BioMed Research International

BioMed Research International / 2021 / Article

Research Article | Open Access

Volume 2021 |Article ID 6668845 |

Shaohua Zhang, Keke Zhang, Wenwen He, Yi Lu, Xiangjia Zhu, "Quantitative Phosphoproteomic Comparison of Lens Proteins in Highly Myopic Cataract and Age-Related Cataract", BioMed Research International, vol. 2021, Article ID 6668845, 18 pages, 2021.

Quantitative Phosphoproteomic Comparison of Lens Proteins in Highly Myopic Cataract and Age-Related Cataract

Academic Editor: Juan M. Bueno
Received09 Dec 2020
Revised23 Mar 2021
Accepted02 Apr 2021
Published10 May 2021


Purpose. To investigate and compare the lens phosphoproteomes in patients with highly myopic cataract (HMC) or age-related cataract (ARC). Methods. In this study, we undertook a comparative phosphoproteome analysis of the lenses from patients with HMC or ARC. Intact lenses from ARC and HMC patients were separated into the cortex and nucleus. After protein digestion, the phosphopeptides were quantitatively analyzed with TiO2 enrichment and liquid chromatography-mass spectrometry. The potential functions of different phosphopeptides were assessed by Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. Results. In total, 522 phosphorylation sites in 164 phosphoproteins were identified. The number of phosphorylation sites was significantly higher in the cortex than in the nucleus, in both ARC and HMC lenses. The differentially phosphorylated peptides in the lens cortex and nucleus in HMC eyes were significantly involved in the glutathione metabolism pathway. The KEGG pathway enrichment analysis indicated that the differences in phosphosignaling mediators between the ARC and HMC lenses were associated with glycolysis and the level of phosphorylated phosphoglycerate kinase 1 was lower in HMC lenses than in ARC lenses. Conclusions. We provide an overview of the differential phosphoproteomes of HMC and ARC lenses that can be used to clarify the molecular mechanisms underlying their different phenotypes.

1. Introduction

Cataract is an age-related degenerative disease and the principal cause of blindness worldwide [1]. It frequently accompanies other eye diseases, such as high myopia, glaucoma, uveitis, and trauma, all of which displaying clinical processes distinct from those of age-related cataract (ARC). Highly myopic cataract (HMC) is more prevalent in Asia than in other regions [24]. Compared with ARC, HMC is characterized by earlier onset and cataract with greater nuclear sclerosis and rapid progression [5, 6], which implies that there are unique pathological processes involved in the development of each type of cataract. However, the underlying molecular differences between ARC and HMC remain unclear.

Lens proteins are some of the most long-lived proteins in the body and are the targets of numerous posttranslational modifications [7]. These modifications, especially phosphorylation, are implicated in the regulation of protein solubility and activities [8]. Previous studies demonstrated the differential expression of phosphorylated proteins in normal and cataractous lenses and suggested that protein phosphorylation affects the occurrence and development of cataract [9]. Given the vital role of phosphorylation in protein denaturation and the intriguing clinical differences between HMC and ARC, the differential phosphoproteomes in these two conditions warrant investigation. Differences in the phosphorylation of lens proteins could help distinguish the phenotypes of HMC and ARC lenses.

In this first comparative study of the differential phosphorylation status of HMC and ARC lenses, we quantified the phosphoproteomes of the lens cortex and nucleus separately. This was followed by Gene Ontology (GO) functional analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the differentially phosphorylated proteins to determine the implications of phosphorylation to the unique phenotype of HMC.

2. Materials and Methods

In this study, we analyzed the phosphoproteomes of HMC and ARC lenses to detect differences between HMC and ARC. The workflow of this study is shown in Figure 1.

2.1. Tissue Collection

In total, three HMC lenses and three age-matched ARC lenses were collected from patients, who provided informed consent before they underwent extracapsular cataract extraction surgery at the Eye and Ear, Nose, and Throat Hospital, Fudan University, Shanghai, China. The research strictly adhered to the tenets of the Declaration of Helsinki and was approved by the Ethics Committee of the Eye and ENT Hospital, Fudan University. Patients with an axial length of ≥26 mm were diagnosed with high myopia. In all lenses, the nucleus and cortex were separated by coring through the visual axis with a 4.5 mm diameter trephine. The lens tissues were divided into four groups: ARC lens cortex (ARC-C), ARC lens nucleus (ARC-N), HMC lens cortex (HMC-C), and HMC lens nucleus (HMC-N).

2.2. Sample Preparation

Each tissue was lysed with 200 μL of lysis buffer (4% sodium dodecyl sulfate, 100 mM dithiothreitol, 150 mM Tris-HCl, pH 8.0), disrupted with agitation using a homogenizer, and boiled for 5 min. The samples were ultrasonicated and boiled again for another 5 min. Insoluble material was removed by centrifugation at 12,000 g for 15 min. The supernatant was then collected for protein digestion. The protein concentrations were quantified with a BCA Protein Kit (Bio-Rad, Shanghai, China).

2.3. Protein Digestion

We digested the proteins in 250 μg of each sample with the filter-aided sample preparation procedure described by Wisniewski et al. [10]. Briefly, 200 μL of uric acid (UA) buffer (8 M urea, 150 mM Tris-HCl, pH 8.0) was used to remove the detergent, dithiothreitol, and other low-molecular weight components with repeated ultrafiltration. To block the reduction of cysteine residues, 100 μL of 0.05 M iodoacetamide in UA buffer was added and the samples were incubated for 20 min in the dark. The filter was washed three times with 100 μL of UA buffer and then twice with 100 μL of 25 mM NH4HCO3. The protein suspension was then digested with 3 μg of trypsin in 40 μL of 25 mM NH4HCO3. The mixture was incubated at 37°C overnight, and the resulting peptides were collected as the filtrate.

2.4. Enrichment of Phosphorylated Peptides with TiO2 Beads

The phosphopeptides were captured according to the TiO2 protocol [11], adapted for label-free quantitative proteomics. The peptides were concentrated with a vacuum concentrator and resuspended in 500 μL of loading buffer (2% glutamic acid, 65% acetonitrile (ACN), and 2% trifluoroacetic acid (TFA)). The TiO2 beads were added and then agitated for 40 min and centrifuged for 1 min at . The supernatant was mixed with another TiO2 bead, resulting in the second beads which were collected as before. The beads were washed sequentially with 50 μL of washing buffer I (30% ACN, 3% TFA) and 50 μL of washing buffer II (80% ACN, 0.3% TFA) three times to remove the remaining unabsorbed material. The phosphopeptides were finally eluted with 50 μL of elution buffer (40% ACN, 15% NH4OH) [12]. The eluates were lyophilized for further analysis.

