This article is mentioned in:

Introduction

N-linked glycosylation is a well-studied protein post-translational modification (PTM) that occurs at the Asn residue in the consensus motif Asn-X-Ser/Thr, where X is any amino acid except Pro (1). N-glycosylation modulates the folding, stability, trafficking and turnover of proteins, especially those of secreted or membrane attached proteins, which are involved in various cell processes such as cell-cell interaction or intracellular signaling (2–4). As N-glycosylation is involved in important cell functions, numerous N-glycosylation sites are evolutionarily conserved (5).

We hypothesized that the losses of certain ancestrally conserved N-glycosylation sites during evolution may have been involved in the acquisition of novel human phenotypes. The loss of N-glycosylation often disrupts the normal function of proteins due to improper folding, trafficking, or activity of the proteins (6,7). A proteome-wide analysis of non-synonymous single-nucleotide variations in the N-glycosylation motifs of human proteins revealed that 259 sites were lost because of missense substitutions, some of which are involved in various diseases (8). Although loss of a glycosylation modification usually results in disadvantageous phenotypes, some losses may be beneficial and fixed in humans during evolution. For example, loss of the glycan moiety N-glycolylneuraminic acid from cell surface proteins by the inactivation of the CMAH gene, encoding CMP-N-acetylneuraminic acid hydroxylase, was associated with the evolution of resistance to a certain type of malaria in early humans, although this loss subsequently led to susceptibility to other pathogens (9,10).

A large number of N-glycosylation sites identified from non-human animals and a suitable bioinformatics procedure are necessary to identify cases where ancestrally conserved N-glycosylation sites were lost during human evolution. An ideal dataset for this analysis is the N-glycoproteome data obtained from mouse tissues and plasma using high-throughput mass spectrometry (11). Previously, a bioinformatics method was used to identify novel gains of N-glycosylation sites during human evolution (12). In the present study, the procedure involved a simple modification to identify losses of ancestral N-glycosylated Asn residues during human evolution following the divergence of the Euarchonta lineage from the Glires lineage. Additionally, a comprehensive literature survey was performed to infer the possible functional outcomes of these changes, especially for human-specific losses.

Materials and methods

Mouse N-glycosylation site data

For the N-linked glycosylation dataset from a non-human proteome, we initially tested mouse data in the UniProt database. However, there were only 419 experimentally verified mouse N-glycosylation sites (as of December 20, 2013). Therefore, mouse N-glycoproteome dataset from Zielinska et al was utilized (11). This dataset consisted of 6,367 N-linked glycosylation sites in 2,352 proteins. Approximately 74% of the sites in the UniProt database were re-identified in this data set.

Mammalian orthologous proteins

Mammalian orthologs of the mouse glycosylated proteins were obtained from the University of California Santa Cruz (UCSC) Genome Browser Database (http://genome.ucsc.edu). The 'CDS FASTA alignment from multiple alignments' data, derived from the 'multiz100way' alignment data prepared from 100 vertebrate genomes (13), were downloaded using the Table Browser tool of the UCSC Genome Browser (14). Orthologous protein sequences from 62 mammalian species were extracted from these alignment datasets. The selected mammalian species included humans, chimpanzees, gorillas, orangutans, gibbons, rhesus macaques, crab-eating macaques, baboons, green monkeys, marmosets, squirrel monkeys, bushbabies, treeshrews, lesser Egyptian jerboas, prairie voles, Chinese hamsters, golden hamsters, mice, rats, naked mole rats, guinea pigs, chinchillas, brush-tailed rats, rabbits, pikas, pigs, alpacas, Bactrian camels, dolphins, killer whales, Tibetan antelopes, cattle such as cows, sheep, and goats, horses, white rhinoceroses, cats, dogs, ferrets, pandas, Pacific walruses, Weddell seals, black flying foxes, megabats, David's myotis bats, microbats, big brown bats, hedgehogs, shrews, star-nosed moles, elephants, cape elephant shrews, manatees, cape golden moles, tenrecs, aardvarks, armadillos, opossums, Tasmanian devils, wallabies and platypuses. Detailed information on species and genome assemblies is available at the UCSC Genome Browser web site (http://hgdownload.cse.ucsc.edu/goldenPath/hg19/multiz100way).

