Across deserts and Andes: a genetic exploration of GlyCam-1 in domestic camelids

Abstract

Background

Lactophorin is a major whey protein in camelid milk that prevents fat globule aggregation and inhibits spontaneous lipolysis. It has also been proposed to play immunological functions, including the prevention of mastitis in lactating animals and the suppression of pathogen replication in the respiratory and gastrointestinal tracts of suckling offspring. In this study, we explore the genetic variation of the glycosylation-dependent cell adhesion molecule-1 (GlyCam-1) gene in camelids, which encodes lactophorin, and examine the functional implications of the identified polymorphisms.

Results

The regulatory regions and the complete gene were sequenced and analysed in the Old World Camelids (OWC, dromedary and bactrian) and New World Camelids (NWC, llama and alpaca) using an integrated approach of molecular techniques and bioinformatics. The GlyCam-1 gene spans 2,567 bp in OWC and 2,504 bp in NWC and consists of 4 exons and 3 introns, highlighting its conserved structure. Sequencing results revealed inter- and intraspecies genetic variation, with 18 polymorphic sites in dromedaries, 7 in bactrian camels, 32 in alpacas, and 40 in llamas. Significant exonic polymorphisms were observed in NWC, potentially affecting gene expression and protein structure. Notably, the p.Glu46Lys and p.Ser134Pro variants in NWC were predicted to have a deleterious effect on protein function. Regulatory region analysis identified SNPs predicted to alter transcription factor binding sites (TFBS). In dromedaries, the g.563C > T substitution was predicted to affect the NF-κB binding, whereas in NWC, g.163A > G was predicted to modify a MZF1 binding site. The first SNP was associated with increased gene expression, and the latter was linked to significantly lower gene expression as a result of dual-luciferase reporter gene assay. Five haplotypes were observed in both NWC, with AGG and AGGG being the most prevalent in alpacas (0.707) and llamas (0.803), respectively; whereas a deviation from Hardy-Weinberg equilibrium was found for the SNP g.163A > G in alpacas.

Conclusions

This study underscores the evolutionary conservation of GlyCam-1, highlights the importance of the found genetic variants and their potential use as candidate markers for future association studies with immune regulation and dairy traits.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12917-025-05196-8.

Background

Milk is one of the primary sources of high-quality proteins in the human diet. Milk proteins are generally divided into two main groups: caseins, which form large colloidal aggregates, and whey (or serum) proteins, which represent the soluble fraction of milk. In camel milk, caseins account for 61.8–88.5% of total proteins, while whey proteins represent only 11.5–38.2%, depending on the study and sample location [1]. However, reported proportions may vary among studies due to differences in analytical techniques, camel breeds, lactation stages, and regional conditions [2, 3]. Despite their lower proportion, whey proteins are important in the food industry for their nutritional value and for their physical, chemical, physiological, functional, and technological properties [4]. Camel whey proteins also represent a potential source of bioactive peptides with health-promoting benefits [4].

Camel milk whey differs from cow milk in that it lacks β-lactoglobulin and contains higher levels of antimicrobial agents [5]. Camel whey contains five main components: serum albumin, immunoglobulin G (IgG), lactoferrin, camel whey basic protein, and α-lactalbumin [6]. Lactophorin is yet another important camel whey protein that in bovine has been named proteose peptone 3 (PP3), whereas in mice and rat have been called glycosylation-dependent cellular adhesion molecule 1 (GlyCam-1) [7, 8]. Lactophorin is also highly- and specifically-expressed in the mammary glands of lactating bovines [9], goats [10], and sheep [11]. However, lactophorin concentration in camel milk is approximately three times more than expressed in cow milk at 954 mg L^-1 vs. 300 mg L^-1, respectively [12].

Recent studies highlight the role of GlyCam-1 gene in traits related to dairy production. In cattle the GlyCam-1 has been located within quantitative trait loci (QTLs) associated with milk protein content, yield, and fat percentage [13]. In sheep, it ranks among the ten most highly expressed genes during peak lactation [11]. In goats, genetic analyses show evidence of positive selection on GlyCam-1 gene relative to milk quality [14], which supports its candidacy as a gene associated with milk production traits [15].

Similar to its homologous counterpart in cow’s milk, the canonical camel GlyCam-1 protein is non-N-glycosylated. It helps prevent fat globule aggregation during their secretion into the alveolar lumen of the lactating udder and inhibits spontaneous lipolysis by lipoprotein lipase [12]. Biochemical and molecular studies of camel GlyCam-1 revealed an N-terminal insertion containing potential O-glycosylation sites [16, 17]. These sites have been predicted and have not yet been confirmed experimentally by proteomic analysis.

According to the authors, GlyCam-1 may have immunological functions, including the prevention of mammary gland infections in lactating animals and inhibition of pathogen replication in the respiratory and gastrointestinal tracts of suckling offspring. The potential post-translational modifications (phosphorylation and O-glycosylation) predicted for camel GlyCam-1 may contribute to functional similarities with bovine PP3, including stabilization of calcium phosphate not bound to casein micelles and participation in iron transport [18, 19]. Camel GlyCam-1 also exhibits susceptibility to bacterial proteolysis [6], which could influence the flavor, texture, and functional properties of dairy products [20]. More recently, camel GlyCam-1 has demonstrated potent anticancer activity, attributed to its unique 8 kDa C-terminal peptides [21].

At the molecular level, the Lactophorin or GlyCam-1 gene has been described in several species, including mice, rats, camels, cattle, buffalo, sheep, goats, horses, and pigs [17, 22]. The exon-intron organization is conserved across species, comprising four exons interrupted by three introns. However, DNA sequence alignment reveals interspecific variations, with sequence identity ranging from 71% to 98% among the species [22].

At a proteomic level, the study of the primary and secondary structures of caprine PP3, bovine PP3 and camel GlyCam-1 showed dissimilarities among the different species in terms of N-terminal heterogeneity and C-terminal amphipathic α-helix [8, 17]. In camels, two non allelic variants A (156 aa) and B (141 aa) have been identified. Variant B is shorter, likely due to alternative splicing of exon 2 due to a weaker computed probability of the acceptor and donor splicing sites [17].

To our knowledge, aside from the study by Kappeler et al. [17], the genetic diversity of GlyCam-1 gene has not been thoroughly investigated in camelids. Recent studies have shown that the genetic variation of camel milk proteins can (1) influence nutritional and technological dairy traits [23–26], (2) have implications for health [27, 28], (3) offer insights into species diversification [26, 29] and (4) contribute to understanding camelid adaptation and resilience [30].

Therefore, the main aim of the present study is to comprehensively investigate the genetic variation of the GlyCam-1 gene and its regulatory regions in Old and New World Camelids. By integrating bioinformatics with functional analyses using gene reporter assays, we aim to enhance our understanding of the molecular mechanisms governing the synthesis of this important whey protein.

Materials and methods

DNA samples

DNA samples previously utilized in earlier studies, obtained from the collections of the University of Turin (Italy), the Justus-Liebig University of Giessen (Germany), and Kuwait University (Kuwait) were used for the sequencing of GlyCam-1 gene and the genotyping of the DNA variants identified through sequencing. In particular, a total of 203 DNA samples were used: 47 Camelus dromedarius (from Tunisia and Kuwait), 53 Lama glama (from Italy and Germany), 95 Vicugna pacos (from Italy and Germany) and 8 Camelus bactrianus (from Germany) (Additional Table 1). DNA was originally isolated from whole blood samples using phenol-chloroform method [31]. Quality and integrity of DNA samples were assessed by NanoDrop ND-1000 Spectrophotometer (Thermo Scientific, Waltham, MA, USA).

PCR amplification and sequencing

A set of primers was designed for the amplification of the entire GlyCam-1 gene in the four camelids, using DNAsis Max 3.0 software (Hitachi) (Table 1). Melting temperatures were chosen in a very narrow range (from 60.0 °C to 60.8 °C) to allow the use of a single standard annealing temperature across primer combinations. During optimization, a gradient PCR was performed for each primer pair to determine the optimal annealing temperature. For genotyping, one primer pair required a slightly lower annealing temperature (Ta = 59.5˚C) based on the gradient results to ensure efficient and specific amplification. Reference genome sequences for C. dromedarius (GenBank ID: NC_087446.1 region 57287209…57284017) and Vicugna pacos (GenBank ID: NW_021964178.1 region 4489048…4490288) were used as templates.