2.5. Liquid Chromatography- (LC-) Electrospray Ionization Tandem Mass Spectrometry (MS/MS) Analysis with Q Exactive™

The peptides from each sample were desalted with a C18 Cartridge (Empore™ SPE Cartridges C18 (standard density), bed I.D. 7 mm, volume 3 mL; Sigma), concentrated with vacuum centrifugation, and equilibrated with 40 μL of 0.1% () TFA. MS experiments were performed on a Q Exactive mass spectrometer coupled to an Easy nLC™ liquid chromatography (Thermo Fisher Scientific). The phosphopeptide extract (5 μg) was injected onto a C18 reversed-phase column (Thermo Scientific EASY-Spray™ Column, 10 cm long, 75 μm I.D., 3 μm resin) in buffer A (2% ACN, 0.1% formic acid) and separated with a linear gradient of buffer B (80% ACN, 0.1% formic acid), at a flow rate of 250 nL/min over 60 min. The most abundant precursor ions from the survey scan (300–1800 m/z) for higher-energy C trap dissociation (HCD) fragmentation. The target value determination was based on predictive automatic gain control. The duration of dynamic exclusion was 25 s. Survey scans were acquired at a resolution of 70,000 at 200 , and the resolution of the HCD spectra was set to 17,500 at 200 . The normalized collision energy was 30 eV, and the underfill ratio, which specifies the minimum percentage of the target value likely to be reached at the maximum fill time, was defined as 0.1%. All MS experiments were performed in triplicate for each sample.

2.6. Sequence Database Searches and Data Analysis

All the raw data were identified with the MaxQuant software (version and screened against the UniProt human database, containing a total of 156,914 entries. The datasets were searched with a mass tolerance of 6 ppm. The search followed the enzymatic cleavage rule: trypsin/P, allowing two missed cleavages; tolerance on a mass measurement of 20 ppm; fixed modification; and carbamidomethylation of cysteines. Protein N-terminal acetylation and methionine oxidation were defined as variable modifications. The cutoff for the global false discovery rate for peptide and protein identification was set to 0.01. Label-free quantification was performed with MaxQuant, as previously described [13]. Protein abundance was calculated based on the normalized spectral protein intensity (label-free quantification (LFQ) intensity). In the quantitative comparison of groups, a protein site was included if it was identified in at least 50% of samples in at least one (HMC or ARC) patient cohort.

2.7. Bioinformatics

The GO program Blast2GO was used to annotate the differentially phosphorylated proteins and to create histograms of the GO annotations, including cell components, biological processes, and molecular functions. The KEGG database (KEGG; was used for pathway annotation. The GO terms and KEGG pathways with computed values < 0.05 were considered significantly enriched.

2.8. Statistical Analysis

To identify quantitative differences in the phosphorylation states between each group, the degree of phosphorylation at the same site was estimated as the difference ratio and a ratio of >2 was considered to indicate overabundant phosphorylation. Conversely, a reduction in -fold was considered to indicate less abundant phosphorylation [14, 15]. Statistical significance was determined with a -test. A value < 0.05 was considered statistically significant.

3. Results

3.1. Phosphoproteome Identification in HMC and ARC Lenses

The clinical information for the lens samples is given in Table S1. In this study, we identified 451 unique phosphopeptides in 164 phosphoproteins from the HMC and ARC lens samples (Table S2). Among the 164 phosphoproteins, 84 contained a single phosphorylation site, 26 contained two phosphorylation sites, and 17 contained three phosphorylation sites (Figure 2(b)). The 522 phosphorylation sites identified included 364 on serine (S), 109 on threonine (T), and 49 on tyrosine (Y) accounting for 69.7%, 20.8%, and 9.4% of the total sites, respectively (Figure 2(c)). Among the 451 phosphopeptides, 250 (55.4%), 113 (25.1%), and 49 (10.9%) had one, two, and three phosphorylation sites, respectively. The other 39 phosphopeptides had more than three phosphorylation sites (Figure 2(d)).

3.2. Identification of the Differential Phosphopeptides

To investigate the quantitative differences in the phosphorylation status of the four groups (ARC-C, ARC-N, HMC-C, and HMC-N), a two-fold change was used as the cutoff to screen for differentially phosphorylated proteins. The level of phosphorylation was higher in the cortex than the nucleus, in both the HMC and ARC lenses. Twenty-six phosphopeptides were hyperphosphorylated in the HMC-C, whereas 10 were underphosphorylated compared with the HMC-N. In the ARC lenses, we detected 104 more abundant phosphopeptides and eight less abundant phosphopeptides in the cortex relative to the nucleus. A comparison of the HMC and ARC lenses revealed that the level of phosphorylation was higher in the HMC-N group than in the ARC-N group. We identified 58 phosphopeptides that were significantly altered in the HMC-C group: 23 were more abundant and 35 were less abundant than in the ARC-C group (Figures 3(a) and 3(c)3(e)). Among the phosphosites identified, 303 (77.3%) were shared by all four groups (Figure 3(b)).

Seventeen phosphosites were exclusively detected in HMC-C, but not in HMC-N, that included cytoskeletal proteins, oxidoreductases, and binding proteins (Table 1). The phosphorylation status of S-formylglutathione hydrolase and glyceraldehyde-3-phosphate dehydrogenase was greater in the HMC-C than in the HMC-N. Meanwhile, in comparisons of ARC-C and ARC-N, a total of 33 phosphosites were exclusively detected in ARC-C and the phosphorylation degree of β-crystallin A3 (phosphorylation at site Y36) was much higher in ARC-C, with a different ratio of 34.04. Other significantly overabundant phosphosites were also detected in phakinin, β-crystallin B, and filensin () (Table 2).

