Process research and development of L-alanyl-L-glutamine, a component of parenteral nutrition

Article abstract of DOI:10.1021/op990079w

A large-scale manufacturing method of L-alanyl-L-glutamine used for a component of parenteral nutrition has been studied. The method consisted of a reaction of D-2-chloro- or D-2-bromopropionic acid with thionyl chloride and Schotten-Baumann reaction with L-glutamine followed by ammonolysis reaction. The intermediate D-2-chloropropionyl-L-glutamine was found to be more stable than its bromo analogue. In the ammonolysis reaction, the former intermediate needed a higher reaction temperature, but the by-products produced had little effect on the quality of the final product. The structures of the by-products were conjectured mainly by mass spectrometry and they were removed by anion resin treatment and recrystallization.

Full text of DOI:10.1021/op990079w

Organic Process Research & Development 2000, 4, 147−152
Process Research and Development of L-Alanyl-L-glutamine, a Component of
Parenteral Nutrition
Takahiro Sano,§ Toru Sugaya,*,§ Kunimi Inoue,§ Sho-ichi Mizutaki,§ Yasuyuki Ono,† and Masaji Kasai§
Sakai Research Laboratories, Kyowa Hakko Kogyo Co., Ltd., 1-1-53 Takasu-cho, Sakai, Osaka 590-8554, Japan, and
Bio-Chemicals DiVision, Kyowa Hakko Kogyo Co., Ltd., 6-1 Ohtemachi, 1-Chome, Chiyoda-ku, Tokyo 100-8185, Japan
Abstract:
the deprotected product with ammonia,5 and the method
which reacts an active ester of Cbz-Ala with nonprotected
L-glutamine and removes the protecting group from the
intermediate compound;6 (2) the method of producing AlaGln
via N-carboxyl anhydride;7 and (3) the method using D-2-
bromopropionyl chloride as a starting compound via an
intermediate compound, D-2-bromopropionyl-L-glutamine.8
The methods (1) using a protecting group need the step
of removing the protecting group from the intermediate
compound, and the operation for the step is complicated.
Therefore, the methods (1) yield AlaGln at higher cost. The
method (2) uses N-carboxyl anhydride of L-alanine without
involving a protecting group. However, significant by-
products are produced, for example, tripeptides, and the yield
of the intended product is low. In addition, it is difficult to
purify the intended product. In the method (3), since an acid
chloride that has a high reactivity towards water is added to
an aqueous solution of L-glutamine, for example, in the
reaction of D-2-bromopropionyl chloride,8 the method suffers
from partial hydrolysis of the acid chloride. Therefore, the
method (3) yields by-products, and the yield of the intended
product is low. Since the produced D-2-bromopropionyl-L-
glutamine is purified by extraction with organic solvent, the
yield of the product is low. In addition, in the method (3),
since the ammonolysis of D-2-bromopropionyl-L-glutamine
is carried out at an elevated temperature, by-products are
considerably enhanced, and the optical purity of the produced
AlaGln is often low.
A large-scale manufacturing method of L-alanyl-L-glutamine
used for a component of parenteral nutrition has been studied.
The method consisted of a reaction of -2-chloro- or -2-
bromopropionic acid with thionyl chloride and Schotten-
Baumann reaction with -glutamine followed by ammonolysis
reaction. The intermediate -2-chloropropionyl- -glutamine was
found to be more stable than its bromo analogue. In the
ammonolysis reaction, the former intermediate needed a higher
reaction temperature, but the by-products produced had little
effect on the quality of the final product. The structures of the
by-products were conjectured mainly by mass spectrometry and
they were removed by anion resin treatment and recrystalli-
zation.
