CA1100953A - PREPARATION OF 1-N¬.omega.-AMINO-.alpha.-HYDROXYALKANOYL| KANAMYCINS - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
An improved process for the preparation of 1-N-[L-(-)-.omega.-amino-.alpha.-hydroxyalkanoyl]kanamycins comprises acylating a polysilylated kanamycin in a substantially anhydrous organic solvent with an acylating derivative an .omega.-amino-.alpha.-hydroxyalkanoic acid.

Description

)9S3 ... ..
~hi~ inven~io~ sela~ ~ to an improved proces~ for ~he p~eparation of compounds o the ~or~nu~ a ~I;!N~2 ~~

~10 --~

o}~ o

2~n ~ ¦ 2 in which R is O~ or ~ and rl i~ an in~ger of f~o~ O ~o 2, OE a nontoxic pha~maceutically acceptable acid addit~ o~ sal~
th~!res:~f t which cumprise~ acylating polysilyla~ed ;cana3nycin P
or B or poly~ilylatPd kanamycin A or B con~ain~g a blockii~g ~r~up other than 5ilyl C~ ~he 6 ~-amln~ moiety, in a ~ubs~antially anhydrou~ orgarlic solv~nt~ with an acylating deri~a iv~ of an acid c~f ~he formula BoHN-C~I2- (CH2 ) n-f H-COOH I I
OH
in which n is an integer o~ from O to 2 and B is an amino-bl~ g group, arld ~ubse*~ently removin~ all bLocking g~oups.

J9~3 .. . . ..

The Xanamycins are well-known antibiotic~, having bean described, for example in th~ ~er~k Xndex, 8~h edi~ion, pp.
537-8, Numerous de~:ivative~ o~ the ~anamycins also ar~ )cnownO
~he stru~tural formulae of lcanamycin~ A and B a~e given l:~elow, along with th~ ~tandard t~usnber~ng syste3n used in the art.
He~einaEter, where readily under~tand~able, the va~ious kanamy ::in d~riva~iv~s will ~a referred to as derivatives of kanamycin A
or B rather than by ~tructural formula, so as to avoid the n~ es~i~y of compari~g comple~c struc:~-ares to de~ermine difference~.

6' .
4 ~ C~2~E~2 ~o~

no Kana~ycin A: R = OH
Kanamyc in 13: R ~ ~2 U~ S. Pat nt 3,781,26~ d1scloss and cla~ns l-~L-l-) yaamino a-hyd~oxybutyryl] }ca~amycin A ~amikacin) and B, as well as ~heir mo3~o-and di-carbobenzyloxy protected deri~ratiYesO For lower and higher homolc~g~ see U. S~ Patents 3,886,i3g and 3,904,597, The compound~
a~:e prepar~d g5~

by acyla~ing a 6'-~p~otected kanamycin A or B with an acylating derivativ~ of an N-protected L-(-) y-amino-a-hydr~xyb~tyric acid, in an aqueou3 ~edium, ollowed by removal of one or both N-pxotecting group3 .
U. S. Patant 3,974,137 discloses and claims a process ~or preparing l-~L~ r-ami~o-~-hydroxybutyxyl~kanamycin A which comp~ises reac~in~ 6'-carbobenzyLoxykana~ycin A with at least three moles of benzaldehyde, a substituted benzaldehyde or pivaldehyde, to produce 6'-N-carboben2yloxykanamycin A containing Schiff base moiatie~ on the 1,3 and 3n-positions, acylatin~ this ~e~ra-protected kanamycin A derivative with the N-hydroxy~uccinLmide e~ter of L~
y-benzyloxycar~onylamino~a~hydroxybutyri~ acid, and subsequen~ly removing the protecting groups.
Belgian Pat~nt 828192 discloses and cla~ms a pro~ess for -;
preparing l-~L-~o~-y ami~o-~-hydroxybutyryl]kanamycin A by prepar~
ation o~ the s~me te~ra~prot~cted kanamy~in ~ derivativa as in U. S. 3,974,137, acylating with the N-hydxoxy-5-norbornene-2 t 3-dicar~oxLmide ester o~ y benzyloxycarbonylamino-a-hydroxy-bu~yric acid, and subsequently removing the prot~cting groups.

The present invent~on provides an Lmproved ~nd commercially at ~acti~e proc~ 3 ~or the preparation of compound~ of formula I. The uce of a polysilylat~d ~an~myGin A or B a~ a starting material gi~e~ high s~lu~lity in the organic sol~e~t system9 thus p~rmitt~.ng reaction at high concentra~ions~ Although the reaction ~q u~ually conducted in ~lution containing about 10-20~ polysilylated ~anamy~in star~ing ~a~erial, excelle~t r~sults hava been obtained at concentrations o~ a~ou~ 50 gm~/lOO ml~ o~
solvent~ .

95~

As with pxior art processes, the present process gives a mixture of acylated Pxoducts- The desired l-N~acylated pro-duct is separated from the other products by chromatography and, if desired, the by~products may be hydrolyzed to the starting kanamycin for recycling. In prior art processes it was found that any 3"-N-acylated material which was produced caused a loss of about an equal amount of the desired l-N-acylated product, due to the great di~ficulty of separating the latter from the former. A particularly desirable feature of the present process is the extremely low amoun-t of undesirable 3"-N-acylated pro-duct which is produced (typically, none is detected).
In preparing l-[L~ -amino-~-hydroxybutyryl]kanamy-cin A, amikacin by various prior art procedures, there is typically also produced the 3"-N-acylated product (BB-Kll ), the

3-N-acylated product (BB-K29), the 6'-N-acylated product (BB-K6) and polyacylated material, as well as unreacted kanamycin A.
Thus, in commercial production of amikacin by acylation of 6'-N-carboben~yloxy kanamycin A in an aqueous medium, followed by removal of the protecting group, we f~und that about 10% of the . ~ .
20 desired amikacin (2. 5 kg. in a 25 kg. batch) usually was lost because of the presence of BB-Kl- as a co-product. When pre-paring amikacin by the present process, BB-Kll typically is not detected -~n the reaction mixture.
; The present invention provides the process for the ~-preparation of a l-N-[~-amino-~-hydroxyalkanoyl]kanamycin A or B having the formula , .~
.~

,..... . . ;

~ O :

R 1 / ~ ~ 2 H ~~ -7\\
HO ~ ~ OH '~ NH

\ /HO CH

NH2 ~ / (I 2)n ~
HO \ / , 2 :
I ~ NH2 ' ~' in which R is OH or NH2 and n is an integer of from 0 to 2, or a nontoxic pharmaceutically acceptable acid addition salt there~
~ of, which comprises acylating polysilylated kanamycin A or B or ; polysilylated kanamycin A or B containing a blocking group other than silyl on the 6'-amino moiety with an acylating derivativ~
of the acid of the formula ~NH~CH2~(CH2)n~l~-CH ~;
II OH

in which n is an integer of from 0 to 2 and B is an amino-block-~ 10 ing group, in a substantially anhydrous organic solvent, and - subsequently removing all blocking groups.
The blocking groups which may be used to protect the ` 6'-amino moiety of the kanamycin and the amino group of the acy-. lating acid (group B in Formula II) are conventional blocking groups for the protection of primary amino groups and are well known to those skilled in the art. Suitable blocking groups include alkoxycarbonyl groups such as t-butoxycarbonyl and t-amyloxycarbonyl; aralkoxycarbonyl groups such as benzyloxy-carbonyl; cycloalkyloxycarbonyl groups such as cyclohexyloxy-carbonyl; haloalkoxycarbonyl groups such as trichloroethoxy-~ 5 carbonyl; acyl groups such as phthaloyl and o-nitrophenoxyacetyl;
and other well-known blocking groups such as the o-nitrophenyl-thio group, the trityl group etc.
The acylating acid of formula II may be in its (+) or (-) isomeric form or a mixture of the two isomers (the d, 1 form), thus producing the corresponding compound o formula I in which the l-N-[~-amino-~-hydroxyalkanoyl] group is in its (~) ~or (R)]
form or its (-) [or (S)] form, or a mixture thereof. Each such isomeric form, and the mixture thereof, is included within the scope of this invention. In one preferred embodiment, the acylating acid of formula II is in its (-) form. In another preferred embodiment the acylating acid of formula II is in its (+) form.
In one embodiment of the invention the starting mate-rial is polysilylated kanamycin A or B (and preferably polysily-lated kanamycin A). In another embodiment the starting materi~l is polysilylated kanamycin A or B (and preferably polysilylated kanamycin A) containing a blocking group other than silyl on the 6'-amino moiety, said blocking group preferably being select-ed from those of the formulae Rl ~ .

,~ CH20C--, CH3-c-0-C- ~
2 C~3 R
'~
~ ~S~
N2 'd,J

~, ", 09~;i3 x3 2 and wherein Rl and R2 are alike or different and each is H, F, Cl, Br, NO2, OH ~lower)alkyl or (lower)alkoxy, and X is Cl, Br, F or ; I, and Y is H, Cl, Br, F or I. The most preferred blocking group i5 the carbobe.nzyloxy group.
In a preferred embodiment of the invention the acylating derivative of the acid of Formula II is an active ester, and preferably its active ester with N-hydroxysuccinimide, .
N-hydroxy-5-norbornene-2,3-dicarboximide or N-hydroxyphthalimide.
In another preferred embodiment the acrylating derivative of - 10 the acid of Formula II is a mixed acid anhydride, and preferably its mixed acid anhydride with pivalic acid, benzoic acid, zobu~ylcarbonic acld or benzy1carbonic acid.

~
' -- 6a -In a most preferred embodiment, this invention relates to the preparation o~ l~N-[L~ y-a~ino-a-hydroxybutyryL]~aa~mycin R or a nontoxic pharmaceutically acceptable acid addition salt ~hereo~, which comprises acylating polysilylated kanamycin A with a mixed acid anhydride of L~ y-~enzyloxycarb~nylamino-a-hydroxy-butyric acid (and pre~erably its mixed acid anhydride with pivalic acid, ben~oic acid, isobutylcarbonic aeid or benzylcarbonic acid) in a substantially anhydrou~ organic solvent, and subseq~ently remo~ing all blocking groups~
In a~ther most pre~erred embodimèn~, thi~ in~ention relate~
to the preparation o~ 1 N~[L~ y-amino-a-hydroxybutyryl~kanamycin A or a nontoxic phasmaceutically acceptable acid addition sal~
therQo, which compri3es aoylating polysilylated kanamycin A

: conta~ning a car~oben~yloxy group on ~he 6'-amino moiety wi~h a maxed a~id anhydride o~ Lo(-)-y-b~n2yloXyca~onylo ; amino-~-hydroxybutyric acid ~nd preferably it~ mixed acid anhydride with pivalic ac~d, benzoic acid, i~obutylcarbonic acid o~ benzylca~bonic acid) in a substantially anhydrou~
organic sol~en~, a~d subsequently removing all blocking group~.
In another mo~t preferred embod ~ ent, kh~s inve~tion : rela~e~ to t~ prepaxation o~ 1-N- 1L~ Y-aminO-a~hYdrOXY~U~YEY1~
Xanamycin A or a nontoxic pharmaceutically acceptable acid addition salt ther~of, which comprise~ acylating poly~lyla~e~ kanamycin A
wi~h an active e~te~ oS L-(-3-y-benzyloxycarbonylæmino-~hydroxy-buty~i~ acid ~and pre~erably its acti~e ester with N-~ydroxyr succinimide, N-hydroxy-5 n~rbornene-2,3-dicarboximide or N hydroxy-phthalimide~ in a sub tantially anhydrous organic ~ol~ent, and subsequently r~mo~ing all blocking groups.
in another most pref~rre~ embodL~nt, this invention relate-~to ~he preparation of l-~-E~ r-amino-a-hydroxybutyryL]kana~
mycin A or a nontoxic pharmaceu~ically acceptable acid addition salt thereof, which ~omprises acyla~iny polysilyla~ed ~anamycin A
containing a carbobenzyloxy group on the 6'-amino moiety with an activ~ es~er o~ L~ y-benzyloxycarbonylamino-a-hydroxybutyric acid (and pre~erably its ackive ester with W-hydroxysuccinimide~
N-hydroxy-5-no~bornene-2,3-dicarboximide or N-hydxoxyphthalLmide) in a substantialLy anhydrous organic solve~t, and sub~equently rem~ving all blocking groups.