Protein namesGene nameAccessionScoreModified sequencePositionFC value

Phosphosites exclusively detected in HMC-C
 Alcohol dehydrogenase 1AADH1AP07327106.43_EIGADLVLQIS(ph)KESPQEIAR_225NaN0.003
 Alcohol dehydrogenase 1AADH1AP07328105.84_S(ph)GGTLVLVGLGSEMTTVPLLHAAIR_266NaN0.039
 Aldehyde dehydrogenase family 1 member A3ALDH1A3P4789588.19_IFINNEWHES(ph)KSGK_43NaN0.026
 Beta-crystallin B2CRYBB2P4332094.45_DSSDFGAPHPQVQS(ph)VRR_186NaN0.010
 Protein 4.1EPB41P1117163.16_KLS(ph)MYGVDLHK_394NaN0.013
 Glutathione reductaseGSRP0039074.12_GHAAFT(ph)SDPKPTIEVSGK_177NaN0.008
 Leukotriene A-4 hydrolaseLTA4HP0996094.85_CS(ph)VDFTR_27NaN0.002
 Mpv17-like protein 2MPV17L2Q567V2106.13_LLS(ph)AGQLLFQGR_14NaN0.004
 cAMP-dependent protein kinase catalytic subunit alphaPRKACAP1761268.108_KGS(ph)EQESVK_11NaN0.003
 Ribose-phosphate diphosphokinasePRPS2A0A140VK4153.865_VAILVDDMADT(ph)CGTICHAADK_228NaN≤0.001
 Tryptophan-tRNA ligase, cytoplasmicWARSP2338199.802_KLS(ph)FDFQ_467NaN≤0.001
Phosphosites exclusively detected in HMC-N
 Carbonic anhydrase 4CA4P2274869.03_EQT(ph)VSM(ox)KDNVR_263NaN0.001
 Carbonyl reductase (NADPH) 1CBR1P1615275.462_LFS(ph)GDVVLTAR_30NaN≤0.001
 Beta-crystallin A4CRYBA4P5367360.564_GEY(ph)PSWDAWGGNTAYPAER_74NaN≤0.001
 Glutathione synthetaseGSSP48637156.67_QIEINTIS(ph)ASFGGLASR_149NaN0.003
Phosphosites with upregulation in HMC-C (selected)
 S-Formylglutathione hydrolaseESDP10768117.37_MSIFGHS(ph)MGGHGALICALK_1494.610.003
 Glyceraldehyde-3-phosphate dehydrogenaseGAPDHP0440667.995_IISNASCT(ph)TNCLAPLAK_1532.580.001
 Alpha-crystallin A chainCRYAAP02489123.19_QS(ph)LFRTVLDSGISEVR_512.410.003
 Fructose-bisphosphate aldolaseALDOAJ3KPS380.229_CQY(ph)VTEK_2082.390.023
Phosphosites with downregulation in HMC-C
 Gamma-crystallin CCRYGCP07315125.77_RGEYPDYQQWM(ox)GLS(ph)DSIR_730.020.036
 Fructose-bisphosphate aldolaseALDOAJ3KPS3132.4_GILAADEST(ph)GSIAKR_370.15≤0.001
 Beta-crystallin A3CRYBA1P05813184.26_VES(ph)GAWIGYEHTSFCGQQFILER_700.450.031
 Gamma-crystallin CCRYGCP07315235.83_VES(ph)GCWMLYERPNYQGQQYLLR_400.480.029

Protein namesGene nameAccessionScoreModified sequencePositionFC value

Phosphosites exclusively detected in ARC-C
Actin, cytoplasmic 1ACTA2P60709111.6_GYS(ph)FTTTAER_199NaN0.006
 Alcohol dehydrogenase 1AADH1AV9HW8993.73_SGGTLVLVGLGS(ph)EMTTVPLLHAAIR_277NaN0.024
 Alcohol dehydrogenase 1AADH1AV9HW8982.89_EIGADLVLQISKES(ph)PQEIAR_228NaN≤0.001
 Alcohol dehydrogenase 1AADH1AV9HW89105.8_S(ph)GGTLVLVGLGSEMTTVPLLHAAIR_266NaN0.001
 Cysteine proteaseATG4DB4DZK064.3_KYS(ph)IFTEKDEILSDVASR_151NaN0.001
 Alpha-crystallin A chainCRYAAP02489142.5_HFSPEDLT(ph)VK_86NaN≤0.001
 Beta-crystallin A2CRYBA2P5367285.46_LLS(ph)DCANVCER_31NaN0.026
 Beta-crystallin A4CRYBA4P5367386.48_GFQYVLECDHHS(ph)GDYK_170NaN0.015
 Beta-crystallin B1CRYBB1P53674162.4_WNTWSS(ph)SYR_129NaN0.016
 Quinone oxidoreductaseCRYZQ0825798.77_AGESVLVHGAS(ph)GGVGLAACQIAR_158NaN≤0.001
 Eukaryotic initiation factor 4A-IIEIF4A2Q14240137.9_GYDVIAQAQS(ph)GTGK_79NaN≤0.001
 Protein 4.1EPB41P1117199.02_QAS(ph)ALIDRPAPHFER_521NaN≤0.001
 Glucose-6-phosphate isomerase (fragment)GPIA0A0A0MTS265.72_ELQAAGKS(ph)PEDLER_470NaN≤0.001
 Heat shock 70 kDa protein 4HSPA4P3493263.71_AFS(ph)DPFVEAEK_76NaN≤0.001
 Glutathione synthetaseGSSP4863768.94_DGY(ph)MPRQYSLQNWEAR_270NaN0.003
 Lactase-like proteinLCTQ6UWM786.29_S(ph)AEQGLEM(ox)SR_311NaN≤0.001
 L-Lactate dehydrogenase A chainLDHCP0033890.61_S(ph)ADTLWGIQK_319NaN0.028
 Protein kinase C and casein kinase substrate in neurons 3PACSIN3D3DQR092.54_LKEVEAS(ph)K_153NaN0.001
 Peroxisome biogenesis factor 10, isoform CRA_bPEX10A0A024R0A487.26_RAS(ph)LEER_281NaN0.006
 Phosphoglycerate mutase 1PGAM1P1866956.72_FSGWYDADLS(ph)PAGHEEAKR_31NaN0.016
 cAMP-dependent protein kinase catalytic subunit alphaPRKACAP1761268.11_KGS(ph)EQESVK_11NaN≤0.001
 Ribose-phosphate diphosphokinasePRPS2A0A140VK4153.87_VAILVDDMADT(ph)CGTICHAADK_228NaN≤0.001
 Glycogen phosphorylasePYGBP0673761.64_RMS(ph)LIEEEGSKR_430NaN0.002
 SEC14-like protein 2SEC14L2O76054122.5_VGDLS(ph)PR_9NaN≤0.001
 Tryptophan-tRNA ligase, cytoplasmicWARSP2338199.8_KLS(ph)FDFQ_467NaN0.002
 Synaptobrevin homolog YKT6YKT6O1549853.03_IDWPVGS(ph)PATIHYPALDGHLSR_114NaN0.010
 14-3-3 proteinYWHAHQ0491751.15_KNS(ph)VVEASEAAYK_145NaN≤0.001
Phosphosites exclusively detected in ARC-N
 Nucleoside triphosphate pyrophosphataseASMTO9567180.98_VVLASAS(ph)PR_21NaN≤0.001
 Alpha-crystallin B chainCRYAAP02511127_LFDQFFGEHLLESDLFPTSTSLS(ph)PFYLRPPSFLR_45NaN0.010
 Crystallin gamma BCRYGBA0A0U3BWM073.78_GQMSELT(ph)DDCLSVQDR_107NaN0.004
 3-Hydroxyanthranilate 3,4-dioxygenaseHAAOP4695278.43_RLS(ph)LAPDDSLLVLAGTSYAWER_247NaN0.007
Phosphosites with upregulation in ARC-C (selected)
 Beta-crystallin A3CRYBA1P0581369.26_ITIY(ph)DQENFQGK_36####≤0.001
 Alpha-crystallin B chainCRYAAP02511177.9_RPFFPFHSPS(ph)R_21####0.049
 Alpha-crystallin B chainCRYAAP02511151.2_LEKDRFS(ph)VNLDVK_76####0.001
Phosphosites with downregulation in ARC-C
 Beta-crystallin A3CRYBA1P05813252.2_RMEFTS(ph)SCPNVSER_500.260.005
 Gamma-crystallin DCRYGDP07320151.2_RGDYADHQQWMGLS(ph)DSVR_730.270.009
 Coactosin-like proteinCOTL1Q1401970.09_FTTGDAMS(ph)KR_710.45≤0.001
 Beta-crystallin A3CRYBA1P05813123.4_WDAWS(ph)GSNAYHIER_1000.500.049