Table 1.

Primer sequences and annealing temperatures for GlyCam-1 amplification, sequencing, and genotyping in Old/New world camelids

Region	SNP		Species	Primer	Sequence (5’ → 3’)	Ta (°C)	Size (bp)	Genotyping
5’-flanking region			OWC	Cd GlyCam-1 5’F	TTATGCAAGACACCTCACC	60.0	771
			NWC	Cd GlyCam-1 Ex1R	GAAGCCAGACTCACCATTA
Promoter-Intron 2			OWC	Cd GlyCam-1 PromF	CATCCGTACCCCGAAATAAAT	60.0	992
			NWC	Cd GlyCam-1 Int2R	GCTTTGCTTTACTTCTTCCTC
Promoter- Exon 3			OWC	Cd GlyCam-1 PromF	CATCCGTACCCCGAAATAAAT	60.0	1825
			NWC	Cd GlyCam-1 Ex3R	AACATCGTCTTTGGAAACCAG
Intron 1-Exon 3			OWC	Cd GlyCam-1 Int1F	CTGGAGACTTTAACCCCTTA	60.0	1009
			NWC	Cd GlyCam-1 Ex3R	AACATCGTCTTTGGAAACCAG
Intron 2–3’-flanking region			OWC	Cd GlyCam-1 Int2F	TCTATTCAAAGGCACCCAG	60.0	1081
			NWC	Cd GlyCam-1 3’endR	ACGGCCGAGATTTCACAAA
Intron 2–3’-flanking region			OWC	Lg GlyCam-1 Int2F	TAACTGGGTAATGACGCGA	60.0	993
			NWC	Cd GlyCam-1 3’endR	ACGGCCGAGATTTCACAAA
Promoter	g.163A > G	M1	NWC	Lg GlyCam-1 promTypF	CGGGGCTGGAATCCCT	65.0	135	PCR-RFLP
				Lg GlyCam-1 promTypR	GTCGTAGGGGCCGGAATA			MvaI
Exon 1	g.563C > T	M2	Dromedary	Cd GlyCam-1 PromF	CATCCGTACCCCGAAATAAAT	60.0	992	GBS*
				Cd GlyCam-1 Int2R	GCTTTGCTTTACTTCTTCCTC
Exon 3	g.2324G > A	M3	Alpaca	VPA GlyCam-1 Ex3 ASFa	GCCTTCCATCTCCAGAA	59.5	159	AS-PCR
	g.2325G > A		Llama	VPA GlyCam-1 Ex3 ASFg	GCCTTCCATCTCCAGAG
				VPA GlyCam-1 Ex3 R	GTGACTTCTTCGGTCTCTTCT
Exon 3	g.2419C > G	M4	Llama			60.0	339	GBS*
				Lg GlyCam-1 Int2F	TAACTGGGTAATGACGCGA
Exon 3	g.2421G > A	M5	Alpaca	VPA GlyCam-1 Ex3 R	GTGACTTCTTCGGTCTCTTCT
	g.2422G > A		Llama

*Genotyping by sequencing

The final PCR reaction volume was 15 µl, comprising 50 ng genomic DNA, 1× PCR buffer (Promega), 2.5 mM of MgCl₂, 0.2 mM dNTP, 0.16 µM of each primer, and 1 U of Taq DNA Polymerase (Promega). Thermal cycling conditions included an initial denaturation at 95 °C for 3 min, followed by 35 cycles at 95 °C for 45 s, then 60 °C for 45 s and 72 °C for 45 s, and a final extension at 72 °C for 5 min using a T100 thermal cycler (Bio-Rad). Amplicons were purified using the NucleoSpin Gel and PCR Cleanup Kit (Machery-Nagel) and sent to Eurofins Genomics (Ebersberg, Germany) for Sanger sequencing.

Genotyping

Llamas and alpacas were genotyped for three shared SNP markers (Tables 1 and 2) located in the GlyCam-1 gene: one in the promoter region (M1: g.163A > G) and two in the exonic regions (M3: g.2324G > A in alpaca and g.2325G > A in llama; M5: g.2421G > A in alpaca and g.2422G > A in llama). Further information per SNP and institution has been reported in the additional Table 1. Genotyping of promoter variant (M1) was carried out using polymerase chain reaction restriction fragment length polymorphism (PCR-RFLP). Conversely, a multiplex PCR-based assay was set up for the discrimination of the M3 in the NWC population. A fragment of 159 bp was amplified in an allele specific (AS) reaction, while a 331 bp fragment from LALBA gene served as a positive control using the following primers (forward) 5’-CCATTGTCTGATCCCTTTTGGAACT-3’ and (reverse) 5’-GGTTACTCACTGTCACAGGAGATG-3’. The PCR mixture was prepared as described by Pauciullo et al. [26]. Sanger sequencing was used to genotype the M5 (shared among NWCs), the M4: g.2419G > C (exon 3, llamas only), and M2: g.563C > T (exon 1, dromedary camels, 5’-Untranslated Region).

Table 2.

Polymorphisms identified in complete GlyCam-1 gene and regulatory regions across four domestic camelid species

