the [M – H] – ions resulted in dominating neutral losses of 18.01 Da (H 2 O) and 30.01 Da ( CH 2 O) ( Morreel et al., 2010b ). Three resinol-type dilignols with 8–8 linkage including syringaresinol ( 60 ), lariciresinol ( 61 ) and a derived hexoside ( 62 ) were found in root exudates. In accordance with the literature, the CID mass spectrum obtained from deprotonated 60 showed the product ions [M – H-CH 3 ] . – and [M – H-CH 2 O] – as well as a 2,5 X – fragment ion at m / z 181.05 ( Morreel et al., 2010b ). The accumulation of glycosylated syringaresinol and lariciresinol in roots of A. thaliana was described ( Matsuda et al., 2010 ; Nakatsubo et al., 2008 ). Based on the fragmentation characteristics observed for dilignols, three trilignols ( 63–65 ) with coniferyl and syringyl alcohol as monomeric untis were tentatively identified based on the fragmentation characteristics observed for corresponding dilignols. 2.4.5.2. Coumarins. The coumarin glucoside scopolin represents another abundant root phenylpropanoid ( Bednarek et al., 2005 ; Kai et al., 2006 ). It was recently shown, that the biosynthesis of coumarins in roots and their exudation can be stimulated inter alia upon iron deprivation indicating that coumarins are involved in the acquisition of iron in A. thaliana ( Fourcroy et al., 2014 ; Schmid et al., 2014 ; Schmidt et al., 2014 ). Esculetin ( 66 ), scopoletin ( 67 ) and its glucosides esculin ( 68 ) and scopolin ( 69 ) were detected in root exudates and identified by authenticated standards. Beside these coumarins which are already known to be exuded from roots ( Fourcroy et al., 2014 ; Schmid et al., 2014 ), further esculetin and scopoletin conjugates were tentatively identified in root exudates on the basis of their mass spectral fragmentation pattern. Scopoletin was found to be conjugated to a hexose–pentose and to a benzoylated hexose–pentose as indicated by a neutral loss of 294.10 Da (C 11 H 18 O 9 ) from the [M – H] – ion of 70 and consecutive neutral losses of 122.04 Da (C 7 H 6 O 2 ) and 294.10 Da from the [M – H] – ion of 71 , respectively. In addition, an 8- O -4-linked coniferylalcohol–scopoletin conjugate ( 72 ) was annotated based on the occurrence of a characteristic neutral loss of 48.02 Da from its [M – H] – ion and characteristic A – and B – fragment ions at m / z 195.07 and m / z 191.03. Compounds 73 and 74 were tentatively identified as 8- O -4-linked cross-coupling products of coniferyl alcohol with esculetin and syringyl alcohol with esculetin, respectively. The presence of an unsubstitued 3- OH group in the esculetin moiety thereby results in the formation of a benzodioxane-type linkage motif. In accordance with the literature ( Morreel et al., 2010b ), CID of deprotonated 73 and 74 resulted in neutral losses of water and formaldehyde from the [M – H] – ion and formation of A – and B – fragment ions at m / z 179.07 and m / z 177.02 for 73 and m / z 209.08 and m / z 177.02 for 74 . In addition, a scopoletin dehydrodimer ( 75 ) with unknown linkage motif and an unknown oligolignol ( 76 ) with four phenylpropanoid units of which one was identified as scopoletin were detected in root exudates.

