The Primary Structure of Bovine Thyrotropin II. THE AMINO ACID SEQUENCES OF THE REDUCED, X-CARBOXI-METHYL a AND /3 CHAINS”

SUMMARY The linear amino acid sequence of the two chains of bovine thyrotropin (TSH-cx and TSH-/3) have been determined except for assignment of some amides. The order of the tryptic peptides of TSH-/3 (described in the preceding publication) was established by studies on the peptides obtained from tryptic hydrolysates of maleylated, reduced, S-carboxymethyl TSH-/3, and from the oxidized chain. The preparations of TSH-/3 exhibited heterogeneity at the COOH terminus; molecules of 113 residues terminate with the tripeptide sequence -Ser-Tyr-Met-COOH while others, of 112 residues terminate with -Ser-Tyr-COOH. Evidence for the heterogeneity of the NH2 terminus of TSH-a was again found in the peptides from its maleylated derivative. The molecular weight, calculated from the amino acid sequences and carbohydrate compositions, of TSH-a preparations is approximately 13,600, that of TSH-fi preparations, 14,700.


SUMMARY
The linear amino acid sequence of the two chains of bovine thyrotropin (TSH-cx and TSH-/3) have been determined except for assignment of some amides.
The order of the tryptic peptides of TSH-/3 (described in the preceding publication) was established by studies on the peptides obtained from tryptic hydrolysates of maleylated, reduced, S-carboxymethyl TSH-/3, and from the oxidized chain. The preparations of TSH-/3 exhibited heterogeneity at the COOH terminus; molecules of 113 residues terminate with the tripeptide sequence -Ser-Tyr-Met-COOH while others, of 112 residues terminate with -Ser-Tyr-COOH. Evidence for the heterogeneity of the NH2 terminus of TSH-a was again found in the peptides from its maleylated derivative. The molecular weight, calculated from the amino acid sequences and carbohydrate compositions, of TSH-a preparations is approximately 13,600, that of TSH-fi preparations, 14,700.
TSH-a! contains two oligosaccharide moieties, TSH-/3 one. Both chains are extremely rich in intrachain disulfide bonds, five in TSH-cr and six in TSH-P; their positions remain to be determined.
There are no interchain disulfide bonds.
The sequences of TSH-a! and -0 do not resemble each other superficially, but several very similar or identical sequences are found in TSH-P and in the hormonespecific chain, CII, of luteinizing hormone, despite marked differences in their amino acid compositions.
The sequence of the latter has been announced recently by Liu ef at. (Res. Commun. Chem. Pafh. Pharmacol., 1, 463 (1970)). The similarities may be a reflection of similar sites on the hormone-specific chains for interactions with the nearly identical subunits, TSH-CJ and CI of luteinizing hormone, which lead to the active hormones.
Bovine thyrotropin has recently been found to consist of two peptide chains (1, 2) one of which is very similar in properties to the CI chain of luteinizing hormone. The preceding paper describes the isolation of the peptides obtained by cleavage of reduced S-carboxymethyl TSH' with either trypsin or cyanogen * This investigation was support,ed by United States Public Health Service Grant CA-2290 from the National Cancer Institute. 1 The abbreviations used are: TSH, thyroid-stimulating hormone (thyrotropin); TSH-a, the first. of two subunits of TSH to UCLA Xchool of .Uedicine, Los Angeles, California $0024 bromide (3). A number of amino acid sequences were also determined and sufficient. information was obtained to assign and order 10 tryptic peptides plus COOH-terminal serine to the a! chain of TSH.
These peptides accounted for the entire amino acid composition of TSH-CY. Twelve remaining tryptic peptides plus a cyanogen bromide fragment of 8 residues and the free homoserine released by reaction with cyanogen bromide accounted for the amino acids of the fl chain. Only one tryptic peptide, Ser-Tpr, could not be assigned to either chain. The two remaining large cyanogen bromide fragments (52 to 53 and 51 to 53 residues, respectively) originated from the p chain and their order was determined by isolation of the overlap tryptic peptide.