Item	Region	Dromedary camel		Bactrian camel		Alpaca		Lama		Genotyped
		Position	Nucleotide	Position	Nucleotide	Position	Nucleotide	Position	Nucleotide	SNP
5’-Flanking region	Promoter	22	C	22	C	22	Y	22	C
		24	T	24	T	24	T	24	Y
		30	T	30	T	29/30	-	29/30	-
		31	C	31	C	29/30	-	29/30	-
		39	T	39	T	37	T	37	Y
		95	A	95	A	93	A	93	R
		153	C	153	C	151	Y	151	C
		165	A	165	A	163	R	163	R	M1
		169	T	169	T	167	C	167	C
		239	T	239	T	237	W	237	A
		246	G	246	G	244	R	244	G
		259–260	CA	259–260	CA	257–258	TG	257–258	TG
		262	G	262	G	260	G	260	R
		270/271	-	270/271	-	269	T	269	T
		292	G	292	G	291	R	291	G
		320	T	320	T	319	C	319	C
		339	T	339	T	338	C	338	Y
		345	A	345	A	344	A	344	R
		347	A	347	A	346	A	346	R
		392	C	392	C	391	T	391	T
		397	A	397	A	396	G	396	G
		415	T	415	T	414	C	414	C
		423	G	423	G	422	G	422	R
		435	C	435	C	434	S	434	C
		453–454	GA	453–454	GA	452–453	VG	452–453	RG
GlyCam-1	Exon1	466	T	466	T	465	C	465	Y
		481	G	481	G	480	A	480	A
		526	T	526	T	525	C	525	C
		550	G	550	G	549	A	549	A
		563	Y	563	T	562	C	562	C	M2
		577	A	577	A	576	C	576	C
		610	G	610	G	609	G	609	R
		634	C	634	C	633	T	633	T
		658	T	658	T	657	C	657	C
		665	A	665	A	664	G	664	G
		668	A	668	A	667	C	667	C
		670	A	670	A	669	G	669	G
		682	T	682	T	681	C	681	C
		705	C	705	C	704	C	704	Y
		708	Y	708	C	707	C	707	C
		731	A	731	C	730	C	730	C
		744	Y	744	C	743	C	743	C
		746	C	746	C	745	T	745	T
	Intron 1	773	A	773	A	772	G	772	G
		790	G	790	G	789	K	789	K
		829	G	829	G	828	A	828	A
		845	Y	845	C	844	C	844	C
		870	G	870	G	869	R	869	R
		882	A	882	A	881	C	881	C
		886	A	886	A	885	R	885	A
		909	G	909	G	908	S	908	S
		915	R	915	G	914	G	914	G
		919–924	CAGGGG	919–924	CAGGGG	917/918	-	917/918	-
		930–931	GG	930–931	GG	923–924	GG	923–924	KR
		947	A	947	A	940	R	940	A
		956	C	956	C	949	S	949	C
		967	T	967	T	960	G	960	G
		974	G	974	G	967	R	967	G
		993	T	993	T	986	C	986	C
		1032	G	1032	R	1025	G	1025	G
		1048	G	1048	G	1041	R	1041	G
		1053	G	1053	R	1046	G	1046	G
		1076	G	1076	G	1069	A	1069	A
		1142	G	1142	G	1135/1136	-	1135	G
		1145	G	1145	G	1137	A	1138	A
		1151	A	1151	A	1143	R	1144	A
		1189	C	1189	C	1181	C	1182	Y
		1190	R	1190	G	1182	G	1183	G
		1207	T	1207	T	1199	C	1200	C
		1252	C	1252	C	1244	T	1245	T
		1268	G	1268	G	1260	A	1261	A
		1300	A	1300	A	1292	G	1293	G
		1320	G	1320	A	1312	R	1313	R
		1331	C	1331	C	1323	C	1324	Y
		1335	G	1335	R	1327	G	1328	G
		1338	G	1338	G	1330	A	1331	A
		1370–1376	GCGCAG	1370–1376	GCGCAG	1362/1363	-	1363/1364	-
		1377–1378	CA	1377–1378	CA	1363–1364	TG	1364–1365	TG
		1384–1386	GGA	1384–1386	GGA	1369/1370	-	1370/1371	-
		1392	Y	1392	T	1374	C	1375	C
	Exon 2	1446	A	1446	A	1429	R	1430	A
		1485	Y	1485	C	1468	C	1469	C
	Intron 2	1507	Y	1507	C	1490	C	1491	C
		1607	G	1607	G	1590	G	1591	R
		1649	T	1649	T	1632	G	1633	G
		1673	T	1673	T	1656	C	1657	C
		1693	G	1693	A	1676	G	1677	G
		1708/1709	-	1708/1709	-	1692	G	1693	G
		1716	C	1716	C	1700	T	1701	Y
		1717	A	1717	G	1701	G	1702	G
		1826	T	1826	T	1810	C	1811	C
		1828	T	1828	T	1812	Y	1813	T
		1843	T	1843	T	1828	C	1829	C
		1854–1910	---	1854–1910	---	1837/1838	-	1838/1839	-
		1873	Y	1873	C	-	-	-	-
		1912	G	1912	G	1839	T	1840	T
		1915	T	1915	T	1842	A	1843	A
		1920	Y	1920	C	1847	C	1848	Y
		1921	G	1921	G	1848	S	1849	S
		1935	T	1935	T	1862	C	1863	C
		1944	T	1944	T	1871	C	1872	C
		1958	G	1958	G	1885	C	1886	C
		2017	T	2017	T	1944	C	1945	C
		2018	A	2018	A	1945	T	1946	T
		2034	T	2034	T	1961	W	1962	A
		2068	G	2068	G	1995	T	1996	T
		2114	C	2114	C	2041	C	2042	Y
		2127	C	2127	C	2054	T	2055	T
		2137	G	2137	G	2064	A	2065	A
		2147	T	2147	T	2074	Y	2075	Y
		2183	G	2183	G	2110	G	2111	R
		2190	T	2190	T	2117	C	2118	Y
		2199	A	2199	A	2126	C	2127	C
		2210	M	2210	A	2137	A	2138	A
		2217	Y	2217	C	2144	C	2145	C
		2254	C	2254	C	2181	T	2182	T
		2255	C	2255	C	2182	C	2183	Y
		2276/2277	-	2276/2277	-	2204	A	2205	A
		2280	G	2280	G	2209	G	2210	R
		2307	A	2307	A	2235	G	2236	G
		2316	C	2316	T	2244	C	2245	C
		2317	G	2317	G	2245	A	2246	A
		2329	C	2329	T	2257	C	2258	C
	Exon 3	2396	G	2396	G	2324	R	2325	R	M3
		2470	T	2470	T	2398	Y	2399	T
		2490	G	2490	G	2418	G	2419	S	M4
		2493	G	2493	G	2421	R	2422	R	M5
		2515	A	2515	A	2443	G	2444	G
		2528	A	2528	A	2456	G	2457	G
		2534	C	2534	C	2462	G	2463	G
		2542	T	2542	T	2470	C	2471	C
	Intron 3	2565	R	2565	G	2493	G	2494	G
		2605	G	2605	G	2533	R	2534	G
		2676	G	2676	G	2604	C	2605	C
		2724	K	2724	G	2652	G	2653	G
		2729	G	2729	G	2656/2657	-	2657/2658	-
		2745	A	2745	A	2672	G	2673	G
		2785	C	2785	S	2712	G	2713	G
	Exon 4	2896	T	2896	T	2823	Y	2824	T
		2916	A	2916	G	2843	A	2844	A
		2928	C	2928	C	2855	C	2856	Y
		2933	G	2933	G	2860	R	2861	R
		2941	C	2941	C	2868	Y	2869	C
		2947	C	2947	C	2874	A	2875	A
		2975	G	2975	G	2902	G	2903	R
		2978	G	2978	G	2905	C	2906	C
		2979	G	2979	G	2906	R	2907	A
		2989	C	2989	C	2916	Y	2917	Y
		2993	G	2993	G	2920	G	2921	R
		2997	C	2997	Y	2924	C	2925	C
		3005	C	3005	C	2932	C	2933	Y
		3008	R	3008	R	2935	A	2936	A
3’-Flanking region		3061	G	3061	S	2988	G	2989	G
		3117	A	3117	A	3044	T	3045/3046	-
		3187	Y	3187	C	3114	C	3114	C

The complete list of primers used for PCR amplification is provided in Table 1.

Gene reporter assay

Independent gene reporter assays by specific pGL3 constructs were performed to evaluate luciferase expression (Promega) in 293 T cells for the M2: g.563C > T found in the dromedary GlyCam-1 5’UTR and the M1: g.163A > G in the GlyCam-1 promoter of NWC. To this aim, the DNA of alpacas (n = 4) and dromedaries (n = 4) opposite homozygous for each SNP variant was amplified with the following primers: 5’-AAAGGTACCCTAATGGCCTTGCTTTCCT-3’ (forward) and 5’-TTTGCTAGCCTTGGCAGGGGCTTTATT-3’ (reverse) in dromedary camels; and 5’-AAAGGTACCTTATGCAAGACACCTCACCTC-3’ (forward) and 5’-TTTGCTAGCAGGCTGTCCTCTAGGCT-3’ (reverse) in alpacas. Bold sequences indicate restriction sites for Kpn I (GGTAC↓C) and Nhe I (G↓CTAGC), added to facilitate cloning, while underlined bases were included to improve cleavage efficiency. Amplified fragments containing the SNPs were cloned into the pGL3 basic vector (Promega) upstream of the reporter gene, generating four constructs: g.563C and g.563T (dromedaries) and g.163A and g.163G (alpacas). Cloning was performed using Single-Use JM109 competent cells (Promega) according to standard manufacturer’s guidelines (Promega). Twelve clones for each construct were PCR analysed for the presence of the insert, and plasmids were then purified with the PureYield™ Plasmid Miniprep System (Promega). Correct DNA insertion and absence of polymerase errors were confirmed by sequencing each plasmid construct with vector-specific primers RV3 (5’-CTAGCAAAATAGGCTGTCCC-3’) and RV4 (5’-GACGATAGTCATGCCCCGCG-3’) following the manufacturer’s protocol (Promega).

The 293T cells, obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA), were cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) (Sigma-Aldrich, USA) supplemented with 2 mM Glutamine, 1% penicillin-streptomycin (Sigma-Aldrich, USA), and 10% fetal bovine serum (FBS) (Euroclone). Cells were incubated at 37 °C in a humidified 5% CO₂ atmosphere.

For luciferase assays, cells were seeded at 1 × 10⁴ cells/well in 24-well plates and transfected at 60–70% confluence. Each well received 200 ng of luciferase reporter construct (g.563C or g.563T for dromedaries; g.163A or g.163G for alpacas, each as a pool deriving from the 4 isolated plasmids) and 20 ng of pRL-TK vector (Renilla luciferase; Promega) using Lipofectamine 3000 (Invitrogen, USA). Reporter activity for the different genotypes was measured after 24 h using a Victor Nivo Multiplate reader (PerkinElmer, Waltham, MA) and the Dual-Glo luciferase assay system (Promega). The pGL3 control vector was transfected as a positive control, whereas the pGL3 basic vector was transfected as a negative control.