D
D
L
D
L
Introduction
Recent studies have suggested important metabolic roles
for L-glutamine and required possible sources of L-glutamine
for parenteral nutrition.1 However, L-glutamine is not in-
cluded in currently available amino acid solutions.2 The
solubility of L-glutamine in water is decreased, and its
stability in water is also low. In addition, L-glutamine may
decompose into toxic products, pyroglutamic acid and
ammonia.1 Therefore, peptides containing L-glutamine have
taken the place of L-glutamine. L-Alanyl-L-glutamine (1;
AlaGln) has higher solubility and higher stability than
L-glutamine and is used as the infusion component.1,2
For the production of AlaGln, there are several known
methods, and the well-known methods are as follows: (1)
methods of using a protecting group, for example, the method
which comprises condensing an N-benzyloxycarbonyl-L-
alanine (Cbz-Ala) with protected L-glutamine in the presence
of N,N′-dicyclohexylcarbodiimide (DCC) and removing the
protecting group from the intermediate,3,4 the method which
comprises condensing Cbz-Ala with protected γ-methyl
L-glutamate in the presence of DCC, removing the protecting
group from the intermediate compound, and further reacting
Results and Discussion
Among the methods described above for manufacturing
AlaGln, the authors selected the method via D-2-halopro-
pionyl-L-glutamine as the best for a large-scale production
from the viewpoint of cost and procedures and started the
process development study (Scheme 1).
In the previous papers,8 the reaction of D-2-bromopro-
pionyl chloride, which was prepared from D-2-bromopropi-
onic acid (2a) and thionyl chloride and used without isolation,
with L-glutamine was carried out in water. There were
problems associated with the decomposition of the acid
chloride as well as the difficulty of the product isolation.
* Corresponding author. Telephone: +81 (722) 23-5545. Fax: +81 (722)
27-7214. E-mail: toru.sugaya@kyowa.co.jp.
§ Sakai Research Laboratories.
† Bio-Chemicals Division.
(1) Abumrad, N. N.; Morse, E. L.; Lochs, H.; Williams, P. E.; Adibi, S. A.
Am. J. Pysiol. 1989, 257, E228 and references therein.
(2) Yoshida, S.; Shirouzu, Y.; Yoshizumi, T.; Ishibashi, N.; Noake, T.; Kaibara,
A.; Yamasaki, K.; Shirouzu, K. JJPEN 1996, 18, 381.
(3) Akabori, S.; Sakakibara, S.; Shimonishi, Y. Bull. Chem. Soc. Jpn. 1961,
34, 739.
(5) Shimonishi, Y. Bull. Chem. Soc. Jpn. 1964, 37, 200.
(6) Katoh, T.; Kurauchi, M. Eur. Pat. 311057, 1989; Chem. Abstr. 1989, 111,
97736.
(7) Moeller, H.; Wallat, S.; Hoeffkes, H.; Giede, K. German Pat. 3201511,
1983; Chem. Abstr. 1983, 99, 181474.
(8) Thierfelder, H.; Cramm, E. Hoppe-Seyler’s Z. Phys. Chem. 1919, 105, 58,
and reviews: Greenstein, J. P.; Winitz, M. Chemistry of the Amino Acid;
John Wiley & Sons: New York, 1961; Vol. 1.
(4) Akabori, S.; Sakakibara, S.; Shimonishi, Y. Bull. Chem. Soc. Jpn. 1962,
35, 1966.
10.1021/op990079w CCC: $19.00 © 2000 American Chemical Society and The Royal Society of Chemistry
Published on Web 03/01/2000
Vol. 4, No. 3, 2000 / Organic Process Research & Development
•
147

Scheme 1
Table 1. Production of 3a and 3ba
substrateb
(scale: g)
yieldc
(%)
purityd
(%)
diatereomeric
excess (% de)
product
Figure 1. Stability of 3a and 3b: diastereomeric excess of 3a
and 3b at pH 10. Temperature, 50 °C and 30 °C. Key to
symbols: 4, 3a, 50 °C; 0, 3a, 30 °C; 2, 3b, 50 °C; 9, 3b, 30 °C.
2a (11.5 g)
2a (880 g)
2b (600 g)
3a
3a
3b
72.4
70.0
81.7
98.1
97.5
99.8
97.1
96.6
99.7
a Reactions were carried out via acid chloride, see Experimental Section.
b Optical purity 2a: 87.4 % ee, 2b: 98.8 % ee. c Isolated yields from 2a and
2b. d HPLC area % of the isolated 3a and 3b.