In anoth~r aspec~, the pres~nt inventlon provid~ poly-silylated A or B con~aining a blocking group othex th~n silyl on the 6'-amino moiety. In a pre~erred em~odLmen~ ~he material i~ polysilylated k~namycin A or B (and praferably paly~ilylated k~namycin A) containing an average ~umber o silyl groups tand prefe~ably t~Lmethylsilyl group~ ~and preferabLy trim~thyl~ilyl groups~ per molecule o~ ~rom 4 to 80 In a~other p~eferred embodiment the material is polysilylat~d kanamycin A or B ~and prefesably poly~ilylated kanamycin A) containing a blocking gr~up other than ~ilyl on the 6'~amino group a~d containing an ave~age ~umber of silyl group~ (and pre~e~ably tr~me~hylcilyl group~ p~r molecule of from 3 ~o 7.
As u~ed her~in and in the cl~Lms~ the term ~nGntoxic~ phar~
2ceutically accaptable acia addition salkn o a compound of Formula I ~eans a mono~ di-, tri- or ~etrasalt formed by the intexaction of one ~olecule of a compound of ~ormula I with 1 4 e~uivalen~3 of a ~ontoxic, ph~rmac~utically accepta~le acid. Included among th~se acids ase ~cetic, hydrochloric, ~ulfuri~, maleic, pho~
phoric, nitric, hydrobromic, ascorbic, mali~ and citric acid, and .those other acids commonly u~ed to make salts o~ amine-containing pharmaceuticals~
Acyla.ion o~ the poly~ilylated kanamycin A or B starting mate~ial (with or wi~hout a biocking group othe~ t~an ~ilyl on the 6'-a~ino moie~y~ may, in general, be conducted in an orsanic solveAt in which the s~arting ma~arial has su~icie~t solubility. These startin~ materials are highly solu~le in most co~nn organic solven~s~ Suitable solvents include for .
example, acetone, diethyl ketone, methyl n-propyl ketone, methyL
i~o~utyl ketone, methyl ethyl ketone, acetonitrile, glyme, ~l(;)V953 diglyme, dioxane, toluene, tetrahydrofuran, cyclohexanone, methylene chloride, chloroform, carbon tetrachloride and mixtures of acetone/butanol or diethyl ketone/bu-tanol. The choice of solvent is dependent on the particular s-tarting materials employ-ed. Ketones, generally, are the preferred solvents. The most advantageous solvent for the particular combination of reactants being utilized can readily be determined by routine experimenta-tion.
Suitable silylating agents for use in preparing the polysilylated kanamycin starting materials utilized herein in-clude those of the formula \ Si / - - NH

\ I R

N - Si H \ R4 m IV \ ~ V

wherein R5, R6 and R7 are selected from the group consisting of hydrogen, halogen, (lower)alkyl, halo(lower)alkyl and phenyl, at least one of the said R5, R6 and R7 groups being other than halo-gen or hydrogen; R4 is (lower)alkyl, m is an integer of 1 to 2 and X is selected from the group consisting of halogen and -N_ R

wherein R is hydrogen or (lower)alkyl and R9 is hydrogen, (lower)alkyl o~

-,~

~L0~9S3 R6 S i ~ ,, R ~ .
in which R5, R6 and R7 are as deEinecl above.

- 9a -~ peci~ic sllyl compou~ds of Fo~mulas IV and V are: trLmethyl-chloro3i}ane, hexamethyldisilazane, triethylchlorosilanç, ~thyl-trichlorosilane, dimethyldichlorosilane, triethylbromo~ilane, tri-n~propylchlo~osilane, methyldiethylchlorosilane, dLmethyl-ethylchlo~o ilane, dimethyl-t-~utylchlorosilane, phenyldLmethyl-bromosilane, b~nzylmethylethylchlorosilane, phenylethylm~thyl-chloro~ilane, triphenylchlorosilane, triphenyl~luorosilane, tri-o-tolvlchlorosilane, tri-p;dLmethylami~ophenylchlorosilane, N-ethyltsiethylsilylamins, hexaathyldisilazane, ~rip~enylsilyamine, tri-n-p~opylsilylamine, tetraethyldLmethyldisilazane, hexaphenyl-disilazan~, hexa p-tolyldi~ilazane~ etc. ~lao u~e~ul Ar~ hexa-alXylcyclotrisilazanes and octa-alXylcyclo~strasila2anes. Other suitable silylating agent~ are silylamides (such as txialkyl-cilylaceta~idQs and bis-trialkyl~ilylacetamides), ~ilylurea~
~such as ~ ethylsilylurea~ and silylureides. Trimethyl-silylLmidazole also may be utilized.

A preferred 5ilyl group is the ~rLmethylsilyl group and preferred silylating agents for introducing ~he trimethylsilyl group are hexamethyldisilazane, bis~trimethylsilyl)acetamid~
and ~rimethylsilylacetamidP. Hexame~hyldisilazane i~ most pref erred, When utilizing ~oly~ilylated kanamycin A or B containing a blocking group other than silyl o~ the 6 ' -amino moiety as a ~tarting material, said s~arting material may be prepared ei~her by polysilylating tha desired 6'-N-blocked kanamycin A or ~, or by introducing tha desired 6~-N-blocking group i~to polysilyla~ed kanamycin A or B.

Method3 for the introduction of silyl groups into organic ~ompounds, including certai~ a~inoglycoside~,are knawn in ~he art~
The polysilyla~ed kanamycins (wi~h or without a bloc~ing gro~p ot~?r than silyl on the 6 '-amino moiety) may be prepared by m~.ods which a-~ known ~ s~, or as described in this specification.

~ 10 --As used herein, the term polysilyl,~ I kanamycin A or B
r~fers to kanamycin A or B containing from two to ten silyl ~OUp3 i~ the molecule. Thu~, the term pc: ly~ilylated kanamycin A or B
`does not include per3ilyla~ed kanamycin A or B, which would COntaLn eleven s~lyl groups in the malecule.
The precise number of ~ilyl groups (or their location) pr~3~n~ in the polysilylated kanamycin star~ing ma~e~ials ~with or without a blockir~g group other th~n 3ily} on the 6 ' -a~ino moiety) is not ~cnown. We have found that both under~ilylation and over~ilylation lower the yield ~ t~e desi~ed product and increas@ the yield of other produ ::~5 ~
In ~he ca a o~ g~o~9 under- or o~ersilylation, lit~le or none o~ the d~ired product may be ~o~med. The degree of silylation which will give the ~reatest yield o de ired produc~ will depend on the particular r~actants b~ing used in the acylat~ on s~ep. The mo~t atYantage~sus degree of silylatior~ u$ing any combination o~ reactan~s ca~ readily b~ determined by rou~ine experi~entation.

.

9~i3 ~

When preparing l-N~ -y-amino~a-hydroxybutyryl]kana-mycin ~ by acylating poly~ilylated ka~amyGin A with tha N-hydroxy-succinimide ester o~ L~ y-benzyloxycarbonylamino-a-hydroxy-butyric acid in acetone solution, we hav~ found that good yields o~ the desired product a~e ob~ained by u~ilizing poly3ilylat~d kanamycin A which has been prepa~:ed by reacting ~rom about ~ to a}~out 5. 5 mo1~3 of hexamethyldisilazane per mole o~ ~canamycin A.
Greater or le~ser amounts of hexamethyldi~ila~ane may be utilized, but the yield o~ desired product in ~he subsaquent acylation s~ep i3 lowered ~igni~icantly~ In the specific p~ocess ~t forth above we pre~er t~ utiliz~ from abou~ 4.5 to about 5.~ mo}es o hexam~thyldi~ilazane per mole o~ ~anamycin in order to obtain maximum yield of product in the acylation step.

: It will be appreciated that ~ach mole of hexamethyldisilazane i~ capable of in~roducing two equiYalents of the trime~hylsilyl group into kanamycin A or s . Bo~A kanamycin ~ and B have a total o~ eleven sites (N~2 and OH groups) which migh~ be silyla~ed, while kanamycin A and B containing a blocking group other than silyl o~ th~ 6'amino ~oiety have a to~al of 10 such si~es.
Thus, 5.5 moles of hexamethyldisi~azane per mole o~ kanamycin A or B could theoretically comple~ely silylate all OH and ~H2 moie~ o~ ~he kanamycin, while ~.0 m~Lec of hexamethyldisilaza~e could comple~ely silylate one mole of kanamycin A or B containing a blocking group other than silyl on ~he 6'-amino moiety. ~owever, w~ believe ~hat such extensive silylation does no~ t~ke plac~

with th0se mclar ratios during reasonable reaction time periods, although higher degrees of silylation are obtained in a given reaction time when a silylation catalyst is added.

Silylation catalysts greatly accelerate the rate of silylation. Suitable silylation catalysts are well known in the art and include inter alia amine sulfates (e.g. kanamycin sul-fate), sulfamic acid, imidazole and trimethylchlorosilane.
Silylation catalysts generally promote a higher degree of silylation than is required in the process of this invention.
However, oversilylated kanamycin A or B can be used as starting material if it is first ~reated with a desilylating agent to reduce the degree of silylation before the acylation reaction is carried out.
Good yields of desired product are obtained when acylating polysilylated kanamycin A prepared using a 5.5:1 molar ratio of hexamethyldisilazane to kanamycin A. However, when kanamycin A silylated with a 7:1 molar ratio of hexamethyldi-silazane (or with a 5.5:1 molar ratio in the presence of a silylation catalyst) was acylated in acetone with the N-hydroxy-succinimide ester of L-(-)-~-benzyloxycarbonylamino-~-hydroxy-butyric acid, less than a 1% yield of the desired product was obtained. However, when this same "oversilylated" kanamycin A
was acylated with the same acylating agent in acetone solution to which water [21 moles water per mole of kanamycin; 2.5%
water (W/V)] had been added as a desilylating agent 1 hour before acylation, a yield of approximately 40% of the desired product was obtained. The same results are obtained if the water is replaced by methanol or other active hydrogen compound capabl~ of effecting desilylation, e.g. ethanol, propanol, butanediol, methyl mercaptan, ethyl mercaptan, phenyl mercaptan, or the like.
Although it is usual to utilize dry solvents when working with silylated materials, we have surprisingly found that, even in the absence of "oversilylation", tne addition of water to the reaction solvent prior to acylation often gives equally good yields, and sometimes gives better yields of desired product than in a dry solvent. In acylation reactions conducted in acetone at the usual concentrations of 10-20~ (W/V) of pol~-silylated kanamycin A, we have found that excellent yields of 1-N-[L-(-)-~-amino-~-hydrox~butyryl]kanamycin A were obtained when adding up to 28 moles of water per mole of polysilylated kana-mycin A; at 20% concentration, 28 moles per mole is approximately 8~ water. With other combinations o~ reactants, even more water may be tolerated or be beneficial. The acylation reaction may be conducted in solvents containing up to about 40% water, although at such high water concentrations one must utilize short ~;~
acylation times in order to avoid excessive desilylation of the ~ -polysilylated kanamycin A or B starting material. Accordingly, as used herein and in the claims, the term "substantially an- ;
hydrous organic solvent" is intended to include solvents contain-ing up to about 25% water. A preferred range is up to about 20 water~ a more preferred range is up to about 8% water, and a most preferred range is up to about 4% water.
As indicated above, the most desirable degree of sily-lation for any combination of acylation reactants may be readily determined by routine experimentation. It is believed that the preferred average number of sil~l groups in the starting mate- `-rial will usually be between 4 and 8 for kanamycin A or B and between 3 and 7 for kanamycin A or B containing a blocking group - other than silyl on the 6'-amino moiety, but this is only theory and is not considered an essential part of this invention.
The duration and temperature of the acylation reaction are not critical. Temperatures in the ran~e of about -30C to - about 100C may be used for reaction times ranging from about one hour up to a day or more. We have found that the reaction usually proceeds well at room temperature and, for convenience and economy, prefer to conduct the reaction at ambient tempera-X ~ 14 -` `\` :

