Copyright 1974. A11 rights reserved
`
`PEPTIDE HORMONES
`
`x 856
`
`Howard S. Tager' and Donald F. Steiner
`Department of Biochemistry, University of Chicago, Chicago, Illinois
`
`CONTENTS
`
`INTRODUCTION
`
`INSULIN
`NONINSULIN INSULIN-LIKE PROTEINS
`GROWTH HORMONE, PROLACTIN, AND PLACENTAL LACTOGEN
`GLYCOPEPTIDE HORMONES
`GLUCAGON AND RELATED HORMONES
`GASTRIN AND RELATED HORMONES
`CORTECOTROPIN, MELANOTIOPIN, AND LIPOTROPIN
`PARATHYROID HORMONE
`CONCLUDING REMARKS
`
`
`
`INTRODUCTION
`
`
`
`509
`510
`514
`515
`517
`522
`525
`526
`528
`530
`
`The last comprehensive review in this series on the biochemistry of peptide hormones
`appeared in 1969 (1). However, various aspects of this subject have been reviewed
`frequently in the Annual Reviews of Biochemistry. Physiology. Pharmacology, and
`Medicine, as well as in other review series including Recent Progress in Hormone
`Research, Hormonal Proteins and Peptides, and Vitamins and Hormones. This volume
`includes a review of peptide hormone binding to cellular constituents and the
`possible relationship of this phenomenon to their biological effects (2). The Atlas
`of Protein Sequence and Structure by Dayhoff2 is also a valuable source for
`comparisons of primary structural relationships among various peptide hormones.
`In view of the breadth of this topic, our discussion is limited to structural and
`biosynthetic studies; peptide hormone secretion and action are not considered. This
`is necessitated by the abundance of new information and concepts in all these
`areas. Rapidly accumulating sequence information on peptide hormones has
`provided interesting new clues to evolutionary and functional interrelationships
`among many hormones. Thus, several groups of related peptide hormones appear
`to have evolved from a relatively small number of ancestral proteins. Likewise,
`
`I Present Address: Department of Biochemistry, Medical College of Ohio, Toledo,
`Ohio 43614.
`2 Published by the National Biochemical Research Foundation, Silver Spring, Md. The
`last comprehensive edition appeared in 1972 (Vol. 5). Supplements are, published annually.
`
`509
`
`MYLAN EXHIBIT - 1027
`Mylan Pharmaceuticals, Inc. v. Bausch Health Ireland, Ltd. - IPR2022-00722
`
`

`

`510
`
`TAGER & STEINER
`
`recent biosynthetic studies of a variety of endocrine peptides indicate that the
`primary gene products differ significantly from the known peptide hormones. In at
`least six cases these peptides appear to be synthesized from larger precursors, and
`multiple molecular forms of the same hormone, representing biosynthetic
`intermediates as well as metabolites, may circulate in the blood. These new
`findings have altered earlier notions regarding the evolution, biosynthesis, and
`active forms of peptide hormones.
`
`INSULIN
`
`The last three to four years have seen considerable progress in elucidating the
`mechanism of insulin biosynthesis and in isolating and characterizing proinsulin
`(3, 4) and related peptide products from various species. It is now firmly
`established that proinsulin is synthesized on ribosomes in the rough endoplasmic
`reticulum of the /3 cells (5-7) and that the precursor is then transferred via an
`energy-dependent process to the Golgi apparatus (8-11), where proteolytic
`to insulin begins (12-15). Conversion continues to about 95%
`conversion
`completion over a period of hours within newly formed secretory granules after
`they have formed by budding from the inner lamallae of the Golgi apparatus
`(7, 13-15). The process may be terminated at this stage by cocrystallization of the
`insulin with the residual proinsulin and intermediate products (16), giving rise to the
`dense crystalline granule inclusion that can be seen by electron microscopy
`within mature granules (17, 18). This small amount of proinsulin retained within
`the granules is secreted subsequently with insulin and has been identified in the
`circulation of man and animals (19-22).
`The conversion process has been studied in greater detail in secretion granule
`fractions isolated from islets prelabeled with radioactive amino acids in vitro
`(7, 12-15). Conversion in these particles exhibits a relatively sharp pH optimum
`at or slightly above pH 6.0 (15, 23). At least two kinds of proteolytic activity appear
`to be required for the conversion of proinsulin to insulin. The first is an endo-
`peptidase with trypsin-like specificity which cleaves on the carboxyl side of the
`pairs of basic residues that link the connecting polypeptide chain to the termini of
`the insulin A and B chains (4, 24). The second is an exopeptidase having specificity
`similar to carboxypeptidase B, which removes the C-terminal basic residues from
`both insulin and the C peptide3 (25). Both kinds of activities have been
`demonstrated in disrupted secretion granule preparations (7, 15), and the endo-
`peptidase activity may be associated with the secretion granule membrane (7).
`Further support for this localization of the proinsulin converting enzymes within
`the /3 cells has been obtained in electron microscopic histochemical studies (26).
`Sufficient amounts of these enzymes for more detailed chemical characterization
`
`3 The arrangement of the mammalian proinsulin polypeptide chain is: N14,-A
`chain • Arg Arg • C-peptide • Lys Arg • B chain-COON. The C peptide thus becomes that
`portion of the connecting peptide sequence, aside from the pairs of basic residues at the
`ends, which is removed in the conversion to insulin (14).
`
`