In this paper we describe the tryptic peptides obtained from the separate cy and /3 chains after reduction, S-carboxymethylation, and blocking of their lysine residues to the action of trypsin by maleylation (4). The ordering of the conventional tryptic peptides and cyanogen bromide fragments in TSH-LU is confirmed and the order of the tryptic peptides from TSH-P is established. The complete amino acid sequences, excluding some assignment of amides, of the reduced X-carboxymethyl derivatives of both TSH-a and -/3 are given. EXPERIMENTAL PROCEDURE Bovine TSH was prepared with the final purification step by countercurrent distribution, and the two chains were separated by gel filtration after their dissociation in 1 M propionic acid as described previously (2). Most of the analytical methods and materials used are described in the preceding paper (3). Mercaptoacetic acid and phenol were added to samples before hydrolysis for amino acid analysis (3,5) if X-carboxymethylcysteine, methionine, or tyrosine were suspected. Most of the Edman degradations described herein were carried out by the method described by Gray in which the successive NHz-terminal amino acids n-ere identified as their dansyl derivatives (6,7). The method is referred to as the dansyl-Edman.
Reduction, X-Carboxyrnethylation, and Maleylation of TSH-CY and +-The individual chains were reduced and S-carboxymethylated as described, with excess reagent removed by gel filtration while protected from light (8). The materials were maleylated by the procedure of Butler et al. (4) The pH was adjusted to 9.0 with 1 N NaOH and, after standing 15 min at room temperature, the reaction mixture was passed through a column (2 x 90 cm) of Sephadex G-25 (coarse) in the presence of 157; NH4HC0s.
The protein was then recovered, salt-free, by freeze-drying for 48 hours.
Oxidation with Performic Acid-Performic acid was prepared by adding 1 volume of 30y0 HZ02 to 9 volumes of 99y0 formic acid and allowing the mixture to stand at room temperature for 1 hour. The solution was chilled in ice. One volume of methanol and solid phenol (I%, w/v) were then added.
Immediately, 8 mg of TSH-fl were dissolved in 0.5 ml of the solution and allowed to remain in the ice bath for 3 hours. The excess reagents and salts vvere removed by the passage through a Sephadex G-25 (coarse) column (1.2 x 40 cm) in the presence of 0.5% NHdHC03 at 4". Although some bubbles formed in the column, complete separation of the protein from lower molecular weight mat.erials was obt,ained.
The protein was recovered by freezedrying.
Protein or peptide concentration was 5 to 10% (w/v) and the reaction time 2 to 4 hours or longer. Fractionations of peptides after hydrolysis are described in subsequent sections.
Assignment of AmidesPeptides were digested with aminopeptidase M (Henley and Company, Kew York). Approximately 30 mpmoles of peptides were dissolved in 1 y0 NH4HC03 and 20 ~1 of aminopeptidase &!I (10,000 milliunits per ml of water) added.
The time of hydrolysis, at 40", was 24 to 48 hours with the longer time used for the large peptides.
The hydrolysates were t,hen taken t,o dryness and analyzed on the amino acid analyzer.
Since both glutamine and asparagine appeared at the serine position, the appearance of a new or larger peak at the serine position as well as the disappearance of aspartic acid or glutamic acid relative to that found after acid hydrolysis permitted identification of amides in several peptides. Nomenclature-Peptides obtained from tryptic digestions of the maleylated chains were designated as T(m) and from oxidized material as T(ox).
The half-cystirres were determilred as the S-carboxymethyl derivative.
The major peptides llecessarp for establishing the order are shown.
Tryptic Peptides jrow~ Maleylated, Reduced, S-Carboxymethyl TX&a--The separation on a column of Sephadex G-50 (fine) of the peptides from the tryptic hydrolysate of the maleylated derivative is shown in Fig. 3. Details of the hydrolysis and separation are given in the legend.
Fractions I, II, and III were passed through the column a second time, Fractions I and III gave symmetrical elution profiles; Fraction II was partialI? separated into two components, Fraction II-1 and II-2 as shown in Fig. 4. Amino acid analysis showed that Fractions I, II-l, and II-2 required further purification which was achieved, with only partial success for Fraction II-2, by paper electrophoresis at pH 6.4. The peptides were eluted with 107, pyridine.