Twelve and 14 phosphosites were exclusively detected in the HMC-C and ARC-C, respectively, when comparing these groups. A protein with a high degree of phosphorylation was γ-crystallin D, which was phosphorylated at Y29 (Table 3). Phosphorylation of the lens cytoskeletal proteins, phakinin and filensin, was significantly lower in the HMC-C than in the ARC-C. In the HMC-N and ARC-N, the predominant differentially expressed phosphosites were found in the α-crystallin B chain, at S21, T170, and S76. Table 4 shows that the proteins with the most abundant phosphorylation included crystallins and structural proteins, particularly β-crystallin B1 and filensin.

Protein namesGene nameAccessionScoreModified sequencePositionFC value

Phosphosites exclusively detected in HMC-C
 Fructose-bisphosphate aldolaseALDOAJ3KPS380.229_CQY(ph)VTEK_208NaN≤0.001
 N-Acetylserotonin O-methyltransferase-like proteinASMTO9567180.979_VVLASAS(ph)PR_21NaN≤0.001
 Alpha-crystallin A chainCRYAAP02489193.97_YRLPSNVDQS(ph)ALSCSLSADGMLTFCGPK_127NaN0.047
 Beta-crystallin B1CRYBB1P53674163.34_WNTWS(ph)SSYR_128NaN0.014
 Beta-crystallin B3CRYBB3P2699869.122_CELS(ph)AECPSLTDSLLEK_42NaN≤0.001
 Gamma-crystallin CCRYGCP07315111.12_SCCLIPQT(ph)VSHR_85NaN0.003
 Glutathione reductase, mitochondrialGSRP0039074.12_GHAAFT(ph)SDPKPTIEVSGK_177NaN0.008
 Mpv17-like protein 2MPV17L2Q567V2106.13_LLS(ph)AGQLLFQGR_14NaN0.004
 Ubiquitin C variant (fragment)UBCQ59EM9122.44_TIT(ph)LEVEPSDTIENVK_30NaN≤0.001
 14-3-3 protein zeta/delta (fragment)YWHABE7EX2955.885_DICNDVLS(ph)LLEK_99NaN0.003
Phosphosites exclusively detected in ARC-C
 Alcohol dehydrogenase 1AADH1AV9HW89124.19_AMGAAQVVVTDLSATRLS(ph)K_211NaN0.048
 Alcohol dehydrogenase 2AADH1AV9HW8982.885_EIGADLVLQISKES(ph)PQEIAR_228NaN≤0.001
 Carbonyl reductase (NADPH) 1CBR1P1615275.462_LFS(ph)GDVVLTAR_30NaN≤0.001
 Alpha-crystallin B chainCRYAAP02511152.88_IPADVDPLTITS(ph)SLSSDGVLTVNGPR_135NaN≤0.001
 Beta-crystallin B1CRYBB1P53674162.38_WNTWSS(ph)SYR_129NaN0.016
 Protein 4.1EPB41P1117199.021_QAS(ph)ALIDRPAPHFER_521NaN≤0.001
 Glucose-6-phosphate isomerase (fragment)GPIA0A0A0MTS265.716_ELQAAGKS(ph)PEDLER_470NaN≤0.001
 Heat shock 70 kDa protein 4HSPA4P3493263.709_AFS(ph)DPFVEAEK_76NaN≤0.001
 L-Lactate dehydrogenase A chainLDHCP0033890.614_S(ph)ADTLWGIQK_319NaN0.028
 Phosphoglycerate mutase 1PGAM1P1866956.72_FSGWYDADLS(ph)PAGHEEAKR_31NaN0.016
 Synaptobrevin homolog YKT6YKT6O1549853.033_IDWPVGS(ph)PATIHYPALDGHLSR_114NaN0.010
 14-3-3 protein etaYWHAHQ0491751.147_KNS(ph)VVEASEAAYK_145NaN≤0.001
Phosphosites with upregulation in HMC-C (selected)
 Gamma-crystallin DCRYGDP07320130.01_HYECSSDHPNLQPY(ph)LSR_2929.250.006
 Alpha-crystallin B chainCRYAAP02511184.11_RPFFPFHS(ph)PSR_1911.990.003
 Alpha-crystallin A chainCRYAAP02489155.17_T(ph)LGPFYPSR_135.08≤0.001
 Quinone oxidoreductase PIG3TP53I3Q53FA7141.54_RGS(ph)LITSLLR_2603.92≤0.001
 Alpha-crystallin A chainCRYAAP02489123.19_QS(ph)LFRTVLDSGISEVR_512.770.013
Phosphosites with downregulation in HMC-C (selected)
 Retinal dehydrogenase 1ALDH1A1P0035278.921_YILGNPLT(ph)PGVTQGPQIDKEQYDK_3370.28≤0.001
 Brain acid soluble protein 1BASP1P80723127.95_AEGAATEEEGT(ph)PK_360.300.008