All transfections were performed in triplicate (pool of 4 isolated plasmids per genotype) and repeated across three independent experiments. In the end, the complete set of 12 isolated plasmids per genotype was used. Firefly luciferase activity was normalized to Renilla luciferase expressed by the vector pRL-TK co-transfected as an internal control. Promoter activity for each construct was calculated as the ratio of normalized luciferase activity to the activity of the pGL3 control vector.

Bioinformatic analysis

DNAsis-Max v3.0 software (Hitachi) was used for SNP discovery, homology searches, sequence comparison and multiple alignments. The same software was additionally used to determine the Isoelectric point (pI) and the molecular weight of the predicted proteins. Interspersed elements were identified using the RepeatMaskers Web Server (open-4.0.9) with Dfam database v3.0 (http://repeatmasker.org/cgi-bin/WEBRepeatMasker, accessed on 21 March 2024). Phosphorylation sites were predicted using NetPhos v3.1 (https://services.healthtech.dtu.dk/services/NetPhos-3.1/, accessed on 5 December 2024), and O-glycosylation sites were predicted via NetOGlyc v4.0 with a score threshold ≥ 0.80 (https://services.healthtech.dtu.dk/services/NetOGlyc-4.0/, accessed on 5 December 2024).

The regulatory regions of the gene were analysed for putative transcription factor binding sites using TFBIND tool (based on TRANSFAC R.3.4) (https://tfbind.hgc.jp, accessed on 25 October 2024) and Alibaba v2.1 software (based on TRANSFAC 4.0) (http://gene-regulation.com/pub/programs/alibaba2/, accessed on 6 March 2024). The presence of consensus splicing signals, polypyrimidine tracts and branchpoints (BPs) were assessed by NNSPLICE ver 0.9 (https://www.fruitfly.org/seq_tools/splice.html, accessed on 10 March 2025). Protein secondary structure was predicted by Jpred v4 software (http://www.compbio.dundee.ac.uk/jpred/, accessed on 6 December 2024) and NetSurfP v2.0 (https://services.healthtech.dtu.dk/services/NetSurfP-2.0/, accessed on 7 December 2024). Functional and structural implications of amino acid changes were evaluated using web tools PROVEAN v1.1.3 (http://provean.jcvi.org/seq_submit.php, accessed on 7 December 2024) and Conserf (https://consurf.tau.ac.il/consurf_index.php, accessed on 7 December 2024). Phylogenetic relationships of GlyCam-1 protein sequences were inferred using the Neighbor-joining method in MEGA7 v7.0.14, applying the Poisson substitution model with pairwise deletion for gaps/missing data and 1,000 bootstrap replicates to assess branch support.

Statistical analysis

Allele frequencies, heterozygosity estimates, and fixation indices were calculated using GENEPOP v4.8.4. Deviations from Hardy–Weinberg equilibrium (HWE) were tested separately for each species using the exact probability test based on the Markov chain method (dememorization = 10,000; batches = 100; iterations per batch = 5,000). Inbreeding coefficient within populations (F_IS) and pairwise fixation index (F_ST) values were estimated following Weir and Cockerham [32] to evaluate genetic diversity and population structure. Linkage disequilibrium (LD) between loci within species was assessed using Haploview ver. 4.2 (https://www.broadinstitute.org/haploview/haploview) with four gamete rule as block LD definition. Haplotype diversity and frequencies were inferred using PHASE v2.1 with standard MCMC settings (1,000 iterations, 100 burn-in, thinning interval = 1). Posterior reconstruction probabilities for the best-supported haplotype pairs were generally high (mostly > 0.9), with occasional lower values for a few individuals; therefore, no additional probability filtering was applied. Haplotype reconstruction was performed independently for llamas and alpacas to ensure species-specific resolution of allelic phase.

Differences in relative luciferase activity (relative light unit - RLU) expressed as mean ± SEM of the different DNA contructs (g.563C vs. g.563T and g.163A vs. g.163G) were determined using one-way ANOVA with two groups carried out with JASP ver 0.19.1. Bonferroni post-hoc comparison between genotypes was applied to assess differences (p ≤ 0.001). Graphical representation (RLU mean ± SEM) that includes also the pGL3 control has been realised by Prism 10 software (ver 10.1.2).

Results

GlyCam-1 gene structure in camelids

The GlyCam-1 gene spans 2,576 base pairs (bp) in OWC and 2,504 bp in NWC, with 818 bp distributed across its four exonic regions: 297 bp in exon 1, 45 bp in exon 2, 225 bp in exon 3, and 251 bp in exon 4. The intronic regions account for about 1,764 and 1,693 bp in OWC and NWC, respectively (Additional file 1). The sequenced segments also include 460/459 bp of the 5’-flanking region in OWC and 157/156 bp of the 3’-flanking region in NWC (Additional file 1). All sequences generated in this study have been submitted to GenBank under the IDs: PV445494, PV445495, PV445496, and PV445497 for Camelus dromedarius, Camelus bactrianus, Vicugna pacos and Lama glama, respectively.

Bioinformatics analysis confirmed that the gene structure is conserved across all species. The Open Reading Frame (ORF) encodes 155 amino acids, including a signal peptide of 18 amino acids in OWC and 15 in NWC (Additional file 2). Analysis of exon-intron boundaries revealed splice junction sequences consistent with the canonical GT/AT rule, with no deviation observed, as well as no deviations were found for branch points and poly-pyrimidine tracts. The translation starts at an ATG codon situated at the 234th nucleotide of exon 1 in both Old and New World Camelids (Additional file 1 and 2). The mature protein (136 amino acids) is encoded from the last 10 nucleotides of exon 1 to the 131 st nucleotide of exon 4. Translation terminates with a TGA stop codon positioned between nucleotides 132 and 134 in exon 4. The polyadenylation signal (AATAAA) is located in the same exon, 95 bp downstream of the stop codon (Additional file 1).

Protein analysis predicted seven putative serine/threonine phosphorylation sites in both OWC and NWC (Additional file 1), and at least 3 glycosylation sites at positions 74, 75 and 83 in OWC and at positions 77, 78, and 82 in NWC (Additional file 1).

Genetic variability in Old and New world camelids

Sanger sequencing evidenced genetic variability across the four camelids. Intraspecies analysis of GlyCam-1 gene sequencing revealed 18 polymorphic sites in dromedaries (15 transitions and 3 transversions), 7 polymorphisms in bactrian camels (5 transitions and 2 transversions), 32 polymorphic sites in alpacas (24 transitions and 8 transversions), and 40 polymorphic sites in llamas (35 transitions and 5 transversions). Interspecies comparison revealed 153 variations among the studied species (Table 2). Notably, the SNPs M3 (g.2324G > A in alpaca and g.2325G > A in llama) and M5 (g.2421G > A in alpaca and g.2422G > A in llama) found in NWC were responsible for the amino acid substitutions p.Glu46Lys and p.Ser77Thr, respectively. The SNP M4: g.2419G > C, specific to llamas, caused the p.Ser78Asn substitution, while g.2823C > T, specific to alpacas, resulted in the p.Ser134Pro changes.

To further evaluate the population-level variability of the GlyCam-1 gene among camelids, genetic diversity indices were calculated. Expected heterozygosity (He) ranged from 0.113 in llamas to 0.489 in dromedaries, while observed heterozygosity (Ho) varied from 0.110 to 0.446 across species (Table 3). Alpacas displayed moderate diversity (Ho = 0.317), whereas llamas showed lower variability (Ho = 0.110). Dromedaries exhibited a single polymorphic locus, but with relatively high Ho (0.446).

Table 3.

Population-level genetic diversity and differentiation in camelids. Expected heterozygosity (He), observed heterozygosity (Ho), local inbreeding coefficient (F_IS), pairwise fixation index (F_ST) and linkage disequilibrium (LD) parameters

Population	N. of typed individuals	N. of typed loci	He	Ho	FIS	FST		LD (r2)
						Alpaca	Llama
Alpaca	95	3 (M1-M3-M5)	0.330	0.317	0.041	-	0.112	M3-M5 (0.944)
Llama	53	4 (M1-M3-M4-M5)	0.113	0.110	0.025	0.112	-	M4-M5 (0.002)
Dromedary	47	1 (M2)	0.489	0.446	−0.096			-
Overall	195	5 (M1-M2-M3-M4-M5)	-	-	0.093

The inbreeding coefficient (F_IS) was generally low, indicating limited deviation from random mating within species (F_IS = 0.093 overall).

