Detection, Quantification, and Total Synthesis of Novel 3-Hydroxykynurenine Glucoside–Derived Metabolites Present in Human Lenses

Abstract

Purpose.: 3-Hydroxykynurenine O-β-D-glucoside (3OHKG) protects the lens from UV damage, and novel related species may act analogously. The aim of this study was to detect, quantify, and elucidate the structures of novel 3-hydroxykynurenine glucoside–derived metabolites present in the human lens.

Methods.: Compounds were detected and quantified by liquid chromatography with tandem mass spectrometry (LC-MS/MS) in 24 human lenses of different ages, of which 22 were normal and two had cataract. Structures of these were confirmed through total synthesis.

Results.: 3OHKG concentrations decreased with age in the lens nuclei, whereas the levels of three novel species, 4-(2-amino-3-hydroxyphenyl)-2-hydroxy-4-oxobutanoic acid O-β-D-glucoside (3OHKG-W), 3-hydroxykynurenine O-β-D-glucoside yellow (3OHKG-Y), and 2-amino-3-hydroxyacetophenone O-β-D-glucoside (AHAG), increased, though to different extents. In contrast, the concentrations present in the cortex of the lens remained constant with age.

Conclusions.: Three novel 3OHKG-derived metabolites have been detected in extracts from human lenses.

Introduction

Kynurenine-based compounds derived from tryptophan, which have strong optical absorbance bands in the 300- to 400-nm wavelength range, are present in the human lens and protect this structure from UV-induced damage by acting as physical quenchers of excited state species generated by incident radiation. ¹ These compounds, which are collectively known as UV filters and occur predominantly in primates, include the abundant species 3-hydroxykynurenine O-β-D-glucoside (3OHKG), 4-(2-amino-3-hydroxyphenyl)-4-oxobutanoic acid O-β-D-glucoside (AHBG), kynurenine, and 3-hydroxykynurenine. ^2,3 3OHKG, 3-hydroxykynurenine, and kynurenine can deaminate at physiological pH to form reactive α,β-unsaturated carbonyls. ⁴ The deaminated form of 3OHKG (3OHKG-D) appears to be a precursor for other UV filters including glutathione and cysteine adducts and the reduced compound AHBG. ^5,6 The deaminated forms of kynurenine and 3-hydroxykynurenine have been reported to form glutathione adducts and reduced compounds that are present at pmol/mg protein levels in the human lens. ⁷ It has been proposed that these UV filter compounds when bound to lens proteins contribute to protein oxidation and cross-linking, as well as the lens discoloration observed in age-related nuclear (ARN) cataractous lenses. ⁸  

Herein we describe the identification and quantification of three novel lens metabolites: 4-(2-amino-3-hydroxyphenyl)-2-hydroxy-4-oxobutanoic acid O-β-D-glucoside (3OHKG-W), 3-hydroxykynurenine O-β-D-glucoside yellow (3OHKG-Y), and 2-amino-3-hydroxyacetophenone O-β-D-glucoside (AHAG). 3OHKG-W and 3OHKG-Y are proposed to derive from deaminated 3OHKG and AHAG from cleavage of 3OHKG. 

Methods

Materials

Twenty-four human lenses aged from 18 to 84 years were obtained from the Sydney Eye Bank after ethical approval (University of Sydney Human Research Ethics Committee #7292). The nucleus and cortex of each lens was extracted separately, first with absolute ethanol and then 80% ethanol, as described previously. ⁹ All chemicals used were of analytical reagent grade or higher. Non-HPLC grade organic solvents were distilled prior to use. Trifluoroacetic acid (TFA, >99%) and formic acid (>98%) were obtained from Sigma-Aldrich (Castle Hill, NSW, Australia). 3OHKG was synthesized as described previously. ¹⁰ Milli-Q water (Millipore, North Ryde, NSW, Australia) was used to prepare all solutions. Thin-layer chromatography (TLC) used reversed phase silica gel 60 RP-18 F254 plates (Merck, Darmstadt, Germany) and a mobile phase of n-butanol/acetic acid/water (12:3:5, vol/vol/vol). Plates were sprayed with ninhydrin and visualized using 254- and 365-nm UV light. 

