FSH glycosylation macroheterogeneity in pituitary and urinary hFSH samples was evaluated by Western blotting. Microheterogeneity in two highly purified urinary and pituitary hFSH preparations was evaluated by nano-electrospray mass spectrometry of peptide-N-glycanase-released oligosaccharides. An age-related loss of hypo-glycosylated hFSH in individual female pituitaries was indicated by progressively reduced abundance of hFSH21 relative to hFSH24. Urinary hFSH was evaluated as a potentially non-invasive indicator of glycoform abundance, as the only way for pituitary FSH to reach the urine is through the blood. Both highly purified and crude postmenopausal urinary hFSH preparations possessed the same amount of hFSH21 as postmenopausal pituitary gland FSH. Considerable microheterogeneity was encountered in both pituitary and urinary hFSH glycan populations, as 84 pituitary hFSH glycan ions were observed as compared with 68 urinary hFSH glycans. The biggest quantitative differences between the two populations were reduced abundance of bisecting GlcNAc-containing and fucosylated glycans, along with sulfated glycans in the urinary hFSH glycan population. The relative abundance of sialic acid and glycan antenna did not rationalize the retarded electrophoretic mobilities of the urinary hFSHβ21- and α-subunit bands relative to the corresponding pituitary hFSH bands, as the most abundant glycans in the former possessed only 2 more branches and the same sialic content as in the latter. Site-specific glycosylation information will probably be necessary.
Follicle-stimulating hormone (FSH) is a member of the glycoprotein hormone family, which is a subfamily of the cystine knot growth factor superfamily [1,2]. The glycoprotein hormones are composed of heterodimeric glycoprotein subunits, a common α-subunit, and a hormone-specific β-subunit. While the α-subunit primary structure is identical for all glycoprotein hormones within the same species, the oligosaccharide populations differ in a hormone-specific manner [3–6]. Characterizing the oligosaccharides released from an α-subunit preparation can identify the hormone from which the subunit was derived [7]. There are 3 to 4 β-subunits in vertebrates, which combine with α-subunit to create either FSH, luteinizing hormone (LH), thyroid-stimulating hormone (TSH), or in primates and equids, chorionic gonadotropin (CG) [8]. As both glycoprotein hormone subunits are cystine knot proteins [9–11] the protein backbone is folded into a series of three loops, two relatively rigid hairpin loops on one side of the knot, designated L1 and L3, and a single, flexible loop on the other side [12], designated L2. Oligosaccharides are attached to all 3 loops in a subunit-specific pattern (Figure 1). FSH subunits possess two potential N-glycosylation sites on each subunit and all four are of the Asn-Xaa-Thr type, which exhibit very efficient carbohydrate attachment [13]. Indeed, the α-subunit is always glycosylated at both sites in all known glycoprotein hormones. Because FSH α and β subunits co-migrate during electrophoresis, it is difficult to detect missing N-glycans in this hormone. FSHβ-specific Western blots have revealed partial glycosylation in equine FSHβ, human FSHβ (hFSH β), rhesus FSH β, and Japanese macaque FSHβ [14–16]. During the past few years, we have studied partially glycosylated hFSH isolated from pituitary extracts, postmenopausal urine, and conditioned tissue culture medium containing recombinant hFSH. Each glycosylation site in hFSH is decorated with a population of N-glycans. When total glycans are removed from reduced, carboxy-methylated FSH subunits, 39–130 glycans are found in mass spectra. We have data from only one glycosylation site, αAsn52, which possessed 29 neutral core ions, and when decorated with various patterns of sialic acid grew to 109 unique glycan structures. Micro heterogeneity can affect electrophoretic mobility, for example, placental hCGα with hybrid and biantennary glycans migrated faster than pituitary hFSHα, with triantennary, biantennary and tetraantennary glycans, which complicated sorting out the hFSH variants that resulted from loss of one or more N-glycans [17].