These problems seemed to be overcome under the Schotten-
Baumann procedures. The reaction of D-2-bromopropionyl
chloride, prepared from 2a, and L-glutamine proceeded fast
in a two-phase condition of toluene-water with alkali. The
produced D-2-bromopropionyl-L-glutamine (3a) crystallized
from the water layer after acidifying. However, on a large
scale, the produced D-2-bromopropionyl-L-glutamine (3a)
showed lower diastereomeric excess9 and contained a larger
amount of impurities than on a small scale (Table 1),
presumably due to the instability of D-2-bromopropionyl-L-
glutamine (3a). Therefore, the authors modified the method
of using D-2-bromopropionyl-L-glutamine (3a) to its chloro
analogue, D-2-chloropropionyl-L-glutamine (3b), because the
chloro substituent was more stable than the bromo substitu-
ent. The starting material, D-2-chloropropionic acid (2b), is
obtained by a method similar to the bromo compound, that
is, synthesis from D-alanine,10 L-lactate,11 or enzymatic
resolution of DL-chloropropionic acid or its ester.12
Figure 2. Stability of 3a and 3b: diastereomeric excess of 3a
and 3b at pH 6. Temperature, 50 °C and 30 °C. Key to symbols:
4, 3a, 50 °C; 0, 3a, 30 °C; 2, 3b, 50 °C; 9, 3b, 30 °C.
The stability of bromo and chloro intermediates 3a,b in
each condition of pH 10 (reaction), pH 6 (neutralization),
pH 2 (crystallization) is shown in Figures 1-5. The
diastereomeric excess of D-2-bromopropionyl-L-glutamine
(3a) remains almost unchanged at pH 10 and 6 at 30 °C;
yet, at 50 °C it decreased at each pH (Figures 1 and 2). The
content, which was measured by HPLC method, decreased
at pH 10 and 6 at 50 °C (Figure 4). At pH 2, the
Figure 3. Stability of 3a and 3b: diastereomeric excess of 3a
and 3b at pH 2. Temperature, 50 °C and 30 °C. Key to symbols:
4, 3a, 50 °C; 0, 3a, 30 °C; 2, 3b, 50 °C; 9, 3b, 30 °C.
diastereomeric excess of D-2-bromopropionyl-L-glutamine
(3a) did not change even at 50 °C (Figure 3), and the content
slightly decreased (Figure 4). The stability at pH 2 was
probably due to insolubility of the crystals.13 These results
were also observed when using the real reaction solution
adjusted to each pH (data are not shown). On the other hand,
D-2-chloropropionyl-L-glutamine (3b) did not show any
decrease both of the diastereomeric excess (Figures 1-3)
and the content (Figure 5) at each pH and temperature.
Therefore, in the manufacture of AlaGln, the use of D-2-
(9) Pure L-isomer of glutamine was used in the reaction, and at this reaction
step the racemization to D-glutamine or the amount of intermediates
containing D-glutamine (D-2-bromopropionyl-D-glutamine, L-2-bromopro-
pionyl-D-glutamine) were not checked. At the final step, D-glutamine-
containing peptides (L-alanlyl-D-glutamine and D-alanyl-D-glutamine) were
monitored by HPLC; however, they were not detected either in the reaction
mixture or in the final product.18 Therefore, the optical purity was represented
by diastereomeric excess % (% de) of D-2-bromopropionyl-L-glutamine to
L-2-bromopropionyl-L-glutamine. In the production of D-2-chloropropionyl-
L-glutamine, the same representation on the optical purity is also used.
(10) Fu, S.-C. J.; Birnbaum, S. M.; Greenstein, J. P. J. Am. Chem. Soc. 1954,
76, 6054.
(11) Gerrard, W.; Richmond, M. J. J. Chem. Soc. 1945, 853.
(12) Kircher, G.; Scollar, M. P.; Kribanov, A. M. J. Am. Chem. Soc. 1985, 107,
7072.
(13) In these experiments, at pH 2 the substrates were not soluble in water, and
the data were obtained by using the undissolved heterogeneous solution.