Computational screening for candidate lost N-glycosylation sites

The total number of mouse N-glycosylation sites in the data set from Zielinska et al was 6,367 (11). The 'multiz100way' alignment data, containing 57,289 alignment sets, were analyzed to identify human and other mammalian orthologs of each of the mouse N-glycosylated proteins (Fig. 1). Ad hoc Perl scripts were used to analyze the data. There were 1,658 orthologous protein datasets containing human and mouse protein sequences. This dataset covered 4,633 mouse N-glycosylation sites. From each dataset, the mammalian sequences were extracted and realigned using MUSCLE (http://www.drive5.com/muscle) (15).

Figure 1 Summary of procedure for identifying loss of ancestral N-glycosylation sites during human evolution. Computational screening and manual inspection were employed to identify the loss of ancestral N-glycosylation sites in human proteins during human evolution.

Each of the positions that aligned with a mouse N-glycosylation site was examined using ad hoc Perl scripts. Sites that were conserved in humans, where the human protein had a consensus N-glycosylation motif, were discarded. Sites where ≥30% non-Euarchonta mammals did not have an Asn residue, indicating a frequent loss in these species, were also discarded. A total of 47 sites in 43 protein alignments were obtained after this computational screening step.

Manual inspection to select lost N-glycosylated Asn residues in the human lineage

As a final step, we manually scrutinized the 47 candidates to identify highly probable instances of N-glycosylation site loss during evolution of the human lineage. In each dataset, the species that had many gaps compared to other mammals were removed. When the mouse sequence utilized from Zielinska et al (11) differed from that of the UCSC database by at least three residues, the case was discarded as the orthology of the aligned proteins could not be guaranteed. We also discarded cases in which the mouse N-glycosylation site did not conform to the canonical sequence, or cases showing low sequence conservation among mammals.

Finally, 40 ancestral N-glycosylation sites in 37 proteins were identified to be lost during human evolution. The human and mouse protein sequences in the UCSC alignment were mapped to UniProt database sequences to utilize the UniProt annotation record. We examined the multiple sequence alignment and the mammalian phylogenetic tree to infer the timing of the loss of the N-glycosylated Asn residue.

Results and Discussion

Identification of N-glycosylation sites lost during human evolution and timing of loss

We applied a bioinformatics procedure previously developed to identify novel N-glycosylation sites during human evolution, with modifications (12). Initially, there were 6,367 experimentally identified mouse N-glycosylation sites from 2,352 proteins in the dataset from Zielinska et al (11) and 57,289 orthologous protein sequence alignments from 62 mammalian species extracted from the UCSC 'multiz100way' data (13,14). These data were analyzed to collect N-glycosylation sites lost during human evolution after the Euarchonta (primates and treeshrews) diverged from the Glires (rodents and rabbits).

As a result, 40 N-glycosylation sites in 37 proteins were identified to have been lost during human evolution (Table I). Of the 37 proteins, three proteins encoded by the ICAM1, LRP2 and MASP2 genes had each lost two N-glycosylation sites (nos. 13 and 14 for ICAM1, 23 and 24 for LRP2, and 27 and 28 for MASP2), and the remaining 34 proteins had lost one site each. Fig. 2 shows the number of N-glycosylation sites that have been lost in each common ancestor along the human lineage: humans, three; humans and chimpanzees, two; African great apes, six; great apes, one; apes, two; catarrhines, three; simians, 19; primates, three; and Euarchonta, one.

Figure 2 Timing of loss of ancestral N-glycosylation sites in the human lineage. The number of lost ancestral N-glycosylation sites is shown on the branch of each corresponding clade.

Table I List of ancestral N-glycosylation sites that were lost during human evolution.

Table I

List of ancestral N-glycosylation sites that were lost during human evolution.

