Lectin Microarray Profiling and Relative Quantification of Glycome Associated With Proteins of Neonatal wt and rd1 Mice Retinae

Abstract

Purpose.: Progressive dynamic, relative quantitative changes were compared in glycans associated with retinal proteins of wild type (wt) and retinal degeneration 1 (rd1) mice during neonatal development and degeneration of retinae.

Methods.: Proteins extracted from retinae of postnatal days 2 (PN2), PN7, and PN14 wt and rd1 mice were labeled with Cy3-fluorescent dye. Glycome of these proteins was quantified relatively by lectin microarray technique. Net fluorescence emitted by individual complexes formed between 45 lectins and Cy3-labeled proteins was measured by evanescent-field fluorescence-assisted microarray reader.

Results.: GlcNAcβ1-oligomer and high-mannose/Manα1-6Man were major glycans associated with the proteins of PN2, PN7, and PN14 wt and rd1 mice retinae. Gal/GalNAc/Man₃-core-bi-/tri-antennary-complex, Sia2-3Galβ1-4GlcNAc, and high-mannose glycans were conjugated mainly to proteins from PN7 rd1 and PN14 wt retinae, respectively. With increasing neonatal age, mannosylated, GlcNAcβ, and sialylated (minor component) glycans were increased, and fucosylated GlcNAc/Galβ glycans were decreased significantly in wt retinal proteins. This trend was less evident in PN14 rd1 retinal proteins. Mouse retina was almost devoid of Siaα2-6 (except WGA bound Sia), Fucα1-2, and Gal/GalNAc-containing glycans. STL reacting GlcNAc oligomers were high in PN2 rd1 retinae.

Conclusions.: Quantitative dynamic, relative variation in high-mannose and GlcNAc glycans, Siaα2-3Galβ1-4GlcNAc associated with proteins from PN2, PN7, and PN14 wt and rd1 mice retinae suggested that these glycans participate in retinal development and degeneration, and may be used as markers for retinal electrophysiologic integrity during transplantation/therapy studies; Siaα2-3Galβ1-4GlcNAc–specific Agrocybe cylindracea lectin and other lectins may be used to enrich/purify retinal ribbon synapse glycoproteins and other glycoproteins including rhodopsin. Further investigations are required.

Introduction

Retinal degeneration 1 (rd1) mouse, an animal model of retinitis pigmentosa similar to that observed in humans, has a mutation in the β-subunit of cGMP phosphodiesterase-6 gene, which elevates the level of cGMP and Ca²⁺ ions in the photoreceptors. These changes modify the expression of approximately 60 other genes involved in the processes of transcription, proliferation, adhesion, and apoptosis of retinal cells. Neonatal rd1 mouse retina shows increased activities of protein kinase C, matrix metallo-proteinases (MMPs), cathepsins, modification and accumulation of fragmented proteoglycans, rapid loss of approximately 90% of the rod photoreceptors between the age of postnatal day 10 (PN10) and PN21, leading to the degeneration of retina. ^1–3  

In mammalian cells, N-/O-linked glycans or O-linked GlcNAc monomers are conjugated respectively to plasma membrane, extracellular matrix (ECM), and nucleocytoplasmic proteins at asparagine/threonine/serine and threonine/serine residues. Collectively, glycans comprise the glycome, which is diverse, cell-/tissue-type specific and influences cellular/metabolic processes. Variation in the glycans gives rise to heterogeneous population of glycoisoforms of proteins with modified structural and functional properties. ^4–10  

Glycans are implicated in the retinal development, regeneration, and physiopathologic processes due to their association with a variety of intra- and extracellular proteins. ^3,9,10 During the last 30 years, static glycans associated with glycoproteins/proteoglycans of retina of different species have been studied qualitatively by lectin-based low throughput immunohistochemical and blotting techniques. Profiling and relative quantification of the dynamic retinal glycome is crucial to understand its role in the regulation of structure and function of retinal proteins. An aberrant glycosylation of such proteins in mutant mice adversely affects the structure and function of retina. ^3–21 This has remained undetermined due to lack of relevant technologies. Since 2005, high throughput lectin microarray technique involving the use of large numbers of lectins of high purity and well defined specificities became available for differential profiling and relative quantification of the glycome associated with complex biological samples. ²² Therefore, relative quantitative, dynamic changes in the glycome associated with the proteins of neonatal PN2, PN7, and PN14 wild type (wt) and rd1 mice retinae were compared by lectin microarray technique ²² and the results were correlated with known retinal proteins and physiopathologic processes. 

Methods

Collection of Neonatal Mice Retinae

Six each of PN2, PN7, and PN14 wt and rd1 mice (total number 36) obtained from the animal colony of Lund University were decapitated. Retinae were dissected out from the enucleated eyes and stored at −80°C until analyzed. The experiments were conducted according to the guidelines of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and with the approval (# M 221-09) of the Swedish National Animal Care and Ethics Committee at Lund University. 