Protein namesGene nameAccessionScoreModified sequencePositionFC value

Phosphosites exclusively detected in HMC-N
 Retinal dehydrogenase 1ALDH1A1P0035278.92_YILGNPLT(ph)PGVTQGPQIDKEQYDK_337NaN0.019
 Fructose-bisphosphate aldolaseALDOAJ3KPS392.46_RTVPPAVTGITFLS(ph)GGQSEEEASINLNAINK_276NaN0.008
 Fructose-bisphosphate aldolaseALDOAJ3KPS380.23_CQY(ph)VTEK_208NaN≤0.001
 Cysteine proteaseATG4DB4DZK064.3_KYS(ph)IFTEKDEILSDVASR_151NaN0.014
 Carbonic anhydraseCA2P0091861.44_EPIS(ph)VSSEQVLK_216NaN0.001
 Alpha-crystallin A chainCRYAAP02489142.5_HFSPEDLT(ph)VK_86NaN≤0.001
 Beta-crystallin A4CRYBA4P5367386.48_GFQYVLECDHHS(ph)GDYK_170NaN≤0.001
 Beta-crystallin A4CRYBA4P5367360.56_GEY(ph)PSWDAWGGNTAYPAER_74NaN≤0.001
 Beta-crystallin B3CRYBB3P2699869.12_CELS(ph)AECPSLTDSLLEK_42NaN≤0.001
 Quinone oxidoreductaseCRYZQ0825798.77_AGESVLVHGAS(ph)GGVGLAACQIAR_158NaN0.011
 Eukaryotic initiation factor 4A-IIEIF4A2Q14240137.9_GYDVIAQAQS(ph)GTGK_79NaN0.001
 S-Formylglutathione hydrolaseESDP10768117.4_MSIFGHS(ph)MGGHGALICALK_149NaN≤0.001
 Glyceraldehyde-3-phosphate dehydrogenaseGAPDHP0440668_IISNASCT(ph)TNCLAPLAK_153NaN0.012
 Glutathione synthetaseGSSP48637156.7_QIEINTIS(ph)ASFGGLASR_149NaN0.003
 Inosine-5-monophosphate dehydrogenaseIMPDH1Q5H9Q6121.6_LVGIVT(ph)SR_234NaN0.009
 Protein kinase C and casein kinase substrate in neurons 3PACSIN3D3DQR092.54_LKEVEAS(ph)K_153NaN0.032
 Peroxisome biogenesis factor 10, isoform CRA_bPEX10A0A024R0A487.26_RAS(ph)LEER_281NaN≤0.001
 Glycogen phosphorylase, liver formPYGBP0673761.64_RMS(ph)LIEEEGSKR_430NaN0.036
 SEC14-like protein 2SEC14L2O76054122.5_VGDLS(ph)PR_9NaN≤0.001
 Ubiquitin C variant (fragment)UBCQ59EM9122.4_TIT(ph)LEVEPSDTIENVK_30NaN0.009
 14-3-3 protein zeta/delta (fragment)YWHABE7EX2955.89_DICNDVLS(ph)LLEK_99NaN≤0.001
Phosphosites exclusively detected in ARC-N
 Alpha-crystallin B chainCRYAAP02511127_LFDQFFGEHLLESDLFPTSTSLS(ph)PFYLRPPSFLR_2NaN0.010
 Beta-crystallin A3CRYBA1P05813103.2_WDAWSGSNAY(ph)HIER_174NaN0.009
 3-Hydroxyanthranilate 3,4-dioxygenaseHAAOP4695278.43_RLS(ph)LAPDDSLLVLAGTSYAWER_160NaN0.007
Protein NDRG1NDRG1Q9259796.54_S(ph)REMQDVDLAEVKPLVEK_56NaN≤0.001
Phosphosites with upregulation in HMC-N (selected)
 Alpha-crystallin B chainCRYAAP02511177.9_RPFFPFHSPS(ph)R_21####0.020
 Alpha-crystallin B chainCRYAAP0251184.3_EEKPAVT(ph)AAPK_170####0.000
 Alpha-crystallin B chainCRYAAP02511151.2_LEKDRFS(ph)VNLDVK_76####0.004
 Beta-crystallin B1CRYBB1P53674154.7_QWHLEGSFPVLAT(ph)EPPK_248####≤0.001
Phosphosites with downregulation in HMC-N
 Phosphoglycerate kinase 1PGK1P0055869.38_AHS(ph)S(ph)MVGVNLPQK_1740.380.005
 Carbonyl reductase (NADPH) 1CBR1P16152228.2_FRS(ph)ETITEEELVGLMNK_1600.470.008
 Carbonyl reductase (NADPH) 1CBR1P16152207.6_GQAAVQQLQAEGLS(ph)PR_560.470.019
 Beta-crystallin SCRYGSP22914190.8_KPIDWGAASPAVQS(ph)FRR_1720.490.016

3.3. Gene Ontology and KEGG Pathway Enrichment Analyses

The potential functions of the phosphoproteins differentially expressed in the different groups were examined by GO analysis. The GO analysis showed that the proteins differentially phosphorylated in the HMC-C and HMC-N were enriched for proteins involved in the cellular compartment, especially in the extracellular exosome, cytoplasm, and plasma. The molecular functions of these proteins were mainly related to the structure of the lens (Figure 4(a)). The GO analysis showed that the differentially phosphorylated phosphopeptides in the ARC lenses were enriched for visual perception, nerve impulse transmission, and protein homodimerization. The differentially phosphorylated phosphoproteins also included cell components, mostly related to the extracellular exosome and cytoplasm, that are involved in the structure of the lens and ATP binding.