`

`PEPTIDE HORMONES
`
`511
`
`have not been obtained from /3 granules, and studies with a variety of proteolytic
`inhibitors have been inconclusive in establishing their relationships to other known
`trypsin-like enzymes or carboxypeptidases (15). A further complication arises from
`the recent finding that a chymotrypsin-like cleavage occurs in rat connecting
`peptides during their proteolytic excision in incubated whole rat islets (27). These
`results suggest that the secretory granules may contain low levels of several kinds of
`proteolytic enzyme activities. Specificity of cleavage thus may be determined as
`much by tertiary structural features of the substrates as by special adaptations of
`the proteases. An enzyme that converts proinsulin has been isolated from whole
`pancreas, but it has not been fully characterized with respect to either its origin in the
`islet tissue of the pancreas or its cleavage specificity and mechanism of action (28).
`In view of a recent report that immunoglobulin chains having extended amino-
`terminal regions are synthesized during in vitro translation of myeloma cell mRNA
`fractions (29), one might inquire whether insulin precursors larger than proinsulin
`also exist. However, aside from some evidence for the expected role of an N-terminal
`residue of methionine in the initiation of proinsulin synthesis in fetal calf pancreas
`(30), no convincing indications of larger precursors have been found. Islet polysomes
`active in proinsulin synthesis appear to be mainly trisomes, a size consistent with the
`expected mRNA length of about 258 nucleotides required to encode the 86-residue
`proinsulin polypeptide (31).
`Comparative studies of insulin biosynthesis in the cod (32) and angler fish (33),
`as well as in such primitive vertebrates as cyclostomes (34), indicate the formation
`and cleavage of a proinsulin similar in size to the mammalian proteins. A require-
`ment for trypsin-like cleavage has been demonstrated for both of the fish proinsulins,
`and an interesting intermediate cleavage form, having an N-terminal tripeptide
`A-chain extension, has been isolated from angler fish islets by Yamaji et al (35).
`A number of reports have appeared on the biosynthesis, isolation, and characteriza-
`tion of intermediate forms of mammalian proinsulins in various species (15, 24, 27,
`36-40).
`The proinsulin C peptide, somewhat analogous to the activation peptide in some
`zymogen proteins, also has become a focus of attention. Due to localization of the
`conversion process within secretion granules, the C peptide accumulates with
`insulin in equimolar amounts (41) and is secreted along with the hormone by
`exocytosis of the granule contents (42). C peptides from nine mammalian and one
`avian species have been isolated and sequenced (41, 43-50). A high rate of mutation
`acceptance—much higher than the rate for insulin and approaching that for the
`fibrinopeptides (49)—as well as the appearance of deletions in more than one
`region of the C-peptide sequence, suggest that structural requirements in this
`portion of proinsulin are less stringent than in the hormonally active portion of
`the molecule. Whether or not biologically important functions other than efficient
`peptide chain folding, sulthydryl oxidation, and specific enzymic cleavage are also
`encoded in this peptide remains unanswered.
`Synthesis of several mammalian C peptides has been accomplished recently by
`classical fragment condensation (51-56). The synthetic porcine C peptide, contain-
`ing all four terminal basic residues, was tested for its ability to promote the
`
`