148
•
Vol. 4, No. 3, 2000 / Organic Process Research & Development

Table 2. Ammonolysis of 3a and 3ba
AlaGln
substrate
conditions
diatereomeric
diatereomeric
excess
(% de)
temp time yieldb purityc
excess
(% de)
(°C)
(h)
(%)
(%)
3a (97.2)
3a (96.6)
3b (96.9)
3b (96.9)
20
40
50
60
20
6
15
9
78.1
78.7
71.2
72.9
99.0
99.2
97.5
98.1
98.9
98.6
97.4
97.5
a Other reaction conditions: 28% aq. NH3 38 mol. equiv to 3a and 3b; 40-
100 g of subatrates were used. b Isolated yields. c HPLC area % of the isolated
AlaGln; before purification.
Figure 4. Stability of 3a: content of 3a at 30 °C and 50 °C.
pH, 2, 6 and 10. Key to symbols: 0, 30 °C, pH 2; 4, 30 °C, pH
6; ], 30 °C, pH 10; 9, 50 °C, pH 2; 2, 50 °C, pH 6; [, 50 °C,
pH 10.
Figure 5. Stability of 3b: content of 3b at 50 °C. pH, 2, 6 and
10. Key to symbols: 0, pH 2; 4, pH 6; ], pH 10.
Figure 6. Structures of by-products.
reaction temperature (60 °C); however, the content of AlaGln
decreased, and several decomposed products appeared. The
chemical structures of the two major decomposition products
were inferred from LC-MS analysis as L-alanyl-L-glutamic
acid (4) and a diketopiperazine 5 derived from AlaGln,
respectively (Figure 6). The former was identified by HPLC
analysis with the purchased standard sample. The diketo-
piperazine formation by the thermal decomposition of the
dipeptide was well-known.14 Therefore, for confirming the
structure of the compound 5 AlaGln was heated in the
differential scanning calorimeter (DSC) apparatus.15 The
thermally decomposed product by being heated until 230 °C
was obtained by washing the DSC cell with water and the
thus obtained compound was correlated with the diketopi-
perazine 5 by HPLC analysis.
chloropropionyl-L-glutamine (3b) was expected to avoid the
fatal decomposition in the intermediate stage. Fortunately,
the starting material D-2-chloropropionic acid (2b) was
obtained with higher optical purity than its bromo analogue
2a (Table 1). Although in the crystallization step the
undesired diastereomers remained in the mother liquor, the
yield and diastereomeric excess of D-2-halopropionyl-L-
glutamine 3a,b depended on the optical purity of the starting
acids 2a,b. The result of the production of D-2-chloropro-
pionyl-L-glutamine (3b) on a large scale is shown in Table
1.
In the next step, the ammonolysis reaction, the reactivity
of D-2-bromopropionyl-L-glutamine (3a) and its chloro
analogue 3b were compared. The reaction using D-2-
bromopropionyl-L-glutamine (3a) and aqueous ammonia was
reported to proceed at an elevated temperature of about 80
°C.8 However, it was found that the reaction progressed fast
at 40 °C and slowly even at room temperature. The reaction
of D-2-chloropropionyl-L-glutamine (3b) needed a higher
temperature (60 °C), and the diastereomeric excess of the
obtained AlaGln was slightly lower than that of the case of
using D-2-bromopropionyl-L-glutamine (3a) (Table 2).
On the other hand, impurities in the reaction mixture with
D-2-chloropropionyl-L-glutamine (3b) were greater, presum-
ably due to the higher reaction temperature. Indeed, in the
test of stability with the isolated AlaGln, the diastereomeric
excess of AlaGln in aqueous ammonia did not change at the
In the ammonolysis reaction of D-2-chloropropionyl-L-
glutamine (3b), another by-product was observed, which was
almost undetected in the reaction of D-2-bromopropionyl-L-
glutamine (3a). This compound was isolated by HPLC
separation, and the chemical structure was proposed to be a
hydrolyzed compound 6 from the mass spectrum. The
(14) Kopple, K. D.; Ghazarian, H. G. J. Org. Chem. 1968, 33, 862.