When the HMC and ARC groups were compared, the differentially phosphorylated proteins were functionally related to the structure of the lens or were crucial for visual perception in the cortical and nuclear regions. Concerning cellular component, the differentially expressed phosphoproteins in the HMC-C and ARC-C were mainly cytoplasmic proteins. However, the differentially phosphorylated proteins in the nuclear regions of the HMC and ARC groups (HMC-N and ARC-N, respectively) were predominantly extracellular exosome proteins (Figures 4(c) and 4(d)).

We also performed KEGG pathway enrichment analysis to identify the biological pathways associated with the differentially phosphorylated proteins. The 20 most abundant enrichment terms with are shown in Figure 5. When the cortex and nucleus of HMC were compared, the most significantly enriched pathway was glutathione metabolism. However, when the HMC and ARC lenses were compared, glycolytic enzymes were most frequently differentially expressed.

As shown in Figure 6, when we compared the HMC and ARC lenses, the differentially phosphorylated proteins were enriched in the glycolysis and glutathione metabolism pathways. The key glycolytic enzyme, phosphoglycerate kinase 1 (PGK1), was the least phosphorylated protein in the HMC lenses. However, glutathione synthetase (GSS) and glutathione-disulfide reductase (GSR), the key enzymes in glutathione synthesis, were hyperphosphorylated in HMC.

4. Discussion

To improve the efficiency of identifying the phosphorylation sites in this study, we used TiO2 enrichment combined with LC-MS/MS. For the first time, two parts of the lens with known histological differences, the lens cortex and nucleus, were compared separately. Proteomic differences were detected in the different regions of the lens, and by quantifying the differences in the phosphorylated proteins between HMC and ARC, we clarified the different pathogeneses in these two phenotypes.

We identified 522 phosphorylation sites in 164 phosphoproteins in this study. Previous studies have reported 73 phosphorylation sites and 32 phosphoproteins in normal and cataractous lenses, using immobilized metal affinity chromatography and nano-LC-coupled MS/MS [15]. α-Crystallin A and α-crystallin B are the most abundantly phosphorylated proteins in the porcine lens [16]. Our data show that besides these two crystallin proteins, beta A3, beta B1, beta B2, beta S, and gamma D crystallins were also phosphorylated at many peptide sites.

In this study, the number of phosphorylation sites was significantly greater in the lens proteins of the cortex than in those of the nuclear region, in both HMC and ARC. One possible explanation to this finding is that the lens epithelium cells immediately adjacent to areas of the cortex are metabolically relatively active and metabolites decrease from the lens cortex toward the lens nucleus. Protein enzymatically phosphorylated in the outer cortex could gradually dephosphorylate nonenzymatically in the metabolically inactive nucleus. Among the phosphoproteins in these groups, a high percentage of the differentially phosphorylated proteins were crystallins and lens structural proteins, including β-crystallin, α-crystallin, phakinin, and filensin.

α-Crystallin is a small heat shock protein that maintains the transparency of the lens. Phosphorylation is considered to change its chaperone activity by inducing a change in the protein’s structure and altering the subunit exchange dynamics [17]. The phosphorylation of α-crystallin B has been shown to regulate the protein’s activity in both lenticular and extralenticular tissues [5, 1823]. The commonest functional modification sites in α-crystallin B are S19, S45, and S59. We detected these phosphopeptides in the lenses of patients with both HMC and ARC. When we compared the S19 site between these two groups, the level of phosphorylation was higher in the HMC-C than in the ARC-C. However, the pathological significance of these proteomic changes requires further analysis. There was a slight difference in the phosphorylation at S59 between the ARC-C and ARC-N and between the HMC-N and ARC-N. The phosphorylation of α-crystallin B at S59 is thought to be associated with actin nucleation and the migration of lens epithelial cells [18]. We detected no difference in the phosphorylation at S45 in any paired comparison of the four groups. Filensin and phakinin are two unique protein components of the lens fibers that assemble to form an intermediate filament, known as the beaded filament [2427]. As previously reported, filensin and phakinin in the lens fiber cells are essential for maintaining the transparency of the lens [17]. As the lens fiber differentiates and with aging, these proteins become the targets of phosphorylation as a posttranslational modification [18]. It has also been reported that the phosphorylation of intermediate filament proteins plays an essential role in regulating the kinetics of these proteins, including their solubility, conversion, and the fiber structure [1921].

In the HMC group, the phosphosites that differed strongly in their phosphorylation between the cortex and nucleus predominantly comprised enzymes involved in glutathione synthesis, including GSS and S-formylglutathione hydrolase. Our KEGG pathway analysis also showed that the largest proportion of phosphoproteins was associated with glutathione metabolism.

GSS and S-formylglutathione hydrolase catalyze key steps in glutathione synthesis. Glutathione is an essential antioxidant that protects the lens from oxidative damage [22]. The level of glutathione synthesis is lower in the cataractous lenses than in the lenses of emmetropic eyes but is lowest in myopic lenses [23]. It has also been demonstrated that eyes with high myopia are susceptible to oxidative damage and are associated with an increased incidence of nuclear cataract (with an adjusted odds ratio of 3.01) [5]. The lens typically exists in a low-oxygen environment [28, 29], and increased exposure to oxygen appears to cause cataract. Previous studies have shown that the degree of vitreous liquefaction is positively correlated with the level of nuclear opacity in the lens after adjustment for age [30]. As a possible mechanism, vitreous liquefaction increases the flow of fluid in the vitreous cavity and allows oxygen to flow from the retina to the lens. In patients with high myopia, vitreous liquefaction often occurs in the early stage of myopia and the severity of this complication increases as myopia worsens [31]. The glutathione content varies between different types of cataract. Subcapsular cataract, with an additional secondary nuclear cataract, shows a particularly rapid reduction in glutathione [32]. As a result, highly myopic eyes are more susceptible to oxidative damage than less myopic eyes, which leads to the formation of nuclear cataract. Consistent with this, our experimental data show that the number of phosphorylated glutathione synthase molecules was significantly higher in the cortex of the HMC lens than in the nucleus.

By comparing the catalogues of differentially phosphorylated protein in the cortical regions of HMC and ARC, results showed that the degree of phosphorylation of GSS and GSR, the key enzymes of glutathione synthesis, was higher in HMC than in ARC. However, the precise roles of GSR and GSS phosphorylation remain unclear.