No.	UniProt ID	Position	Sequencea	Clade	Gene	Protein
1	ABCA1_HUMAN	1499	LPPPQRKQNTADILQDLTGRNISDYLVKTYV	Simians	ABCA1	ATP-binding cassette sub-family A member 1
	ABCA1_MOUSE	1499	LPPPQRKQKTADILQNLTGRNISDYLVKTYV
2	ADAM9_HUMAN	636	TKCGAGKICRNFQCVDASVLNYDCDVQKKCH	Primates	ADAM9	Disintegrin and metalloproteinase domain-containing protein 9
	ADAM9_MOUSE	636	TKCDAGKICRNFQCVNASVLNYDCDIQGKCH
3	ASM3A_HUMAN	367	QYYLNLTEANLKGESIWKLEYILTQTYDIED	Simians	SMPDL3A	Acid sphingomyelinase-like phosphodiesterase 3a
	ASM3A_MOUSE	364	QYYLNLTEANLKGESNWTLEYVLTQAYSVAD
4	C4BPA_HUMAN	67	PPTLSFAAPM-DITLTETRFKTGTTLKYTCL	Simians	C4BPA	C4b-binding protein α chain
	C4BPA_MOUSE	74	PPAIPNALPA-D–VNRTDFESHTTLKYECL
5	CAD13_HUMAN	489	GPVFYPDPMMVTRQEDLSVGSVLLTVNATDP	African great apes	CDH13	Cadherin-13
	CAD13_MOUSE	489	GPVFYPDPMMVTKQENISVGSVLLTVNATDP
6	CBG_HUMAN	224	QPFDLASTREENFYVDETTVVKVPMMLQSST	Simians	SERPINA6	Corticosteroid-binding globulin
	CBG_MOUSE	217	LPFSPENTREEDFYVNETSTVKVPMMVQSGN
7	CELR1_HUMAN	2140	QVDGARALQLVRALRSATQHTGTLFGNDVRT	Humans	CELSR1	Cadherin EGF LAG seven-pass G-type receptor 1
	CELR1_MOUSE	2155	RMDGNRSLRLAKALRNATQGNSTLFGNDVRT
8	CPN2_HUMAN	311	LTHNQLETVAEGTFAHLSNLRSLMLSYNAIT	Primates	CPN2	Carboxypeptidase N subunit 2
	CPN2_MOUSE	311	LSYNQLETIPEGAFTNLSRLVSLTLSHNAIT
9	CSF1R_HUMAN	493	EHNQTYECRAHNSVGSGSWAFIPISAGAHTH	Simians	CSF1R	Macrophage colony-stimulating factor 1 receptor
	CSF1R_MOUSE	491	KHNMTYFCKTHNSVGNSSQYFRAVSLGQSKQ
10	CTL4_HUMAN	198	TNV–TPPALPGITNDTTIQQGISGLIDSLN	Euarchonta	SLC44A4	Choline transporter-like protein 4
	CTL4_MOUSE	196	PNI–TLPEDLRI-NNTTVSNGISGLLDSIN
11	DCC_HUMAN	60	EPSDAVTMRGGNVLLDCSAESDRGVPVIKWK	Primates	DCC	Netrin receptor DCC
	DCC_MOUSE	60	EPSDAVTMRGGNVLLNCSAESDRGVPVIKWK
12	FETUA_HUMAN	99	TLETTCHVLDPTPVARCSVRQLKEHAVEGDC	Catarrhines	AHSG	α-2-HS-glycoprotein
	FETUA_MOUSE	99	TLETTCHALDPTPLANCSVRQLTEHAVEGDC
13	ICAM1_HUMAN	359	GVPAQPLGPRAQLLLKATPEDNGRSFSCSAT	Simians	ICAM1	Intercellular adhesion molecule 1
	ICAM1_MOUSE	362	GVEPRPPTPQVQFTLNASSEDHKRSFFCSAA
14	ICAM1_HUMAN	47	SPSKVILPRGGSVLVTCSTSCDQPKLLGIET	African great apes	ICAM1	Intercellular adhesion molecule 1
	ICAM1_MOUSE	47	HPREAFLPQGGSVQVNCSSSCKEDLSLGLET
15	IGSF5_HUMAN	160	FIPSVNLVVAENEPCEVTCLPSHWTRLPDIS	Simians	IGSF5	Immunoglobulin superfamily member 5
	IGSF5_MOUSE	146	NIPSNNLIVTEGEPCNVTCYAVGWTSLPDIS
16	ITB5_HUMAN	479	GCSVGLEPNSARCNGSGTYVCGLCECSPGYL	Simians	ITGB5	Integrin β-5
	ITB5_MOUSE	479	GCSTGL-PNSARCSGNGTYTCGLCECDPGYL
17	LAMA1_HUMAN	1337	IKASYGQGLQQSRISDISMEVGRKAEKLHPE	Simians	LAMA1	Laminin subunit α-1
	LAMA1_MOUSE	1344	IKASYGQGLQQSRIANISMEVGRKAVELPAE
18	LAMA2_HUMAN	923	DAVDAKNCQPCRCNAGGSFSEVCHSQTGQCE	Great apes	LAMA2	Laminin subunit α-2
	LAMA2_MOUSE	919	DAVNAKNCQPCRCNINGSFSEICHTRTGQCE
19	LAMP5_HUMAN	102	IALTRGAEVKGRCGHSQSELQVFWVDRAYAL	Simians	LAMP5	Lysosome-associated membrane glycoprotein 5
	LAMP5_MOUSE	102	ISLTRGAEVKGHCGHNESELEVFWVDHAYTL
20	LAT3_HUMAN	54	ILKNEGFYSSTCPAESSTNTTQDEQRRWPGC	African great apes	SLC43A1	Large neutral amino acids transporter small subunit 3
	LAT3_MOUSE	54	MLKKEGFYSSLCPAENRTNTTQDEQHQWTSC