The amino acid compositions of the above fractions together with those of the other peptides from the maleylated derivative :ar'e The values of phenylalanine and proline are low, most likely caused by the lack of the NHz-terminal sequence Phe-Pro in some molecules and in agreement with the isolation of orT-1 and la and crCNBr-1 and la (3). Peptide crT(m)-la has the same composition with the exception of Residues 1 through 6 and must have resulted from a chymotryptic-like cleavage at the Phe-Thr bond at position 6. The compositions of the other  fragments agree perfectly with the order of the tryptic peptides established in the preceding paper (3). It should be noted crT-6, which is placed within aT(m)-3, gave a poor analysis for methionine ( (3), Table I). The finding of 2 residues of methionine in aT(m)-3 confirms the ordering of &NBr-3 and aCNBr-4 (3). The summation of the residues in the four primary tryptic peptides from the maleylated derivative is in agreement with the composition of the intact chain and the yields are close to theory.
As in the case of the cyanogen bromide fragments, the compositions of the large maleylated peptides are sufficiently distinctive to permit unequivocal assignment of the smaller tryptic peptides within them.
Sequence of TSH-a--In the preceding paper (3), the complete sequences of Peptides otT-2, otT-4, aT-5, otT-7, aT-8, aT-9, and ~LT-11 are given and the sequence of Peptide crT-6 was deduced from the partial sequences of cyanogen bromide fragments In this and subsequent tables the symbol -indicates that the sequence was determined by the dansyl method after each step of the Edman degradation.
A solid l i ne under the sequence indicates the residues positioned by sequence studies; a dashed l i ne signifies residues known to be present by composition. CMCys is S-carboxymethylcysteine.
sequence I arCNBr-3 and aCNBr-4. The remainder of the TSH-or sequence was determined by enzymatic digestions of the larger tryptic peptides from the reduced, S-carboxymethyl derivative, both before and after its maleylation.
Peptide &-l-It was found that the peak which emerged from columns of P-10 directly after the glycopeptide fraction of a tryptic hydrolysate of both chains contained only Peptides aT-1 and @T-l2 (5, Fig. 1). The two peptides separated during electrophoresis on paper at pH 1.9. Peptide CLT-1 (0.5 pmole) was hydrolyzed with chymotrypsin (50 pg) in 0.1 ml of 0.126 M NH,HCOs at pH 8.0 and 40" for 4 hours. The reaction mixture was spotted on paper and the peptides were isolated after electrophoresis at pH 4.8. They appeared as C-2, C-l, and C-3 in the order of anode to cathode, and their compositions as well as sequences, determined by the dansyl-Edman, are shown in Table II. The order of the chymotryptic peptides is established because the entire tryptic peptide has NHz-terminal phenylalanine.
Peptide C-2 must have originated from Peptide aT-la which is missing its ?JHz-terminal Phe-Pro sequence.
Peptide olT-S-The sequence of the first 7 residues from the NH2 terminus and of 7 residues from the COOH terminus of Peptide aT-3 was described in the previous paper (3). The NH2terminal sequence was obtained by the Edman degradation (3 ,  Table IV) and the COOH-terminal sequence was deduced from a study of CSBr fragments of TSH (3 , Table XIII).
Their compositions and the sequence of Th-5 are shown in Table III.
These data, along with the data of the hydrolysis of aCKBr-2 with carboxypeptidase A, complete the sequence of LOT-3. The 2nd residue of proline is placed by difference at the COOH terminus of Th-3 and Th-4.
The NHz-terminal sequence was determined by the dansyl-Edman technique. The first 3 residues agreed with previous data from Peptide crT-10 (3, Table V). The 4th residue, histidine, could not be identified by the dansyl-Edman.
By the difference in amino acid composition between crT(m)-4 and crT-10 (3 , Table I), and because both peptides have the same NHz-terminal sequence, serine is placed at the COOH terminus of aT(m)l.
Hydrolysis of Peptide aT(m)-4 (0.7 pmole) with chymotrypsin (210 pg) in 0.5 ml of 0.06 M NH4HC0a at pH 8.0 and 40" for 2 hours yielded three peptides and free tyrosine. They were separated on a column of Sephadex G-25 as shown in Fig. 5~. Fraction I was the glycopeptide C-l ; Fraction II was a tripeptide C-4 and Fraction IV consisted of free tyrosine.