Pairwise F_ST comparisons revealed moderate genetic differentiation among the NWC (F_{ST =} 0.112), indicating that approximately 11.2% of the observed genetic variance is attributable to differences between these two populations. Significant linkage disequilibrium was detected between loci M3–M5 in alpacas (r² = 0.944), whereas the very low squared correlation coefficient (r² = 0.002) in llamas suggests that the four markers are basically inherited independently and no LD exists (Table 3).

Effect of non-synonymous SNPs on protein structure and function

Missense variants in GlyCam-1 gene found in llamas and alpacas were analyzed for their impact on protein function and structure. The p.Glu46 > Lys substitution (SNP M3) altered the predicted isoelectric point (pI), increasing it from 4.84 to 5.09, and caused a slight shift in molecular weight (15341.35 Da vs. 15340.41 Da). The other three variants (p.Ser77Thr, p.Ser78Asn, and p.Ser134Pro) did not affect the pI but increased the molecular weight to 15355.38 Da, 15368.38 Da, and 15351.39 Da, respectively. All four SNPs altered the Absolute Surface Accessibility (ASA), especially at position 78 (88.89 vs. 127.79) and 134 (91.52 vs. 103.66). Minor changes in helix-coil probabilities were also noted at position 134 (0.713 vs. 0.946) (Table 4). Regarding functional implications, the variants p.Glu46Lys, found to change a conserved amino acid (Additional file 1), and p.Ser134Pro were predicted to have deleterious effects (Additional file 3), while the other variants were considered neutral.

Table 4.

Effect of GlyCam-1 gene SNPs on relative and absolute surface accessibility and secondary structure probabilities

SNP	Amino acid	Position	RSA	ASA	Probability for α-helix	Probability for β-strand	Probability for coil
g.2324G > A*	Glu (E)	46	0.662	147.654	0.074	0.039	0.887
	Lys (K)		0.636	150.092	0.071	0.039	0.891
g.2419G > C⁑	Ser (S)	77	0.538	83.380	0.036	0.026	0.938
	Thr (T)		0.527	90.670	0.041	0.035	0.923
g.2421G > A*	Ser (S)	78	0.574	88.894	0.042	0.037	0.921
	Asn (N)		0.655	127.798	0.037	0.012	0.950
g.2823C > T†	Ser (S)	134	0.781	91.528	0.283	0.004	0.713
	Pro (P)		0.731	103.668	0.053	0.002	0.946

^*Positions refer to Vicugna pacos GlyCam-1 sequence. The same effect has been recorded for the SNPs g.2325G > A and g.2422G > A found in llama

^⁑Position refers to Lama glama GlyCam-1 sequence

^†Positions refer to Vicugna pacos GlyCam-1 sequence

GlyCam-1 gene regulatory region

The regulatory regions of the GlyCam-1 gene were sequenced and characterised in OWC and NWC, transcription factor binding sites (TFBS) were predicted. A total of 179 TFBS were conserved among all four camelids, including elements commonly found in the promoter regions of genes associated with milk production and protein expression (Table 5).

Table 5.

Shared and genotype-specific transcription factors in the 5’-flanking region of GlyCam-1 in camelids

Transcription factor	Consensus sequence	Signal sequence	Strand	Score	Position
					C. dromedarius	C. bactrianus	L. glama	V. pacos
CEBP-β	NKNTTGCNYAAYNN	TGCTTGATAAATAA	(+)	0.845	−421	−421	−422	−422
GATA	NGATAAGNMNN	TGATAAATAAC	(+)	0.927	−417	−417	−418	−418
GR	NNNNNNCNNTNTGTNCTNN	AAATAACACACACCTCACA	(-)	0.776	−413	−413	−414	−414
EGR1	WTGCGTGGGCGK	ATGCCTGGCAGG	(+)	0.774	−348	−348	−349	−349
STAT	TTCCCRKAA	TGCCTGGCA	(+)	0.795	−347	−347	−348	−348
SP1	GRGGCRGGGW	GGGGCTGGAA	(+)	0.899	−337	−337	−338	−338
GATA-3	NNGATARNG	CTTTATCTG	(-)	0.908	−283	−283	−284	−284
GRE	GGTACAANNTGTYCTK	GAGGACAGCCTAGCCC	(-)	0.827	−267	−267	−268	−268
OCT1	NNGAATATKCANNNN	TTCTTTATTCATTCA	(+)	0.787	−251	−251	−252	−252
AP1	NTGASTCAG	TTTATTCAT	(-)	0.848	−248	−248	−249	−249
YY1	NNNNNCCATNTWNNNWN	GATAAACAATGCAAGTA	(-)	0.776	−179	−179	−179	−179
ER	NNARGNNANNNTGACCYNN	TAAAGCAGGTGTAACCCAG	(+)	0.767	−164	−164	−164	−164
NF-KB	NGGGACTTTCCA	GTGGAAGTCTCA	(-)	0.779	−140	−140	−140	−140
CEBP-α	NNATTRCNNAANNN	TAATTAAATCATCA	(+)	0.823	−106	−106	−106	−106
HNF3B	NNNTRTTTRYTY	ACACAAATAAAA	(-)	0.892	−93	−93	−93	−93
TATA box	STATAAAWRNNNNNN	AAATAAAAGTAAAAT	(+)	0.812	−89	−89	−89	−89
NF-KB⁂	GGGAMTTYCC	TGGAATGTCC	(+)	0.817	96	-	-	-
MZF-1*	KNNNKAGGGGNAA	CTTCCACTCCTGA	(-)	0.816	-	-	−309	−309
AP-2α†	MKCCCSCNGGCG	CTCCCGGGGAAG	(-)	0.913	-	-	−126	-
CEBP-α⁑	NNATTRCNNAANNN	CAAAATAAATA	(+)	0.727	-	-	+ 150	-
SP1¶	NGGGGGCGGGGYN	TATTCCGGCCCCT	(-)	0.782	-	-	-	−222
AP-2α•	MKCCCSCNGGCG	AGCCTGGGGCAG	(+)	0.845	-	-	-	−43

Only TFBS commonly found in milk-related genes are shown, along with their consensus motifs. The sense strand (5’ to 3’) is denoted by “+”, and the complementary strand (3’ to 5’) is represented by “-”. Negative numbering indicates the nucleotide distance from the start of exon 1

^⁂TF genotype dependent by SNP g.563C > T and loss with allele g.563T

^*TF genotype dependent by SNP g.163A > G and loss with allele g.163G

^†TF genotype dependent by SNP g.422G > A and loss with allele g.422 A

^⁑TF genotype dependent by SNP g.609G > A and created with the allele g.609 A

^¶TF genotype dependent by SNP g.244G > A and loss with the allele g.244 A

^•TF genotype dependent by SNP g.434C > G and loss with the allele g.434G

In the dromedaries, the M2: g.563C > T transition was predicted to alter an NF-κB binding site. No polymorphisms were found in the bactrian regulatory region. In contrast, six shared SNPs were identified in the llama and alpaca promoter regions.

In llamas, the SNPs M1:g.163A > G and g.422G > A were predicted to alter MZF-1 and AP-2α TFBS, respectively, while g.609G > A may create a new c/EBPα site. In alpacas, the same M1 was detected, along with g.244G > A and g.434C > G, predicted to affect Sp1 and AP-2α TFBS, respectively (Table 5).

Genotyping and haplotype analysis

The SNPs with biological relevance M1 (found in the promoter and shared between llamas and alpacas), M2 (in the dromedary 5’-UTR of exon 1), and M3, M4 and M5 (responsible for amino acid change) in NWC, were genotyped using different methods.

A PCR-RFLP assay was developed to genotype M1 in 82 SAC using MvaI endonuclease. Digestion of the 135 bp PCR product yielded two fragments (95 and 40 bp) for AA homozygotes, undigested fragments for GG homozygotes, and three fragments (135, 95, and 40 bp) for heterozygous AG (Fig. 1). On the contrary, M3 did not alter any restriction sites, so an allele-specific PCR method (AS-PCR) was developed to genotype 148 NWC (Fig. 2).