Nuclear Magnetic Resonance (NMR) Spectroscopy

¹H, ¹³C, correlation spectroscopy (COSY), HSQC (¹H–¹³C heteronuclear single quantum correlation), and HMBC (¹H–¹³C heteronuclear multiple bond correlation) data were obtained using a Bruker Avance 400 spectrometer (¹H, 400 MHz; ¹³C, 100 MHz; Bruker Biosciences, Preston, VIC, Australia) at 25°C. Compounds were dissolved in D₂O or CD₃OD, with undeuterated residual compounds used as reference. Resonances are quoted in parts per million (ppm), and coupling constants (J) are given in Hertz. 

Liquid Chromatography–Mass Spectrometry (LC-MS)

Lens extracts (see above) were centrifuged through a 10-kDa cutoff filter to remove high molecular mass compounds, lyophilized, and resuspended in 50 μL 1% (vol/vol) acetonitrile/water. Samples were separated using a Thermo Fischer Scientific (North Ryde, NSW, Australia) Surveyor HPLC system with a Phenomenex (Lane Cove, NSW, Australia) Synergi 4 μm Fusion-RP 80 column (150 × 2.0 mm × 80 Å; 30°C) coupled to a Thermo Fischer Scientific LCQ Deca XP Max ion trap with electrospray ionization (ESI) in positive ion mode. Samples were eluted at a flow rate of 0.2 mL/min using a gradient system with 100% buffer A (5 minutes) before a linear increase to 40% buffer B over 40 minutes, then increasing to 95% B over 2 minutes, washing with 95% B for 1 minute before returning to 100% A over 2 minutes, and re-equilibration with 100% A for 14 minutes (buffer A: 0.05% [vol/vol] TFA in water; buffer B: 100% acetonitrile). Eluted compounds were detected by UV absorbance at 360 nm, and MS selected reaction monitoring (SRM) of expected ion fragments (Table 1). Mass spectrometer conditions were electrospray needle voltage 4.5 kV, sheath gas flow rate 42, sweep gas flow rate 24, and a capillary temperature of 275°C. The sheath and sweep gases were N₂, while the collision gas was helium. High resolution mass spectrometry (HRMS/MS) data were obtained on a Thermo Fischer Scientific LTQ FT Ultra Hybrid Mass Spectrometer (Mark Raftery, University of New South Wales, Sydney, Australia). 

Table 1.

 

LC-MS Retention Times, m/z M+H⁺, and m/z M+H⁺ of Fragment Ion Following Loss of Glucose

Peak	Expected Compound	Retention, min	m/z M+H+	m/z M+H+ -Glucoside
1	3OHKG	4.1	387.1	225.1
2	3OHKG-W	16.7	388.1	226.0
3	Unknown 3	19.3	314.1	152.3
4	3OHKG-Y	20.1	370.1	208.0
5	AHAG	21.9	314.1	152.1
6	AHBG	22.6	372.1	210.0
7	3OHKG-D	23.2	370.1	208.0

UV-Vis Absorbance Spectroscopy

A Shimadzu (Rydalmere, NSW, Australia) SPD-M20A Photodiode array detector was used to record ultraviolet-visible (UV-Vis) spectra. 

Synthesis and Characterization of 3OHKG-Y, 3OHKG-W, and 3OHKG-D

3OHKG-Y, 3OHKG-W, and 3OHKG-D were synthesized from 3OHKG using a modified method of Tokuyama et al. ¹¹ 3OHKG (100 mg, 0.26 mmol) was dissolved in NaHCO₃ (15 mL, 0.60 M, pH 8.25, sparged with argon), then heated to reflux for 4 hours under argon in the dark. Reaction progress was monitored by TLC (R_f 0.89, reversed phase [RP], 10% acetonitrile). The solution was acidified to ∼pH 6 with 1 M HCl and then lyophilized. Salts were removed using a Waters (Rydalmere, NSW, Australia) C18 RP Sep-Pak column using water/0.05% TFA, followed by 5% acetonitrile/0.05% TFA. Samples were purified by preparative reversed phase HPLC (250 × 21.2 mm × 80 Å Phenomenex RP column) using gradient elution (flow rate 8.0 mL/min) starting at 0% B (10 minutes), 0-40% B (80 minutes), 40-95% (4 minutes), 95% (2 minutes), 95-0% (4 minutes), 0% (30 minutes), with buffers A and B as stated above. During preparative scale HPLC separation, 3OHKG-W eluted as two close adjoining peaks (diastereomers, ∼1:1), while 3OHKG-Y eluted as a shouldered peak. ¹H and ¹³C NMR data of the purified 3OHKG-Y confirmed the presence of two diastereomers (1:1). ¹H NMR data for 3OHKG-D determined that the E isomer was synthesized. 