96	Unknown	–	326 /	C12 H 16 O 6	[M + Na] +	279.08	1	239 ([M + H-H O] + ), 207, 179, 147, 119 2
339
97	C 14 H 28 O 5 Malonyl-Hex	I	404	C23 H40 O 13	[M-H-CO] – 2	479.25	– a	479 , 437, 291, 161
98	C 18 H 19 N 5 O 7 Sulfate	I	234	C18 H 19 N 5 O 10 S	[M – H] –	496.08	6	496, 374, 345, 294, 223, 150, 121, 97
99	Unknown secondary amine	I	63	C 6 H 13 N	[M + Na] +	122.09	1	100 ([M +H] + ), 83, 55
100	Unknown	–	234	C20 H 28 O 11	[M – H] –	443.15	6	443 , 291, 151, 145, 125, 107, 101
101	Unknown	–	182	C21 H28 O 12	[M – H] –	471.15	6	471 , 291, 179, 163, 145, 143, 125, 123, 119, 113, 107, 101
102	Unknown	–	260	C11 H 10 O 6	[M – H] –	237.04	3	237, 222, 193, 178, 149, 134
103	Unknown	–	573	C 28 H44 O 4	[M + H] +	445.33	3	445 , 427, 409, 391, 353, 339, 325, 313, 311, 285, 283, 271, 269, 215

Pent, pentoside; Hex, hexoside; DeoxyHex, deoxyhexoside; DHBA, dihydroxybenzoic acid; JA, jasmonic acid; G, coniferyl alcohol; S, sinapyl alcohol; FA , ferulic acid; lysoPC, lysophosphatidylcholine. 1 verification level: S, chromatographic and mass spectral behavior as observed for an authenticated standard; I, putative structure deduced from interpretation of elemental composition, collision-induced dissociation mass spectra and online H / D exchange experiments; L, compound reported for A. thaliana in the literature, for references see Supplementary Table 2. 2 number of exchangeable protons as determined by on-line H/ D exchange: a number of exchangeable protons could not be unambigously determined due to low intensity, b compound does not form any quasi-molecular ion, number of exchangeable protons determined for the [M + H-NH] + ion. For detailed information see Supplementary Table 3. 3 3 for details see Supplementary Table 4. Fig. 3. Molecular structures of identified and putatively annotated compounds detected in root exudates of Arabidopsis thaliana . For annotation levels see Table 1. Hex Hexosyl; Pent Pentosyl. 2.4.5.3. Hydroxycinnamic acid conjugates and flavonoids. Although phenolic acids represent well-known constituents of root exudates ( Lanoue et al., 2010 ), free hydroxycinnamic and hydroxybenzoic acids could not be detected in root exudates of A. thaliana using the established experimental system. Due to the low process efficiencies found for cinnamic and salicylic acid (vide supra), it remains unclear, whether A. thalina roots do not exude these compounds under sterile conditions or their detection is hampered by the sample preparation protocol. However, three monolignol-ferulic acid conjugates ( 77 , 78 , 79 ) and a ferulic acid dimer ( 80 ) were identified. Interestingly, two of them ( 77 , 79 ) were detected as sulfate conjugates as indicated by a neutral loss of 79.96 Da from the [M – H] – ions. For the 8- O -4 cross-coupled coniferyl alcohol–ferulic acid conjugate 77 a characteristic neutral loss of 48.02 Da from the [M – H-SO 3 ] – ion was registered together with A – and B – fragments at m / z 195.06 and m / z 193.05. As observed for phenylcoumaran-type dilignols, the CID mass spectra of the 8–5 cross-coupled coniferyl alcohol–ferulic acid conjugates 78 and 79 are characterized by neutral losses of water and formaldehyde from the [M – H] – or [M – H-SO 3 ] – ion, respectively. Another sulfate conjugate was compound 81 . Calculation of the elemental composition of 81 suggested the presence of three nitrogen atoms. The CID mass spectrum obtained from [M+H] + of 81 revealed neutral losses of 17.03 Da (NH 3 ), 74.08 Da (C 3 H 10 N 2 ), 88.10 Da (C 4 H 12 N 2 ) and 145.16 Da (C 7 H 19 N 3 ) from the [M+H-SO 3 ] + ion, the latter resulting in a fragment ion at m / z 291.07 with the elemental composition of C 18 H 11 O 4 + . This suggests that 81 represents a sulfated cyclic didehydro-di(coumaroyl)spermidine conjugate. The Arabidopsis genome contains a gene ( At 2g 25150 ) encoding a functional coumaroyl-CoA:spermidine