ture. However, for maximum yields and selective acylation, we prefer to conduct the acylation at from about 0 to 5.
Acylation of the l-amino moiety of the polysilylated kanamycin A or B (with or without a blocking group other than silyl on the 6'-amino moiety) may be conducted with any acylat-ing derivative of the acid of Formula II which i5 known in the art to be suitable for the acrylation of a primary amino group.
Examples of suitable acylating derivatives of the free acid in-clude the corresponding acid anhydrides, mixed anhydrides, e.g.
alkoxyformic anhydrides, acid halides, acid azides, active esters and active thioesters. The free acid may be coupled with the polysilylated kanamycin starting material after first reacting said free acid with N,N'-dimethylchloroformininium chloride ~cf. Great Britain 1,008,170 and Novak and Weichet, Experientia XXI, 6, 360 (1965)~ or by the use of an N,N'-carbonyl~
diimidazole or an N,N'-carbonylditriazole [cf. South African - 14a -~l~U953 Speci~ication 63/2684 ] or a car~odi~mide reage~ te~pecially N, -Nl-dicyclohexylcarbodiimide, N,N'-diisopropylcasbodiLmid~ or N-cyclohexyl-N'~ morpholinoethyl)~arbodiimide: cf. Shee~an and ~e~s, J~AoC~S~ ~ 77~ 19~7 (1955)~, or ~f an alkynylamine reagent [c~. R. Buijle and H. G. Viehe, Angew. Chem~ International Editi~n~
3, 582, ~19~4)1 ar o~ an isoxa~olium ~al~ reagen~ ~c~. R. B. W~od-ward, R. A. Olofson and Ho Mayer, J. Ame~. Chem. Soc., 83, 1010 tl961)1, O~ Of a k~tenLmine seagent td . C. L. StQven~ and M.
Eo Munk, JO Amer. Chem. Soc., 80, 4065 ~1958)l or of hexachloro-cyclo~riphospha~riazine o~ hexabromocyclotriph~phat~iazine ~U. S. Pat. No~ 3j651,050) or of diphenylphosphoryl azide~DPA; J. Ame~. Chem~.Soc., 94, 6203-6205 (1972)l or of diethy~phosphoryl cyanide [DEPC; Tetrahedron Le~ers No. 18, pp. 1595-1598 (1973)~ or of diphenyl phosphite ~Tetrahedron Letter~ No. 49, pp. 5047-5050 ~1972)]. Ano~her equivalent of th~ acid iS a corresponding azolide, i.e., an amide o~ the corr~sponding acid whos~ amide nitrogen is a me~ber o~ a quasiaromatic fiv member~d ring con~aining at leas~ two rogen atom~, i.eO ~ ~ida~ole, pyrazole~ he triazoles, be~zi~ida~ole, beri~ot~iazole and their ub~tituted dexivatives.
As will be appreciated by those skilled in the ar~, it some~mes ~ay be desi:rable or necessary to pr~tect the hydroxyl group of the ?cyla~ing deri~a~ive o~ the acid of Formu~a $I ~ e . g . when ~talizing ~cyla~ing deriYa . iYe~ such as an acid halide. Pro-tection of the hydroxyl group may b~ ~c~omplished by mQans known in the ~rt, a. g . by use o~ a car~benzyloxy group p ~y acatylatios~
by silylation, or She like.
A~t~r co~ple~ion of th~ acylatio~ reac~ion, all bloc3ci~lg gxoups ar~ re~ov~d by ~thod~ known ~ se to yield the desired product of ~onaula I. The Cilyl groups may, for example, ~e~dily be ~moved by hyd~olysis with watex, prefexably a~ low pH.
Blockin~ group ~ of the 0 ~ ~ ~ 3 acylating de~ivative o~ the acid o~ Formula I~, and the blocking g~oup o~ the 6'-ami~o ~oiety of the palysilylated kanamyci~
s~arting mat~rial (if pres~nt) may also be removed by known m2thods. Thus, a t-butoxycar~onyl group ~ay ~e ~emoved by the us~ of formic acid, a ca~obenzyloxy group by catalytic hydro-qenation, a 2-hydroxy-l-naphthcarbonyl group by aaid hydroly~is, a t~ichlo~oethoxycar~onyl gr~up by treatment with ~i~c dust in glacial acetac acid, the phthaloyL ~roup by treatme~t with hy~razine hydrate in e~hanol under hea~ing, e~c.

Yi~lds o~ p~oduct were deteEmined by variou~ methods.
After removal of all bloaking groups and chromatog~aphy on a CG-S~ (N~4~ column, the yield o~ amikacin could be determined by is~la~ion of th~ crys~alli~e solid from the approp~ia~e fraction~ or by microbiol~gi~al assay (turbidimetric or plate) o~ the appropriate ~ractions. Another t~ch~ique which we utilized was high pero~ance liquid chromatography of the un~educed acylation ~ixture, i.e. the aqueous solution ohtained aftex hydrolysi~ of the silyl groups and removal o~ orga~ic ~olvent but before hyd~o~enolysis to remove the re~aini~ blo~ki~g group(s~. This a~say wa~ no~ a direct assay for amiXacin or BB-K29 but for ~he corresponding mono- or di~ Locked compounds.
The inst~ument utilized was a Waters Associates ALC/GPC 244 high pressurA iiquid chromatograph with a Waters Associates Mode~
440 absorbance detector and a 30 cm x 3 . 9 mm i.d . ~BoNdap~k C-l~
colu~n,- und er the oll~wing conditions:

~obila Phas~: 25~ 2~propanol 75~ d.0lM ~odiu~ acetate p~ 4.0 Flow Rate: 1 ml./minut~
~etectcr: UV at 254 Nm.
Sensitivity: 0~04 AUFS
Diluent: DMSO
Injected Amount: 5 ~l Concentration: 10 mg./ml.

Chast spe~d varied, but 2 minu~es~inch wa~ typical. The a~o~e condition~ gave W traces with p~aks which were easy to mea~ure quantit~ 21y. Th~ r~sul~-~ o~ the above analysen are re~erred ~o in ~he specification a~ ~LPC assays~
In order to avoid the repetition of complex chemical nam ~
tha following abbreviation~ axe somet~mes utilized i~ this spec-i~ication.

AHBA L-~ amino-a hydroxybu~yric acid B~ N-Car~oben2yloY.y derivative of AXBA

~ONB N-~y~roxy-5-norbornene-2,3-dicarboxImide NAE N-hydroxy;5~nor~ornene-2~3-dicarboximide (or B~A-'ONB') acti~ated es~r of BHBA

~ONS N-hydroxysuccinLmide SAE Nrhydroxysuccinimide actiYated ester of ~or ~HBA- ' ONS ' ) - B~BA

DCC dicyclohexylcarbsdiLmide DCU dicyclohexylurea ~MDS hexamethyldisilazzne BSA bis(~rime~hylsilyl)acetamide MS~ trimethylsilylacetamide - - 17 ~

TFA trifluoroacetyl ~-~OC tert. butyloxycarbonyl ~ icalite~ is a trademark of the Great ~akes CaIbon Corporation ~or diatomaceous earth.
~ Amberlite CG-SOR i~ a Trademark of the Rohm ~ Haas Co.
~or the chxomatographic grade of a weakly acid ~ationic Q~change resin of the car~oxylic-pol~methacrylic type.
~ -Bondapak" is a Trademark of Waters Associates for a sexie-~ of high performanc~ L~quid chromatography columns.
All temperatures herein are glven in degrees centigrade.
As uesd harein, the terms "(lower)alkyl" and n (lower~alkoxy"
~efer ~o alkyl or alXoxy groups containing from 1 to six.carbon atomsO

Description of the Preerred Embodiments .~ .

6'-N-Carbob~nzyloxykanamycin ~ ~15 g., 24.24 m. moles~ was slurried in 90 ml. o~ dry acetonitrile and heatad to reflux under a nitrogen atmo~phe~ examethyldisilazane ~17.5 g., 108.48 m~
mole3) wa~ added slowly o~r 30 minutes, and ~he resulting solu~ion was refluxed for 24 hour Aft~r rem~val of ~e solven~ in vacuo (40~ and complete dsying under vacuum (10 mm), 27.9 g. of a white, æm~rphou~ solid was obtained t90.71~ calcula~ed as 6'-N-Carbobenzy-loxyk namycin A (Silyl~g].

This solid was dissolved in 150 m}. of dry diethyl ketone at 23. L~ benzyloxyc~rbonylamin~-a-hydroxybutyrie acid N-hydroxy-5-nor~ornene-2,3-dicarboximide ester ~N~E~ (11.05 g.~
26.67 m. mole~) dissolved in 100 ml. o~ dry diethyl ketone at 23 was added slowly wi~h good agitation o~er 1/2 hour. The solution was stirr~d a~ 23 for 78 hours. The yellow, clear solution (pH
7.0) was dilut~d with 100 ml. of water. The pH o the mixture was adju~ted to 2.8 (3N HCl~ and s~irred vigorous~y at 23 for 15 minu~es. The aqueou~ phasa was separated, and the organic phase was extracted with 50 ml. of pH 2.8 water, The combined aqueous frac~ion weæe washed with 50 ml~ o ethyl ace~a~e. The solu~ion was placed in a 500 ml. Parr kot~le) toge~her wl~h 5 g, of 5% palladium on casbon catalyst (En~lhard) and reduced at 50 psi ~2 for 2 hours at 23 . ~h~ mix~ure was filtered through a pad o~ Dicalite which was then washed wi~h an addi~ional 30 ml.
of water. The colorless filtra~.e was concentrated in vacuo (40-45) to 50 ml. The solution was charged on a 5 x 100 cm C~-50 ~NH4+) ion exchange column. Af~er wa5hin~ with 1000 ml. o~
water, unreacted kanamycin ~, 3-~L~ y-amino~a-~ydroxybutyryl]-kanam~cin A (BB-X291 and amikacin w~re eluted with 0.5~ ammonium hydroxide. Polyacyl material was recovared with 3N ammonium hydroxide. Bioassay, thi~ layer chromatography and optical rotation were used to monitor the pro~res~ of alution. The volume and observed optical rotation of each fraction of elua~e, a~ well as the weight and percent yield of solid isolated from each ~raction ~y evaporation to dryne~s, ~xe summarized below:
Volume a W~ight Material (ml) 5~ ~ % Yield ~anamycin A 1000 ~O.L15 00989 9.15 BB-K29 1750 +0.24 4~37 32.0 Amikacin 2000 ~0.31 6020 47.4 Polyacyls 900 ~0.032 00288 2.0 The ~pent diethyl Xetone layer was ~how~ by high perfon~ance liquid ~hromatography to contain an addi~ional 3-~ amikacin ~

The crude amikacin (6.20 gms.) was di.ssolved in 20 ml. of water and dilut~d with 2Q ml. of methanol, aAd 20 mlO of isopropanol wa~ added to induce crystallization. Th2re was obtained ~.0 g~.
~45.8Yo) o~ c~ystalline amikacin.

~, .

Amikacin by Selective Acylation o~ Pol~ ~rimethYLsilY11 6'-~-._ . .. . .

Poly~t~imothylsilyl) 6'-N-~arbobenzyloxy kana A prepared a~ in Example 1 ~103 g., .0~1 moles, calculated as 6'-N-Carbo-benzyloxykanamycin A (Silyl)9) was di~solved in 10~ ml. of dry aceton~ at 23. ~ y-benzylo~ycarbonylamino-a hydroxybutyric acid N-hydroxy-5-norbornen2-2,3-dicarboxlmide ester (NAE) (35.2 g., ,085 mole~ di~olved in 180 ml~ of dry acetone at 23 wa~
added slowly with good agitation to the solution of poly(trimethyl-5ilyl) 6'-~-Carbobe~zyloxykanamycin A over a 15 minu~P period.
The.qolution was sti~ed at 23 for 20 hours under a nitrogen atmoephe~e~ The pale yellow, clear solution ~pH 7~2) was dilut~d with 100 ml. of water~ The pH o~ the ~ixtura was adjusted to 2.5 (3N HCl) and s~irring continued at 23 for lS mi~ut85. Aceton~
was remov~d usin~ s~ea~;ejector vacuum a~ about 35. The solu~ion was placed i~ a 500 ml. Parr bot~le, to~ether with 10 ~. of 5~
palladium on- carbon cat~lyst (Engelhard) and reduced a~ 40 psi H2 for ~ hour~ ~t 23. The mixtu~a was ~iltered through a pad o~
diato~aceous ea~th which was then washed with an addi~io~al 50 ~1 of water. A~ter eoncent~ation to approx$~ately 1/3 volume, ~he soluti4n ~pH 6.9-7.2) was charged on a 6 x 110 cm~ CG-50 ~NH~) ion exchange column and eluted with a stapwise gradient from ~2 to 0.6 N a~monium hydroxide to reco~er amikacin. An automatic palar-imeter was used to monito~ the progress of elution. Combinations were made on the b3,si~ of thin layer chromatography evaluation.
~he combined amiXacin fxa~tion were concen~rated to 25-30% so}ids.
The solution ~

was diluted with an equal volume o methanol, followed by two volumes of isop~opanol to i~duce crystalli~a~ion. There was re~ered 18~2 g. (40~) o~ crystalline a~ikaCin-The recovery of 12% k~namycin A, 40% BB-K29 and 5% poly-acyLated kanamycin gave a material balance of 97~.