Analytical data (HRMS/MS, ¹H and ¹³C NMR, UV-Vis) for 3OHKG-Y, 3OHKG-W, and 3OHKG-D are reported as Supplementary Material. 

Synthesis of AHAG

AHAG was synthesized as described previously, ¹² and analytical data are provided as Supplementary Material. 

Extraction Efficiency

Method 1 (Wood and Truscott Method).

Bovine lens tissue, 100 to 200 mg, was mixed with an equivalent mass of 500 μg/mL aqueous solutions of each of 3OHKG, 3OHKG-W, 3OHKG-Y, and AHAG standards. The mixtures were homogenized and extracted (2 × 150 μL absolute ethanol), then centrifuged, filtered, lyophilized, resuspended into 150 μL 1% (vol/vol) acetonitrile/water, and analyzed by RP LC-MS as described above, with quantification based on UV-Vis absorption of the eluted compounds. All tests were performed in triplicate. ¹³  

Method 2 (Inoue and Satoh Method).

Bovine lens tissue (100–200 mg) was mixed with an equivalent mass of 500 μg/mL aqueous 3OHKG standard. The samples were then extracted with 5% KOH in 80% ethanol/water solution, with aliquots removed at 0, 8, 24, 32, and 48 hours. The aliquots were then centrifuged to remove proteins, neutralized with 1 M HCl, and analyzed by RP LC-MS as described above. ¹⁴  

Results

Twenty-four human lenses from donors aged 18 to 84 years were obtained from the Sydney Eye Bank. Twenty-two of these donors had normal lenses (with four being a second lens from the same donor), and two donors had ARN cataract. Each lens was separated into nucleus and cortex, and the UV filters were extracted as described previously. ¹³  

Mass Spectrometric (LC-MS) Analysis of Lens Extracts

The lens extracts were centrifuged to remove high molecular mass compounds, lyophilized, and resuspended in acetonitrile/water before analysis. Materials present in the extracts were separated and analyzed by RP LC-MS/MS (ESI+) with data for the nuclear and cortex regions of the normal and cataractous lenses examined separately (Fig. 1, Table 1; structures in Figs. 2, 4). ESI analysis in positive mode confirmed the presence of 3OHKG (m/z 387.1, peak 1; Fig. 1a), along with protonated molecular ions of m/z 388.1 (peak 2; Fig. 1b), m/z 314.1 (peak 3; Fig. 1c), m/z 370.1 (peak 4; Fig. 1d), m/z 314.1 (peak 5; Fig. 1c), m/z 372.1 (peak 6; not shown), and m/z 370.1 (peak 7; Fig. 1d). These peaks also showed fragment ions consistent with the loss of a glucose residue (162.0 Da). 

Figure 1

 

LC-MS chromatograms of novel UV filter metabolites (a) 3OHKG, m/z 387.1, (b) 3OHKG-W, m/z 388.1, (c) AHAG, m/z 314.1, and (d) 3OHKG-Y and 3OHKG-D, both m/z 370.1, in lens samples that are unspiked (above) or spiked with synthetic standards (below). Peak numbers correlate with text and Table 1. Insets show expanded peaks of interest in unspiked samples.

Figure 2

 

UV-Vis profiles of synthesized standards.

Peak 1 (3OHKG).

Loss of the glucose moiety from 3OHKG (m/z 387.1) produced a fragment ion with m/z 225.1 and a stronger m/z 208.0 fragment due to further loss of ammonia (17.0 Da). The main fragment ion observed at m/z 370.1 corresponds to the mass of 3OHKG-Y and 3OHKG-D, which are formed by loss of ammonia (17.0 Da) from 3OHKG. Two minor fragment ions with m/z 314.1 (AHAG) and m/z 388.1 (3OHKG-W) were also observed. 

Peak 2 (3OHKG-W).

Peak 2 exhibited a protonated molecular ion of 1 Da greater mass than 3OHKG (m/z 388.1), which is consistent with the β-hydroxy analogue of 3OHKG (3OHKG-W). A m/z 226.0 fragment ion eventuates upon loss of the glucoside, with a weaker-intensity m/z 208.0 fragment ion due to a further loss of the water. 

Peaks 3 and 5 (Unknown 3 and AHAG).