Retinal Proteins: Extraction, Estimation, Labeling With Cy3 Fluorescent Dye

Four retinae from two each of PN2, PN7, and PN14 wt and rd1 mice were pooled to form one sample. Thus, a panel of 18 samples, three each from PN2, PN7, and PN14 wt and rd1 mice retinae, was formed. Retinal proteins were extracted by sonication in 500 μL phosphate buffered saline containing 1% each of a cocktail of EDTA-free protease inhibitors and Triton X-100 detergent (PBST) to extract proteins, including the membrane-bound ones. 

Total protein content of the retinal extracts was estimated at 562 nm by using Micro-BCA protein assay reagent kit (PIERCE #23235; Thermo Scientific, Rockford, IL) and a multi-well plate reader (SpectraMax M5; Molecular Devices, Sunnyvale, CA). Retinal extracts were diluted with PBST to get a protein concentration of 50 μg mL⁻¹. A 1.25 μg aliquot of proteins in 25 μL volume of retinal extracts was mixed with 100 μg monoreactive Cy3 fluorescent dye (#PA23011; GE Healthcare, Sykesville, MD) and incubated in dark for two hours at room temperature. Excess of free Cy3 fluorescent dye was removed by gel filtration on 0.5 mL Zeta desalt spin columns (PIERCE #89882; Thermo Scientific). 

Lectin Microarray Profiling and Relative Quantification of Retinal Glycome

Identification and relative quantitative profiling of glycans associated with PN2, PN7, and PN14 wt and rd1 mice retinal proteins was done on LecChip Ver. 1.0 (GP BioSciences Ltd., Yokohama, Japan) by lectin microarray technique supported by Bioinformatics tools. ²² A LecChip Ver. 1.0 (GP BioSciences Ltd.) has seven wells, each with 45 lectins (Table 1) immobilized in a pattern of 9 × 5 columns spotted in triplicate. These lectins specifically react with a wide range of N-/O-glycans linked to proteins. For nomenclature, abbreviation, source, and basic carbohydrate specificities of lectins, see the Lectin Frontier Database (available in the public domain at http://riodb.ibase.aist.go.jp/rcmg/glycodb/LectinSearch and Ref. 23). 

Table 1. 

 

Abbreviations and Name of Sources of Lectins

Lectins appreciably reactive with the retinal glycans*
STL	Solanum tuberosum	UDA	Urtica dioica
DSA	Datura stramonium	LEL	Lycipersicon esculentum
NPA	Narcissus pseudonarcissus	WGA	Triticum aestivum
HPA	Helix pomatia	Jacalin	Artocarpus integliforia
AOL	Aspergillus oryzae	ACA	Amaranthus caudatus
GSL-IB4	Griffonia simplicifolia	LCA	Lens culinaris
GNA	Galanthus nivalis	AAL	Aleuria aurantia
PTL-I	Psophocarpus tetragonolobus	PHA-E	Phaseolus vulgaris
ConA	Canavalia ensiformis	HHL	Hippeastrum hybrid
Calsepa	Calystegia sepium	ACG	Agrocybe cylindracea
RCA-120	Ricinus communis	PNA	Arachis hypogaea
TxLC-I	Tulipa gesneriana	TJA-II	Tricosanthes japonica
Lectins negligibly reactive with the retinal glycans*
MAL	Maackia amurensis	MAH	Maackia amurensis
SNA	Sambucus nigra	SSA	Sambucus sieboldiana
TJA-I	Tricosanthes japonica	LTL	Lotus tetragonolobus
PSA	Pisum sativum	UEA-I	Ulex europaeus
DBA	Dolichos biflorus	VVA	Vicia fava
EEL	Euonymus europaeus	WFA	Wisteria floribunda
MPA	Maclura pomifera	ECA	Erythrina cristagalli
BPL	Bauhinia purpurea	ABA	Agaricus bisporus
PWM	Phytolacca americana	PHA-L	Phaseolus vulgaris
GSL-IA4	Griffonia simplicifolia	GSL-II	Griffonia simplicifolia
SBA	Glycine max

 Adapted from study of Hirabayashi J et al., ²³ Lectin Frontier Database, available in the public domain at http://riodb.ibase.aist.go.jp/rcmg/glycodb/LectinSearch.

*  Lectin grouping is based on results of this study.

Lectin-Ligand Binding Curves.