Using In Situ Blocking A o ~ ~

~ anamycin A ~ree base ~18 g. actiYity~ 37.15 m~ moles) was slurried in 200 ml. of dry ac~tonitrile and heated to reflux.
~exam~thyldisilazane ~29.8 g., 184~6 m. moies) wa~ added over 30 minut~s a~d the mixture was stirred at re~lux for 78 h~urs to give a light yellow cle~ solu~ion. Removal o~ the ~olvent under vacuum left an amorphous solid ~e~idue (43 gm., 94~) ~calculated as kanamyci~ A ~silyl)l~]~

B. ~

- p-~B~nzyloxycarbonyloxy)benzoic a~id (5,56 g., 20043 m.
moles) wa~ slurried in 50 ml. of dry ace~onitrile at 23a. N,0-bis ~r~methyL~ilyl acetamide ~8.4 g., 41~37 m~ mole~ wa~ added with good ~tirrin~. The ~olution was held for 30 minute~ at 23 ~, and then added over 3 hours wi~h Yigor~us stirring ~o a solutio~ of poLy(trimetAyl~ilyl~anamycin A (21~5 g~, 17.83 ~.
mole, calculated as the (silyl)lQ oompound) in 75 ml. of dry aceto~itrile at 23. The mix was stirred for 4 houxs, the solven~ ~

, ~3L0~53 was removed in vacuo (40~3, and the oily ~es~due was dis~olved ... ...
in 50 ml. o~ dry aceton~ at 23CC.

L~ y-benzyloxycarbonylamino-a-hydroxybutyric acid N-hydroxy 5-norbornene-2,3-dicarboximide e~ter ~N~E) (~.55 g.
20.63 m. moleY) in 30 ml. of ~c~tone was added to the above ~olution over a period of 5 minutes. The mixture was held at ~3C for 78 hour~, The solution was dilu~ed with 100 ml. o~
water and the pH ~7.0) lowered to 2.5 ~6N HCl). The mixture was placed in a S00 ml. Parr hottle together wi~h 3 g. of S~
palladium cn carbon catalyst ~Engelhard) and r~uc~d at 40 psi ~2 ~os 2 hours at 23. The mixture wa~ filtered through a pad o~ diatomaceous ear~h which was then washed with 20 ml. o~ water.
The combined ~iltrate and washings (168 ml.) were determined by microbiological assay against E. ~oli to contain approx~ma~ely 11,400 ~g/ml. (19g yield) of amiXacin.

Pre~aration of 1 N~L~ Y-Amino-a-hYdroxybutyr~l~kanamycin A

A.

A su pension of 10 g. ~20.6 ~. moles) kanamycin A in 100 ml~
of dry ac~tonitrile and 25 ml, ~119 m. ~oles3 1,1,1,3,3,3~hexa-methyldisilazane was re~lux2d for 72 hours. ~ clear light yellow solu~ion r~sul~ed. The solution was stripped ~o drynes~ n vacuo at 30-40C. There was obtained 21.3 g~ of poly(trimethylsilyl~
Xanamycin A a~ a light tan amorphous powder [85~ yield calculated as kanamycin A (silyl)l~].

9~3 B, ~

To a solution of 2.4 g~ (2.0 m. mole~) of poly~tr~methyl-silyl) ~anamycin A in 30 ml. of dry acetone was addsd 51OW~
2.0 m. moles of L~ r ~benzyloxycarbQnylamino-a hydroxybutyric acid N-hydroxy-5-norbornene-2,3-dicarboximide ester ~NAE~ in 10 ml. o~ dry acetone at 0-5C. The reaction mixture was sti~red at 23C for a week and then ~tripped to dxyne~s in ~acuo at a bath temperature o~ 30-40C. Water ~60 ml,) w~s then added to th~ residua, followed by 70 ml. o~ methanol to obtain a ~olutionO
The s~lution wa3 acldi~ied wi~h 3N HCl to pH 2. 0 and then reduced at 50 psi H2 ~or 2 hou~s, using 500 mg of 5% palladium on car~on catalyst. The material was filtered, and the combined filtrate and wa~hings were determined by microbiological assay again~t _ to contain a 29.4~ yield of amikacin Pr~aratLon of AmiXacin by Sele~tive ~-AcYlation of Polytr~meth~l-~.

I~ SummarY
Silylation of 6l-W-carbobenzoxy ~ana A in acetoni~rile using hexamathyldisilaza~e (HMDS) ffords the 6'-N-carbobenzoxy Rana ~
tsilyl)g intermediate ~. This silyla~ed Xana A is readily soluble i~ most organic solvents. Acylation w~th NAE in anhydrous a~etone at 23 using a 5% molar exce~ o~ NAE xelative to 6'-N Cbz Kana A
input-af~orded a mixture con~aini~g only Cbz dexivatives o amiXacin and BB K29, ~ome unre~cted Kana A and so~e polyacyl ma~rial. No B~-Kll was de~ec~able in any of ~hese studie~.
Elution of a~ acetone acylation mix, a~ter re~uction and workup, from a CG-50 (NH4l) colu~n u~ing an a~mOnium hydroxide gradient afforded i901a~ed yield~ of pure amikacin in the 40% r~nge~

II. Equations OH OH
HO /~ OH ~/
~0 ,~ /1' 5l6'NEI
~2N~/ H2N~L ' C-O
O C~
6 ' -N-C~z Kana A
C26H42013N4 t618.65) + ~CEI33 3Si-NH-Si ~CH3) 3 H~DS ~161.4) ~ ¦ CH3CN

/'~O.R OH ~2~
/1' 5,6 NH

o NH3 C~

R -- Si~CH3)3 C~z Kar~a ~

C53H114013N4Si9 ~1268 . 3) .

OH ~_ . CbzNH (CH2) 2-CH-COOH ~ HO~ ~ + DCC:
~ (Z~6) B~3A (253.4) HONB (179.2) ~Ace~one o OH () 11 DCU ~ CbzHN~CE~2) ~-1H-C~O-N~
(2;~4 ~ 3 ,~

Q) NAE (414.6) diCb2aznikac in 8S4) , + (~9 ~ diCbzBB--K2 9 23 . ~

6 'Cbz 1, 3~diBHBA-~tana A
~ _ +
5 % PdJC
~'~:~z Rana A

~}cac~ . + BB-K29 + 1, 3~diAH~ Kana A ~ ~;ana A
(585.62) (722.76~ ~484.5) ~CG-50 (NH4~) AmiX~cin 1:~00953 III. Materials ~ Vol~ ml. Mole~
6 '-~-Cbz ~ana A 50 .081 HMDS 58.~ 76.5 .365 Acetonitrile 300 BHB~ 2105 .085 HONB 15.2 .085 DCC 17.48 .085 Acetone 260 CG-50tN~4+) 3000 Methanol As required IPA As required IV. Safety 6'-N-C~z ~ana A - No direct information ava~1-able. Avoid dust contact.
Acetonitrile o Treat as a cyanide. Avoid breathing vapors. May caus2 . skin irritation.
H~xamethyldisila2an2 -~HMDS) Irritant, handle with care.
_~ .
6'-N-Cbz Kana A ~Silyl)g - No direct information avail-----------_----------- able, handle with care.
BHB~ - Toxicity is not established.
Avoid exposure to solids~
HONB - Toxicity unknown. Use pre-caution in handling.
DCC - A severe skin and eye irritant~
Avoid inhala~ion of mist or vapors. Toxic~
Acetone - Flammable. Inhalation may produce headache~ fatigue, excitement, bronchial irritation, and, in large amount5, narcosis.

~0~39~

No direct in~ormation avail~
able; always handled direc~ly as solution in aceton~.
Methanol o ~lammable. Poisoning may occur rom inge~tion, inhal~^
ation or percutaneous ab~orp-tion ~, - Flammable . Inge~tion or inhalatiort of large quantities of the vap~r may cau~e headache, dizzine~s, m~nt:al depre~sion, ` vomiting, narc~ is.
ATNnonium h~roxide - Toxic vapc~rs. Wear mask, avoid contact with l~quid.
CG-SO 5NH4~) - No toxi~ty data available~
__ handle with care.

V. Proc A, ~ =~_~ ~9 [6'-N-Cbx Kana A ~Silyl)g]
.
1. Slurry 50 g. of 6'-N-carbobenzyloxykanamycin A
(KF C4%) in 300 ml. of aceton~trile (KF C0.01~). Bring to reflux 574 ) maintaining a stream of dry nitrogen through the slurry.

2. Add ~lowly over a 30 minute period 75.8 ml. hexa-methyldisila~ane (HMDS)~ Complete solution will occur with evolution o a~monia ga~.

3. C~ntinue reXluxing for 18~20 hours under a nitrogen purge.

4. Concentra~e the clear, ligh~ yellow solu~ion under vacuu~ (bath ~emp. 40-50) to . foamy solid. Yield~ of the ~ilylg compound a 9- 92g. 530-94% Theory).

NOTE~ For future reference; in other solvent studies this solid is normally not i301ated but used direc~ly for the acylation.

B. Pre~aration o~ N-h~drox -S-norbornene-~,3-dicarboximide acid (NAE3 ___ 1. Di~solve 2105 g. of L-t-)-r~carbobenzyloxyamino-~-hydroxybutyric acid (B~BA) in L00 ml. of dry acetone at 23 followed by 15.2 g. of N-hydroxy~S-norbornene-2,3-dicar~oxlmide (~ONB). A complete solution will result.

2. ~ver 30 minutes add a solution of 17~48 g. o~
dicy~Lohexyl~arkodiimide ~DCC) in 50 m~. of a~etone with agitatlon.
The temperature will rise to approxLmately 40 during the addi~ion with precipitation of dicyclohexylurea ~DCU).

3. Agi~ate the slurry for 3 4 hours allowing the temperature to equilibrate to 23 25~.

4. Remo~e ~he urea derivative by fil~ra~icn; wash the cake with 30 ml. acetone. Save the filtrate plus washings ror the acylation step below.

C~ Acylation of 6~-N-Cbz Kana A (Silyl)9~

1~ ~issol~e ~he 6'-N-Cbz Ka~a A ~silyl)9 .isolated in Part A, Step 4 in 100 ml of dry acetone at 23-24.

2. With good agitation s~owly add the NAE solution prepared in Part B over a 15 minute period. The temperature wiLl gradually rise to approximately 40. Allow the solu~ion to equili-brate to 23 and continue stirring for 18-20 hours under a nitrogen .tmo~phereO

3. Add 100 ml. of water and lower the pH ~6.9-7.2) to 2~2-2.5 with 6N hydrochloric acid. Agitate for lS minutes at 23.

395~

(NOTE: A second layer may form this does not pres~n~ a problem in the workup~.

4. Remove acetone under vacuum at a bath temp~rature o~ 30-35. Transfer the concentrate ~o a suitable hydrogenation ve~sal ~prepurged with nitrogen). Add 10 g. 5~ palladium on carbon catalyst, and hydrogenate at 40 psi for 2 3 hours.

5. Filter the mixture through a Dicalite pad, washing the hydrogena~ion v~sal and cake with an additional 5Q ml. water.

6. Concentrate the filtrate plu~ wash to approximately 1/3 volume (50 ml.) under vacuum at 40-45.

7. Check ~h~ pH. It should be i~ the xange 6 . 9-7 ~ 2 .
If not, adjust with lN ammonium hydroxide. Charge the mixtur~ on a CG~50 ~NH4+) colu~n (6 x 110 cm)7

8. Wash the column with 1000 ml. of deionized water.
Then elute with 0~5-0.6N ammonium hydroxide using an automatic polarimeter to monitor the progre~s o~ elu~ion. The order of elu~ion is as follows:

Residual Kana A - > BB-K29 - ~ AmiXacin~
No BB-Kll was detected in any o~ our acylation workups. Polyacyl material i.e. the 1,3-diA~BA analog of Kana A, i5 recovered by washing the column with 3N ammonium hydroxide~

9. Combine ~he amikacin frac~ions and concen~rate to 25-30~ soiids. Dilute with 1 volume of methanol, and seed with amikaci~ crystals.