In our comparison of the two lens tissues in HMC and ARC, the number of phosphopeptides was much higher in the HMC-N than in the ARC-N, which may be associated with the severity of nuclear cataract in patients with high myopia. Truscott [33] proposed that there is a barrier to the transport of metabolites within the lens. This barrier may increase the half-lives of reactive molecules, thus promoting the posttranslational modification of proteins in the nucleus, and may also prevent an adequate flux of antioxidants reaching the lens interior, thus allowing the oxidation of the nuclear components. Other authors have suggested that a common underlying mechanism in the pathology of cortical and nuclear cataract is the failure of the microcirculatory system to regulate the cell volume in the lens cortex or to deliver antioxidants to the lens nucleus [34]. Therefore, we suggest that the nuclear region of the cataractous lens may be a meaningful target region for the posttranslational modification of proteins. The HMC lens may be the best model to study the transfer of antioxidants to the nucleus through the barrier.

When we examined the glycolysis and metabolic pathways, which were enriched in differentially phosphorylated proteins in HMC and ARC, we found that the degree of phosphorylation of PGK1 was lower in HMC-C and HMC-N than in the ARC-C and ARC-N, respectively. A previous study showed that phosphorylation of PGK1 reduces its activity, thus reducing the glycolytic activity [35]. Glycolysis is the main source of energy generation in the unique low-oxygen environment of the eye. A reduction in energy metabolism impairs the activity of NA+-K+-ATPase in the lens, and the cascade reaction leads to an imbalance in lens homeostasis.

5. Conclusion

In summary, we analyzed the phosphoproteomes of the cortex and nucleus of HMC and ARC lenses, while considering the clinical features of the lenses. We found significant differences in the extent of protein phosphorylation and the types of proteins phosphorylated between different regions of the lens. Our results will be valuable for the future investigation of the molecular characteristics and pathological pathways underlying HMC and ARC.

Data Availability

The data supporting the findings of this study are available within the article and its supplementary materials.

Conflicts of Interest

The authors have no relevant interests to declare.

Authors’ Contributions

Shaohua Zhang and Keke Zhang contributed equally to this work.


This work was supported by grants from the National Natural Science Foundation of China (grants 81870642, 81970780, 81700819, 81470613, and 81670835), the Science and Research Fund of the Shanghai Health and Family Planning Commission (grant 20184Y0014), the Shanghai High Myopia Study Group, and the Shanghai Talent Development Fund (grant 201604).

Supplementary Materials

Supplementary 1. Table S1. The clinical information for the lens samples.

Supplementary 2. Table S2. Phosphopeptide annotation in HMC and ARC lenses.