Retinal proteins extracted from PN14 wt and rd1 mice, and labeled with Cy3 fluorescent dye were diluted serially with Probing Solution (GP BioSciences Ltd.) to get protein concentration of 2000, 1000, 500, 250, 125, 62.5, and 31.25 ng mL⁻¹. A 100 μL aliquot of diluted sample was loaded into a prewashed well of the LecChip Ver. 1.0 (GP BioSciences Ltd.) and incubated for 17 hours at 20°C. Net fluorescence intensities of lectin-ligand complexes were measured four times by the GlycoStation Reader 1200 (GP BioSciences Ltd.). Images of the fluorescent lectin microarrays were acquired by using the evanescent-field fluorescence scanner (GlycoStation Reader 1200; GP BioSciences Ltd.) with an exposure time of 133 msec and Camera Gain of 85, 95, 105, 115, and 125. Results of the measurements were analyzed with help of the GlycoStation Tools Pro Suite 1.5 (GP BioSciences Ltd.). To expand the dynamic range of the measured data, a gain merging method was applied. ²⁴ Binding curves between each of the 45 lectins and 31.25 to 2000 ng mL⁻¹ concentrations of proteins from PN14 wt and rd1 retinae were determined. Protein concentration of 62.5 ng mL⁻¹ was selected to determine optimum signal intensity, without signal saturation for lectin-ligand binding. 

Profiling and Relative Quantification of Retinal Glycome.

For differential profiling and relative quantification of glycans, proteins extracted from triplicate samples of PN2, PN7, and PN14 wt and rd1 retinae were labeled as before, and serially diluted. A 100 μL aliquot of each sample, providing 62.5 ng mL⁻¹ proteins, was applied per well of the LecChip Ver. 1.0 (GP BioSciences Ltd.) and processed as described above for lectin-ligand binding curve studies. Net proportional intensities of the fluorescence signals (given as arbitrary units [AU]), emitted by the complexes formed between a lectin and specific glycan associated with the retinal protein(s), were considered to represent relative quantities of glycans. 

Statistical Analysis

Mean ± SEM values of proteins and fluorescent intensities of lectin-ligand complexes of PN2, PN7, and PN14 wt and rd1 retinal proteins were compared for statistical significance of the differences. Bars represent mean ± SEM values. One-way ANOVA and Fisher's protected least significant differences post hoc comparisons were made (StatView Software; SAS, Chicago, IL). P > 0.05 (nonsignificant), P ≤ 0.05 (significant), P ≤ 0.01 (very significant), and P ≤ 0.001 (highly significant) were denoted. 

Results

The rd1 mouse model of retinitis pigmentosa shows rapid degeneration of retina due to loss of approximately 90% rod photoreceptors between the age of PN10 and PN21. ^1–3 Retinal proteins and glycans from wt and rd1 mice were analyzed and compared at three time points, namely PN2, PN7, and PN14. They represent stages of progressive changes in neonatal development of wt and rd1 retinae except that PN14 rd1 represents the disease-associated stage when rapid degeneration of rd1 mouse retina still is in progress. Compared to the static qualitative changes as reviewed in the Introduction, this study identified different types of glycans that show dynamic and relative quantitative changes in their nature, and suggest a relationship between glycans and development of neonatal wt and rd1 mice retinae; and to a limited extent between glycans and retinal degeneration, especially at neonatal age of PN2 and PN7 rd1 mice. 

Retinal Protein Content and Glycome Profile

Protein content (mean ± SEM) of PN2, PN7, and PN14 wt (up to PN14) and rd1 (up to PN7) mice retinae was significantly (P ≤ 0.05) increased with age. However, protein content in PN14 rd1 mice retinae was significantly (P ≤ 0.05) lower than that in PN14 wt retinae (Fig. 1). 

Figure 1

 

Total protein (mean ± SEM μg mL⁻¹) content of retinal extracts from PN2, PN7, and PN14 wt and rd1 mice was increased significantly (P ≤ 0.05) with increasing age. However, total protein content of PN14 rd1 extracts was significantly lower than that of PN14 wt extracts.

Lectin-Glycan Binding and Related Curves.

Of 45 lectins, 24 lectins showed appreciable binding to a variety of glycans associated with the proteins extracted from PN14 wt and rd1 mice retinae. Binding reaction of PN14 wt and rd1 retinal proteins with the remaining 21 lectins was negligible. Binding properties (Fig. 2) and binding curves (Fig. 3) of retinal protein glycans with lectins are retina specific. 

Figure 2

 

Representative scans of reactions between Cy3-labeled proteins extracted from PN2, PN7, and PN14 wt and rd1 mice retinae and 45 lectins show age- and strain-dependent differences in lectin binding. Cy3-labeled BSA marker was spotted in the four vertical spots on the left.

Figure 3

 

Binding curve (net fluorescence intensities) studies of the 45 lectins with 31.25 to 2000 ng mL⁻¹ concentration of Cy3-labeled proteins from PN14 wt (red circles) and rd1 (green circles) mice retinae showed appreciable reactivity with 24 lectins, and reactivity with the remaining 21 lectins was negligible.

Differential Profiling of the Glycome.