10. Add slowly over 2 hour~ 2 volumes of isopropanol ~IPA) with good stirring~ and crystallize at 23 fox 6-8 hours.

11. Filter the solid, wa~h with 50 ml. of 1:1:2 w~er/methanolfIPA mixture, and finally with 2S ml. IPA.

12. Dry in a ~racuum o~rer at 40 for 12-16 hours.
YiQld: 17.3 19.0 g. ~38~42%) O~ amikacin having the ollowing propertie:
TLC
CHC13 methanol - NH,IOH - water (1 : 4 : 2 : 1 ), 5 x 2 0 cm.
~ilica gal plates ~rom Quantum Industries -- one zone a~
detected with ninhydrin ~ 0 . 4 ) .

s..-iei. ~t.e-OI

t~l H~0 0.1M NH40H 0.1~ H2S04 589 + 101.6 + 101.9 + 103.5 C s 1.0%

13. The recovery of BB-X29 in this system was also 39042~, r2sidual Kana A 10~14% and 1,3-di AHBA-Kana A approxlmately 5% to give a m~terial balance >95%.

. Exam~le 6 ~ _ ~ trime~ch~l-_ilyl Kana A in Anhydrous Ace one I. ~

Silylation of Kana A ' base ' in acetonitrile uslng he~a-me~hyldisilazane (HMI:~S) yielded pol~r~rimethylsllyl Kana A.
The extent of silylation i~ as ye~ uncer~ain, but for the time being is assumed to be Kana A ~Sily})10. Poly~ilylat:ed Kana. A
i3 readily solu~ n most organic sol~rents. Acylation with SAE
in anhydrous acetone a~ 23 using a 1:1 molar ratio of SAE
relative to Kana A irlput afforded a mixt~re containing bz deri~atL~e5 of amikacin and BBK29, u ually in the r tio 2-3/1:
BB~6 ~app~oximately 5 896~, unreacted Kana A (15-2~%) and sc~rQe poly~cyl ma~rial (approximakely 5-10%)" Again, a~ wa~ ~e~
in our previous work on the acyla~ion of p~lytr~nethylsilyl 6 ' -N-C:arbob~n~oxy ~an~ Ar no BBKll was detec~ed in any of the~e ~xperimentsl. Reduction and worl~-up of an acetorle acylation mix, followed by chromatcgraphy on a CG-50- (N}~g+~ column using 0.5 N a~moniu~ hydroxide, ~rorde~ i~olated crystalline amiXacin the 34-39% range.

l~V~)~S3 II . ~g~, A, ~ ~o~

Xana A ' base ' C18~36011N4 ~484 . Sl~

~CH3 ) 3 Si-N~ 5~ ~CX3 ~ 3 H~S (161. 4 ) C~3CN

OR Og 0~ R
~ ~o OR ¦ / 3 N}lR ~N~3 : R = Si ~CH3) 3 ~ Kana A (Silyl)l~

C48}~116~311N4silo (1206 . 35) .

t E~
2 ~ 2 CH-COOH ~ HO N I ~ DCC

3HBA N-EIOS ( 2 0 6 . 3 ) 253 . 4 ) tllS . 9 ) ~`
EtOAc OE~ O ~
DCU + Cb2NH ~CH2 ) ;~CH-C-O-N

(~4.3) SAE
(3S0 . 33 ~

~z Amikacin (720) Cbz BB-K2 9 Acets~ne 0 ~ Cbz BB-K6 +
Kana A
l-~I2 -- Polyacyls (PrhmariLy 1, 3-diBHBA-Kan~ A~
- ' ~
Amikacin ~ K2 9 ~ BB K6 ~ 1. 3-diAHBA-Kana A ~ Kana A
. (585.62~ - (722.76) CG-50 (NH4~) ,~
.. . ~ / .
Amikaci~

III. MateriaLs . _ ~ Vol., ml. Moles __ __ }tana A ' base ' 50 .103 HMDS (Sp. gr. 0~774) 86.6~ .537 Acetoni~rile 6 0 0 SAE 35. 03 .10 Acetone 8 5 0 CG-50 ~NH46~ 3000 Methanol As required IPA As required IV. ~

l~ana A ' base ' - Known drug - usual caution advised .

Kana A (Silyl) 10 ~ No direct information avail-able, handle with care.

Other materials ~ See Example S

V~ Proc A. Preparation of Kana A (Silyl) 10 ~

1.. 51urry 50 g. of Kana A 'base' (K~? 2.5 - 3.5%) in 500 ml. of acetonitrile (KF c0.01~). Bring to reflux (74) main-t~ining a stream of dry nitrogen ~hrough the slurry.

2. Add slowly over a 3a minute pe:riod 112 ml. hexamethyl-di~ilazane (~MD5~. Comple~e solution will occur withirl 4-5 hours with evolution of anunonia gas~

3. Continue refluxing for 22-26 hours under a nitrogen purge .

-~{~ss3 4. Concentrate the clear faint yellow solu~ion undervacuum (40) to a syrupy residue. Flush with an additional 100 ml. acetonitrile, and dry completely under high vacuum for 3-6 hours. Yields of whitish amorphous solid are 109-115 g. (90-9S~ o~ theo~y, calculated as Kana A ~Silyl)10).

s .

1. Dissolve 100 g o L(~)-u-benzyloxycarbonylamino-~-hydroxybutyric acid ~BHBA) and 45.38 g of N-hydroxysuccinimide ~N-HOS) in 1300 ml of ethyl acetate (KF ~0.05~) with s~irring at 23C.
2. Dissolve 81.29 g of dicyclohexylcarbodiimide (DCC) in 40G ml.of ethyl acetate (KP ~0.05%) at 23C~ With good agitation add this solu~ion over 30 minutes to step 1 solution.
The temperature will rise ~o - 40-42C with concurrent precipitation of dicyclohexylurea IDCU~.
Agita~e the slurry 3-4 hours allowing the temperature to equili-bra~e ~o 23~C.

3. Filter the DCU; wash the cake with 2$0 mL of ethy~
acetate ~F ~0~05%). Discard the DCU cake. Save the filtrate and wa~hes.
. Concentrate the filtrate plus washes to ~ 500 ml.
(in vacuo at 30-35C)~ Some produc~ will crystalli7e out.

1~ 953 5. Transfer the concentrate to a suitable vessel and add with vigor~us agitation 100 ml.of heptane. If necessaryJ add seed crystal~ of SAE. Crystalliza~ion will begin almost immediately. Agitate the ~lurry for 30 mi~sutes at 23 C .
6. Add, over 30 minute~, 400 ml.of heptane and agitate the ~lurry 4-5 hour~ at 23C.

7. Filter and wash the cake with 200 ml.o~ ~:1 heptane/-ethyl acet~te followad by 100 ml.of heptane.

8. Dry ln a vacuum oven at 30-35C ~or 18-20 hours.
Yield i3 110.1-131.4 g (80 95~).
MP - 119-120 with softening at 114 (Corr.).
TLC - 4 acetone:l2 benzene:l CH3CO2H - Detec~ion 1% aqueous KMO4.
~f - O.7 for SAE; 0.2 BHBA on 2 x 10 cm presc~red silica gel plates from Analtech IncO

.C. Acylation of Kana A (Silyl)10.
~.
1. Dissolve the Kana A (Silyl)10 isolated in Part A, Step 4 in 500 mL dry acetone at 23C.

2. With good agitation add rapidly the SAE prepared in Part B (35.03 g) as a 10% solution in dry acetone over a ~-10 minute period. The tempera~ure will xise approximately 5~
Allow the solution to equilibrate to 23, and continue stirring for 18 29 hours.

3~ The light orange, clear solution is dilu~ed with 400 ml.of water, and th~ pH (7.0 7.5) lowered ~o 2.2-2.5 with ~10~953 3 N hydrochloric acid. The clear solution i~ now a~itated at 23~ for 15030 minutes.
4. Aceton~ i~ remo~ed under vacuum at a bath temper ature of 30-35~ (a small ~mount of material may separ~te at this point, but presents no problem). Tran~er the ooncentrate to a sui~able hydrogenation vessel. Add 10 g 5~ palladium on carbon catalyst, and hydrogenate a~ 50 p~ for ~-3 hours.
5. Filter the mixture throu~h a Dicalite pad, and wash the hydrogenation ves~el and cake with an ad~itional 2 X 50 ml, water.

6 ~ Concentrate the f iltrate plus washin~s ~o approxi~
mately 1/3 volume (150^165 m ) under vaauum at 40 45~

7 . The pH at this point is in the range 6 . 0-7 . 0 .
The mixture is charged on a C::-50 (NH4+) column (6 X 110 .cm) .

8. Wash the column with 1000 ml. o~ deion~zed water, Elute with 0. 5 N ammonium hydroxide us ng an autom~tic polarimeter ~o moni~or ~he progress of elution~, The order of elution is as follows:
Re~idual ~ana A ~ BB-K6 ~ BB-K29 ~ Amika~:in.
No BB-K11 was de~ected in ~ny of our experi~en~s.

9. Combine the amikacin ~rac~ions and ~on~:entrate to 25-30% solids. DLlute with 1 volume methanol, and seed with amikacin crystals.

9~3 10. Add slowly o~er 2 hours 2 volumes of IPA w~th good stirring and crystalli~e at 23 for 6-8 hour~.
11. Filter the solid, wash with 35 m~ of 1:1:2 water~methanol/IPA, and finally with 35 ml.IPA.

12. Dry in a vacuum o~er at 40 ~or 16-24 hours.
Yield: 19.91-22.84 g ~34-39~) IR, PMR and CMR ~pectral dat~ in addition to specif ic rotation were completely consistent for the desised structuxe.

C~C13Jme~ha~ol~N~4O~water (1:4:2:1) S x 20 om. ~ilica g~l plates from Quantum Industries ~ 1 Zone ~mikaoin having R~ ~0.4 (Ninhydrin Detection)O
, Pre~aration or Am~kacin by Acylation of Poly(~rimethylsilyl) 6'-of Pivalic Acid and BHBA

A ~

BHBA (5.066 gm., 20.0 m moles), BSA (4.068 gm., 20~0 m moles) and triethylamine (~.116 g, 2200 m moles~ were dissolved in 200 ml. of sieve d~ied te~rahydrofuranO The solution was re~luxed ~or 2 1/4 hours and then ch~lled to -10Co Pivaloyl chloride ~2.412 gm., 20.0 m moles) was added over a period o 2-3 minutes~
with stirring, and stirring was continued for 2 hours at lO~C.
The temperature was then ~llowe~ to climb to 23 C .

1~ 53 Poly(trime~hyl~ilyl) 61-N-Cb~ Xana A prepared as in Example 1 (5.454 gm., 4.~7 m moles, calculated as 6'-Chz Kana A t~ilY1)9) was di~solved in 50 ml. dry (~olecular ~i~ve) ~e~rahydrofuran at 23C~ One-half of tha ~olution o~ mixed anhydride prepa~ed in step A, above, (10. n m mol~) was added o~er a period of twen~y minutes, with stirring, and stirr~ng was continu2d for 7 days.

Water (100 m~ was then added to the reaction mix~ure, and the p~ (5.4) was adjus~ed to 2.0 with 3M H~504.
Stirring wa~ continued for 1 hour and the solutlon was extracted with ethyl acetate. Polyacylated material began to crystallize, so ~he reaction mixture was filtered~ After drying over P~05, the recottered solids weighed 0.702 gms. The extraction of the reaction mixture was continued for a total of 4 X 75 ml. of e~hyl acetate, after which the excess ~thyl aceta~e was stripped from the aqueous layer. An aliquot of the aqueous solution was subj ected to as~ay by ~2LC . The resulting cur~e indicated a 2 6 . 4 % yi~ld of di-Cbz amiXac:in.
:, The aqueous layer was hen hydrogen~ted i~ a Parr apparatus at 50 p . s . i . ~I2 pressure for ~wo hours, using 0 . 5 gm.
1096 Pd on carbon catalys~ . The material was f iltered, and the combined ~iltrate and washi~gs were determined against E coll to contain a 31. 2% yield of amikacin. Amikacin/BB -K29 ra~io ap~rox-~nately g-10/1; traces o~ polyacyl and unreacted Kana A present.

`` ~ll~V~53 '~

kac:in ~Y t~e Ac~lation . Anhydrous Solvent Poly(trime~hyl~ilyl) Kana A prepared as in Example 3 (2.40 gm., 2.0 m mole~, calculated a~ Xana A (silyl)10) was dissolved in 20 ml. o~ acetone which had bee~ dried with a mola~ular sieve. The ~olu~ion was stirred at 23C and a sol~ion of SAE ~0.701 qm., 2.0 m moles) in 10 ml. o~ sieve dried acetone was a~ded over a period of 10 seco~ds. Stirring wa~ continued at 23C for 22 hours. Water ~50 ml.) was added and th~ pH (7.5) was adjus~ed to 2.5. The aceton~ was stripped in vacuo at 40C
and ~he aqueous solution was then reduced at 51 p~s.i. H2 pressure at 23C for two hours, utilizing 1.0 gm of 10~ Pd on carbon as catalyst. Micro~iological assay showed a 31~ 4~ yield o amikacin~

B. water Added to Solvent ., , . ~ _ _ Step ~, above, was repeated, except that 1.0 ml. (56 m moles) of wat~r was added to the poly~trimethylsilyl) Kana A
solution, and stirred for 15 minu~es, prior to a~ylation with SAE. Microbiological assay showed a 33.80~ yield o~ amikacin.