`

`512
`
`TAGER & STEINER
`
`recombination of insulin A and B chains in vitro, but it failed to influence the
`yield (57). Synthetic porcine and bovine C peptides cross-react well with antibodies
`directed against the corresponding natural proinsulins or C peptides, and frag-
`ments of these peptides have been successfully utilized to study the antigenic
`determinants in this region of the proinsulin molecule (53, 54, 58).
`Both proinsulin and the C peptide have been detected in the circulation of man
`and other species by means of specific immunoassays (for a review see 59). The
`level of proinsulin rises slowly after a glucose load in normal subjects but does not
`exceed 20% of the total insulin-like immunoreactive material. Although abnormal
`proportions or absolute concentrations of proinsulin have been found in obesity,
`chronic renal failure, and patients with severe hypokalemia, the major diagnostic
`significance of elevated proinsulin levels has been in detecting patients with /3 cell
`tumors (60). The immunoassayable C-peptide levels have been shown by Rubenstein
`and co-workers to mirror changes in insulin levels and thus provide a means for
`evaluating endogenous insulin production in diabetic individuals in whom insulin
`antibodies and administered animal insulins invalidate direct insulin measurements
`(61, 62). The handling of proinsulin and C peptide in vivo differs significantly from
`that of insulin (63, 64). This factor must be taken into consideration when
`interpreting changes in peripheral blood levels of immunoreactive insulin and in
`estimating the relative biological potency of proinsulin by means of in vivo blood
`glucose-lowering assays. The biological activity of proinsulin on fat cells (65) and
`muscle tissue (66) in vitro is about 3-5% that of insulin. Higher activity is usually
`found in vivo, ranging from 20-30%, a result attributable to the slower turnover
`rate of proinsulin rather than to any proteolytic conversion of proinsulin to insulin
`in the circulation or tissues (59). The difference in turnover rates of proinsulin and
`insulin is due largely to the relatively greater uptake and degradation of insulin by
`the liver, a major site of insulin destruction in the intact organism (67). The precise
`enzymic mechanism of insulin degradation remains controversial despite a recent
`renewal of interest in this problem (68-71). The studies of Varandani and co-
`workers have brought forth new evidence implicating an initial step of reductive
`cleavage in this process (71-73).
`The recent elucidation of the three-dimensional structure of insulin, initially at a
`resolution of 2.8 A (74, 75) and with recent refinements at 1.9 A (76) by Hodgkin
`and her co-workers, represents an important breakthrough in the study of peptide
`hormone structure. It is beyond the scope of this review to describe this structure
`in detail or the growing literature on the chemical and immunological properties and
`their structural correlations in normal, modified, or synthetic insulin molecules.
`Several recent reviews are cited (76—80). Modifications include the introduction of
`various substituent groups on the amino, histidyl, or carboxyl groups of insulin
`(76, 81-86), selective reduction and substitution of disulfide bonds (87), and the
`introduction of intramolecular crosslinks between the Al and Bl, or Al and B29
`(t-lysine) amino groups (79, 88, 89). The latter group of derivatives, especially the
`series linked between Al and B29 by dicarboxylic acids (79), are of added interest,
`since these bridges simulate the naturally occurring connecting polypeptide in
`proinsulin. Adipoylinsulin, in which adipic acid serves to crosslink the amino
`
`