Fig. 1.

Genotyping of the SNP M1: g.163A > G in the GlyCam-1 promoter of NWC by MvaI PCR-RFLP. Lane 1, homozygous AA; lane 2, heterozygous AG; lane 3, homozygous GG. The L lane is GeneRuler 50 bp DNA Ladder (Thermo Fisher Scientific, Waltham, MA)

Fig. 2.

Genotyping of SNP M3 in the GlyCam-1 exon 3 of NWC using AS-PCR. The SNP is located at g.2324 G > A in alpacas and g.2325 G > A in llamas. The 331 bp fragment is common to all samples and represents the positive control (LALBA gene). Lanes 1, 3 and 5 show amplicons obtained with the specific forward primer for adenine. Lanes 2, 4 and 6 show amplicons obtained with the specific forward primer for guanine. The L lane is GeneRuler 50 bp DNA Ladder (Thermo Fisher Scientific, Waltham, MA)

The SNP M2 in dromedaries and the SNPs M4 and M5 in NWC were genotyped by Sanger sequencing. The former was unique to llamas, while the latter was found in both NWC (Fig. 3).

Fig. 3.

Sequencing chromatograms showing three genotypes of SNP M5 in GlyCam-1, responsible for p.Ser78Asn. The SNP is located at g.2421 G > A in alpacas and g.2422 G > A in llamas

Genotype distributions and allele frequencies are reported in Table 6. Hardy–Weinberg equilibrium was assessed separately for each species using exact tests. All loci were in equilibrium except M1 in alpacas (g.163A > G, p = 0.0033). At least 16 different alleles were identified, responsible for eight distinct protein isoforms in NWC (Additional Table 2).

Table 6.

Genotypes and allele frequencies of SNPs M1-M5 in the GlyCam-1 gene across camelids. HWE was computed using the Markov chain method exact test (GENEPOP v4.8.4)

Region	SNP	Amino acid	Species	Genotype distribution				Allele frequency		HWE exact p-value
Promoter	M1			AA	AG	GG	Total	A	G
	g.163A > G		Alpacas	43	4	3	50	0.90	0.10	0.0033
			Llamas	27	4	1	32	0.90	0.10	0.2265
Exon 1	M2			CC	CT	TT	Total	C	T
	g.563C > T		Dromedary	4	23	20	47	0.33	0.67	0.7404
Exon 3	M3	p.Glu46 > Lys		AA	AG	GG	Total	A	G
	g.2324 G > A		Alpacas	5	35	55	95	0.24	0.76	1.0000
	g.2325 G > A		Llamas	0	5	48	53	0.05	0.95	1.0000
Exon 3	M4	p.Ser77 > Thr		CC	CG	GG	Total	C	G
	g.2419G > C		Llamas	0	9	44	53	0.08	0.92	1.0000
Exon 3	M5	p.Ser78 > Asn		AA	AG	GG	Total	A	G
	g.2421 G > A		Alpacas	5	37	53	95	0.25	0.75	0.7872
	g.2422 G > A		Llamas	0	2	51	53	0.02	0.98	1.0000

Haplotype AGG was the most frequent in alpacas (0.707), whereas AGGG was the most frequent in llamas (0.803), with other frequencies shown in Table 7.

Table 7.

Haplotype frequencies in the GlyCam-1 gene across new world camelids estimated using PHASE v2.1. Data are based on successfully genotyped animals (Alpaca n = 95; Llama n = 53)

Species	Haplotype				Frequency	Standard Error
	M1	M3	M4	M5
Alpacas (n = 95)	A	A	-	A	0.224	0.010
	A	G	-	A	0.010	0.002
	A	G	-	G	0.707	0.014
	G	A	-	A	0.018	0.009
	G	G	-	G	0.041	0.014
Llamas (n = 53)	A	G	G	G	0.803	0.022
	A	G	G	A	0.016	0.004
	A	G	C	G	0.086	0.001
	A	A	G	G	0.035	0.009
	G	G	G	G	0.060	0.021

Interspersed elements

The promoter region of GlyCam-1 contains two interspersed elements: one shared across all camelids and another specific to dromedaries and bactrians. The first is a LINE (L2a subfamily) measuring 187 bp in OWC and 188 bp in NWC. The second is a SINE (MIRs), 122 bp long, and specific to dromedary and bactrian camels. Three additional interspersed elements were identified in introns. A 220 bp L2a LINE is shared across all camelids, but it is 9 bp shorter in NWC. A 38 bp simple repeat (AGGGGAC)_n was unique to C. dromedarius. In intron 2, a 56 bp DNA element (TcMar-Tigger) was found in llamas and alpacas but not in OWC (Table 8). The overall proportion of interspersed elements within GlyCam-1 was 17.76% in dromedaries, 16.57% in bactrians, and 14.58% in llamas and alpacas.

Table 8.

Position and type of interspersed elements in GlyCam-1 gene of camelids compared to Bos Taurus

	Camelus dromedarius GenBank ID : PV445494			Camelus bactrianus GenBank ID : PV445495			Lama glama GenBank ID : PV445497			Vicugna pacos GenBank ID : PV445496			Bos taurus GenBank ID: X83391.1
	Position		Name	Position		Name	Position		Name	Position		Name	Position		Name
	Location	Nucleotide		Location	Nucleotide		Location	Nucleotide		Location	Nucleotide		Location	Nucleotide
SINEs
MIRs	Promoter	1/122	MIRb	Promoter	1/122	MIRb							Promoter	1/174	MIRb
													Intron 1	1678/1951	Bov-A2
													Intron 3	2385/3526	Putative SINE
													3’-Flanking	4027/4263	Bov-tA
LINEs
LINE2	Promoter	220/406	L2a	Promoter	220/406	L2a	Promoter	218/405	L2a	Promoter	218/405	L2a
	Intron 1	1154/1373	L2a	Intron 1	1154/1373	L2a	Intron 1	1147/1357	L2a	Intron 1	1146/1356	L2a	Intron 1	1533/1691	L2a
DNA elements
TcMar-Tigger							Intron 2	1818/1873	Tigger19a	Intron 2	1817/1872	Tigger19a
Simple repeats
													Promoter	437/473	A-rich
	Intron 1	912/949	(AGGGGAC)n
													Intron 1	1937/1956	(AGC)n
													Intron 2	2553/2590	(AC)n
													3’-Flanking	4241/4267	(ACTT)n

Phylogenetic analysis

Phylogenetic analysis using GlyCam-1 protein sequences from this study and UniProt entries revealed clear evolutionary patterns. NWC clustered together and were close related to OWC with C. dromedarius forming a distinct subgroup. C. bactrianus showed closer affinity to NWC (Additional file 4). Ruminants (Bos taurus, Bubalus bubalis, Capra hircus, and Ovis aries) formed a close group, while rodents and primates were more distantly related, with domestic pigs and horses occupying an intermediate position (Additional file 4).

Gene reporter assay

The impact of M2 (g.563C > T) and M1 (g.163A > G) on GlyCam-1 5’-UTR and promoter activity were investigated in dromedaries and alpacas, respectively, using luciferase reporter assays. In dromedaries, the T allele increased promoter activity nearly two-fold compared to the C allele (0.042 ± 0.010 vs 0.025 ± 0.004), though the difference was not statistically significant (p = 0.102) (Fig. 4) (mean difference of 0.017 and − 0.004–0.038 of 95% CI). In alpacas, the A allele of M1 was associated with significantly higher luciferase expression (0.425 ± 0.032) than the G variant (0.029 ± 0.004), p < 0.001 (Fig. 5), with a mean difference of 0.396 and [0.332–0.461] of 95% CI for mean difference. Therefore, the expression of the g.163A is 13.7-fold higher than that of g.163G.

Fig. 4.

Luciferase assay assessing the effect of SNP M2 (g.563C > T) on GlyCam-1 5′UTR activity in dromedaries. Relative luciferase activities (relative light units) are expressed as mean ± SE. Genotype g.563C (0.025 ± 0.004). Genotype g.563T (0.042 ± 0.010). No significant difference was observed (NS: p > 0.05)

Fig. 5.