Example g -A~
BH~3A ~1 , 2~7 gm~, 5 . 0 m moles) and ~1-trimethyl~ilyl-ace~mide (MSA) tl.313 gm., 10.0 m moles) ~n 20 ml. o sieve dried acetone was ~tirred at 23 JC r and triethylamine ~?EA) ~0.70 ml., 5.0 m mol~s) were added~ The mixture was refluxed under a N2 atmosphere ~or 2 1/2 hours. The mixture was cooled to -2û~C and isobutylchloroo~nate (0.751 gm., 0-713 ml~., 5.50 m mole-~) was added. Triethylamine hydrochloride inunediately began to separa~e. The mixtu~e was stirred ~or 1 hour at -20C~, B ,. A~ ~
Poiy(trimethylsilyl) 6l-N-C:bz Kana A prepared a~ in Example 1 (6.215 gm., 4O9 m mole~, c lculated as the ~silyl)g compound) was dissolved in 20 ml. of ~ieve dried ac~tone, with stirring, at 23C. The solution was cooled to -~0C and the cold mixed anhydride ~olution from step A was slowly added ove;r a period of 30 minutes. The reaction mixture was s~irred for an additional 1 1/2 hours at -20C and ~hen for 17 hours at 23C.
The re~ction mixture was then poured in~o 150 ml" of water at 23C: wi~h ~tirring, the p}~ (7.753 was adjus~ed to 2.5 wi~h 3N
HCl, ar~d stir~ing was con~inued ~or 15 minutes . Aceto~e was then 3tripped in vacuo at 40C~ An aliquot sf the resul~ing aqueous soLution was subjec~ed to assay by HPLC. The ~:esulting CUnr8 indica~ed ~ 34.33~6 yi~ld of di-Cbz amikacin~

~Q9S3 The main portion o~ the aqueous ~olutlon was redtlced at 50 p.s.i. H;~ pressure at 23C fox 3 1/4 hours, util~zing ~ . O gms of Pd/C catalyqt . The catalyst was resnoved by f iltration a~d the combilled f il~rate and wa~hing~ were determi ned ~y micro-biological assay against E . colL to contain a 35 . 0% yield o~
amikacin .

N~bz Xana ~ in ~-Pentanone ;^

Poly ~trimethyLsilyl) 6 '-N-C~z Xana prepared as in Exa~ple 1 (30 gm., 23.65 m moles, calculated as 6'-N~bz Kana A
~ilyl) g) dis~ol~ed in 100 ml. sieve dried 3 pentanone was stirred a~ 23C, and NAE (26.02 m moles, 10~6 excess) was added over a period of 4a minutes. Stirring was conl:inued for 113 hours at 23C and the mixture was then added to 250 ml~, water with vigorous stirr~ng. The pH (7.3) was adjusted to 2.5 with 3N HCl, the mixtu:ce was stirred ~or an add itional 3 0 minutes, and the 3 ~pentanone was stripped in va~uo at 4 0 C . Tha aqueou~
solution was extracted with 4 x 100 ml. of ethyl acetate. An aliquot oi~ the aqueous solution was thesl subjected to assay by HPLC. Th~ resulting cus~re i~dicated a 46.12% yield o~ di-Cbz amikacin .

The main portion o~ the aques~us reac~ion mix~uré was reduced a~ 51.0 p.s.i. ~2 pressure at 23C for 2 1/2 hours, utili2ing 3 . O gms . of 1096 Pd/C catalystO Microbiological assay o~ an aliqus:~t of the combined f iltrate and washings indicated a 40.249~ yield of amikacinO The mzin portio~ o~ the reduced aqueous reaction mixture was then concentrated iD vacuo at 405C ~0 . -- ~3 --approx ~ately lOd ml. and fractionated on a CG ~0 (N~+) ion exchange column (4 inches x 4 feet, containing approx~mately 10 liter~ of resin~. The aqueous ~olutlon was charged on the column, the column was wa~hed with 5 liters of wa~er, and the ma~erial wa~ eluted with 0~5N NH40H t~ollowed by 3N NH40H to elut2 polyac:ylated product~) . Polar~matry o~ the fractions showed the presence of a 42.7~ yield o~ ~ikacin. a 12.0% y~eld of ~nreacted kanamycin A, a }2.4~ yield of polyacylated material and a 23~2% yield o~ 8B-K29.

- ~ .

.

; A~ Poly(trLmethylsilyl) 6' N-Cbz Kana A prepared as in Example 1 ~2.537 gmO9 2.0 m moles, calculated as 6'-N-Cbz Rana A ~sily~39) in 300 ml. dry cy~lohexanone was acy}ated for 2~
hours at 23C with an NAE 501ution in dry cyclohexanone (10.8 ml. of 0.1944 m mclafml. ~olu~ion, 2.10 m mole)0 The reaetion mix~ure was then added to 150 ml. of water, with ~tirring, and the pH (5.6) was adjusted to 2.5 with 3N HClo The cyclohexano~e was stripped in va uo at 40C and an aliquo~ of the r~mainin~
aqueous phase was taken ~or assay by HPLC.
The mai~ portion o~ ~he aqueous phase was reduced under 50 p.s.i.
H2 pressure for 3 hours at 23C, u~ing 1JO gm of 10% Pd/C ca~alyst.
The catalyst was removed by fil~ration and the comhined ~ rate a~d washings were assayed microbiologically for amikacin.

B. ~eaction A, above, was repeated~ excep~ that the acylation wa~ continued ~or 115 hours inst~ad of 20 hours.

~4 -1~ 953 Yield$

~miXacin ~L n~bid~e~ric . Platt~
E~action A 49.18~ 42.87~ 39.16S
React~cn B 56~17~ 55.39% 3~.45%

~ .

A. Example 11 A was repeated except tha~ dry te~rahydrofuran was utili2ed as solv~nt instead of dry cyclohexanone.
B. Example 11, B was repea~ed except tha~ dry ~et:rahydrofuran was utilized as solY~nt instead of dry cyclohexanone.

Y~eld s ~n ~) ~ Plaee Reaction ~ 29.27*~ 28.,34~s 28~,189~
R~actio~ B 33.39% 21.. 5296 28,,63S

A. Exampl~ 11 A was repeatad except that the a~ylation was continuad for 44 hcur~ utili2ing dry dioxane a~ the solvent.
B. Example 11 B was repeat~d except that the acyl ation w~
continued for 18 1~2 hour~ utilizing dry dioxane a the ~olveIlt.

.
Yields ~ ~n (di-Cbz hniXacin) ~rbidimdtric ~late R~actioll ~39.1~% 43.~7% 33s3~%
Reaction E~42.82~ 2205S9~ 33,37~

:

N-Cbz Kana A ir~ Anhydrous Die~h 1 ketone at 75~C

~ To a stir~ed solution of poly [~r~methylsilyl ) 6 ' ~ Cbz gana A prepared as in Exampl~ 1 (20537 gm. ~ 2.0 m mole~, cal-culat2d as 6 '-N-Cbz ~ana A ~silyl) g) in 32 ml. sieYe dried di~thyl ketone at 75C wa~ added a ~olution of N~E 110.8 mle, o~ 0.1944 m moles/ml. of diethyl 3c~one~ 2.15 m mole~ over a period of 15 minu~es. Stirring wa~ ~ontinued a~ 75C for an additional 3 hour~ aftsr which the mix~ure wa~ poured into 150 ml,. of water. ~he p~ wa3 ad ju~ted to 2. 8 with 3N HCl and the diethyl 3ce~one was s~ripped in vacuo at 40C. HPLC as~ay _ 46 --a~ an aliquo~ of the aqueous phase indic:ated a 3g.189~ yield o~ di~bz amikacin.
~ he main portion of t}~e aqueous phase wa~ reduced under 49.8 p~-~.i. H2 pr~3sur~ fc~r 3 1~4 hours at 23C, usi~g 1.0 g~ of PdJC eatalys~. The cataly~t was re~o~ret by filtration and ~he combined ~iltrate and washings were assayed micro- -~iologically for amikacin., Turbid~netric assay showed 27.84%.
yield and Plate assay ~howed 28. 6% yield.

ExamDle }5 Pre aration o~ AmiXacin Dy t:he ~cylati~n o Poly~tr~methyl~ilyl) Xana A With NAE a~ 0-5 A~ter Back ~IYd~olvsis Nith Water A .

Kanamycin A (10 gm of 97.6% purity, 20.14 m moles) in 100 ml of sieve-d~ied acel:snitrile was brough~ to reflux under a nitrogen a~nosphe~e. A mixtwre of HMDS (22.76 gm, 141 m molP, 7 moles E~er mole o kanamycin A) and l'MCS ~1 mll 0.856 gm, 7~,88 m mole~) was - added to the ref luxirlg reaction mixture over a period of 10 minutes,. Re~lwc was con~inued for 4-3/4 hours and the m~xture wa~ then cooled, conc~ntrated ln ~racuo to a yellow viscous syrup and dried under high vacuum for . 2 hou~ ,, The yield of produc~
wa~ 23.8 gms (97 .9%, ca}culated as kanamycin A (5ilyl) 10 B. ~
Poly(trimethylsilyl) kanamycin A (23.8 gTns~ 2û.14 m moles) prepæed in ~tep A abo~re was dissolved in 250 ml of ~ieve-dried acetone at 23 and then cooled to 0 5 . Water (3 ~ 63 ml~ 201. 4 m moles~ 1;0 moles per mole of polysilylated kanamycin A) was added, with stirriny, and the mixture was allowed to stand under ~V9S3 -moderate vacuum for 30 minu~es. NA~ (13.133 m moles, 0.95 moles per mole or polysilylated kanamycin A) in 108.3 ml of acetone was then added over a period o~ <1 minutQ. The mixture was stirred at 0-5 for 1 hour, diluted with water, the pH
adjusted to 2.5, and the acetone was then r~moved in vacuo.
The aqueous solution was then reduced at 50 p.s.i. EI2 pressure a~ 23 for 2-1/2 hours using 2.0 gm~ of 10% Pd on carbon as a catalyst. The reduced reaction mixture wa9 filtered throu~h Dicalite, concentrated to ca. 100 ml Ln vacuo at 40 and then charg~d on CG~50~H4~) column (6 liters resin, 5 x 100 cm). It was washed with wate~ and ~hen eluted with O.6N-l.ON-3N NH~QH.
There was obtained 60.25~ amikacin, 4.37% BB-K6, 4.35% BB-R2~, 2~.47% kanæmycin A and 2.18% polyacyls.

. .

A. ~

6'-N-Cbz kanamycin A (20.0 gm, 32~4 m moles) in 200 ml of sieve-dried acetonitrile wa8 brouslt to reflux under a nitrogen at~5phere. H~DS ~47.3 ml, 226.8 m mo es, 7 moles per mole o 6'-N-Cbz kana A) was dded over a 10 minute pexiod and re~lux wa~ ~-ontinued for 20 hours. ~h~ mixture was then csQled, concen~rat~d _ vacuo, and dried under high vacuum for 2 hour~

to give 3g.1 gms of white amorphous solid ~95.4~ yield, calculated as 6'-N-Cb2 kana A (silyl)9).

`~ ~lOQ9S,3 B. ~cylation Poly[trlmethylsilyl) 6'-N-Cbz kana A (39.L gm, 32.4 m molea) psepared in step A a~ove was di5solved in 400 ml of dry ac:etone, with st~rring, at 23. Methanol (6.6 rnl, 162 m moles, 5 moleS
per mole of polysilylated 6 '-N-Cbz lcana A) was added and the mix~ure was stirred at 23 ~or 1 hour under a 9t~0ng nitrogen purge. The mixture wa~ ~ooled to 0~5 and a s.olution of 5AE
~11.35 gm, 32,4 m moles) in 120 ml.o~ pre~cooled, dry ace~one was added. The mixture was stirred for an additional 3 hours at 0-5 and then placed in a 4 cold room for 1 week, Water (300 ml) was added, the p~ was ad justed to 2. 0, the mixture was stirred for 1 hour, and the acetone was then stripped in vacuo. The resultant aqueous solution was reduced at 54.0 p.s.i. H2 pressure for 17 hours at 23 utilizing 3 . O gm of 109c Pd on carbon as ca~alyst. It was then filtered through Dicali~e, concentra~ed in vacuo to 75 100 ml, charge on a CG-50 (NH +) column and eluted _. 4 with wa~OE and 0. 6N NH40H . These was obtained 52 . 5296 amikacin,

14,.5% BB-K29, 19.69~ kanamyGin A and 1.7.1% polyacyls.