(15) Thermogravimetry and differential thermal analysis of AlaGln showed an
endothermic peak at 229 °C with loss of 1 mol of water and then gradual
decomposition with loss of weight. (TG-DTA conditions; MAC Science
TG-DTA 2000, sample amount: ∼5 mg, heating rate: 10 °C/min) The
endothermic peak was supposed to be the formation of the diketopiperazine.
The same thermal analysis pattern was observed with L-alanyl-L-leucine
(endothermic peak at 215 °C) and glycyl-L-glutamine (at 216 °C).
Vol. 4, No. 3, 2000 / Organic Process Research & Development
•
149

Scheme 2
Figure 7. Relation between the amount of ammonia and
production of AlaGln: Reaction temperature, 60 °C. Key to
symbols: 0, yield; 4, content; ], diastereomeric excess.
structure and the stereochemistry were confirmed by com-
paring with the synthetic product from L-2-acetoxypropionic
acid. While the intermediate was not detected, it was
suspected that the hydroxy compound 6 was produced by
the intramolecular substitution reaction of the carboxylate
anion affording a cyclization intermediate and followed by
ammonolysis. Therefore, it finally retained the inverse
stereochemistry at the OH-substituted position (Scheme 2).
In the ammonolysis reaction, the method using D-2-
bromopropionyl-L-glutamine (3a) seems superior than the
chloro analogue regarding the yield and gentle reaction
condition. However, from the viewpoint of the ease of
obtaining the precursor compound mentioned in the previous
synthetic step and of cost performance,16 the authors selected
the method of using D-2-chloropropionyl-L-glutamine (3b)
as the more profitable method for manufacturing AlaGln and
optimized the ammonolysis procedures. Typically, the reac-
tion was carried out under heating D-2-chloropropionyl-L-
glutamine (3b) in 28% aqueous ammonia of 38 equiv based
on D-2-chloropropionyl-L-glutamine (3b) in a pressure vessel
and continued until the amount of D-2-chloropropionyl-L-
glutamine (3b) decreased to 3% with HPLC monitoring.
First, the reaction temperature was surveyed. The reaction
at 60 °C finished in 9 h, and at 50 °C it took 15 h. The
produced AlaGln was isolated after concentration of the
reaction mixture followed by acidification and crystallization
from aqueous methanol. The diastereomeric excess and the
amount of by-products in the reaction mixture were almost
the same at 60 °C and 50 °C, and the results of the reaction
were shown in Table 2. In the reaction at 60 °C, the amount
of ammonia was decreased gradually from 30 equiv to 5
equiv. The relation between the amount of ammonia and the
yield and the diastereomeric excess of the isolated AlaGln
are indicated in Figure 7.
Figure 8. Relation between the amount of ammonia and by-
products 4, 5, 6, and 7 in the reaction mixture. Reaction
temperature, 60 °C. Key to symbols: 0, 4; 4, 5; ], 6; O, 7.
low (40%). The amounts of impurities in the reaction mixture
increased as the amount of ammonia decreased (Figure 8).
Two by-products significantly increased as the quantity of
ammonia decreased. One of these by-products had already
been identified as compound 5. The other was inferred from
LC-MS to be a dimeric product 717(Figure 6). The quantity
of two further by-products, 4 and 6, hardly changed. The
compound 4 was observed at no more than 0.4% (HPLC
area) even in the reaction with 5 equiv of ammonia.
While most of the by-products produced in the reaction
were removed by crystallization, a small portion of dimeric
compound 7 as well as another by-product 4 still remained
in the crystals. Those compounds have an acidic character;
therefore, a process of passing through an anion resin was
adapted as a purification step combined with recrystallization.