Twenty-five different types of glycans, low in Sialic acid, were identified from wt and rd1 (Fig. 4) mice retinal proteins by 24 lectins in the microarray. Of the 24 lectins 16, namely STL, UDA, DSA, LEL, NPA, WGA, HPA, Jacalin, AOL, ACA, GSL-I4B, LCA, GNA, AAL, PTL-I, and PHA-E, reacting with retinal glycans showed net fluorescence intensities ranging between approximately 900 and approximately 50 to 75 AU (Fig. 4). These retinal glycans are listed here, in a decreasing order of their lectin-ligand fluorescence intensities (Table 2). GlcNAc oligomer glycans recognized by STL and complex type N-glycan having outer galactose and bisecting GlcNAc recognized by PHA-E were, respectively, the most abundant (∼900 AU) and least abundant (∼50–75 AU) glycans associated with the proteins from wt and rd1 mice retinae (Table 2). Overall, there is wide variation in the nature (Table 2) and levels of glycans associated with retinal proteins of PN2, PN7, and PN14 wt and rd1 (Fig. 4) mice. 

Figure 4

 

Relative quantification (mean ± SEM) of fluorescence intensities of Lectin microarray profiles of the glycans associated with proteins extracted from PN2, PN7, and PN14 wt (top) and rd1 (bottom) mice retinae show diversity and varying quantitative profiles of glycans.

Table 2. 

 

Lectin Microarray Profile of the Glycome Associated With the Retinal Proteins of PN2, PN7, and PN14 wt and rd1 Mice Showed Following Glycans in Both Types of Retinae and Suggested Similarities in the Nature of Their Glycans

Glycans in Quantitatively Decreasing Order	(AU)	Lectin
GlcNAc oligomers, Oligosaccharide with GlcNAc, MurNAc	(∼900)	STL
GlcNAcβ1-4GlcNAc, Mixtures of Man5 to Man9*		UDA
(GlcNAcβ1-4)n, Galβ1-4GlcNAc†		DSA
GlcNAc trimers/tetramers†		LEL
High-Man, Manα1-6Man		NPA
Sia, GlcNAc*	(∼300)	WGA
α-linked terminal GalNAc†		HPA
Galβ1-3GalNAc, GalNAc*		Jacalin
Fucα1-6GlcNAc(core Fuc)†		AOL
Galβ1-3GalNAc		ACA
α-linked Gal	(∼150)	GSL-IB4
Fucα1-6GlcNAc, α-D-Glc, α-D-Man†		LCA
High-Man, Manα1-3Man*		GNA
Fucα1-6GlcNAc, Fucα1-3(Galβ1-4)GlcNAc†		AAL
α-linked terminal GalNAc		PTL-I
Complex type N-glycan with outer Gal and bisecting GlcNAc*	(∼50–75)	PHA-E

 Net fluorescence intensities of the lectin-glycan complexes (Fig. 4) are listed in decreasing order and ranged between approximately 900 AU for STL and approximately 50 to 75 AU for PHA-E ligands.

*  Fluorescence intensities increased with increasing age of wt and rd1 mice.

†  Fluorescence intensities decreased with increasing age of both wt and rd1 mice.

Siaα2-3Galβ1-4GlcNAc, mannosylated glycans; and galactosylated glycans, recognized by two groups of lectins, namely ACG, Con A, HHL, Calsepa; and TxLC-I, RCA 120, PNA, TJA-II, were associated, respectively, with the retinal proteins of PN14 wt and PN7 rd1 (Fig. 4) mice. Fluorescence intensities of the above-referred glycans bound by Con A, HHL, TxLC-I, RCA-120, PNA TJA-II were up to approximately 50 AU, and those bound by ACG and Calsepa were up to approximately 100 AU (Table 3). 

Table 3. 

 

Lectin Microarray Profile of the Glycome Associated With the Retinal Proteins of PN2, PN7, and PN14 wt and rd1 Mice Showed That the Following Glycans Were Associated Mainly With the Proteins From PN14 wt and PN7 rd1 Retinae

Glycans Conjugated to Retinal Proteins	Lectin
Mainly conjugated to retinal proteins of PN14 wt mice
High-Man, Manα1-6(Manα1-3)Man*	Con A
High-Man, Manα1-3Man, Manα1-6Man*	HHL
Mannose, maltose†	Calsepa
Siaα2-3Galβ1-4GlcNAc†	ACG
Mainly conjugated to retinal proteins of PN7 rd1 mice
Galβ1-4GlcNAc*	RCA-120
Galβ1-3GalNAc*	PNA
Man3 core bi-/tri-antennary complex type  N-glycan, GalNAc*	TxLC-I
Fucα1-2Galβ1-> or GalNAcβ1-> groups at  nonreducing end*	TJA-II

*  Net fluorescence intensities of the lectin-glycan complexes (Fig. 4) were ≤50.