`

`PEPTIDE HORMONES
`
`513
`
`groups of residues Al and B29, has been reduced and reoxidized in vitro under
`conditions suitable for proinsulin reoxidation with comparable yields ranging up to
`75% (90). These results indicate that the role of the connecting peptide in promoting
`correct pairing of half-cystine residues in proinsulin can be played by a non-
`peptide molecular prosthesis.
`None of the insulin derivatives prepared to date have exhibited higher biological
`activity than insulin itself. As with proinsulin, assays often indicate higher activity
`in vivo than in vitro (79), suggesting that many of these analogs may accumulate
`to a greater extent in the blood as a result of their decreased susceptibility to
`degradation. This cannot be the case with the identical turkey and chicken insulins
`which exhibit 2-4 times higher activity than bovine insulin in several in vitro
`bioassay systems (91). Weitzel and co-workers suggest that this heightened activity
`may result from enhanced receptor binding due to the substitutions of histidine
`for alanine and asparagine for serine at positions A8 and A9, respectively.
`On the other hand, substitution of the B5 histidine by alanine in a synthetic bovine
`insulin led to lower biological activity (92). Duck insulin may help to clarify the role
`of substitutions in the A8-10 region in enhancing biological activity of chicken and
`turkey insulins. This avian species has glutamic acid at position A8 and proline at
`A10, and the B chain differs only at position 30 where threonine replaces alanine
`(93). Bioassay results have not yet been reported.
`In contrast, guinea pig insulin, which differs from porcine insulin at 17 positions
`(49), displays significantly lower biological activity in other mammals as well as in
`guinea pigs (94). In addition to altered molecular topography due to the amino
`acid changes, guinea pig insulin does not bind zinc and does not form dimers or
`higher polymers in solution (94). Associated with this rather drastic change in
`properties is the replacement by asparagine of the BI0 histidine residue, which
`coordinates with zinc in two zinc porcine insulin crystals (74). By contrast, the two
`insulins in the mouse, an old-world rodent, closely resemble other mammalian
`insulins and are identical to those of the rat (95).
`Another insulin of considerable interest has been isolated from the islet organs of
`a primitive jawless vertebrate, the Atlantic hagfish (34, 96), an animal belonging
`to one of two extant orders of the cyclostomes, which are believed to have
`diverged from the gnathostomes about 600 million years ago (97, 98). Hagfish insulin
`has about 10% of the activity of bovine insulin in mammalian systems (96). About
`40% of its amino acid residues differ from those found in mammalian insulins,
`including replacement of the zinc-binding BI0 histidine residue by aspartic acid (99).
`Nevertheless, it forms large tetragonal crystals at pH 6.0 in the absence of zinc (34).
`Almost all of the invariant residues in the known gnathostomian insulins (49),
`including the half-cystines, are conserved in hagfish insulin.
`Studies with selectively degraded or synthetic insulins indicate that the absence
`of the Al glycine amino group (100) or the AI glycine residue (79, 81) results in a
`loss of biological activity. In the absence of the A1-4 tetrapeptide sequence (101)
`or the C-terminal A21 asparagine residue (102, 103), biological activity is either
`absent or extremely low. Although C-terminal shortening of the B chain up to
`residue 27 does not affect the biological activity (104, 105), further deletions toward
`
`

`

`514
`
`TAGER & STEINER
`
`residue 22 (102) progressively reduce biological activity (104), suggesting that the
`region B22-26 may play an important role in biological activity (106) as
`well as in dimerization (76). Replacement of the cysteine residues providing the
`intrachain disulfide loop, between A6-11 with two alanine residues results in a
`product with about 10% of normal biological activity in vivo (107). X-ray diffraction
`studies are required to differentiate purely local from more generalized topo-
`graphical changes in these altered molecules, and thus delineate more precisely
`the region(s) required for cell binding and biological function.
`The three-dimensional structure of proinsulin has evoked considerable interest
`ever since the prohormone was first isolated. Although it has been crystallized
`successfully in several forms by Low and co-workers (108, 109), the crystals have
`not been of sufficient quality to permit extensive data collection. Preliminary X-ray
`diffraction analyses, however, confirm the results from other physical studies
`indicating that proinsulin aggregates similarly to insulin to form dieters and
`hexamers under appropriate conditions (110, 111). The hypothesis that the insulin
`moiety of proinsulin has essentially the same conformation as insulin is borne out by
`similarities in optical rotary dispersion (ORD) and circular dichroism (CD) spectra,
`by the titration behavior of the tyrosine residues of both peptides (112, 113), and by
`the high degree of immunological cross-reactivity of proinsulin with an tisera to insulin
`(114, 115). Little evidence for the existence of ordered structure in the connecting
`peptide portion of proinsulin has been obtained (112). The peptide is believed to be
`folded in some manner over the external surface of the insulin monomers in proinsulin
`hexamers, spanning the 8-10 A gap between the C terminus of the B chain and the
`N terminus of the A chain (16). Proinsulin and insulin evidently can form mixed
`polymers, which may account for the tendency of the prohormone to crystallize with
`insulin during the commercial preparation of insulin (16).
`In addition to the studies on proinsulin mentioned above, recent structural
`investigations on insulin and various derivatives of insulin have utilized a variety of
`physical probes, including ORD and CD (103, 112, 116), Raman spectroscopy (117),
`electron paramagnetic resonance spectroscopy (118), and infrared spectroscopy
`(119). Recent evidence on the structure of insulin fibrils produced at elevated
`temperatures in acid solutions indicates that these are cross-a structures having a
`uniform cross-section of 29 x 47 A, and consisting of flattened insulin monomers
`packed heterologously in layers 4.7 A thick in the direction of the fibril axis (120).
`
`NONINSULIN INSULIN-LIKE PROTEINS
`
`It has become increasingly clear that proteins other than insulin can share many of
`its biological actions and may have additional physiological functions as well. One
`of these is the nonsuppressible insulin-like activity (NSILA) of plasma that has been
`studied extensively by Froesch and co-workers (121, 122). Although it circulates
`normally as an inactive larger complex, a soluble active substituent of molecular
`weight about 7500 can be separated under acidic conditions. This single-chain
`peptide can compete weakly but effectively with insulin for receptor sites in
`membrane preparations (123, 124), and it reproduces all the known biological
`
`