Two peaks with m/z 314.1 (peaks 3 and 5) were observed, each having a mass consistent with the cleavage of the 3OHKG side chain at carbon 3 via a reverse Aldol reaction. Both peaks also displayed a fragment ion with m/z 152.3. Spiking of the samples with authentic AHAG confirmed that AHAG coeluted at peak 5. The species corresponding to peak 3 (Unknown 3) gave a stronger signal for a m/z 152.3 fragment ion than AHAG, but displayed no UV absorption at 360 nm. Thus, peak 3 is believed to correspond to either a contaminant or another small molecule present in the human lens that does not act as a UV filter compound. 

Peaks 4 and 7 (3OHKG-Y and 3OHKG-D).

The observation of two species with m/z 370.1 (peaks 4 and 7) is consistent with the presence of the deaminated 3OHKG-D species. 3OHKG-D (assigned to peak 7 through spiking experiments) can isomerize through a ring-closing Michael addition, resulting in peak 4 m/z 370.1 (3OHKG-Y). Both compounds gave a fragment ion with m/z 208.0. 

3OHKG-Y was differentiated from 3OHKG-D as it coeluted with a compound having m/z 324.1, consistent with thermal loss of formic acid (46.0 Da) from 3OHKG-Y in the heated source of the mass spectrometer. In contrast, 3OHKG-D does not readily lose its carboxylic acid and hence does not exhibit a coeluting species corresponding to this thermal decomposition. 

Peak 6 (AHBG).

Peak 6 (not shown) had a mass (m/z 372.1) and UV profile consistent with the structure of the known UV filter AHBG, with a fragment ion of m/z 210.0 due to the loss of the glucoside. 

Total Synthesis of Novel UV Filter Metabolites

As the quantities of the lens UV filter metabolites were too low for unambiguous structure elucidation by NMR, the proposed compounds 3OHKG-W, 3OHKG-Y, and AHAG were synthesized for comparison with the lens-derived compounds (Fig. 2). The structures of the synthesized compounds were confirmed by NMR, HRMS, and UV-Vis spectroscopy and were consistent with the above LC-MS structural assignments. Although UV-Vis spectra could not be obtained for these compounds from the lens samples, due to their low concentrations, such data were obtained for synthetic 3OHKG-W, 3OHKG-Y, and AHAG, with each of these having significant absorption maxima in the 300- to 420-nm region. The presence of 3OHKG-D in cataract human lenses has been reported previously. ¹⁵  

Quantification of UV Filter Metabolites in Human Lenses

The three new lens metabolites (3OHKG-Y, 3OHKG-W, and AHAG) and two known metabolites (3OHKG and 3OHKG-D) were quantified using LC-MS/MS detection. 

The compounds were detected by SRM of the loss of the glucoside ([M+H]⁺ → [M-glucoside+H]⁺), and the peak intensities were compared to standard curves generated from the synthetic samples. 

Normal Lenses.

All of the new metabolites were found in normal lenses, with 3OHKG-Y, 3OHKG-W, and AHAG being present at concentrations in the pmol/mg of dry protein range (Table 2). The concentrations of these compounds were correlated with the age of the lens (Fig. 3). In the normal nucleus, 3OHKG decreased in concentration with age of the lens in a statistically significant manner (Table 2). The concentration of 3OHKG approximately halved in the nucleus of lenses aged from 18 to 84 years, consistent with previous reports. ^8,16 The other compounds examined increased significantly in concentration with lens age, with the greatest increase occurring with AHAG and 3OHKG-Y. 3OHKG-W showed a trend toward an increase with age, but this was not statistically significant. 3OHKG-D was detected in low concentrations in some but not all lenses, with no statistically significant changes in concentration detected. 

Figure 3

 

Correlation plots of the concentration of novel metabolites in the nucleus (▪, solid line) and cortex (Δ, dashed line) of human lenses with age of the donor.

Table 2.