21	LCAP_HUMAN	447	NWGLLTFREETLLYDSNTSSMADRKLVTKII	Humans and chimpanzees	LNPEP	Leucyl-cystinyl aminopeptidase
	LCAP_MOUSE	447	NWGLLTFREETLLYDNATSSVADRKLVTKII
22	LPP2_HUMAN	156	SVYVQLEKVCRGNPADVTEARLSFYSGHSSF	Simians	PPAP2C	Lipid phosphate phosphohydrolase 2
	LPP2_MOUSE	155	SGYVQLE-VCRGSPANVTEARLSFYSGHSSF
23	LRP2_HUMAN	1450	SLLLLVASQNKIIADSVTSQVHNIYSLVENG	Catarrhines	LRP2	Low-density lipoprotein receptor-related protein 2
	LRP2_MOUSE	1451	NLLLVVASRDKIIMDNITAHTHNIYSLVQDV
24	LRP2_HUMAN	3838	CLDASDEADCPTRFPDGAYCQATMFECKNHV	Simians	LRP2	Low-density lipoprotein receptor-related protein 2
	LRP2_MOUSE	3840	CLDASDESACPTRFPNGTYCPAAMFECKNHV
25	LYAM1_HUMAN	226	THPLGNFSFSSQCAFSCSEGTNLTGIEETTC	African great apes	SELL	L-selectin
	LYAM1_MOUSE	226	IHPLGNFSFQSKCAFNCSEGRELLGTAETQC
26	MA2B1_HUMAN	345	KNLDKLIRLVNAQQAKGSSVHVLYSTPACYL	Simians	MAN2B1	Lysosomal α-mannosidase
	MA2B1_MOUSE	345	KNMDKLIRLVNAQQVNGSLVHVLYSTPTCYL
27	MASP2_HUMAN	103	TLCGQESTDTERAPGKDTFYSLGSSLDITFR	African great apes	MASP2	Mannan-binding lectin serine protease 2
	MASP2_MOUSE	103	TLCGQESTDTEQAPGNDTFYSLGPSLKVTFH
28	MASP2_HUMAN	642	DSCRGDSGGALVFLDSETERWFVGGIVSWGS	Apes	MASP2	Mannan-binding lectin serine protease 2
	MASP2_MOUSE	641	DSCRGDSGGALVFLDNETQRWFVGGIVSWGS
29	MERTK_HUMAN	97	QVTSVESKPLPPLAFKHTVGHIILSEHKGVK	Simians	MERTK	Tyrosine-protein kinase Mer
	MERTK_MOUSE	91	QVTSTASKLLPPVAFNHTIGHIVLSEHKNVK
30	MET_HUMAN	358	FGVFAQSKPDSAEPMDRSAMCAFPIKYVNDF	Simians	MET	Hepatocyte growth factor receptor
	MET_MOUSE	357	FGVFAQSKPDSAEPVNRSAVCAFPIKYVNDF
31	PTPRB_HUMAN	709	VRECSFSSLTPGRLYTVTITTRSGKYENHSF	Simians	PTPRB	Receptor-type tyrosine-protein phosphatase β
	PTPRB_MOUSE	710	VSECSFSSLTPGRLYNVTVTTKSGNYASHSF
32	PTPRF_HUMAN	950	AWDPPVLAERNGRIISYTVVFRDINSQQELQ	Simians	PTPRF	Receptor-type tyrosine-protein phosphatase F
	PTPRF_MOUSE	941	TWDPPVLAERNGHITNYTVVYRDINSQLELQ
33	SIAT9_HUMAN	280	LFKSVDFNWLQAMVKKETLPFWVRLFFWKQV	Humans	ST3GAL5	Lactosylceramide α-2,3-sialyltransferase
	SIAT9_MOUSE	279	LFKSVDFKWLQAMVKNESLPFWVRLFFWKQV
34	ST14_HUMAN	489	WADCTDHSDELNCSCDAGHQFTCKNKFCKPL	Catarrhines	ST14	Suppressor of tumorigenicity 14 protein
	ST14_MOUSE	489	WADCPDYSDERYCRCNATHQFTCKNQFCKPL
35	STAB2_HUMAN	63	LNLGVKCPDGYTMITSGSVGVRDCRYTFEVR	Apes	STAB2	Stabilin-2
	STAB2_MOUSE	71	VNIAVKCPDGYIKITNGTVGVRDCRYSLKIQ
36	SUSD2_HUMAN	703	FCNFDVAATGSLSTGTATRVAHQLHQRRMQS	African great apes	SUSD2	Sushi domain-containing protein 2
	SUSD2_MOUSE	700	FCILDVMSTGSSSVGNATRIAHQLHQHRLKS
37	TMM62_HUMAN	384	SGPIFVLKWNPRNYSSGTHNIEVIVQDSAGR	Simians	TMEM62	Transmembrane protein 62
	TMM62_MOUSE	384	SGPIFILKWNPRNYSNGTHTIEVFVQDSAGR
38	VGFR3_HUMAN	582	ELLEGQPVLLSCQADSYKYEHLRWYRLNLST	Simians	FLT4	Vascular endothelial growth factor receptor 3
	VGFR3_MOUSE	582	DPLEGQSVRLSCRADNYTYEHLRWYRLNLST
39	VNN1_HUMAN	146	NSIYVVANIGDKKPCDTSDPQCPPDGRYQYN	Humans and chimpanzees	VNN1	Pantetheinase
	VNN1_MOUSE	148	NSIYVVANMGDKKPCNTSDSHCPPDGRFQYN
40	VSI10_HUMAN	100	ATSLHIESLSLGDEGIYTCQEILNVTQWFQV	Humans	VSIG10	V-set and immunoglobulin domain-containing protein 10
	VSI10_MOUSE	121	AGALRIEALRLEDDGNYTCQEVLNETHWFPV