`

`PEPTIDE HORMONES
`
`515
`
`effects of insulin (125, 126). As suggested first by Hall, recent evidence points to the
`probable identity of this substance with somatomedin, the new designation (127) for
`the well-known serum sulfation factor believed to be causally associated with the
`growth-promoting actions of growth hormone in vivo (124, 128-130). Although
`insulin, somatomedin, and perhaps the submaxillary gland nerve growth factor,
`which has some structural similarities to proinsulin (131), all produce basically
`similar anabolic and mitogenic effects, it is clear that nonidentical receptors with
`sharply differing binding specificities for these peptides exist in their various target
`tissues (124). Whether or not lowered cyclic AMP levels (130) can account for the
`spectrum of responses elicited by these distinctive groups of peptides is a question
`that requires much further study.
`
`GROWTH HORMONE, PROLACTIN, AND PLACENTAL
`LACTOGEN
`
`The primary structures of mammalian growth hormones (GH)' have been actively
`investigated in recent years. The ovine hormone contains 189-191 residues (132, 133).
`The primary structure of bovine GH is practically identical to that of the ovine
`hormone (132-134). Partial sequences of porcine (135) and equine GH (136-138)
`suggest that as a group the ovine, bovine, porcine, and equine hormones vary by less
`than 5% of their structure. The peptide chain of human GH is about the same length
`as that of the ovine hormone, but less than two thirds of its residues are homologous
`with ovine or bovine GH (139-142). This difference in structure probably accounts
`for the lack of bovine GH activity in man (143). The mammalian growth hormones
`contain two disulfide bridges: one spans 110 residues and joins cysteine residues
`53 1 64 ; the other spans only six residues and joins residues 181 and 189.
`Two other peptide hormones show remarkable structural similarities with GH.
`The chorionic hormone, placental lactogen, contains 191 residues, 85% of which are
`homologous with those of human GH (140, 144, 145). This near identity in primary
`structure is reflected by a similar placement of disulfide bridges and some similarities
`in biological activity. Since the placental lactogen appears to possess a portion of
`the activity of human GH in promoting growth, Li and his associates have suggested
`that it be called chorionic somatomammotropin (146). The other hormone
`exhibiting structural homology with GH is pituitary prolactin. The complete
`primary structures of both the porcine and ovine prolactins are now known (147-
`149). The bovine hormone is probably nearly identical to the ovine (150). Despite
`earlier uncertainty as to the existence of a unique prolactin in the primate pituitary,
`the monkey (151) and human (152) hormones have recently been isolated, and the N-
`terminal sequence of the human hormone was determined by Niall et al (153).
`
`4 The abbreviations used in this article are the following: GH, growth hormone
`(somatotrophic hormone); LH, luteinizing hormone (interstitial cell-stimulating hormone);
`FSH, follicle-stimulating hormone; TSH, thyroid-stimulating hormone; HCG, human
`chorionic gonadotropin; ACTH, adrenocorticotrophic hormone (corticotropin); MSH,
`melanocyte-stimulating hormone (melanotropin); LPH, lipotrophic hormone (lipotropin);
`and PTH, parathyroid hormone.
`
`