Fraction III contained two components with the same composition (Tyr , His, Lys, Ser), one of which gave a light pink color upon reaction with ninhydrincollidine rather than purple.
The two were separated by electrophoresis at pH 6.4; the sequence of the darker staining material was determined.
The positioning of maleyllysine as the 2nd residue from the COOH terminus of crT(m)l was based on the specificity of tryptic cleavage and the hydrolysis of crT-10 with carboxypeptidases A and B (3, Table V  well as the NHz-terminal tyrosine of C-3 (Table IV).
Hydroly-was not attempted and the fraction was treated with subtilisin sis of C-l with carboxypeptidase A allowed tyrosine to be placed (54 pg) in 0.5 ml of 0.06 M NHIHCOa, pH 8.0, at 40" for 4 hours. at the 5th residue.
The reaction mixture was then fractionated on Sephadex G-25. The remaining portion of Peptide C-l (0.65 pmole) was then The elution profile is shown in Fig. 5c. Fraction I shown in this treated with thermolysin.
Thermolysin (200 pg) in 0.5 ml of figure contained glycopeptides with COOH-terminal hetero-0.06 M NH4HC03, pH 8.0, at 40" for 22 hours did not cause geneity as indicated by its amino acid composition.
Fraction II significant hydrolysis. Another 500 1.18 of thermolysin were showed five ninhydrin-positive spots upon paper electrophoresis then added and the reaction was continued for 18 hours at 40". at pH 3.6. Only four of these contained sufficient material for The products were separated on Sephadex G-25 as shown in analysis.
The material nearest the anode contained Peptides Fig. 5b. Fraction II contains Peptide Th-2 which was purified S-4 and S-5, which were then separated by chromatography in on paper by electrophoresis at pH 3.6. Its composition, as well Solvent I (3). The second, third, and fourth spots were S-2, as sequence, is shown in Table IV. Fraction I contains both S-8, and S-7, respectively. Their compositions as well as some Peptide Th-1 and undegraded C-l as indicated by the partial of their sequences are also shown in Table IV. The overlaps residues of threonine, S-carboxymethylcysteine, serine, and of these subtilisin peptides along with the thermolysin peptide, tyrosine (Table IV).
Further separation of these two peptides Th-2, gave another 6 residues of the sequence. In order to   Fig. 3. confirm that the 4th residue from the NHz-terminus was histidine, Fraction I (0.60 pmole) of Fig. 5c was further hydrolyzed with subtilisin (50 I.cg, 8 hours; another 50 pg, 9 hours) at 40" in 0.5 ml of 0.06 M NH4HC03, pH 8.0. The reaction mixture was separated on a column of Sephadex G-50 as shown in Fig. 5d. Apparently Sephadex G-50 did not give any fractionation of the glycopeptides which contained different numbers of amino acid residues. Fraction I consisted of glycopeptides S-l and S-3 which were separated by paper electrophoresis at pH 3.6. Their compositions are shown in Table IV. Hydrolysis of Peptide S-l with carboxypeptidases -1 and B showed a Thr-Glu sequence at the COOH terminus of S-l. By difference, histidine was placed NHz-terminal to the Thr-Glu sequence.
The position of the glutamic acid was also confirmed by the composition of dipeptide S-6 which was isolated from Fraction II of Fig. 5d after paper electrophoresis at pH 3.6. Peptides S-2 and S-7 were also found in this fraction.
The complete amino acid sequence of crT(m)-4 was thus established.
Tryptic Peptides from Maleylated, Reduced, S-carboxymethyl T&Y-P-The separation on a column of Sephadex G-50 (fine) of the peptides from the tryptic hydrolysate of derivatized TSH-P is shown in Fig. 6. Details are given in the legend. Amino acid analyses showed that, except for Fractions III, VII, and VIII, the fractions did not require further purification.
Their compositions are given in Table V together with their assignments as /3T(m)-1, etc., which are based on their compositions.
Fraction III was further purified by paper electrophoresis at pH 6.4; two components were seen. Fraction 111-l was assigned as PT(m)-3b and the other had the same composition as Fraction I. Electrophoresis at the same pH revealed 9 ninhydrin-positive components to be present in Fraction VII but only three contained significant amounts of peptide material.