†  Net fluorescence intensities of the lectin-glycan complexes (Fig. 4) were approximately 100 AU.

Intensities of fluorescence emitted by glycans Fucα1-2-, Fucα1-3-, Galα1-3-, Galβ1-3-/GalNAc, and Siaα2-6- (linked in different arrangements) bound by the remaining 21 lectins, were negligible (≤10 AU, Table 4) after reaction with proteins from PN2, PN7, and PN14 wt and rd1 mice retinae. 

Table 4. 

 

Lectin Microarray Profile of the Glycome Associated With the Retinal Proteins of PN2, PN7, and PN14 wt and rd1 Mice Showed Negligible Levels of the Following Sialylated, Galactosylated, and Fucosylated Glycans

Glycans Imperceptible in wt and rd1 Retinal Proteins	Lectin
Siaα2-3Galβ1-4GlcNAc	MAL
Siaα2-3Galβ1-3(Siaα2-6)GalNAc	MAH
Siaα2-6Gal/GalNAc	SNA
Siaα2-6Gal/GalNAc	SSA
Siaα2-6Gal/GalNAc	TJA-I
Fucα1-3(Galβ1-4)GlcNAc, Fucα1-2Galβ1-4GlcNAc	LTL
Fucα1-6GlcNAc, α-D-Glc, α-D-Man	PSA
Fucα1-2 Galβ1-4GlcNAc	UEA-I
Agalactosylated tri-/tetra-antennary N-glycans, GlcNAc	GSL-II
Tri-/tetra-antennary complex type N-glycan	PHA-L
α- or β-linked terminal GalNAc, GalNAcα1-3Gal	SBA
α-linked terminal GalNAc, GalNAcα1-3Gal	VVA
Galα1-3Gal, blood group B antigen	EEL
α-linked GalNAc	GSL-IA4
GalNAcα1-3GalNAc, blood group A antigen	DBA
Galβ1-4GlcNAc	ECA
GalNAcβ1-4GlcNAc, Galβ1-3(−6)GalNAc	WFA
Galβ1-3GalNAc, GalNAc	MPA
Galβ1-3GalNAc, GalNAc	BPL
Galβ1-3 GalNAc	ABA
(GlcNAcα1-4)n	PWM

 Net fluorescence intensities of the lectin-glycans complexes were ≤10 AU.

Relative Quantitative Changes in the Glycome During Retinal Development.

GlcNAcβ1-4 (tri-, tetra-, and oligomers)–containing glycans associated with the proteins from neonatally developing PN2, PN7, and PN14 wt and rd1 (Fig. 4) retinae reacted maximally (>400 AU) and in decreasing order with the lectins STL, UDA, DSA, LEL, and NPA (Table 2). Net fluorescence intensity of GlcNAcβ1-oligomer; and high-mannose, Manα1-6Man glycans, respectively, reacting with STL and NPA (Fig. 4) was high but nonsignificantly variable with the increasing neonatal age of wt and rd1 mice retinae. 

Fluorescence intensity of glycans (GlcNAcβ1-4)_n, GalNAcβ1-4GlcNAc; and Fucα1-6GlcNAc-with-core-fucose, respectively, reacting with DSA and AOL (Fig. 4) was significantly decreased (Table 5), with an increase in neonatal age of both wt and rd1 mice retinae (except PN14 rd1 retinae). GlcNAc trimer/tetramer glycans associated with wt retinal proteins reacting with LEL were decreased nonsignificantly (Table 5). However, such decrease in glycans conjugated to proteins of PN14 rd1 (Fig. 4) retinae was significant (Table 5). 

Table 5. 

 

Lectin Microarray Profile of the Glycome Associated With the Retinal Proteins of PN2, PN7, and PN14 wt and rd1 Mice Showed Age-Dependent Significant Quantitative Changes in Fluorescence Intensities of Lectin-Ligand Complexes

	PN2-7	PN2-14	PN7-14	PN2-7	PN2-14	PN7-14
Glycans increased with neonatal age
Complex type N-glycan outer Gal and bisecting GlcNAc (PHA-E)	NS	NS	NS	*	NS	NS
High-Man, Manα1-3Man (GNA)	†	†	NS	†	†	NS
Man, maltose (Calsepa)	–	–	‡	–	–	NS
GlcNAcβ1-4GlcNAc, Mixtures of Man5 to Man9 (UDA)	‡	‡	†	NS	†	NS
Galβ1-3GalNAc, GalNAc (Jacalin)	NS	*	NS	†	†	NS
Sia, GlcNAc (WGA)	‡	†	NS	NS	†	NS
Glycans decreased with neonatal age
Fucα1-6GlcNAc, α-D-Glc, α-D-Man (LCA)	†	‡	†	NS	NS	NS
Fucα1-6GlcNAc (core Fuc) (AOL)	†	‡	‡	†	†	NS
Fucα1-6GlcNAc, Fucα1-3(Galβ1-4)GlcNAc (AAL)	NS	*	*	NS	NS	NS
(GlcNAcβ1-4)n, Galβ1-4GlcNAc (DSA)	†	‡	‡	‡	‡	NS
GlcNAc tri-/tetramers (LEL)	NS	NS	NS	NS	†	NS
α-linked terminal GalNAc (HPA)	‡	‡	‡	NS	*	NS

 Abbreviated names of Lectins are given within parentheses. NS, nonsignificant.