 

Concentrations of Novel and Known Lens Metabolites in Human Lens Nucleus and Cortex Extracts, and Correlation of Data With Age

Compound	Concentration Range	Pearson Coefficient	P (2-Tailed)	Significance
Nucleus
3OHKG	0.09–4.16 nmol/mg	−0.5960	0.0021	Yes
3OHKG-W	0.36–5.60 pmol/mg	0.2059	0.3705	No
3OHKG-Y	0.31–4.08 pmol/mg	0.5336	0.0072	Yes
AHAG	0.28–3.41 pmol/mg	0.5997	0.0025	Yes
3OHKG-D	0.06–0.71 pmol/mg	0.0806	0.7505	No
Cortex
3OHKG	0.05–5.61 nmol/mg	−0.4867	0.0159	Yes
3OHKG-W	0.23–1.63 pmol/mg	−0.3233	0.1778	No
3OHKG-Y	0.18–1.81 pmol/mg	−0.2843	0.1758	No
AHAG	0.08–1.14 pmol/mg	−0.0886	0.6965	No
3OHKG-D	0.04–0.13 pmol/mg	−0.6777	0.0654	No

In the normal lens cortex samples, the concentrations of most of the tested compounds did not change in a statistically significant manner. 3OHKG was an exception that showed a decrease with age, though these levels varied markedly, particularly in the younger lens samples. 

Cataract Lenses.

Two cataract lenses were also examined; larger numbers of intact cataract lenses were unavailable due to changes in clinical management and practice. These cataract lenses were found to have similar concentrations of the various compounds in both the nucleus and the cortex. Low levels of 3OHKG were detected in the lens nuclei; these values were 32% of the average 3OHKG concentration of normal lenses (2.01 nmol/mg normal lens protein versus 0.64 nmol/mg cataract lens protein), while the 3OHKG concentrations detected in cataract cortex were 28% of those detected in normal lens cortex (1.63 nmol/mg normal lens protein versus 0.46 nmol/mg cataract lens protein). 3OHKG-Y and AHAG concentrations in cataract lenses were similar to those of normal lenses. For both cataract lenses, the concentration of 3OHKG-W was close to the detection limit and therefore not quantified. 3OHKG-D was observed only at a low level in one of the cataract lenses. 

Extraction Efficiency

The levels of the UV filters and UV filter-derived compounds reported here are dependent on the efficient extraction of these metabolites from the lens samples. This was therefore assessed by two independent methods using bovine lens tissue extracted with either ethanol (method 1) or 5% KOH in 80% ethanol/water (method 2). 

Method 1 (Wood and Truscott Method ¹³ ).

Extraction efficiencies for 3OHKG, AHAG, 3OHKG-Y, and 3OHKG-W were determined; data for 3OHKG-D were not obtained as this proved to be too unstable for quantification. Reversed phase LC-MS data for 3OHKG did not show any evidence for breakdown and conversion to other UV filters (such as AHAG), and an extraction efficiency of 57 ± 4% (mean ± standard deviation) was determined. For the other species, the recovery efficiencies were AHAG, 79 ± 15%; 3OHKG-Y, 68 ± 14%; and 3OHKG-W, 41 ± 12%. These data indicate that the detection of these species in the lens samples is not a result of artifacts arising from the extraction method. All data presented (except for 3OHKG-D) in Figure 3 have been corrected with the corresponding average extraction efficiency. 

Method 2 (Inoue and Satoh Method ¹⁴ ).

Under the basic conditions of this extraction method, the 3OHKG concentration decreased and that of AHAG increased, over time. After 48 hours, ∼10% of the starting 3OHKG was still detectable by RP LC-MS, with AHAG the only other detectable product. These data indicate that 3OHKG can be converted to AHAG under basic conditions but that this is a slow process. 

Discussion

Three novel lens metabolites, 3OHKG-W, 3OHKG-Y, and AHAG, were identified and quantified in human lenses varying in age from 18 to 84 years. The synthetic compounds had high-resolution masses and fragmentation patterns consistent with the structures postulated for the materials identified in the lens extracts. The UV-Vis spectra determined for the synthetic compounds were similar to those for known UV filter compounds. ¹⁵  

In the nucleus all three metabolites increased in quantity with age, with AHAG showing a 6.5-fold increase between ages ∼20 and ∼68 (0.42–2.71 pmol/mg lens protein). 3OHKG-Y showed a 3.3-fold increase over the same age period (0.69–2.26 pmol/mg lens protein), and 3OHKG-W increased 2.8-fold (0.71–1.98 pmol/mg lens protein). The only compound that decreased in concentration over the same age period was 3OHKG (2.54–1.72 nmol/mg protein). This decrease in 3OHKG concentration levels is consistent with other reports. ¹⁷ 3OHKG-Y and 3OHKG-W are postulated to arise from a common intermediate, 3OHKG-D ([7] in Fig. 4). Both 3OHKG-Y and 3OHKG-W were found to increase in concentration with age, though 3OHKG-W is found at a higher average concentration in young lenses. In contrast, 3OHKG-Y was present at higher levels in old lenses (>57 years of age). These differences with age were not observed in the lens cortex, where the concentrations of 3OHKG-W, 3OHKG-Y, and AHAG decreased moderately with age. Overall, concentrations of each compound in the cortex were, in general, 2- to 3-fold lower than in the nucleus, and these levels did not vary dramatically with age, with the concentrations remaining at the levels detected in young lenses. 