Their compositions are also given in Table V and show that the three can be designated as PT(m)-1 (Residues 1 through 9, Fig. 2), @T(m)-3a (Residues 14 through 18) which arose from a partial split of a Tyr-Cys(Cm) bond and PT(m)-4a (Residues 46 through 55) which originated from a partial split of a Tyr-Ala bond.
Peptide /3T(m)-1 was the main component of Fraction VI which gave no ninhydrin color. This is in agreement with subsequent data that showed /3T(m)-1 to be derived from the NH&terminal sequence of TSH-fi.
Its KHz-terminal group was blocked by the maleyl group; this leaves no free NH2 groups available for reaction with ninhydrin.
It is important to note that /3T(m)-1 contains 2 methionine residues which strengthens the previous preliminary assignments of the cyanogen bromide fragments @CNBr-1 and PCNBr-2) of TSH-fi (3).
Fraction VIII was found to contain four ninhydrin-positive components, three of them in significant amounts.
These were assigned as PT(m)-2, /3T(m)-5a (Residues 56 through 59) resulting from a partial split of a Tyr-Lys bond, and PT(m)-6 which also arises from a chymotryptic-like split at the Tyr-Phe bond between Residues 74 and 75. The basis of assignment was their compositions (Table V) and comparisons with the cyanogen bromide fragments of TSH-P and the tryptic peptides derived from them (3, Fig. 5). The summation of the numbers of residues in the primary peptides (Table V) again was in good agreement with the composition of the entire chain. As subsequently shown, the partial residue of methionine in Peptide @T(m)-7 is significant.
Peptide PT-4, the tryptic glycopeptide of this chain, must be at the COOH terminus because it contains the arginine, thus the order of these two peptides is established. This order was confirmed by examination of a chymotryptic hydrolysate of the cyanogen bromide fragment, PCNBr-3 as shown later in Table VII.
When added to PT(m)-3, the compositions of PT(m)-2, and /IT(m)-4 account for 46 residues, 10 through 55, and includes all 4 leucine residues of TSH-/3 and one of the three histidines.
The other two histidines are in PCNBr-4. Thus these three tryptic peptides of the maleylated derivative constitute the major portion of the cyanogen bromide fragment @XBr-3 (49 residues), and their order is subsequently shown by the overlaps of chymotryptic peptides from PCIC'Br-3. The composition of Peptide PT(m)-5 (Residues 56 through 69), which contains a methionine, shows that Peptide /3T-8 (the overlap tryptic peptide between the two large cganogen bromide fragments, PCNBr-3 and PCNBr-4) is connected to Peptide @T-9. The reason for placing /3T-9 at the COOH terminus of Peptide @T(m)-5 is because j?T-9 has an arginine residue at its COOH terminus and is in cyanogen bromide fragment PCNBr-4 which has a portion of /3T-8 at its NHS-terminal sequence. ,f3T(m)-5 also overlaps PCNBr-3 and ,&CNBr-4 and, based on the sequence of /3T-8 (Asp-Phe-Met'-Tyr-Lys (3)), the 14 residues of @T(m)-5 less Asp-Phe-hIet plus the compositions of @T(m)-6 and PT(m)-7 (a total of 55 residues) match the composition PCNBr-4 with exception of a serine, a tyrosine, and some methionine. These discrepancies are caused by heterogeneity at the COOH terminus of TSH-@ as described below.
The order of PT(m)-6 and PT(m)-7 was subsequently determined from the amino acid composition of a chymotryptic peptide from PCNBr-4. Ordering of Peptides within TSH-P-The data obtained, so far, from tryptic digests of reduced S-carboxymethyl TSH derivatives with or without maleylation, and from CNBr cleavage of reduced X-carboxymethyl TSH, are insufficient to order all the primary tryptic peptides of TSH-/3. The difficulty is that there are many chymotryptic-like cleavages (Met-His at Residues 9 and 10, Tyr-Cys(Cm) at Residues 18 and 19, Tyr-Phe at Residues 74 and 75) which occurred during tryptic hydrolysis of the derivatives.