*   P ≤ 0.05.

†   P ≤ 0.01.

‡   P ≤ 0.001.

The α-linked terminal GalNAc glycans reacting with HPA (Fig. 4) were significantly decreased in wt retinal proteins, whereas in rd1 retinal proteins such decrease was significant only at PN14 stage (Table 5). Fucα1-6GlcNAc, α-D-glucose, α-D-mannose; and Fucα1-6GlcNAc, Fucα1-3(Galβ1-4)GlcNAc glycans, respectively, reacting with LCA and AAL were decreased with neonatal age in wt and rd1 retinal proteins (Fig. 4), but this trend was significant only in wt retinal proteins (Table 5). 

Comparison of net fluorescence intensities (≤100 AU) of the glycans associated with proteins of neonatal wt and degenerating rd1 retinae showed that mannose; high-mannose, Manα1-6(Manα1-3)Man; high-mannose, Manα1-3Man, Manα1-6Man; and Siaα2-3Galβ1-4GlcNAc glycans, respectively, reacting with Calsepa, ConA, HHL, and ACG (Fig. 4) were present mainly in the proteins from PN14 wt mice retinae (Table 3). Mannose glycans of retinal proteins reacting with Calsepa (Table 3) and GlcNAcβ1-4GlcNAc, Man₅ to Man₉ oligomers reacting with UDA were increased significantly with increasing neonatal age of wt mice (Table 5). However, in rd1 retinal proteins the latter glycans were increased significantly only at PN14 stage (Table 5). High-mannose, Manα1-3Man; and sialic acid, GlcNAc glycans, respectively, reacting with GNA and WGA, and conjugated to wt and rd1 retinal proteins, were increased with age (Fig. 4), but the increase was more significant in wt retinal proteins (Table 5). GalNAcβ1-3GalNAc, GalNAc; and complex type N-glycan with outer galactose and bisecting GlcNAc glycans in wt and rd1 retinal proteins, respectively, reacting with Jacalin and PHA-E were increased with neonatal age (Fig. 4), but the increase was significant only in rd1 retinal proteins (Table 5). 

Relative Quantitative Changes in the Glycome During Retinal Degeneration.

Galβ1-4GlcNAc; Galβ1-3GalNAc; Man₃ core, bi-/tri-antennary complex type N-glycan, GalNAc; and Fucα1-2Galβ1 > or GalNAcβ1- > groups at nonreducing end, respectively, reacting with RCA-120, PNA, TxLC-I, and TJA-II (≤ 50 AU, Table 3) were present mainly in retinal proteins of PN7 rd1 (Fig. 4) mice. 

GlcNAc oligomers in proteins of PN2 rd1 mice retinae reacting with STL were significantly higher (Table 6). As compared to decreasing trend of Fucα1-6GlcNAc, Fucα1-3(Galβ1-4)GlcNAc; and (GlcNAcβ1-4)_n, Galβ1-4GlcNAc glycans in wt proteins, respectively, reacting with AAL and DSA (Fig. 4), the rd1 retinal proteins showed significantly higher level of these glycans (Table 6). 

Table 6. 

 

Lectin Microarray Profile of the Glycome Associated With the Retinal Proteins of PN2, PN7, and PN14 wt and rd1 Mice Showed Significant Quantitative Changes in the Fluorescence Intensities of Lectin-Ligand Complexes, Mainly Formed by rd1 Retinal Proteins

Modification in the Intensities* of Lectin-Glycans Complexes	Lectin
Increased in PN2 rd1 retinal proteins
GlcNAc oligomers, Oligosaccharide with GlcNAc,  MurNAc†	STL
Unlike decrease in PN14 wt retinal proteins, it increased in PN14 rd1
Fucα1-6GlcNAc, Fucα1-3(Galβ1-4)GlcNAc†	AAL
(GlcNAcβ1-4)n, Galβ1-4GlcNAc‡	DSA
Decreased in PN7 rd1 retinal proteins
Fucα1-6GlcNAc (core Fuc)§	AOL
(GlcNAcβ1-4)n, Galβ1-4GlcNAc‡	DSA
High-Man, Manα1-6Man†	NPA
Decreased in PN14 rd1 retinal proteins
Man, Maltose§‖	Calsepa
High-Man, Manα1-3Man, Manα1-6Man‡‖	HHL

*  Fluorescence intensities.