`

`516
`
`TAGER & STEINER
`
`Human amniotic fluid is a promising source for human prolactin (154). Ovine
`prolactin contains 198 amino acid residues and three disulfide bridges. Two of these
`appear to be homologous with the disulfide bridges of GH. The other spans a six-
`residue sequence near the N terminus of the hormone and is also present in human
`prolactin (153).
`Although the amino acid sequences of ovine GH and ovine prolactin show less
`than 30% homology, many of the apparent amino acid interchanges appear to be
`relatively conservative when viewed either by side-chain functionality or by the
`number of nucleotide base changes necessary to alter the appropriate codons (140,
`147). A common evolutionary origin of prolactin, placental lactogen, and growth
`hormone has been suggested (140). Even more provocative is the suggestion by
`Niall and co-workers that these peptide hormones contain internal structural
`homology (140). On this basis, they propose that the three hormones may have
`arisen from repeated tandem duplications of a gene coding for a relatively small
`peptide (140).
`The partial functional similarities of GH and prolactin in mammals are
`paralleled by their similar actions in lower vertebrates (155). Ovine prolactin and
`ovine GH are both potent somatotropins in a variety of such species, including
`teleost fishes and reptiles. The growth hormone present in the pituitaries of modern
`teleosts is apparently ineffective in mammals, however (156). The pituitaries of birds,
`reptiles, and amphibians appear to contain separate prolactin-like and growth
`hormone-like hormones (155). The biological significance of these forms, especially
`the prolactins, is still unknown. The recently isolated growth hormones from duck
`and turtle have remarkably similar amino acid compositions which differ only
`slightly from that of human GH (157).
`Prolactin, placental lactogen, and GH also share some features of secondary
`structure, as determined by CD, fluorescence, and titration behavior (158-160).
`Each of the three hormones contains about 50% a helix and all undergo similar
`structural transitions in acid or base. Aloj & Edelhoch have found that human GH
`is less easily denatured than bovine GH or ovine prolactin by treatment with acid
`or urea (160). The single tryptophan residue in placental lactogen appears to be
`more exposed to the solvent than is the corresponding residue in human GH (159).
`Likewise, this tryptophan residue appears to be more exposed in human than in
`bovine GH (161, 162). Cambiaso and co-workers interpret rates of proton
`exchange as indicating that the structure of human GH is more open than that of the
`bovine hormone (163). Thus relatively small structural or conformational changes
`may account for the lack of activity of bovine GH in man.
`Controversy continues to surround the question of the biological activity of
`various fragments of growth hormone. Bornstein and co-workers claim to have
`isolated two peptides having opposing biological activities after limited proteolysis
`of human GH (164). A C-terminal, 25-residue fragment of the hormone is reported
`to be diabetogenic, while a smaller N-terminal fragment of the hormone appears to
`have an insulin-like effect (164, cf 140). However, Schwartz has reported that a
`cyanogen bromide fragment of human GH containing the C-terminal 21 residues
`lacks the inhibitory actions of Bornstein's 25-residue fragment on several glycolytic
`enzymes in vitro (165).
`
`