Although the kind of the anion resin was not surveyed, a
common anion resin, WA-30 (Mitsubishi Chemical, Co),
gave a good result (Table 3; details are described in
Experimental Section). As a final step, an activated charcoal
treatment of the eluent through the resin after being neutral-
ized and a crystallization from ethanol were adopted from
the consideration of decolorization and removal of undesir-
While the reaction with 5 equiv of ammonia was slow
and did not finish within 15 h, the amount of ammonia was
found to be adequate with 10 equiv. The diastereomeric
excess of the isolated AlaGln was unchanged in each
experiment even at 5 equiv of ammonia. In that case, the
unreacted D-2-chloropropionyl-L-glutamine (3b) was re-
moved in the crystallization step, and the quality of the
isolated AlaGln was excellent; however, the yield was too
(17) To support the suspected structure of the dimeric compound 7, the following
confirmation was carried out: when AlaGln and D-2-chloropropionyl-L-
glutamine (3b) was heated with sodium carbonate in water at 100 °C, the
product had the same retention time of HPLC analysis and showed the same
MS spectrum with the compound 7. Therefore, the compound 7 was
supposed to be derived from the reaction of the produced AlaGln and another
molecular of 3b under ammonia-deficient reaction conditions.
(16) When D-2-bromo and D-2-chloropropionic acid (2a,b) are prepared from
L-lactate in the well-known method, the cost depends on SOBr2 and SOCl2.
The former is more expensive than the latter. (cf. SOBr2: ¥17,000/100 g,
SOCl2: ¥14,900/100 mL (163 g) from the Aldrich catalog, 1998-1999,
Japan).
150
•
Vol. 4, No. 3, 2000 / Organic Process Research & Development

Table 3. Purification of AlaGln
UV 210 nm; retention times, 1: 15.0 min, isomer (D-alanyl-
L-glutamine): 18.4 min, 4: 26.3 min, 5: 6.0 min, 6: 4.0
min, 7: 34.0 min.
HPLC area (%)b
1
4
5
6
7
isomer
LC-MS analyses: HPLC conditions: column, Lichrosorb-
NH2 (GL Science); eluent, 0.05 mol/L AcONH4 and MeCN
(35:65), detection UV 210 nm, retention times, 4: 44.7 min,
5: 3.3 min, 7: 51.0 min. MS conditions: ion mode, ESI
positive; scan range, m/z 30-1000; source temp. 120 °C.
Found (m/z) L-alanyl-L-glutamic acid (4) C8H14N2O5: 219
(M + H)+, (2S,5S)-2-(2-carbamoylethyl)-5-methylpiperazine-
3,6-dione (5) C8H13N3O3: 200 (M + H)+, (2S,4S)-3-aza-
2,4-dimethyl-1,5-pentanedioyl-di-L-glutamine (7) C16H27-
N5O8: 418 (M + H)+.
before purification 97.73 0.12 0.02 0.06 0.12 1.28
resin eluenta
recrystall.
98.53 N.D. 0.01 0.05 N.D. 1.32
99.80 N.D. N.D. N.D. N.D. 0.19
a Collected fractionsts of resin treatment, see Experimental Section. b N.D.:
Not Detected (less than 0.01 %).
able biological contamination such as endotoxins. The use
of ethanol was also adopted from a wider allowance of the
residual solvent than methanol.
Synthesis of D-2-Chloropropionyl-L-glutamine (3b). To
D-2-chloropropionic acid (600 g, 5.53 mol; 98.8% ee), was
added SOCl2 (724 g, 6.08 mol) at 65 °C (Caution: the
scrubber should be completely prepared for remoVing
poisonous gas!) and heated further for 1 h at 85 °C. The
reaction mixture was cooled to room temperature to give
745 g of the oily product. The mixture (722 g) was diluted
with toluene (400 mL) to make a toluene solution of D-2-
chloropropionyl chloride (5.36 mol). A mixture of L-
glutamine (784 g, 5.36 mol), H2O (300 mL), and toluene
(150 mL) was cooled to 0-5 °C and stirred, and then to the
cooled solution was added 5 mol/L aqueous NaOH (1000
mL). To the solution, were added concurrently a toluene
solution of D-2-chloropropionyl chloride described above and
5 mol/L aqueous NaOH below 10 °C while maintaining the
pH of the reaction solution at pH 10. The reaction mixture
was stirred further 1 h at 10 °C. To the mixture, was added
concentrated HCl (30 mL) at room temperature to adjust the
pH to 6. Then, toluene was removed by extraction. To the
water layer, was added concentrated HCl (410 mL) at room
temperature to adjust the pH to 2. The precipitated crystals
were filtered and dried in vacuo to afford 3b: 1038 g (4.38
mol; diastereomeric excess 99.7% de; yield 81.7% from D-2-
chloropropionic acid); mp 153 °C dec (recrystallized twice
from H2O); [R]20D -9.9 ° (c ) 5, H2O); 1H NMR (DMSO-
d6) δ ) 1.54 (3H, d, J ) 6.6 Hz, CH3), 1.70-2.10 (2H, m,
CH2CH2CONH2), 2.14 (2H, t, J ) 7.1 Hz, CH2CH2CONH2),
4.13-4.23 (1H, m, CHNH-), 4.59 (1H, q, J ) 6.7 Hz,
CHCl), 6.82 (1H, s, CONH), 7.37 (1H, s, CONH), 8.60 (1H,
d, J ) 7.7 Hz, -NH-); IR (KBr) υ ) 1738, 1662 cm-1;
HRESIMS calcd for C8H1335ClN2O4 m/z 235.0486 (M - H)+,
found 235.0494.