†   P ≤ 0.05.

‡   P ≤ 0.01.

§   P ≤ 0.001.

 ‖ Glycan not detected in PN2 retinal extracts.

(GlcNAcβ1-4)_n, Galβ1-4GlcNAc; and Fucα1-6GlcNAc(core-fucose)–containing glycans, respectively, reacting with DSA and AOL were significantly lower in PN7 rd1 retinal proteins (Table 6). Mannose, high-mannose; Manα1-3Man, Manα1-6Man containing glycans, respectively, reacting with Calsepa; NPA, and HHL (Fig. 4) were significantly lower in PN7 rd1 and PN14 rd1 retinal proteins (Table 6). 

Discussion

Glycans participate in protein folding, quality control, targeting to cell organelles, receptor regulation, cell signaling, metabolism, adhesion, differentiation, proliferation, apoptosis, and inflammation. Unlike template driven biosynthesis of proteins and nucleic acids, glycans are secondary gene products. Diversity of the glycome and glycoisoforms of glycoproteins depend on the following factors; namely, distribution, concentration, and activities of membrane-bound glycosyl transferases and glycosidases; availability of nucleotide sugar donors; and glycoprotein transit time through the endoplasmic reticulum. ^9,10,25–36  

Glycans of several glycoproteins associated with static retinal cells and ECM, namely opsins, interphotoreceptor retinal binding glycoprotein, α-dystroglycan—a component of the dystrophin-glycoprotein complex, mucin-like glycoproteins, proteinases, namely MMPs and serine proteinases, sialoglycoproteins associated with cones and rods (SPACR), and SPACR covalently linked to chondroitin sulfate (SPACRCAN) have been profiled only qualitatively. ^{7,14–21,37–42} Most of the proteins possessing or interacting with the glycans observed during this study are yet to be identified and characterized/quantified by glycoproteomics, glycan arrays and other related technologies. Meanwhile, the present study compares the relative quantitative changes in the glycome profiles of retinal proteins during progressive neonatal development of PN2, PN7, and PN14 wt retina, and of disease-associated degeneration of rd1 mice retinae. 

Lectin-glycan binding curves and glycan profiling studies showed that GlcNAcβ1-4GlcNAc, GlcNAc trimers/tetramers, GlcNAc oligomers, (GlcNAcβ1-4)_n, Manα1-6Man, mixtures of Man₅ to Man₉, high-Man, and Galβ1-4GlcNAc glycans constitute the major glycan components associated with the proteins of developing PN2, PN7, and PN14 wt and rd1 mice retinae. Nature of the glycome of wt and rd1 retinal proteins is retina-specific and similar during development, but varies quantitatively with neonatal age. Age-dependent down-regulation of GlcNAc glycans, as shown by decreased binding of DSA and LEL to the glycans associated with the proteins of PN14 wt and rd1 mice retinae, indicated that, in some respects, the process of trimming of these glycans in wt and rd1 mice retinae was similar during neonatal development. 

Almost negligible levels of Fucα1-2-, Fucα1-3-, Galα1-3-, Galβ1-3-/GalNAc, and Siaα2-6–linked glycans were observed in the glycome of wt and rd1 mice retinal proteins, suggesting that the mouse retina is deficient in enzymes required for the biosynthesis of such glycans. Observation of significantly higher level of STL-specific glycan GlcNAc oligomers in PN2 rd1 suggested that the glycan profile begins to change at an early stage of development of neonatal mouse retina, as reported previously. ³ Present quantitative studies showed that the nature of retinal glycans is diverse, dynamic, and retina-specific, and varies with the stage of development of wt and of degenerating rd1 mice retinae. 

Different glycans associated with retinal proteins were either decreased or increased with increasing neonatal age of wt and rd1 retinae, and to a limited extent in PN14 rd1 retina undergoing rapid degeneration. A decrease in the levels of most of the glycans during neonatal retinal development or degeneration indicated trimming of glycans, possibly by an age-dependent increase in the activities of specific glycosidases. Similarly, an increase in the levels of only a few glycans indicated extension of these glycans, which can be attributed to an increase in the activities of specific glycosyl transferases, but is a limited process. Such changes in the glycome of proteins of wt and rd1 mice retinae possibly generate glycoisoforms ^18,37,43 of retinal proteins, which bring about stage-specific changes in the structure and function of the retina during its development and degeneration. 