Abstract Follicle-stimulating hormone (FSH) is a key endocrine regulator of ovarian function. FSH is secreted as 2 macroglycosylation variants: partially glycosylated FSH (FSH21/18) and fully glycosylated FSH (FSH24). FSH21/18 is more potent than FSH24 at binding to and activating the FSH receptor (R). The ratio of FSH21/18:FSH24 has been shown to change with age, with FSH21/18 predominant at reproductive prime, and FSH24 predominant during perimenopause/menopause. How these FSH glycosylation variants modulate ovarian follicle functions remains largely unknown. The aim of this study was to investigate the effect of FSH glycosylation variants of pre-antral follicle function. Pre-antral follicles were isolated from 3- to 5-week-old C57BL/6 mice and treated ±10 ng/mL FSH21/18, FSH24, a ratio of 80:20 FSH21/18:FSH24 (to mimic reproductive prime), 50:50 FSH21/18:FSH24 (perimenopause), or 20:80 FSH21/18:FSH24 (menopause) for up to 96 hours. FSH21/18 and 80:20 FSH21/18:FSH24 increased follicle growth, in comparison with control, contrasting with FSH24 and 20:80 FSH21/18:FSH24. Survival rates were decreased in follicles treated with FSH24 or 20:80 FSH21/18:FSH24, with follicles undergoing basement membrane rupture and oocyte extrusion, increased Caspase3 gene and protein expression, and decreased markers of cell proliferation in FSH24 or 20:80 FSH21/18:FSH24–treated follicles. Moreover, this correlated with differential regulation of key genes modulating follicular functions. Pharmacological inhibitors of key FSH signal pathways suggests FSH21/18 and FSH24 initiate different FSHR signal pathway activation, which may determine their differential effects on follicle growth and survival. These data suggest that the nature of FSH glycosylation modulates the follicular cellular environment to regulate follicle growth and survival and may underpin the increasing ovarian resistance to FSH observed during aging.
We reconstituted ovine (o) LH alpha from its amino- and carboxyl-terminal fragments obtained as follows. oLH alpha was nicked at Arg46-Ser47 with Arg-C protease. Nicked oLH alpha disulfide bonds were broken by sulfitolysis, and its N-terminal peptide and C-terminal glycopeptide were separated by Sephacryl S-200 chromatography. Both fragments were mixed, reduced, and reoxidized. Reoxidation products were chromatographed on Sephacryl S-200, and an alpha-monomer fraction was recovered. The putative nicked alpha-monomer fraction was reassociated with native oLH beta, and the resulting oLH derivative was isolated by S-200 chromatography with a reduced yield of 11% (intact subunits yield, 67% oLH). This preparation was 2.6% as active as oLH in a LH receptor binding assay. Two additional oLH derivatives were prepared. Cleavage at alpha Arg46-Ser47 alone, followed by reassociation with native oLH beta, produced Arg-C-nicked oLH alpha:oLH beta (14% yield) that was 3.3% as active as native oLH. Reduction-reoxidation of Arg-C-nicked oLH alpha followed by reassociation with oLH beta produced reduced reoxidized-Arg-C-nicked oLH alpha:oLH beta (11% yield) that was 1.8% as active as oLH. These results indicated that the nicked oLH alpha monomer had been reconstituted from its N- and C-terminal fragments.