Summary
Process research and development of AlaGln have been
performed for large-scale manufacture. From the viewpoint
of the stability and the cost, the method via D-2-chloropro-
pionyl-L-glutamine (3b) was adopted. In the following
ammonolysis reaction, the structures of the by-products were
inferred mainly from mass spectrometry, and a method
removing these compounds by passing through the anion
resin was developed. From the studies described here, a
manufacturing method of AlaGln used for a component of
parenteral nutrition has been accomplished.
Experimental Section
1
General. H NMR spectra were recorded at 300 MHz
on a Bruker AC-300 spectrometer, and signals are given in
ppm using TMS as an internal standard. IR spectra were
recorded on a Shimadzu FTIR-4300 spectrophotometer.
Optical rotation was measured with Jasco P-1020 polarim-
eter. HRMS were recorded on a Micromass LCT or a JEOL
LMS SX-102 mass spectrometer. LC-MS spectra were
recorded on Micromass Qauatro mass spectrometer. D-2-
Bromopropionic acid (2a; (R)-2-bromopropionic acid) was
purchased from Osaka Synthetic Chemical Laboratories Inc.
and D-2-chloropropionic acid (2b; (R)-2-chloropropionic
acid) was purchased from Nippon Fine Chemical Co., Ltd.
For HPLC analysis of the diastereomeric excess,9 L-alanyl-
D-glutamine and D-alanyl-L-glutamine were purchased from
Bachem AG, and D-alanyl-D-glutamine was synthesized.8
Other reagents and solvents are of commercial quality.
HPLC analyses: Purity and optical purity of 3a: column,
YMC-Pack ODS-AQ313 (YMC), eluent: 0.01 mol/L KH2-
PO4; detection, UV 210 nm; retention times, 3a: 31.0 min,
isomer (L-2-bromopropionyl-L-glutamine): 37.5 min. Purity
of 3b: column, Shim-Pack CLC-ODS (Shimadzu); eluent,
aqueous solution of 0.01 mol/L KH2PO4 and 0.01 mol/L
sodium octanesulfonic acid adjusted to pH 2.5 by H3PO4:
MeOH (100:1); detection, UV 210 nm; retention time, 3b:
6.3 min. Optical purity of 3b: column, YMC-Pack ODS-
AQ313; eluent 0.01 mol/L KH2PO4; retention times, 3b: 25.1
min, isomer (L-2-chloropropionyl-L-glutamine): 27.9 min.
Purity and diastereomeric excess of AlaGln(1):18 column,
TSK gel ODS-120 T (Tosoh); eluent, aqueous solution of
0.01 mol/L KH2PO4 and 0.01 mol/L sodium octanesulfonic
acid adjusted to pH 2.5 by H3PO4:MeOH (100:1); detection,
Synthesis of L-Alanyl-L-glutamine (1). D-2-Chloropro-
pionyl-L-glutamine (3b) (60.0 g, 0.24 mol; 96.9% de) and
28% aqueous NH3 (600 mL) were put into a 1 L glass
autoclave with mechanical stirring at room temperature. The
resulting solution was allowed to stand at 60 °C for 8 h.