Sialic acid (possibly a terminal one, with penultimate galactose), detected by WGA, was a minor component (∼300 AU) of glycans in PN2, PN7, and PN14 wt and rd1 retinal proteins, and was increased significantly in PN14 wt retinal proteins. Compared to wt retinal proteins, lesser increase in sialic acid and nonsignificant decrease in α-linked terminal GalNAc glycans in rd1 retinal proteins indicated that, in absence of terminal sialic acid, the penultimate galactose moiety (of rd1 retinal protein glycans) was externalized and predisposed the retinal cells to apoptosis. ³  

Net biosynthesis of mouse retinal proteins was increased with the neonatal age; however, the protein content in PN14 rd1 retina was lower than that in PN14 wt retina due to proteolysis by increased activity of proteinases, namely MMPs and cathepsins. ^3,44 MMP-9 is a glycoprotein present on the cell surface and ECM, and has varying proportion of O-/N-linked galactosylated glycans. Presence of relatively fewer galactosylated N-glycans generates free and more active glycoisoforms of MMP-9, which degrade the ECM. ⁴³ Relatively higher levels of galactosylated glycans in rd1 retinal proteins observed during this study suggested that higher activity of MMP-9 reported previously in rd1 retinal proteins ³ possibly is due to the presence of such glycoisoforms of MMP-9. 

Rhodopsin, a glycoprotein, constitutes more than 90% of the total proteins in the disc membranes of rod outer segment (ROS), and undergoes trimming of galactose and sialic acid residues during transport from Golgi apparatus to the ROS. Rhodopsin is conjugated to high-mannose glycans, namely Man₃GlcNAc₃, along with transient and varying levels of galactosylation and/or sialylation. However, in vitro studies have shown that rd mouse retina has relatively higher galactosylation of rhodopsin. ^{18,19,45–48} Present in vivo study also showed higher galactoseβ-1 glycan (recognized by lectins RCA-120, TJA-II, and PNA) levels in the PN7 rd1 retinal proteins and lower levels of high-mannose glycans in PN14 rd1 retinal proteins (recognized by lectins Con A, HHL, and Calsepa). These results suggested that there is an aberrant trimming of galactose moiety of glycans conjugated to rd1 retinal proteins, namely the major protein rhodopsin. Such aberrations, possibly due to decreased activity of galactosidases, may be responsible for mislocalization and degradation of rhodopsin in rd1 mouse photoreceptors as has been reported in previous studies with rds mouse, rd1 mouse, and RCS rat. ^1,3,18,44,49  

Significantly higher levels of high-mannose glycans bound by lectins HHL, Con A, and Calsepa; and of Siaα2-3Galβ1-4GlcNAc bound by lectin ACG were associated mainly with PN14 wt retinal proteins. Omori et al. reported that Siaα2-3Galβ1-4GlcNAc conjugated to α-dystroglycan component of dystrophin glycoprotein complex maintains the electrophysiologic integrity of the ribbon synapse between photoreceptors, bipolar-, and horizontal-cells of wt mouse retina. ⁷ Therefore, higher proportion of Siaα2-3Galβ1-4GlcNAc in PN14 wt retinal proteins compared to those in PN14 rd1 retina appears to be a marker for determining the integrity of the ribbon synapse junctions. 

From the comparative lectin microarray profiles of the glycome of retinal proteins, the following conclusions were deduced: that the nature of glycans in PN2, PN7, and PN14 wt and rd1 retinal proteins is diverse, dynamic, retina-specific, and correlates quantitatively with neonatal retinal development and, to a limited extent, with retinal degeneration; the quantification of retinal glycans, namely Siaα2-3Galβ1-4GlcNAc, could be used as a diagnostic marker for evaluating the retinal electrophysiologic integrity during transplantation and therapeutic studies; and Siaα2-3Galβ1-4GlcNAc-specific Agrocybe cylindracea lectin (ACG) and few other lectins may be used to enrich/purify retinal ribbon synapse glycoproteins and other glycoproteins including rhodopsin. However, more studies are required to confirm these aspects and to correlate further changes in the glycome with the structure and function of retinal proteins and retina. 

Acknowledgments

The author thanks Sten Andréasson for infrastructural facilities; Masao Yamada, Kyoko Yokota, and Ryoko Sawada, GP BioSciences Ltd., Yokohama (Japan) for lectin microarray analyses; Per Ekström for providing mice; Birgitta Klefbohm for collecting retinae; Poonam Ahuja Jensen and Sanjay Ahuja for statistical analyses, preparation of figures, and database searches; and Leif Johnson for reorganizing Figure 3. 

Supported by Stiftelsen Kronprinsessan Margaretas Arbetsnämnd för Synskadade (KMA), 645 40 Strängnäs, and Stiftelsen för Synskadade i.f.d. Malmöhus län, 205 20 Malmö, Sweden. 

Disclosure: S. Ahuja, None 

Footnotes

 Supported by The Stiftelsen Kronprinsessan Margaretas Arbetsnämnd för synskadade (KMA, Sweden).

Footnotes

 Disclosure: S. Ahuja, None