  1. D. Pascolini and S. P. Mariotti, “Global estimates of visual impairment: 2010,” The British Journal of Ophthalmology, vol. 96, no. 5, pp. 614–618, 2012. View at: Publisher Site | Google Scholar
  2. T. Y. Wong, P. J. Foster, J. Hee et al., “Prevalence and risk factors for refractive errors in adult Chinese in Singapore,” Investigative Ophthalmology & Visual Science, vol. 41, no. 9, pp. 2486–2494, 2000. View at: Google Scholar
  3. S. Vitale, L. Ellwein, M. F. Cotch, Ferris FL 3rd, and R. Sperduto, “Prevalence of refractive error in the United States, 1999-2004,” Archives of Ophthalmology, vol. 126, no. 8, pp. 1111–1119, 2008. View at: Publisher Site | Google Scholar
  4. J. Sun, J. Zhou, P. Zhao et al., “High prevalence of myopia and high myopia in 5060 Chinese University students in Shanghai,” Investigative Ophthalmology & Visual Science, vol. 53, no. 12, pp. 7504–7509, 2012. View at: Publisher Site | Google Scholar
  5. G. L. Kanthan, P. Mitchell, E. Rochtchina, R. G. Cumming, and J. J. Wang, “Myopia and the long-term incidence of cataract and cataract surgery: the Blue Mountains Eye Study,” Clinical and Experimental Ophthalmology, vol. 42, no. 4, pp. 347–353, 2014. View at: Publisher Site | Google Scholar
  6. C. W. Pan, P. Y. Boey, C. Y. Cheng et al., “Myopia, axial length, and age-related cataract: the Singapore Malay Eye Study,” Investigative Ophthalmology & Visual Science, vol. 54, no. 7, pp. 4498–4502, 2013. View at: Publisher Site | Google Scholar
  7. Z. Kyselova, “Mass spectrometry-based proteomics approaches applied in cataract research,” Mass Spectrometry Reviews, vol. 30, no. 6, pp. 1173–1184, 2011. View at: Publisher Site | Google Scholar
  8. E. Thornell and A. Aquilina, “Regulation of alphaA- and alphaB-crystallins via phosphorylation in cellular homeostasis,” Cellular and Molecular Life Sciences, vol. 72, no. 21, pp. 4127–4137, 2015. View at: Publisher Site | Google Scholar
  9. C. H. Huang, Y. T. Wang, C. F. Tsai, Y. J. Chen, J. S. Lee, and S. H. Chiou, “Phosphoproteomics characterization of novel phosphorylated sites of lens proteins from normal and cataractous human eye lenses,” Molecular Vision, vol. 17, pp. 186–198, 2011. View at: Google Scholar
  10. J. R. Wiśniewski, A. Zougman, N. Nagaraj, and M. Mann, “Universal sample preparation method for proteome analysis,” Nature Methods, vol. 6, no. 5, pp. 359–362, 2009. View at: Publisher Site | Google Scholar
  11. A. Montoya, L. Beltran, P. Casado, J.-C. Rodríguez-Prados, and P. R. Cutillas, “Characterization of a TiO2 enrichment method for label- free quantitative phosphoproteomics,” Methods, vol. 54, no. 4, pp. 370–378, 2011. View at: Publisher Site | Google Scholar
  12. M. R. Larsen, T. E. Thingholm, O. N. Jensen, P. Roepstorff, and T. J. D. Jørgensen, “Highly selective enrichment of phosphorylated peptides from peptide mixtures using titanium dioxide microcolumns,” Molecular & Cellular Proteomics, vol. 4, no. 7, pp. 873–886, 2005. View at: Publisher Site | Google Scholar
  13. C. A. Luber, J. Cox, H. Lauterbach et al., “Quantitative proteomics reveals subset-specific viral recognition in dendritic cells,” Immunity, vol. 32, no. 2, pp. 279–289, 2010. View at: Publisher Site | Google Scholar
  14. E. J. Soderblom, M. Philipp, J. W. Thompson, M. G. Caron, and M. A. Moseley, “Quantitative label-free phosphoproteomics strategy for multifaceted experimental designs,” Analytical Chemistry, vol. 83, no. 10, pp. 3758–3764, 2011. View at: Publisher Site | Google Scholar
  15. Q. Yao, H. Li, B. Q. Liu, X. Y. Huang, and L. Guo, “SUMOylation-regulated protein phosphorylation, evidence from quantitative phosphoproteomics analyses,” The Journal of Biological Chemistry, vol. 286, no. 31, pp. 27342–27349, 2011. View at: Publisher Site | Google Scholar
  16. S.-H. Chiou, C.-H. Huang, I.-L. Lee et al., “Identification of in vivo phosphorylation sites of lens proteins from porcine eye lenses by a gel-free phosphoproteomics approach,” Molecular Vision, vol. 16, no. 35, pp. 294–302, 2010. View at: Google Scholar
  17. M. Oka, H. Kudo, N. Sugama, Y. Asami, and M. Takehana, “The function of filensin and phakinin in lens transparency,” Molecular Vision, vol. 14, pp. 815–822, 2008. View at: Google Scholar
  18. M. Ireland and H. Maisel, “Phosphorylation of chick lens proteins,” Current Eye Research, vol. 3, no. 7, pp. 961–968, 1984. View at: Publisher Site | Google Scholar
  19. N. O. Ku, J. Liao, C. F. Chou, and M. B. Omary, “Implications of intermediate filament protein phosphorylation,” Cancer Metastasis Reviews, vol. 15, no. 4, pp. 429–444, 1996. View at: Publisher Site | Google Scholar
  20. M. B. Omary, N. O. Ku, J. Liao, and D. Price, “Keratin modifications and solubility properties in epithelial cells and in vitro,” Sub-Cellular Biochemistry, vol. 31, pp. 105–140, 1998. View at: Google Scholar
  21. M. B. Omary, N. O. Ku, G. Z. Tao, D. M. Toivola, and J. Liao, ““Heads and tails” of intermediate filament phosphorylation: multiple sites and functional insights,” Trends in Biochemical Sciences, vol. 31, no. 7, pp. 383–394, 2006. View at: Publisher Site | Google Scholar
  22. A. Kamei, “Glutathione levels of the human crystalline lens in aging and its antioxidant effect against the oxidation of lens proteins,” Biological & Pharmaceutical Bulletin, vol. 16, no. 9, pp. 870–875, 1993. View at: Publisher Site | Google Scholar
  23. T. Micelli-Ferrari, G. Vendemiale, I. Grattagliano et al., “Role of lipid peroxidation in the pathogenesis of myopic and senile cataract,” The British Journal of Ophthalmology, vol. 80, no. 9, pp. 840–843, 1996. View at: Publisher Site | Google Scholar
  24. H. Maisel and M. M. Perry, “Electron microscope observations on some structural proteins of the chick lens,” Experimental Eye Research, vol. 14, no. 1, pp. 7–12, 1972. View at: Publisher Site | Google Scholar
  25. A. Merdes, M. Brunkener, H. Horstmann, and S. D. Georgatos, “Filensin: a new vimentin-binding, polymerization-competent, and membrane-associated protein of the lens fiber cell,” The Journal of Cell Biology, vol. 115, no. 2, pp. 397–410, 1991. View at: Publisher Site | Google Scholar
  26. A. Merdes, F. Gounari, and S. D. Georgatos, “The 47-kD lens-specific protein phakinin is a tailless intermediate filament protein and an assembly partner of filensin,” The Journal of Cell Biology, vol. 123, no. 6, pp. 1507–1516, 1993. View at: Publisher Site | Google Scholar
  27. J. M. Carter, A. M. Hutcheson, and R. A. Quinlan, “In vitro studies on the assembly properties of the lens proteins CP49, CP115: coassembly with alpha-crystallin but not with vimentin,” Experimental Eye Research, vol. 60, no. 2, pp. 181–192, 1995. View at: Publisher Site | Google Scholar
  28. B. M. Palmquist, B. Philipson, and P. O. Barr, “Nuclear cataract and myopia during hyperbaric oxygen therapy,” The British Journal of Ophthalmology, vol. 68, no. 2, pp. 113–117, 1984. View at: Publisher Site | Google Scholar
  29. F. Boscia, I. Grattagliano, G. Vendemiale, T. Micelli-Ferrari, and E. Altomare, “Protein oxidation and lens opacity in humans,” Investigative Ophthalmology & Visual Science, vol. 41, no. 9, pp. 2461–2465, 2000. View at: Google Scholar
  30. G. J. Harocopos, Y. B. Shui, M. McKinnon, N. M. Holekamp, M. O. Gordon, and D. C. Beebe, “Importance of vitreous liquefaction in age-related cataract,” Investigative Ophthalmology & Visual Science, vol. 45, no. 1, pp. 77–85, 2004. View at: Publisher Site | Google Scholar
  31. H. Morita, M. Funata, and T. Tokoro, “A clinical study of the development of posterior vitreous detachment in high myopia,” Retina, vol. 15, no. 2, pp. 117–124, 1995. View at: Publisher Site | Google Scholar
  32. H. Pau, P. Graf, and H. Sies, “Glutathione levels in human lens - regional distribution in different forms of cataract,” Experimental Eye Research, vol. 50, no. 1, pp. 17–20, 1990. View at: Publisher Site | Google Scholar
  33. R. J. Truscott, “Age-related nuclear cataract: a lens transport problem,” Ophthalmic Research, vol. 32, no. 5, pp. 185–194, 2000. View at: Publisher Site | Google Scholar
  34. P. J. Donaldson, A. C. Grey, B. Maceo Heilman, J. C. Lim, and E. Vaghefi, “The physiological optics of the lens,” Progress in Retinal and Eye Research, vol. 56, pp. e1–e24, 2017. View at: Publisher Site | Google Scholar
  35. X. Qian, X. Li, Z. Shi et al., “PTEN suppresses glycolysis by dephosphorylating and inhibiting autophosphorylated PGK1,” Molecular Cell, vol. 76, no. 3, pp. 516–527.e7, 2019. View at: Publisher Site | Google Scholar

Copyright © 2021 Shaohua Zhang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder

Related articles

Article of the Year Award: Outstanding research contributions of 2020, as selected by our Chief Editors. Read the winning articles.