The cyanogen bromide fragments are also relatively large; two of them @CNBr-3 and PCNBr-4) account for 101 residues. The position of PCNBr-1 and /KXBr-2 relative to PCNBr-3 was obtained through the hydrolysis of oxidized TSH-/3 with trypsin.
Data concerning the heterogeneity of the COOH terminus of TSH-P were also obtained from this hydrolysate. The ordering of tryptic peptides within the CNBr fragments was then studied by hydrolysis of ,ENBr-3 and PT-(m)-7 with chymotrypsin.
Peptide /W(oz)-l-- Fig.  7 shows the fractionation of the tryptic hydrolysate of TSH oxidized with performic acid, and the desired peptide, which should have the proposed Met-His bond intact (3, Fig. 5) because of conversion of the methionine to its sulfone, was found in Fraction III.
This fraction contained two peptides which gave positive Sakaguchi reactions (for arginine) ; they were separated by electrophoresis at pH 6.4. The less acidic peptide had the expected composition (Residues 1 through 13) as shown in Table VI. Table VI also shows the relationship of this peptide to CNBr fragments and tryptic peptides from TSH-fi with and without maleylation.
The order of primary tryptic peptides present in PT(ox)-1 is based on the specificity of trypsin and detection of KHz-terminal phenylalanine in intact /3T(ox)-1.
The summation of all alanine seen in the amino acid composition of PCNBr-3 was the isolated chymotryptic peptides, except Peptide C-5 which because of contamination by peptides rich in these four amino resulted from a partial split of Peptide C-4, equals that of the acids. No peptides that would not fit into the sequence have tryptic peptides present in BCNBr-3 (3 , Table VIII). This is been isolated from any chymotryptic and tryptic hydrolysates. further evidence that the excess lysine, serine, proline, and From Table VII Dansyl-Edman a Upon paper electrophoresis at pII 6.4, free phenylalanine and methionine appeared at the same position. Their presence was shown by analyzing the whole fraction, without acid hydrolysis, on an amino acid analyser.
tides overlap the tryptic peptides, except in one location (the showed no detectable NHz-terminal residue; in most cases, we Tyr-Cys(Cm) sequence, between PT-3 and PT-4 as well as C-l were not able to identify histidine as its dansyl derivative. and C-Z). The overlap at this pknt was obtained by the
It should be noted that this fragment has been established from the composition of PT(m)-5 (Table V)  and the position of /IT-10 by the isolation of an arginine-containing peptide from a chymotryptic digest of PCNBr-4. These data and the positioning of /3T-11, 12 and 14a (3 , Table III) from PT(m)-7 are described below.
Peptide PCNBr-4-C-2-In the separation of the cganogen bromide fragments of the entire hormone, PCNBr-4 cochromatographed with orCNBr-4 (3, Fig. 3) and the recoveries after their separation on paper were low. Accordingly, the entire fraction (3, Fraction V, Fig. 3 The resulting peptides were separated by electrophoresis at pH 6.4. Peptide PCNBr-4-C-2 was the only chymotryptic peptide to contain arginine. It should be noted that crCNBr-4 did not contain arginine. The peptide was, therefore, easily located by the Sakaguchi reaction. Staining with ninhydrin indicated that the peptide was overlapped by a second peptide which was removed by electrophoresis at pH 1.9. The amino acid composition of Peptide PCNBr-4-C-2 and its position in the sequence established the order of three tryptic peptides, PT(m)-5, PT(m)-6, and PT(m)-7, as shown in Table VIII.
The other chymotryptic peptides of the hydrolysate were difficult to fractionate, and the remaining order of the primary tryptic peptides was deduced from a chymotryptic hydrolysate of @T(m)-7, described below. Chymotryptic Peptides from Peptide pT(m)-7-Peptide PT(m)-7 (0.1 pmole) was hydrolyzed with chymotrypsin (50 pg) in 0.1 ml of 0.06 M NH4HC03, pH 8.0, at 40" for 2s hours. The reaction mixture was spotted on paper and the resulting peptides were separated by electrophoresis at pH 6.4. They migrated in order from the anode as C-2, C-4, C-3, and C-l plus C-5. Their compositions, which establish the order of the tryptic peptides, are shown in Table IX. COOH Terminus of TXH-B-A partial residue of methionine had been found in Peptide PT(m)-7 (Fraction I, Table V). This was shown to result from heterogeneity at the COOH terminus.