~ ' A. ~ CS in Acetoni~

Kanamycin A ~4 . 88 gm, 10 . 07 m mole) wa~ suspended in 100 ml of sie~e dried acetonitrile with stirring at 23. Tc: th~ stirred suspension was added ~etramethylguanidine (T~iG) (16.234 3Tn, 140~.98 m moles, 14 moles per mole of kanamycin A). The mix~ure was heated to reflux and TMCS ~15.32 gm, 140.98 m moles, 14 moles per mole of kanamycin A) was added over a 15 minute period. A.
_ 49 --white precipitate of TMG- HCl fo:rmed after about one-half of the TMCS had been added. The mixture was cooled to room temperature, concentrated to a tacky residue and dried under hi~h vacutlm ~or 2 hours. The solid was tri~urated with dry THF ~100 ml), and the in~oluble TMG HCl was filtered off and washed with 5 x 20 ml por~ions of THF. The combined filtrate and wa~hing8 were c:oncent~ated ln vacuo at 40 ~o a tacky residue and dried under high vacuum ~or 2 hours. There was obtained 10 . 64 gms of ~ lisht cream tacXy residue ~87.6~ yield, calcula~ed as kanamycin A (silyl)10).

B. ~
Poly (trimethylsilyl) kanamycin A (10 . ~4 gm, lQ . 07 m moles) prepared in step A above was dissolved in 110 ml of sieve-dried acetoneJ, with stirring, at 23 and the solution was cooled to 0-5. Wat~ (1.81 ml, 100.7 m moles, 10 moles per mole of poly ~ilylated kana A) was added a~d the solution was s~irred for 30 minutes under moderate vacuum. SAE ~3.70 gm, 10.57 mf moles, 5%
excess) in 40 ml o~ pre-cooled dry acetone was added over a pexiod of ~1 minu~e, and the m~xture was s~irred for orle hour. The mixture was worked up by ~he general procedure in Example lÇB to give ca. 50~ amikacin, ca. 10% ~B~K2~, 5~8/e~ BB-K6, ca. 20% kana~aycin A and 5-89~ po}yacyls.

Pre aration of Pol ~trimeth lsi,~
~DS
Xanamycin A ~L0.0 gms, 20.64 m moles) was suspende~ in 100 ml sieve-dried ~reshly distilled pyridine at 23. A nitrogen purge was started and the suspellsion was brought to rerl~x. ~IDS (17.33 gms, 107.3~ m moles, 5.2 moles per mole of kanamycin A) was added ~ S3 over a period of la minutes and the mixture was ref luxed for 19 hours. It was then cooled to room ~emperature, concentrated ln vacuo to a light yellow-gold syrup, and dried under high vacuum to a white amorphous solid. There was o~tained 22.1 gms (~2.6g6 yield, calculated a~ kanamycin A (silyl)10).

~.

chloro~

Kanamy~in A ~5.0 gm9 of 97.6~ purity, 10.07 m mole~) was suspended in 100 ml of sieve-dried acetonitrile at 23. Triethyl-amine (TE~) (33.8 ml, 24,5 gm, 241.7 m moles) was added and the suspension was brought to ref lux . A solution of trichloroethyl-silane (23.7 ml, 21.3 sm, 140.98 m moles~ in 25 ml dry acetonitri1e was added ov~r a 20 minute period. Reflux w~s continued for an additional 7 houxs and the mix~ure wa~ cooled to room temperature, whereupon long fizle ne~le~ of TEA EICl separa~ed out. The mixture was allowed to s~and at room temperature for ca . 16 hours ~ concen-trated in v cu_ a~ . 40 ~o a ~acky solid and dried for 2 hours under high ~7acuu~n to a deep orange tacky solid. The solid was triturated with 100 ml dry T~F at 23 and ~he insoluble TEA^HCl was filtered off, ~a~hed with 5 x 20 ml of THF, and driPd to gi~re 16.0 gms of TEA-HCl. The combined fil~rat~ and washin~s were concen.rated in ~tacu~ to a solid ;and dried urlder high vacuum for 2 hours. There was o~tained 1903 gms of poly(triethylsilyl) kanamycin A as a deep orange Yi~Cous syrup.

~ V9~3 Example 20 PreDaration of Pol ~trLmethvlsil 1) Ka _ _ Trimethylsilylurea Xanamycin ~ ~10.0 gm of 99.7% puri~y, 20,58 m moles) waa suspended in 200 ml of sieve-dri~d Acetonitxile, with stirring, at 23. To the suspension was added bis-tr.~methylsilylurea ~BS~) ~29,45 gms, 144~01 m moles, 7 moles per mole of kanamycin), and t~ mixture`was brought to reflux under a nikrogen a~nosphere.
~e~lux was continued ~or 17 hours and the reaction mixture was then cooled to room temperature. A small amount of insoluble material present was removed by filtration, washed with 3 x 10 ml portions o~ acetonitrile and dxied ~1~1381 ~ms). . Infrared showed this to be BSU plus a small amount of unreacted ~anamycin A. The combined filtrate and washings were cooled at 4 fox 16 hours.
Additional solid sepa~ated, was recovered as above, (7~8 gm~) ~nd was shown by infrared to be BSU plus urea. The ligh~ yellow filtrate and washings were concentrated ln vacuo at 40 and dried under high vacuum ~o give 27.0 gm of a white solid which was partly tac~y and partially fine needle-like crystals. The solid was treated with }50 ml of heptane at 23, tha insoluble portion was removed by filtration, washed with 2 x 50 ml por~ions of heptane .
and dried, to give 6~0 gms of whi~e needles (shown by infrared to be BSU plu3 urea). The combined fil~rtte and washings we~e oncentrated in vacuo at 40 and driPd under high vacuum for 2 _ __ houxs to yive 20.4 ~ms of white needles, the infrared spectrum of which was ~ypical for polysilylated kanamycin A. Calculattons showed ~he product to contain an average of 7.22 ~rimethylsilyl group~.

(J953 Preoaration of ~nikacin by the Acylation o Per (trimethYlsilYl~

A~
~ an~mycin A (10.0 ~m, 20.639 m mole~) was suspend~d in 100 ml of siev~-dried acetonitrile, with stirring, at 23. The suspension was brough~ to ref lux under a nitrogen purge and H~S
~3.322 gms, 144.5 m moles, 7 mole~ per mole of kana~myci-n A) wa~
added over a p~riod o ten m~nutesO Reflux was continued for 16 hours and the mixture was ~h~n cooIed to room temperature, concentra~ed in vacuo and dried for 2 hours under high vacuum.
There was obtained 24.3 gm of a whi~e, tacky residue ~92.1% yield, calculated as kanamycin A (silyl) B . A~S~
Per (trLmethylsilyl) kanamycin A (24 .3 gm) prepared in step A
above was dissolved in 240 ml of sieve-dried ac~tone, with stirring~
at 23~. To this solution wa. added 1,3-butanediol (9.25 ml, 103.2 m mole, 5 mol~s per mole of per(trime~hylsilyl) kanamycin A. The mix~ure was stirred at 23 ~or ? hours under a nitrogen purge and ~hen cooled at 0-5. S~E (7~23 gm, 20.64 m moLes) in 7Q ml of pre-cooled ac~tone was added over a period of abou~ 1 minute.
The mixture was stirred at 0-5 for 3 hours and then allowed to _ stand in a 4 cold room for ca. 16 hours. Water (200 ml) was added, tha pH wa~ adjusted to 2.5 and the clear yellow solution was stirr~d at ~.~ for 30 minut~s. The acetone was stripped in uo and ~he aqueau~ ~olution was reduc~d at 55~0 p.s.i. H2 pressur~o at 23 for 2 hours using 3 ~ 0 gm of 10% Pd on carbon as catalyst. ?he reduced solution was filtered through Dicalite and ~ 53 ~1~0~953 chromatographed as in Example 16B to give 47,50% amiXacin, 5.87~h BB-K29, 7.32% BB-K6, 24.26% kanamycin A and 7.41~ polyacyls.

Example 22 re~ration of Amikacin by the Acylat~on o Pol~(trLmethvLsil ~
amic Acid Catal~st To a re~luxing mixture of kanamycin A tS.0 gm, 10.32 m moles) in S0 ml of sieve-dried tetrahydrofuran (THF) were addPd sulfamic acid (100 mg) and HMDS ~12.32 gm, 76~33 m moles). Th~ mixture was re1uxed ~or 18 hours, with complete solution occurring af~qr 6 hours~ The solution was cooled to 23, treated with O,1 ml ~f water nd held at 23 ~or 30 minutes. ~solution of SAE ~3.61 gm, 1~.3 m moles) in 36 ml of THF was added over a period o 30 minutas.
~ter stirring for 3 hours, the mixture was diluted with 100 ml of water and the p~ was adjusted to 2,2 with 10% H~S04, It ~as stirred for 30 minutes at 23 and then concentra-ted in vacuo to remove T~F.
The resulting aqueous ~olution was reducad at 50 p.s.i. H2 p~essu e for ~ hours at 23 using 10% Pd on carbon as a catalyst. The reduced solution was filtered thro~gh Diclaite and the solids were washed with water. The combined filtrate and washings (lS0 ml) were deter~nined by microbiological assay against E. coli ~o contain 1225 mcg/ml (31~5~/o activity yield) of amikacin.

~.
.

~ _~.

AHBA

A. Preparati~n of DicarbobenzYloxY L-(-) a-Amino--hYdroxvbu'yric Dicarboban~yloxy L- ~-) a~amino-c~-hydxox:ybutyric acid (8 gmr . _ .

20.65 m moles) and N hydroxysuccinLmide ~2.37 gm, 20.65 m mole~) were dissolved in 50 ml of dry acetone at 2~. ~icyclohexyl-carbodiimide (4.25 gm, 20.65 m moles) dissolved in 20 ml of dry a~etone was added and the total was agitated at 23 for 2 hours, Dicyclohexylurea wa3 ~iltered o~, the filter cake was washed ~it~ 10 ml o~ dry acetone, and the ~iltrate and washings were combined .

B. Ac~,latlon Poly(trimethylsllyl) kan~nycin A, prepared according o the general procedure of ~xample 21 rom lO.O gms (20.639 m moles~
of kanamycin A, wa~ dissolved in 100 ml of dry acetone. The olution was cooled to 0-5, 3 . 7 ml of deionized water was added , and the solution was stirred at 05 ~or 30 minutes under moderate vacuum.

To this ~olution was added the ~olution of the di-Cbz-blocked acylating agent prepared in step A, and the mix~ure was ~irred at 0-5 for 30 minutes. The mixture was diluted with water, the pH was adjusted to 2O2 and ~he acetone was removed in vacuo.
The a~ueous solution was reduced ~y the general procedure of Example 22 and ~hen f iltered t~rough ~icalite . Chromatosraphy showed 40-45% amikacin, ca. 10% BB~K29, a 1:race of BB K6, ca. 30%
kanamycin A and a small amount of p~lyacyls.

E ~

~5~

Kanamyc:in A (11 gm, 22 . 7 m moles) and 100 mg of imidazole were hea~ed to re1ux in 100 ml of sieve-dried acetonitrile, under a nitroq2n pu~ge, ~SDS ~ 4~ gm, 114.5 m molesj, 5 moles per mole )l)9S3 ;-- of kanamycin A) was added over a period o 3 0 minutPs and t~e mixture was refluxed for 20 hours. Complets solution accurred in ca . 2-1~2 hours . Th~ golution was cooled to 23 ~ and the solvent was r~no~red ln vacuo to leave 21. 6 gms of poly ~trimethyl-silyl) kanamycin A as a foamy r~sidue ~93.1% yield, calculated as kanamycin (silyl) :~.1) .