Relative luciferase activity of the GlyCam-1 promoter variants M1 g.163A and g.163G in NWC. Relative luciferase activities (relative light unit) are expressed as mean ± SE. Genotype g.163A (0.425 ± 0.032) vs. g.163G (0.029 ± 0.004) showed a significant difference for p < 0.0001

Discussion

A multi-technical approach, including sequencing, bioinformatics, and luciferase reporter assays was used to fully characterize the GlyCam-1 gene, which encodes the lactophorin whey protein in both Old and New World camelids. The analysis was extended to the 5- and 3-end flanking regulatory regions.

Sequencing of the gene revealed four exons separated by three introns, highlighting a conserved gene structure across C. dromedarius (GenBank ID: PV445494), C. bactrianus (GenBank ID: PV445495), L. glama (GenBank ID: PV445497), and V. pacos (GenBank ID: PV445496). The exon lengths were identical in all four camelid species, consistent with the findings of Kappeler et al. [17] for C. dromedarius. A recent study by Sabry and Moussa [22] further confirmed the conserved structure of this gene in B. bubalis, B. taurus, O. aries, E. caballus and S. scrofa, as well as in C. hircus, as noted by Le Provost et al. [10]. The GlyCam-1 gene is characterized by a very short exon 2, flanked by two long introns. This exon is likely to be alternatively spliced out during mRNA maturation, resulting in the shorter protein variant B. Indeed, Kappeler et al. [17] reported low splicing probability at the donor and acceptor sites of exon 2. With regard to intron lengths, slight variations were observed compared to the findings of Kappeler et al. [17] for C. dromedarius. The first two introns showed minor size differences, while intron 3 was the same length in both dromedary and bactrian camels. In llamas and alpacas, intron 2 was shorter due to a 55 bp deletion.

At the protein level, glycosylation is one of the most common post-translational modifications. The canonical GlyCam-1 protein, as previously characterized in cows and camels, is non-N-glycosylated, meaning it lacks carbohydrate attachments on asparagine residues. However, potential O-glycosylation sites, which occur on serine and threonine residues, have been reported in camel GlyCam-1 [16]. Our bioinformatic analysis predicted three putative O-glycosylation sites in OWC (positions 74, 75, and 83) and three in NWC (positions 77, 78, and 82) on the mature protein. These predictions are consistent with those of Kappeler et al. [17], who also identified three potential O-glycosylation sites in dromedary GlyCam-1 (Thr16, Thr81, and Thr90) using computational tools, although with lower confidence scores.

Sequencing of the GlyCam-1 gene in the four camelid species revealed inter- and intraspecies genetic polymorphism, with greater diversity observed in NWC (Table 2). This observation aligns with previous studies on other genes - such as LALBA [26], CSN1S1, CSN2, CSN1S2, and CSN3 [25, 33] - which also reported higher genetic variability in NWC compared to OWC. This disparity likely reflects the historical population bottlenecks that significantly reduced genetic diversity in OWC [34].

Among OWC, GlyCam-1 in the bactrian camel exhibited the lowest variability, with only seven polymorphic sites identified. In contrast, the dromedary camel displayed greater intraspecies variation. In NWCs, alpacas, and llamas presented 32 and 40 polymorphic sites, respectively. Among these, four SNPs resulted in amino acid changes: M3 (g.2324G > A; g.2325G > A) and M5 (g.2421G > A; g.2422G > A) were shared between alpacas and llamas; M4 (g.2419G > C) was specific to llamas; and g.2823C > T was unique to alpacas (Table 2).

To better understand the functional impact of these four non-synonymous SNPs g.2324G > A (M3), g.2419G > C (M4), g.2421G > A (M5), and g.2823C > T, we assessed their effects on protein structure and function using various bioinformatic tools. Analyses included predictions of changes in isoelectric point (pI), molecular weight, and secondary structure. For the first SNP, M3, which results in a glutamic acid (negatively charged) to lysine (positively charged) substitution, this amino acid change can, in certain cases, affect the structure or function of the protein due to the alteration in charge properties [35]. In our study, the p.Glu46Lys variant occurs at a highly conserved site and was predicted to be deleterious (Additional file 3 A), as conserved residues often contribute to protein stability and folding [36]. Similar Glu-to-Lys substitutions have been shown to impair protein function. For example, a substitution in the dimerization domain of HapR gene disrupted DNA binding [37]. A comparable change at position 3838 in the AtMDN1 gene induced pleiotropic developmental effects and functional impairment [38]. This was also evidenced by Carson et al. [39], in which p.Glu542Lys and p.Glu545Lys substitutions in the p110α subunit of PI3K disrupted its inhibitory interaction with the regulatory subunit p85, resulting in a constitutive activation of PI3K and uncontrolled cell growth and survival. Since PI3K is a kinase involved in protein phosphorylation - another key post-translation modification - these findings are particularly relevant, even though GlyCam-1 is functionally distinct. Despite differences in function, the charge-altering variants observed in this study may still affect GlyCam-1 protein stability or its interactions, potentially influencing its biological role.

Concerning the SNPs M4 and M5, responsible for p.Ser77Thr and p.Ser78Asn respectively, genotyping results indicated that the G allele, which encodes for serine in both cases, was more frequent (0.88 on average, Table 6). Serine has been reported as a mutational hotspot, likely due to its genetic coding properties and physicochemical characteristics. As Creixelle et al. [40] report, serine is unique in being encoded by six codons that are distributed across two distinct codon groups, making it more accessible via mutation from other amino acids. In addition, its moderate physicochemical properties make it less likely to be eliminated by natural selection [40].

From a functional standpoint, such substitutions are particularly relevant in milk-related genes, as serine is one of the three amino acids susceptible to phosphorylation in milk. According to Kelly et al. [41], substitutions that add or remove serine residues from the primary structure can alter the degree of phosphorylation, potentially affecting the structure and functional properties of milk proteins. In the case of αs1-casein, the degree of phosphorylation has been shown to alter the technological properties of milk, highlighting the importance of serine residues and post-translational modifications in downstream applications [42].

On a functional level, the two aforementioned variants were predicted to be neutral, with no significant impact on protein function. As reported by Prabantu et al. [43], changes in the amino acid sequence can alter protein structure without necessarily affecting its function. This observation was also stressed by Bromberg and Rost [44], who found that in approximately one-third of mutants, changes in protein stability and function were uncorrelated, highlighting the complex relationship between structure and activity.

Our results similarly showed that, although the two identified amino acid substitutions affected molecular weight, relative solvent accessibility, and the probabilities of coil and alpha-helix formation, they did not significantly impair protein function.

The fourth SNP (g.2823C > T), specific to alpacas, resulted in a p.Ser134Pro substitution. This change led to a marked decrease in alpha-helix probability (from 28.3% to 5.3%) and a corresponding increase in coil structure probability (from 71.3% for serine to 94.6% for proline), indicating a minor alteration in the secondary structure of the protein. This shift can be attributed to the unique structural properties of proline, which is widely recognized as a helix breaker [45]. Proline lacks an amide proton, a key participant in the hydrogen bonding network that stabilizes alpha-helices. As a result, proline residues are typically confined to the initial positions of helices, since their presence beyond position four often introduces a pronounced deviation in the helical structure [46, 47]. This characteristic explains the observed shift toward a coil conformation in the mutant protein.

The analysis was extended to the regulatory regions to explore the potential functional impact of promoter variants on gene expression. We conducted dual-luciferase reporter gene assays for two SNPs: M2 (g.563C > T) identified in dromedary camels and M1 (g.163A > G), shared among NWC. The luciferase assays were performed in 293T cells, which may differ in transcription factor composition from mammary epithelial cells. Although the 293T cell line has been successfully used in similar studies in camels and other livestock [26, 48–52], the observed allele-specific effects for M1 and M2 may not fully reflect in vivo expression in the mammary gland that requires further investigation.

In the 5’-untranslated region (UTR) of the dromedary camel, the g.563C > T SNP, predicted to alter an NF-κB binding site and showed a non-significant trend toward altered NF-κB mediated reporter activity (p = 0.102; Fig. 4). NF-κB is known to play a dual role in the mammary gland; it is essential for normal gland development and it is also activated during infections, such as mastitis. Its upregulation has been linked to mammary gland involution, significant reductions in milk production, and neutrophilic inflammation, hallmarks of mastitis [53, 54].