a The N-glycosylation motif. Bold, N-X-S/T in mouse protein.

Of the 37 N-glycosylation sites that were lost in the human lineage since the divergence of the Euarchonta and the Glires, three events occurred in human proteins after the divergence of humans and chimpanzees (Table I, nos. 7, 33 and 40 and Fig. 3). The residue positions for these human-specific losses are Ser-2140 in cadherin EGF LAG seven-pass G-type receptor 1 encoded by the CELSR1 gene, Lys-280 in lactosylceramide α-2,3-sialyltransferase encoded by the ST3GAL5 gene, and Ile-100 in the V-set and immunoglobulin domain-containing protein 10 encoded by the VSIG10 gene.

Figure 3 Human-specific losses of ancestral N-glycosylation sites. The ancestral N-glycosylation sites and the surrounding regions of (A) CELSR1 Ser-2140, (B) ST3GAL5 Lys-280 and (C) VSIG10 Ile-100 are presented. The ancestral N-glycosylation consensus sequences are highlighted in grey, and corresponding sequences that lost the consensus, in black. The adjacent conserved N-glycosylation site Asn-108 in VSIG10 is indicated by plus signs (+++). The residues that are identical to those in the human sequence are indicated by dots (.). Dashes (−) denote alignment gaps. In some species, the sequences were not determined. hum, humans; hac, humans and chimpanzees; aga, African great apes; gra, great apes; ape, apes; cat, catarrhines; sim, simians; pri, primates; eua, Euarchonta; gli, Glires; lau, Laurasiatheria; afr, Afrotheria; xen, Xenarthra; mar, Marsupialia; and mon, Monotremata.

Human-specific loss of N-glycosylation at the amino acid position 2140 of CELSR1

The human cadherin EGF LAG seven-pass G-type receptor 1 or CELSR1, encoded by the CELSR1 gene, is a heavily glycosylated protein with 20 glycosylation sites (http://www.uniprot.org/uniprot/Q9NYQ6). Sequence comparison revealed that an ancestrally conserved glycosylation site at position 2140 was altered from Asn to Ser in humans following the human-chimpanzee divergence (Fig. 3A). The other mammals examined have a conserved Asn residue, conforming to the N-glycosylation motif consensus.