First, free methionine was obtained in low yield from the chymotryptic hydrolysate of /3T(m)-7. Second, both a dipeptide, Ser-Tyr @T(ox)-I3a) and a tripeptide Ser-Tyr-MetO @T(ox)-13) were found in the tryptic hydrolysate of oxidized TSH-/3. Peptide PT(os)-13 was isolated from Fraction VI, Fig. 7 by paper electrophoresis at pH 1.9 and Peptide PT(ox)-13a was t,he sole component of Fraction VII, Fig. 7. Their amino acid compositions together with their relationships to the COOH terminus of Peptide /3T(m)-7 are also given in Table IX.
These data agree with the results of the hydrolysis of reduced, X-carboxymethyl TSH-L?' with carboxypeptidase A, in which the amount of methionine released was much less than that of tyrosine (2 , Table II).
Only trace amounts of homoserine and its lactone were found in the analysis of PCNBr-4 (3) which indicates that most of this particular preparation did not terminate with methionine.
The The sequence of the NHz-terminal nonapeptide, which was not isolated (3), is known from the sequences of PCNBr-1 and PCNBr-2.
The sequence of the COOH-terminal tripeptide is shown in Table IX. The remaining primary tryptic peptides (PT.4 and PT-12) were determined by further hydrolysis with proteolytic enzymes.
Peptide pT(m)-SbGlycopeptide PT(m)-3b has the same composition as Peptide /3T4 and was used for completion of the sequence in this region. The sequence of 8 residues from the NHz terminus of Peptide PT-4 had been determined with the danysl-Edman and 2 residues from the COOH terminus with carboxypeptidases A and B (3 , Table VI).
Four additional residues from the NH2 terminus were determined with Peptide PT(m)-3b as shown in Table X. To confirm the sequence, 0.1 pmole of @T(m)-3b was hydrolyzed with thermolysin (100 pg) in 0.1 ml of 0.126 in NH4HC03, pH 8.0, at 40" for 2 hours. The reaction mixture was spotted on paper and the resulting peptides were separated by electrophoresis at pH 6.5. They migrated from the anode as Th-2, Th-1, and Th-3.
Their yields and compositions are shown in Table X. The amino acid compositions of Th-3 and /3T-4a gave the COOH-terminal sequence while the results of the dansyl-Edman of Th-2 confirmed the results obtained with parent peptide (Table X).
The order of C-l and C-2 is established because of the NHz-terminal X-carboxymethylcysteine in intact PT-12. Assignment of Amide Groups-Most of the amides in TSH-a and TSH-/3 were assigned based on data obtained by enzymatic hydrolysis of peptides to free amino acids. The results are summarized in Table XII. Five amides in each chain have not been assigned. The peptides containing these unassigned residues were resistant to hydrolysis by aminopeptidase 11.

DISCUSSION
The data in this investigation, in which TSH-a: and TSH-6 were studied separately, confirm the overlaps established for TSH-a: in the preceding paper (3). The ease of separation and high yields of peptides from the maleylated, reduced, S-carboxymethyl chains, from oxidized TSH-cr and, as shown in the following paper (9), from reduced, X-carboxymethyl CI chain of LH emphasizes the usefulness of gel filtration as the initial step in fractionation.
In retrospect, a more efficient approach would have been to carry out all studies on the separated chains. It should be noted that while limited amounts of material and the lack of recognition of the subunit nature of TSH until recently (1, 2) have necessitated combining data from studies on fragments from the undissociated hormone and the separate chains, no peptides have been isolated whose composition or sequence is incompatible with the structures given in Figs. 1 and 2. The distribution of methionine and arginine residues in both chains together with the relatively low numbers of such residues as leucine, isoleucine, histidine, and glycine greatly facilitated ordering of small tryptic peptides within larger fragments, in many cases simply by determination of the latter's composition. The most difficult portion of the sequence was the region in TSH-a! (Residues 83 through 95) adjacent to the carbohydrate unit at Residue 82. Two of the five tyrosines, three of the halfcysteines, and all three of the histidines of TSH-(I! are in this region. Much of it was relatively resistant to proteolysis and the histidine at Residue 83 could not be assigned unequivocally on the basis of the Edman degradation.