Although NF-κB is typically considered a transcriptional activator, it can also function as a repressor [55, 56]. NF-κB can negatively regulate cell adhesion molecules such as ICAM-1 and VCAM-1 through miRNA mediated or epigenetic mechanisms [57]. As GlyCam-1 is also a cell adhesion molecule, a similar repressive mechanism may contribute to its reduced expression in the presence of NF-κB, although the impact of the g.563C > T variant on GlyCam-1 expression has to be confirmed in in vivo studies.

GlyCam-1 is expressed in lactating mammary glands and contributes to immune defense and pathogen trapping. It has been found to be downregulated during mastitis caused by Streptococcus uberis and coagulase-negative staphylococci (CNS) infections [58, 59]. Notably, in cattle, GlyCam-1 is located within a quantitative trait locus (QTL) linked to somatic cell score (SCS) [58]. While these results suggest that GlyCam-1 is a candidate gene for mastitis-related traits, direct genotype–phenotype association studies are required before any implications for breeding or selection can be drawn.

Luciferase expression was also influenced by the M1 (g.163A > G) in NWC. Specifically, the A allele, predicted to create a myeloid zinc finger protein 1 (MZF1) TFBS, exhibited significantly higher expression levels (13.7-fold) than the G allele (p < 0.001). MZF1, a member of the Kruppel family of zinc finger proteins, functions as a bi-directional regulator, activating gene expression in hematopoietic cells and repressing it in non-hematopoietic cells [60, 61].

Our findings are consistent with previous studies demonstrating that SNPs disrupting MZF1 binding can alter gene expression. For example, a C > T substitution (ss974768522) in the promoter of the HAL gene disrupted MZF1 binding, and led to transcriptional changes [62]. Similarly, in the bovine SAA2 gene, the SNP c.17 C > G altered the MZF1 binding site significantly impacting promoter activity and gene expression in a genotype-dependent manner, with phenotypic consequences on milk traits [63]. In our study, the observed downregulation of GlyCam-1 expression in one genotype - validated through dual-luciferase reporter assays - may reflect a similar regulatory mechanism, where modulation of MZF1 binding influences transcriptional activity. It should be noted that the MZF1 binding site is predicted in silico and the allele-specific differences in luciferase activity suggested a potential regulatory role to be further validated in in-vivo expression studies.

The effects of other identified SNPs in the regulatory region of the GlyCam-1 gene were also investigated using various bioinformatic analyses. Among NWC, llamas exhibited greater polymorphism in the promoter region compared to alpacas. Among the llamas’ SNPs, g.609G > A was found to create transcription factor binding site for c/EBP-α. This mutation may have an impact on gene regulation.

In alpacas, two SNPs (g.244G > A and g.434C > G) negatively affected the presence of a Sp1 and AP-2α TFBS, respectively. Sp1 is known to play a key role in regulating milk fat synthesis [64] and has been shown to influence LALBA promoter activity [26]. For the AP-2, its involvement in transcriptional regulations has been demonstrated in several studies. For example, the formation of a novel binding site in the promoter region of the β-lactoglobulin gene has been associated with allele-specific expression differences [65]. Similarly, a modified AP-2 site in the bovine lactoferrin promoter has been linked to increased somatic cell scores and shorter calving intervals, underlining its functional significance [66]. In this context, the potential formation of an AP-2 site in the GlyCam-1 promoter may influence gene expression, although further experimental validation is needed to clarify its precise role in transcriptional regulation.

Additionally, bioinformatics analyses were conducted to identify the transcription factors conserved across the four species. The GlyCam-1 regulatory region was found to contain several elements associated with milk protein regulation, including c/EBP-β, c/EBP-α, OCT1, and AP1 [25, 26, 29]. Other conserved elements, identified - such as GR (glucocorticoid receptor) and STAT - have previously been reported in the proximal promoter region of the GlyCam-1 gene in cows, camels, mice, and humans [10]. Transcription factors previously detected in the 5’-flanking regions of camel milk genes, including AP-2, Sp1, GATA-binding proteins, and YY1 (Yin Yang 1), were also identified in this study, supporting their putative roles in transcriptional control mechanisms [67].

The interspersed elements within GlyCam-1 gene were also examined. Transposable elements (TEs), which make up significant portions of eukaryotic genomes, play key roles in genome evolution and gene regulation [68]. In camelids, repetitive sequences account for about one-third of the genome [69], highlighting their importance in gene architecture. In line with Kappeler et al. [17], who identified a 220 bp LINE-2 (L2a) element in intron 1 of dromedary camels, we detected this element in all four camelid species. Interestingly, the element was 9 bp shorter in llamas and alpacas while it retained its full length in dromedary and bactrian camels. Furthermore, two interspersed elements found in bactrian camels and llamas/alpacas were absent in dromedaries, consistent with Kappeler et al. [17].

A broader comparison with bovine PP3, summarized in Table 8, revealed that the camelid GlyCam-1 lacked several features found in bovines, such as SINE Bov-A2 in intron 1, a putative SINE in intron 3, a SINE Bov-tA in the 3’-flanking region, and four additional simple repeats [70]. However, the two LINE-2 identified in our study in the promoter and intron 1 of camelids were also present in bovine PP3, exhibiting 88% and 77% identity, respectively.

Overall, our findings indicate that retroelements are more prevalent than DNA transposons in the GlyCam-1 gene, consistent with previous reports on camelids [68]. That study also highlighted the higher frequency of mammalian-wide interspersed repeats (MIRs) within the SINE family, which aligns with our observations. The variation in interspersed elements between Old World (Camelini) and New world (Lamini) camelids likely stems from their evolutionary divergence. Camelini diverged from Lamini approximately 17 million years ago and exhibit 50% species-specific elements, compared to 33% in NWC. This pattern supports the hypothesis that these elements were inserted after the divergence of the two tribes. Furthermore, our phylogenetic analysis of GlyCam-1 protein sequences (Additional file 4) reveals distinct clustering of Camelini and Lamini species, with high bootstrap support, further corroborating these evolutionary relationships.

Conclusions

This study presents the first comprehensive characterization of the entire GlyCam-1 gene in camelids, revealing a conserved structure alongside higher genetic diversity in NWC. Nonsynonymous SNPs were identified in the exons of llamas and alpacas. Specifically, the SNPs M3 (g.2324G > A in alpacas; g.2325G > A in llamas) and g.2823C > T, resulting in amino acids substitutions p.Glu46Lys and p.Ser134Pro, respectively, were found to potentially affect protein function and interactions. In the dromedary camel, analysis of the regulatory region revealed the presence of M2 (g.563C > T), predicted to alter a NF-κB transcription factor binding site, whereas in NWC, M1 (g.163A > G) was predicted to create a MZF1 binding site. Both M1 and M2 were tested in an in-vitro assay, however, only M1 showed a statistically significant effect on gene expression. Overall, our findings highlight the genetic diversity and functional potential of GlyCam-1 in camelids. These results provide a foundation for future studies investigating the role of GlyCam-1 in lactation, immune function, and the potential regulation of gene expression relevant to disease resistance, including passive immunity transfer from mother to offspring.

Authors’ contributions

A.P. conceived and designed the experiments. N.L., A.P., H.A-H, G.L., C.V. and S.M. performed the experiments. N.L., G.L., G.C., Y.Z. and A.P. analyzed the data. G.G. oversaw the statistical evaluation of data. A.P. contributed reagents, materials, and analysis tools. N.L. and A.P. wrote the article. All authors read, revised and approved the final manuscript.

Funding

This research was financially supported by the project Generous (Grant Number 20200223 – PAUA_RIC_N_COMP_20_01), funded by the King Baudouin Foundation United States (KBFUS).

Data availability

All sequences generated in this study have been submitted to GenBank under accession numbers PV445494, PV445495, PV445496, and PV445497 for Camelus dromedarius, Camelus bactrianus, Vicugna pacos, and Lama glama, respectively. Sequences are also available to reviewers as optional related files.

Declarations

Ethics approval and consent to participate

No live animals were used in this study. Sampling was carried out for previous studies under full bioethical approval, and the same DNA samples, held in institutional collections at the University of Turin, the Justus-Liebig University of Giessen, and Kuwait University were used for the present research.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.