FSH plays a central role in regulating reproductive cycles by stimulating ovarian follicle growth, cell differentiation, and steroidogenic activity. FSH is a heterodimeric glycoprotein composed of two non-covalently associated, dissimilar subunits designated α and β. The α subunit is common to FSH and three other glycoprotein hormones--LH, TSH, and CG--and always possesses two N-glycans. The hormone-specific FSHβ subunit often lacks one or more glycans. Human FSHβ appears to be N-glycosylated in an all-or-none manner, producing two major variants, the classic 4-glycan, tetra-glycosylated hFSH and a novel, 2-glycan, di-glycosylated hFSH. In vitro receptor-binding assays indicate the latter is 10- to 25-fold more active than the former. The abundance of these glycoforms as determined by Western blot analysis of heterodimeric FSH isolated from individual pituitary glands appears to vary as a function of age and physiological state. In the present study, we examined day-to-day changes in urinary hFSH glycoform abundance. Accomplishing this required some changes in our experimental procedures. Western blots using anti-hFSHβ monoclonal antibody RFSH20 require 1 μg hFSH, which is suitable for the amounts available in human pituitary glands. However, only 100-200 ng FSH can be recovered from first void daily urine samples. Assessment with anti-hFSHβ peptide monoclonal antibody P03 or polyclonal antiserum W521, could detect 100 ng hormone, but the ratios of the 21- and 24-kDa bands differed from those determined with RFSH20. As glycoforms can be separated by size using high performance gel filtration, we developed a column-based assay system. When applied to the FSH heterodimer fraction isolated from a male pituitary, the same glycoform ratio was obtained by RFSH20 Western blotting and Superdex 75 gel filtration monitored at both 210 and 280 nm. Analysis of daily urinary FSH immunopurified from three subjects showed day-to-day variation in glycoform abundance. Previous pituitary FSH glycoform abundance determinations, which provide a snapshot of hormonal variation, had suggested changes in glycoform abundance occurred during the cycle. The glycoform abundance pattern in a 27-year-old woman consisted of low abundance of di-glycosylated hFSH early in the cycle, which changed to higher abundance on day 8, five days before the LH surge. This was followed by a pattern of progressive decrease in abundance over a 2- to 3-day period, an abrupt increase in abundance followed by progressive decline until the end of the cycle. In a 43-year-old woman, di-glycosylated hFSH was high early in the cycle, but progressively decreased until the day of the LH surge, when it abruptly rose, then became less abundant for the rest of the cycle. In a third woman, with no obvious LH surge, the abundance of di-glycosylated hFSH started low, rose to a maximum on day 12, and declined progressively until the last day of the cycle on day 19. These results confirm that day-to-day variations in FSH glycoform abundance take place and suggest the pattern of glycoform abundance may contribute to age-related changes in fertility. Supported by NIH P01 AG029531. (poster)
Residues 121–149 of equine LHβ (eLHβ) were removed by a simple mild acid treatment procedure. The modified eLHβ, des(121–149)eLHβ, was isolated by gel permeation chromatography on Sephacryl S-200. Recombination of des(121–149)eLHβ with eLHα and ovine LHα (oLHα) produced LH derivatives as efficiently as recombination with native eLHβ. In rat testicular LH radioligand assay systems employed in this study the potencies of the resulting LH preparations were, in order of decreasing potency: des(121–149)eLHβ:eLHα hybrid > eLH > eLHα + β > oLH > des(121–149)eLHβ:oLHα > oLHα + eLHβ (1:0.82:0.67:0.15:0.02:0.006, eLH tracer; 1:0.88:0.67:0.21:0.02:0.006, hCG tracer). In a horse testicular LH radioligand assay with eLH tracer, only the equine LH derivatives were active, and the order of potencies was the same: des(121–149)eLHβ:eLHα hybrid > eLH > eLHα + β (1:0.58:0.46). In a rat testicular Leydig cell steroidogenesis assay, eLH was the most active preparation, but the relative potencies of the other preparations remained unchanged: eLH > des(121–149)eLHβ:eLHα > eLHα + P > oLH > des(121–149)eLHβ:eLHα > oLHa + eLHβ (1:0.61:0.55:0.27:0.004:0.003). We have previously reported that the hybrid consisting of native eLHβ and oLHα was inactive (<1%) in LH receptor and steroidogenesis assays. The data reported herein confirm this observation and demonstrate that the absence of LH activity in eLHβ:oLHα cannot be attributed to the C-terminal extension on eLHβ, since the des(121–149)eLHβ:oLHα hybrid LH is also inactive. Examination of the intrinsic FSH activity of eLH in both rat and chicken testicular FSH radioligand assays produced the following results; eLH, recombined eLH subunits, and des(121–149)eLHβ:eLHα were all of the same potency (13% and 0.9% as active as eFSH in rats and chickens, respectively). We conclude that the C-terminal extension on eLH and eCG β- subunits is not involved in subunit association, LH receptor binding, or FSH receptor binding. The derivative des(121–149)eLHβ:eLHα provides a model compound that may be useful in determining the role, if any, of the glycoprotein hormone Cterminal extension that appears to have arisen independently at least twice in mammalian evolution. (Endocrinology124: 379–387,1989)