Figure 4

 

Proposed mode of formation of 3OHKG-W [2], 3OHKG-Y [4], AHAG [5], and 3OHKG-D [7] from 3OHKG [1]. Compound numbers correlate with peak numbers, text, and Table 1.

AHAG is postulated to arise from a reverse Aldol reaction of 3OHKG, via cleavage of the amino acid side chain between the second and third carbon atoms. ¹⁸ AHAG has been reported previously as a UV filter present in cataract lenses; however, the detection of this species may be due, at least in part, to the basic extraction conditions used (5% KOH in 80% ethanol), which accelerate side-chain cleavage in 3OHKG. Thus the extraction method used previously ¹⁴ results in diminishing levels of 3OHKG over time, with almost complete conversion to AHAG after 48 hours. In the current study we provide data consistent with the presence of AHAG in normal and cataract lenses, as the extraction method employed did not give rise to any detectable artifactual generation of AHAG from 3OHKG. In contrast to the other novel lens metabolites reported here, AHAG appears to be relatively stable; nevertheless, the concentrations detected in the lens were low, likely due to the deamination of 3OHKG as a more favored reaction than the reverse Aldol reaction that yields AHAG. 3OHKG-D was observed in only some of the normal lens samples, both in the nucleus and cortex, and only in very low quantities in one of the cataract lenses tested. No significant changes in the concentration of 3OHKG-D over time were detected in the normal lenses. This is consistent with the observation that 3OHKG-D is highly reactive, thereby precluding its accumulation. Likewise, in having an absorbance maxima at 409 nm, 3OHKG-D mainly absorbs blue light. This may be another factor contributing to its greater instability. 

While 3OHKG-W, 3OHKG-Y, and AHAG all have absorption maxima in the 300- to 420-nm region, they are present in low concentrations compared to the known lens UV filters. Additionally, along with 3OHKG-D being unstable, there is evidence that 3OHKG-Y is photochemically active, with Zelentsova et al. ¹⁹ showing that the structurally similar kynurenine yellow is reactive upon UV irradiation. Thus, 3OHKG-W, 3OHKG-Y, and AHAG most likely do not function as significant UV filter compounds in the lens, but are intermediate products of 3OHKG decomposition. 

In conclusion, three novel lens metabolites, 3OHKG-Y, 3OHKG-W, and AHAG, have been identified and quantified in 24 human lenses. They are proposed to be derived from the major UV filter 3OHKG. These compounds were present at pmol/mg dry lens protein concentrations and at lower levels than 3OHKG (which is present at nmol/mg lens protein). ¹⁷ Although these metabolites are present at low concentrations, the data obtained are consistent with their presence in both normal and cataractous lenses. The concentrations of these compounds may reflect their reactivity when compared to other species, with this potentially resulting in additional products and/or binding to lens proteins. In the nucleus, 3OHKG-Y and AHAG increased in concentration with age, and higher levels were detected in cataract lenses, though additional lens analyses are required to confirm the latter observation. These metabolites, with the exception of 3OHKG-D, have UV absorbance profiles similar to those of the abundant UV filter compound 3OHKG and other known UV filters. ¹³ The findings of these novel 3OHKG-derived lens metabolites extends our understanding of the chemistry of the human lens, especially with regard to the major human lens UV filter compound, 3OHKG, and of how the concentrations of these novel species alter with both subject age and the presence of cataract. 

Supplementary Materials

docx S1 

Acknowledgments

We thank Mark Raftery of the University of New South Wales for the HRMS/MS data. 

Supported in part by a Macquarie University MQRES Scholarship, a National Health and Medical Research Council (NHMRC) grant (1008667), and an Australia Research Council (ARC) Centre of Excellence grant (CE0561607). Roger J. W. Truscott was an NHMRC Senior Research Fellow. 

Disclosure: N.A. Gad, None; J. Mizdrak, None; D.I. Pattison, None; M.J. Davies, None; R.J.W. Truscott, None; J.F. Jamie, None