The reaction mixture was cooled to room temperature and
concentrated to ∼150 mL in vacuo, and then MeOH was
added dropwise to the residue at room temperature. The
(18) The optical purity of AlaGln(1) was calculated as diastereomeric excess %
(% de) of 1 to D-alanyl-L-glutamine from the HPLC analysis described in
the Experimental Section. Other isomers were confirmed not to be contained
in the final product by the following HLPC conditions: column, Crown-
Pack CR (+) and/or (-) (Daicel); eluent, 0.1% aqueous HClO4 (pH 2);
detection, UV 210 nm.
Vol. 4, No. 3, 2000 / Organic Process Research & Development
•
151

precipitated crystals were filtered and dried in vacuo to afford
35.4 g (diastereomeric excess 97.6% de; yield 69.0%) of a
crude product. A solution of the crude product (50.0 g) in
H2O (100 mL) was passed through an anion resin column
packed with WA-30 (50 mL), and H2O was charged
continuously (SV ) 1). Fractions containing AlaGln were
collected ( ∼250 mL) and concentrated in vacuo until 100
mL. The solution was treated with activated carbon (0.5 g)
and filtered. The filtrate was heated to 50 °C, and EtOH (50
mL) was added dropwise to the solution at 50 °C. The
solution was gradually cooled to room temperature, and the
precipitated crystals were taken out by filtration and dried
in vacuo to afford the purified L-alanyl-L-glutamine (1): 42.5
(30 mL), and the fractions containing the desired compound
were collected. The collected fractions were passed through
a cation resin column packed with SK-1B (30 mL), and the
fractions were combined and concentrated in vacuo to afford
a white solid of 6: 0.95 g (57.3%); [R]20D -7.9 ° (c ) 1.25,
1
H2O); H NMR (DMSO-d6) δ ) 1.28 (d, J ) 6.8 Hz,
3H,CH3), 1.80-2.00 (m, 2H, CH2CH2CONH2), 2.14 (t, J )
7.6 Hz, 2H, CH2CH2CONH2), 4.06 (q, J ) 6.6 Hz, 1H,
CHOH), 4.28-4.32 (m, 1H, CHNH-), 6.81 (s, 1H, CONH),
7.20 (s, 1H, CONH), 7.41 (s, 1H, CONH), 7.54 (s, 1H,
CONH), 7.77 (d, J ) 8.0 Hz, 1H, -NH-); IR (KBr) υ )
3400, 1665 cm-1. A portion of this solid was purified further
by HPLC separation for identification: HRFABMS calcd
for C8H15N3O4 m/z 218.1141 (M + H)+, found 218.1115.
g; mp 217 °C dec (lit.5 215-216 °C dec); [R]20 -3.4 ° (c
D
) 10, 1 mol/L HCl).
Synthesis of L-2-Hydroxypropionyl-L-glutamic acid
Diamide (6). To a solution of L-2-acetoxypropionic acid (1.0
g, 7.6 mmol) in THF (30 mL) was added 5 mL of THF
solution of DCC (1.6 g, 7.6 mmol) at 0 °C and then a solution
of L-glutamic acid diamide (1,1 g, 7.6 mmol) and Et3N (0.8
g, 7.6 mmol) in water (5 mL) was added to the cooled
solution. The mixture was stirred at 0 °C for 3 h, and the
precipitated white solid was filtered. To the filtrate was added
28% aqueous ammonia, and the mixture was stirred at 25
°C for 2 h, and then the reaction mixture was concentrated
to half volume in vacuo. The concentrated solution was
passed through an anion resin column packed with WA-30
Acknowledgment
The authors thank Dr. T. Yasuzawa and Mr. M. Taka-
shima in the Pharmaceutical Research Institute of Kyowa
Hakko Kogyo Co., Ltd. for LC-MS analyses, and we also
thank Messrs. Y. Mimura, T. Matsumura, K. Inoue, Y.
Nakaguchi, A. Sakaguchi, H. Shinmura, and Y. Yamada for
their collaboration.
Received for review September 3, 1999.
OP990079W
152
•
Vol. 4, No. 3, 2000 / Organic Process Research & Development