The CELSR1 protein is a member of the flamingo cadherin protein family, which are proteins located at the plasma membrane with seven transmembrane domains (16,17). It has nine cadherin domains, seven epidermal growth factor-like repeats and two laminin A G-type repeats. This gene is highly expressed during mouse embryonic development, especially in the central nervous system (16,17). Mutations in this protein were reported to cause neural tube defects and caudal agenesis in humans (18,19). Therefore, CELSR1 may play an important role in contact-mediated signaling during nervous system formation in early embryogenesis. CELSR1 also plays an important role in the development of other organs, such as lung branching morphogenesis (20), intraluminal valve formation in lymphatic vessels (21), and hair follicle polarization and orientation (22).

Therefore, changes in the CELSR1 protein may be involved in the evolution of the nervous system, lung, lymphatic system, or hair patterns. However, a probable direct phenotypic consequence of the loss of the N-glycosylation site at position 2140 in humans remains to be determined.

Human-specific loss of N-glycosylation at the amino acid position 280 of ST3GAL5

The human lactosylceramide α-2,3-sialyltransferase, encoded by the ST3GAL5 gene, which is also known as ganglioside GM3 synthase or sialyltransferase 9 (SIAT9), has three N-glycosylation sites (http://www.uniprot.org/uniprot/Q9UNP4). A sequence comparison revealed that the human protein lost a conserved N-glycosylation site at 280 (Asn to Lys) following the human-chimpanzee divergence (Fig. 3B). All of the other mammals analyzed, except three, have the N-glycosylation consensus sequence at this site. A loss of the N-glycosylation consensus motif was also identified in guinea pigs, chinchillas, and brush-tailed rats (also known as degus), which have a Gly residue instead of Asn at the corresponding position. The three species belong to the rodent clade Caviomorpha (23), suggesting that the Asn-to-Gly change occurred in an ancestor of the three mammals.

The ST3GAL5 gene encodes a sialyltransferase, a type II membrane protein that catalyzes the formation of GM3, a glycosphingolipid enriched in neural tissue, by adding sialic acid to lactosylceramide (24,25). GM3 is known to participate in the induction of cell differentiation, modulation of cell proliferation, and integrin-mediated cell adhesion.

Mutations in this gene are associated with several neurological disorders, such as Amish infantile epilepsy syndrome (26), Salt and Pepper syndrome characterized by severe intellectual disability, epilepsy, scoliosis, choreoathetosis, dysmorphic facial features and altered dermal pigmentation (25), or disruption of the structural integrity and function of cochlear hair cells (27). Therefore, the ST3GAL5 enzyme is crucial for normal neural development and function. The loss of an ancestrally conserved N-glycosylation site may be associated with a novel phenotype in the nervous system and function in humans, which may be demonstrated by molecular functional analysis.

Human-specific loss of N-glycosylation at position 100 of VSIG10

The VSIG10 gene encodes for V-set and immunoglobulin domain-containing protein 10. The human VSIG10 protein has nine N-glycosylation sites (http://www.uniprot.org/uniprot/Q8N0Z9). In the present study, we found that this protein lost an ancestrally conserved site at position 121, specifically, an Asn-to-Ile mutation abolished the N-glycosylation consensus (Fig. 3C). Of note, the consensus motif was also independently lost in squirrels and chinchillas. VSIG10 is a single-pass type I membrane protein containing a V-set domain, two immunoglobulin domains, and an I-set domain, which is present in cell adhesion molecules. No known molecular or biological function of VSIG10 has been reported.

In conclusion, we have identified 40 cases for loss of ancestrally conserved N-glycosylation sites, three of which are human-specific. Two human-specific losses occurred in the CELSR1 and ST3GAL5 proteins, which play indispensable roles in the normal development and function of the nervous systems. This finding suggests that the loss of N-glycosylation sites in these proteins may be associated with the evolution of human cognitive function. We suggest that a loss of ancestrally conserved N-glycosylation sites may result in the evolution of novel phenotypes, and the cases identified in the present study may serve as immediate targets for functional analyses to elucidate the molecular basis for an explanation of human phenotype evolution.

Acknowledgments

This study was supported by the National Research Foundation of Korea (NRF) grant (NRF-2012R1A1B3001513) funded by the Ministry of Education, Science and Technology, Republic of Korea.

References