`

`PEPTIDE HORMONES
`
`517
`
`In the case of bovine GH, a 37-residue tryptic fragment derived from the large
`disulfide loop of the hormone appears to have some growth-promoting activity,
`even in man (166, 167). This peptide as isolated by Sonenberg et al is partially
`homologous with a similar sequence of human GH and may include the active core
`of the hormone. Another fragment of human or bovine GH, consisting of residues
`5-117 and containing a single residue of aminoethylcysteine, exhibits some of the
`biological properties of the intact hormone (168), but does not stimulate skeletal
`growth. Synthetic sequences of human GH representing residues 87-123 and
`residues 124-155 appear to have growth-promoting activity in rats (169). In the
`course of studies of the biological activities of GH fragments, several incorrect
`sequences have been synthesized (164, 170, cf 140 and 169). That these peptide
`analogs possess some biological activity suggests that considerable latitude is
`allowed in the recognition of active forms.
`Several recent reports suggest that circulating human GH consists of GH itself and
`a considerably larger form, which usually makes up 10-30% of the plasma
`immunoreactivity. Since this large GI-1 is dissociated to GH by urea or by various
`physical treatments, it may represent a GH aggregate or GH bound to a larger
`protein, rather than a precursor. Such large forms of immunoreactive GH are also
`found in human, rat, and dog pituitary glands (171-173). The physiological
`significance of the large form is not known. A similar large form of placental
`lactogen has been detected in human pregnancy serum and in extracts of human
`placenta (174)..
`
`GLYCOPEPTIDE HORMONES
`
`Since the discovery that the treatment of ovine luteinizing hormone (LH) with dilute
`acid results in the dissociation of the hormone into dissimilar subunits (175, 176),
`considerable effort has been expended in isolating this glycopeptide hormone from
`a variety of sources. The dissociated subunits have been separated from each other
`by countercurrent distribution (177) or by various chromatographic procedures
`(178). The a subunits of ovine (179, 180), porcine (181), and bovine (182-184) LH
`contain 96 amino acid residues, whereas the a subunit of human LH appears to
`contain only 89 residues (185). Carbohydrate is attached at asparagine residues 56
`and 82 of the three nonprimate subunits and at the homologous asparagine
`residues 49 and 75 of the human subunit.
`The primary structures of ovine LH-a as reported by Sairam, Papkofi•' & Li (180)
`and by Liu and associates (179) differ only by an inversion at positions 88 and 89.
`The former group proposes the sequence Cys-Ser, whereas the latter suggests Ser-Cys.
`The homologous sequence reported for human LH-a is Ser-Cys (185), but the
`sequence reported for bovine LH-a and HCG-a is Cys-Ser (182, 183, 186). The
`a subunit of ovine LH appears to be markedly heterogeneous at its N-terminus, due
`to deletion of up to eight of its N-terminal residues (179). An apparent 7-residue
`deletion at the N terminus of human LH-a, when compared to ovine LH-a, has
`been well documented (185, 187, 188). Whether this is a true deletion or an
`artifact due to proteolytic degradation during isolation is uncertain.
`A comparison of primary structures shows the a subunits of bovine and ovine LH
`
`

`

`518
`
`TAGER & STEINER
`
`to be identical, while the porcine subunit differs from these by only 5, and the
`human by 23, residues. Many amino acid sequences within the subunits from
`different species remain invariant, the longest blocks of identity being 14, 11, and 8
`residues. The 10 cysteine residues and their homologous positions also remain
`unchanged.
`The carbohydrate content of LH has been difficult to assess. The total carbohydrate
`and the relative proportions of hexoses, fucose, hexosamines, and sialic acid appear
`to differ among the human (188, 189), ovine (190), bovine (191), and equine
`hormones (192). The equine hormone appears to be unusually high in sialic acid.
`Alterations in the carbohydrate content, particularly in sialic acid, have been
`proposed as an explanation for microheterogeneity in human LH (193).
`The primary structure of the. /3 subunit of ovine LH as determined by Liu et al
`(194) differs from that of Sairam et al (195) in amidation at residues 10 and 62
`and also in the sequence following residue 50. The former group proposes the
`sequence Pro-Met-Pro whereas the latter group suggests Pro-Pro-Met-Pro. More
`recent evidence suggests that the latter sequence With 120 residues is correct (182).
`The N-terminal serine residue appears to be acylated and carbohydrate is attached
`at the asparagine residue at position 13. The bovine (196) and ovine (194) LH-/3
`subunits appear to be nearly identical. The primary structure of the /3 subunit of
`porcine LH differs from that of ovine LH at about 15 of 120 positions (197). Large
`blocks of /3 subunits of the ovine, porcine, and bovine hormones have identical
`amino acid sequences, the longer ones containing 21, 18, and 16 residues.
`Recently, two reports of the primary structure of the /3 subunit of human LH
`have appeared. Both Shome & Parlow (198) and Closset, Hennen & Lequin (199)
`report that the subunit contains 115 residues and that the N terminus is serine and
`the C terminus glycine. Aside from four other amino acid differences between the
`two structures, Shome & Parlow do not report the existence of methionine 43
`and proline 55 in their sequence. The total length of the subunit is maintained,
`however, by differences in the C-terminal sequence following the last cysteine residue.
`Shome & Parlow report seven additional residues (198), whereas Closset et al report
`only five (199). Both groups suggest, however, that human LH-/3 contains 12
`cysteine residues as do the porcine, ovine, and bovine subuni