coumaroyltransferase ( Luo et al., 2009 ). This gene was shown to exhibit a root specific expression pattern and its overexpression resulted in the accumulation of di(coumaroyl)spermidine and a related hexoside in roots ( Luo et al., 2009 ). Compound 81 could represent a metabolite which might be synthesized from di(coumaroyl)spermidine by intramolecular oxidative phenol coupling and subsequent sulfation. Although flavonol glycosides are absent in dark-grown roots in A. thaliana ( Bednarek et al., 2005 ; Hemm et al., 2004 ), a kaempferol triglycoside ( 82 ), which was found to be accumulating in rosette leaves ( Yonekura-Sakakibara et al., 2008 ) was identified in root exudates. 2.4.6. Fatty acid derivatives Among the more hydrophobic compounds of the semipolar root exudate fraction three oxygenated C18 fatty acids ( 83–85 ) were detected. For compounds 83 and 84 on-line H/D exchange experiments revealed the presence of four exchangeable protons indicating that both fatty acid derivatives are functionalized by three hydroxyl groups. The CID mass spectra obtained from the [M – H] – ion of 83 and 84 showed fragment ions at m / z 229.14 (C 12 H 21 O 4 – ) and m / z 171.10 (C 9 H 15 O 3 – ) suggesting that both fatty acid derivatives are hydroxylated in position 12 and/or 13 and 9 and/or 10. In a recent study, 9,12,13-trihydroxy-10,15-octadecadienoic acid and 9,12,13-trihydroxy-10-dodecenoic acid were identified as infection markers in the leaf apoplast washing fluid of A. thaliana following infection by Verticillium longisporum ( Floerl et al., 2012 ) . Authenticated standards of both compounds exhibited identical chromatographic and mass spectral properties as observed for 83 and 84 . The elemental composition of the fatty acid derivative 85 indicated four oxygen atoms and four double bond equivalents, while on-line H/D exchange experiments revealed the presence of only two exchangeable protons. Unfortunately, the CID mass spectrum obtained from deprotonated 85 showed only unspecific neutral losses of water and carbon dioxide for which reason 85 could not be characterized further. In contrast, the CID mass spectrum obtained from [M+H] + of 86 exhibited characteristic fragments at m / z 184.07 (C 5 H 15 NO 4 P + ) and m / z 104.11 (C 5 H 15 NO + 4 ) indicating the presence of a phosphorylcholine moiety. The occurrence of fragments ions at m / z 187.10 (C 9 H 15 O 4 – ), m / z 169.09 (C 9 H 13 O 3 – ) and m / z 125.10 (C 8 H 13 O – ) in the CID mass spectrum obtained from deprotonated 86 was in accordance with a lysophosphatidylcholine of an oxygenated C9 fatty acid. Since on-line H/D-exchange experiments revealed the presence of two exchangeable protons 86 is postulated to be an azelaoyl lysophosphatidylcholine. Azelaic acid itself was shown to be produced by oxidative fragmentation of esterified 18:2 and 18:3 fatty acid species and proposed to be a general marker for lipid peroxidation ( Zoeller et al., 2012 ). In addition, this dicarboxylic acid was identified as pathogen inducible metabolite from the vascular sap of A. thaliana and shown to confer local and systemic resistance against the pathogen Pseudomonas syringae ( Jung et al., 2009 ) . 2.4.7. Unknown components Among the compounds which were reproducibly detected in root exudates of A. thaliana , 17 could not or only partially be structurally characterized. Thereof, 11 compounds were annotated as glycoconjugates with unknown aglycone based on characteristic neutral losses and fragment ions. Among them, five hexose conjugates ( 87 , 91 , 92 , 93 , 95 ), three malonyl–hexose conjugates ( 88 , 94 , 97 ), a hexose–deoxyhexose conjugate ( 89 ) and a malonyl–hexose– deoxyhexose conjugate ( 90 ) were detected. Beside the three sulfated phenylpropanoids and the sulfated jasmonic acid derivative, another sulfate conjugate ( 98 ) was found.