The major problem encountered with the dansyl-Edman was with histidine. In most instances, when this residue was exposed as a new NH2 terminus by the Edman degradation, the dansyl histidine then formed may have been partially destroyed during hydrolysis and was not detected after subsequent thin layer chromatography. After the next step of degradation, the resulting NHz-terminal residue was always detected without ambiguity.
Other technical problems encountered were the apparently complete cleavage by trypsin of the Met-His bond at positions 9 to 10 of TSH-fi and the heterogeneity of the COOH terminus of the same chain. Both were overcome by isolation of the appropriate peptides from a tryptic hydrolysate of oxidized TSH-fi which established both the Met-Met sequence and showed that some molecules must terminate with methionine rather than tyrosine.
The dipeptide, Ser-Tyr, was the only original tryptic peptide (3) not recognized to be in any of the cyanogen bromide fragments; it should have been obtained from the tryptic hydrolysate of /?CNBr-4 (3 , Table IX).
Only 2.5 pmoles of TSH were used in the cyanogen bromide work and there was some uncertainty in the composition of this large fragment, although it was obtained in excellent yield. explains the partial residue of methionine released by treatment of either intact TSH or TSH-/3 with carboxypeptidases (2, 10). The data do not preclude the possibility that some /3 chain molecules terminate at lysine 110. The NHz-terminal and the COOH-terminal heterogeneity of TSH-a and -6. respectively is probably caused by proteolysis (see Reference 3), and it is likely that, in the structure of the intact hormone, these termini are exposed to the environment.
The hormone specific chain of LH, CII, also shows heterogeneity at its COOH terminus (II). In both TSH-a and -/3, partially blocked NH*-terminal residues of phenylalanine may also be present (2). The approximate molecular weights of the two chains, calculated from the sequences and carbohydrate compositions, are 13,600 (TSH-(Y) and 14,700 (TSH-/?).
Approximation is necessary because of the terminal heterogeneities of the protein portions, the oligosaccharide portions, which contain partial residues of fucose and galactose and incomplete assignment of amide groups.
Completion of the latter, determination of the disulfide bonds and of the structure of the oligosaccharides remain. It is likely that unequivocal assignments of all amides will be difKcult because the polymorphism of TSH seen in gel electrophoresis may result from partial amidation at some positions. A few peptides with the same composition but different electrophoretic mobilities were isolated although lack of material has not yet permitted detailed investigation of all such peptides (3). These include the two tryptic glycopeptides of TSH-a which each migrate as two spots during electrophoresis (5). With the exception of heterogeneity at the chain termini, no other evidence for amino acid substitutions, which would be a reflection of allotypes such as found in carboxypeptidase A (12,13) and bovine /3-lactoglobulin (14), has been found thus far. With respect to the oligosaccharides, it is not known whether the heterogeneity at their chain termini, clearly present because of the partial residues of fucose and galactose, extends further into their structures.
All the galactose was found to be in one oligosaccharide (5), that assigned to Residue 82 in TSH-(Y.
TSH and LH are sulfur-rich proteins; the chains of TSH contain 5 and 6 disutide bonds, respectively, in molecules whose amino acid portions have molecular weights of about 10,800 and 13,100 and they must be extremely highly cross-linked.
There is no evidence for free sulfhydryl groups.
In TSH-(r, two pairs of half-cystine residues occupy adjacent positions, one other is separated only by a residue of histidine.
The half-cystines of TSH$ are more scattered.
Other points of interest in the sequences are that the same type of linkage, Asn-X-Thr is present as found in many other glycoproteins (compilations have recently been published (15,16)), if one assumes that the TSH oligosaccharides are attached to the expected asparagine residues, 56 and 82 in TSH-cu, and 23 in TSH-0.
Indirect supporting evidence has been found for two of these positions (5). In the third case, position 82 in TSH-cu, the data of Fig. 5 and Table IV show that the linkage cannot be to serine 89 and amino acid analyses of the material not extract,ed by organic solvents after three steps of the Edman degradation of (YT-10