~ 13B-X26) ~y the Ac lation o~ Pol (trimet am cin B
With SA~S

Ao _ .L~'~

Xanamycirl B (25 gm, 51.7 m moles) in 250 ml o~ sieve-dried acetor~ ile was heated to ref lux under a stream o~ nitrogen.
HMDS (S2.3 ~m, 385.81 m moles, 7.5 moles per mole of kanamycin B) was added over a period of 30 mi~utes foLlowed by 1 ml of TMCS
as catalys~. The mixture was refluxed for 21 hours with complete solution after 1 hour. The solven~ was then removed in vacuQ at 60 and the oily residue was held at 6û under high vacuum fox 3 hour~. There was obtained 53 ~ O gm o~ poly (tximethylsilyl) kanamycin B (85.2% yield, caloulated as kanamycin B (silyl) 10) .

B. A ~ ation The pcly ( tr~methylsilyl~ kanamycin B prepared in s~ep A above (53 . Q ~n) w~q dissolved in 500 ml of dry ac:e~one at 0-5, methanol (20.~ rnl) w3~ added, and the mixture was s~irred in vacuo for 30 minutss 2t 0-5. A colution of SA.E (18.1 gm, 51.67 m moles) in 200 ml of pre-cooled dry acetone ~as added over a period of less than 1 minu~s and the Tni~ture was stirred for 30 minu~es at 0-5.

, _ .

The mixture was worked up acco~ding to the general procedure of Example 22 and then load~d on a column o CG -S0 ~NH4+) ( 8 x 120 cm) .
It was eluted with ar, N}I4OH gradient from 0.6N to 3N. There was obtained 38~ of BB-~26, 5% of the corxesponding 6 '-N-acylated kanamycin B tBB-K~2), 10% of the corresponding 3~ acyla~ed kana-mycin B ~BB-K46) 14.63% l~anamycin B and a small amoun~ o poly-acylatecl kana~sycin B.

~ .

[~
Kanamycin A (19.5 gm, 40.246 m moles) and kanams~cin A sulfate (0.5 gmj 0.858 m mole) [total - 20.0 gm, 41.0 m molesl in 200 ml of sieve-dried acetorlitrile was brought to ref lux . E~DS ( 60 L 3 ml, 287.7 m moles, 7 moles per ~nole or^ kanamycin A) was slowly added and ~he mix~lare was r~fluxed for 28 hours. It was then stripped ~o dryness on a rotAry evaporator and dried under steam ir jector vacuum., There was obtained 47.5 gms of poly~trimethyl-silyl) kanamycin P. as a pale yellow oil (95.82~ yield, calculated as kanamycin A ~silyl) 1 Ester A. - _~_ N-~C~=~

To a su~pe~ion of AHBP~ (5.0 gm, 42 m ms:~les) in 100 ml T~F
was ad~ed tri~luoroacet.ic anhydride (40 gm, 191 m moles), with _ 57 --~ . ~

~lQ(~53 stirring, over a 10 minute period. The solution was stirred for 18 hours at 23 and then concentrated to dryness in vacuo at 50.
The residue wa~ dissolved ir~ 100 ml o~ aqueous methanol (1:1) and stirred for 1 hour. It was then ~oncentrated to dryness in va::uo and redissolved in 50 ml H 0. The aqueous solution was _. 2 extracted with 3 x 50 ml portions of MIBX and, after dryin~ over Na2SO,l,l, the extract was concentrated to an oil. Traces o~ solvent were remo~red by adding a~l dis tilling o~f 4 ml o water . On standing the oil changed to a waxy, crystalline solid ~2.5 ~m, 28~ yi~ld.
The N-tri~luoroacetyl ~H~A t2.4 sm, 11.3 m moles) was dissalved in 50 ml dry acetone and N hydroxysuccin~nide (1.30 gm, 11.31 m.moles) wa added to ~he solution. A solution of dicyclo-hexylcarbs~di~mide t2.33 gm) in 20 ml of dry acetone was slowly added . The reaction mixture was stirred f os 2 hours at 23 and the precipitated dicyclohexylurea was remo~red by iltration and washed with a small amount of acetone . The combined f iltrate and washings ~a solution of th ~-hydroxysuccinimide ester of N-trifluc~roace~yI ~BA) was u~ilized in ~he next step wi~hou~
isolation .

B. Ag~
To a solution of poly (tx~ethylsilyl) kanarnycin A prepared as in Examp}~ 26 (11.31 ~ mole~) in 54 ml of acetone was added 2,0 ml (113.4 m moles) o~ water, and the mix~ure was s~irred in ~racuo at 0-5 for 30 minutes. The N-hydroxysuccinimide ester of -- . ... . .
~-tri~luoroacetyl AIIBA prepared in step A above (11. 31 m ~oles) was added to the mix~ure and i~ was then held a~ 5 for 1 hour.
The p~I was then ad jus~ o ca . 2 . O wi~h 20~ ~I2S04, the mixture wa-~ sti~red for 30 minutes and the p~I was then raised to ca . 6 . O
with NH40~. The mix~ure was then stripped to dryr~ess in a rotary ~V9~3 evaporator to give 14.4 gm of a tacky of-white solid. The solid was dissolved in 100 ml of water, thP pH was raised from 5.5 to 11.0 with lON NH40H and th~ solution was heated in an oil bath at 70 for 1 hour. The p~I (9.5) was then lowered to 7.0 with HCl, the so~ution was polish filtered to remove a small amount of insolubles and the fiLte~ was washed with water. The oombined ~iltrate and washings (188 ml) was applie~ to a CG-SO (NII4~) column (8 x 90 cm), wa~hed with 2 liters of water and eluted with a ~3~401~ gradient ( O . 6N-l .ON-concentrated) . ~here was obtained 28,9% amikaci~, s~!O B~-K6, 5.7% BR-~29, ~3~8~o kanamycin A, 3.25 polyacyls plus 14.3% o~ an unknown material which was in the first ~raction off the column .

Xanamycin A ~ith t-Butyloxycarbonyl Blocked AHBA N~Hydroxysuccin~mlde Ester A. Prepara-tion of t-BOC AHB~ and Conversion to its N-Hydroxysuccinimide Es~er A solution of A~IBA ~S.O grn, 42 m mole~ in lOO ml of water and 20 ml oi~ ace~one was adju-qted ~o p~ lO with 10N ~laOH. Over a period of 3 4 minutes was added ll. 6 gm (53 m moles) of di-t-butyl dicarbonate, and the solution was s~irred for 35 minutes while maintaining ~he pH a~ 10 by ~he periodic addition of lON NaOH. ~he acetone wa~
removed i~ acuo and the aqueous phase was washed with 40 ml o ethyl acetate. The pH of the aqueous s~lution was lowered to 2.3 ~ith 3N HCl and it was then extrac.ed with 3 x 3 0 mL of MIBK. The combined MIBK Pxtrac~s were dried over Na2S04 and concentrated ~o a clea~ oi.l.y r~sidue f8.2 sm, 89~

The t-BOC-A~BA (4.25 gm, 19.4 m moles) was dissolved in 50 ml of acetone and N-hydroxysuccinimide ~2.23 gm, 19.4 m moLes) was added . A solution of dicyclohexylcarbodiimide ( 4 . 00 gm 19 . 4 m moles) in 20 ml of acetone was slowly added and the mixture was stirred for 2 hours at 23. The precipitated dicyclohexylurea was removed by filtration and was washed with a small amount o~
acetone. The combined filtrat~ and washings (a solution of ~he N-hydroxysuccinLmide ester of t-BOC-AHBA) was utilized in the next step without isolation.

B. Acylation To a solution of poly(trimethylsilyl) kanamycin A prepared as i~ Example 26 (41.28 m moles) in 94 ml of aceto~e was added 3.5 ml ~194 m moles) of water, and the mixture was stirred ln vacuo at 0 5 or 30 minutes. Th~ N-hydroxysuccin~mide es~er of t-BOC-AHBA prepared in step A abo~e (19.~ m moles) was add~d and the mixture was ~llowed to ~tand at 5 for 1 hour. Water ~200 ml) wa~ added and th~ pH (7.~) was lowered to 2. a with 20%
H2S04. After 30 minutes stirring the pH was raised to ca. 6.0 with N~40~ and the mix~ure was stripped to dryness ln vacuo to give 36.3 gms of a golden oil. The oil was dissolved in 200 ml o trifluoroacetic acid, allowed to stand for 15 ~inutes and stripped to dryness in a rotary svaporator. The oil was washed with water and the w~ter was flashed off. Concentrated ~H40H was added to pH 6~0 and w s ~lashed off. The ~esulting solid was dissol~ed in water, filtered, and the ~ilter washed with watar. The combined filtrate and ~ashings (259 ml) wer~ loaded on a CG-50 ~N~I4+) column (8 x 92 cm~, washed with 4 liters of watsr and eluted with an NH40~ ~ra~ient '0.6~-L.ON-concentrated~. Thera w~s obtained ~O.3~%
amikacin, 4.5~0 BB-K6, 8.32% BB-K29, 30.50% kanamycin A and 7.43%
c~ls .

The ~en~ral procedure of Example 1 is repeated, except that the 6'-N-carbobenzyloxykanamycin A used ther~in is replace~
by an equLmolar weight of ~'-N-carbo~enzyl~xykanamycin B~ a~d there is thereby p~oduced l~N-~L-(-)-y-amino-a-hydraxybutyryl]-kanam~cin B .

Examnle 30 The general procedure of Example 1 is repeated except t~at the L-(~-y-benzyloxycar~onylamino-~-hydroxybutyric acid N-hydroxy-S-norbornene-2,3-dicarboximide es~er used therein is replaced by L-~-)-B benzyloxycarbonylamino-a-hydroxypropionic acid N-hyd~oxy-5~norbornene-~,3-dicarboximide ester and L~ benzyloxycarbonylamino~-hydroxyvaleric acid N-hydroxyo 5-nor~ornene-2,3-dica~boximide ester, respect.ively, a~d ther~ is thereby produced 1-N- 1 L (-~ -B-amino-~ hydroxypr~pionyl~kanamyci~ A and L~ amino ~-hydro~y~aleryl~kanamycin A, respec~ively~ -5;~

~e~

The generaL procedure of Example 25 is repeated except that the L~ y-~en~yloxycarbonylamino-a-hydroxybutyric acid N-hydxoxy-ester used therein is replac~d by L~ benzyloxycarbonylamino--hydroxy propionic acid N-hydroxy-succinimide ester and L~ rbenzyloxycax~onylamino~a~ droxyvaleric acid ~ hydroxy-~uccinimide ester~ r~spectively, and ~here is thereby producad 1-N-1L~ amino-a~hydroxypropionyl~kanamycin B and l-N-~L (~ amino-a-hydroxyvaleryllkanamyciD B, respectively,

Claims (13)

1. The process for the preparation of a 1-N-[.omega.-amino-.alpha.-hydroxyalkanoyl]kanamycin A or B having the formula I
in which R is OH or NH2 and n is an integer of from 0 to 2, or a nontoxic pharmaceutically acceptable acid addition salt thereof which comprises acylating polysilylated kanamycin A or B or poly-silylated kanamycin A or B containing a blocking group other than silyl on the 6'-amino moiety with an acylating derivative of the acid of the formula in which n is an integer of from 0 to 2 and B is an amino-blocking group, in a substantially anhydrous organic solvent, and subsequently removing all blocking groups.

2. The process of Claim 1 wherein the acylating derivative of the acid is an active ester or a mixed acid anhydride.

3. The process of Claim 1 wherein the amino blocking group of the acylating derivative of the acid is selected from those of the formulae , , , , , and , wherein R1 and R2 are alike or different and each is H, F, Cl, Br, NO2, OH (lower)alkyl or (lower)alkoxy, and X is Cl, Br, F
or I and Y is H, Cl, Br, F or I.

4. The process of Claim 2 wherein the amino-blocking group of the acylating derivative of the acid is selected from those of the formulae , , , , , and , wherein R1 and R2 are alike or different and each is H, F, Cl, Br, NO2, OH, (lower)alkyl or (lower)alkoxy, and X is Cl, Br, F or I and Y is H, Cl, Br, F or I.

5. The process as claimed in Claim 1, 2 or 3 wherein the acylating derivative of the acid is its active ester with N-hydroxysuccinimide, N-hydroxy-5-norbornene-2,3-dicarboximide or N-hydroxyphthalimide.

6. The process as claimed in Claim 1, 2 or 3 wherein the acylating derivative of the acid is its active ester with N-hydroxysuccinimide, N-hydroxy-5-norbornene-2,3-dicarboximide or N-hydroxyphthalimide, and wherein the amino-blocking group of the acylating derivative of the acid is a carbobenzyloxy group.

7. The process of Claim 3 or 4 wherein the acylating derivative of the acid is its mixed acid anhydride with pivalic acid, benzoic acid, isobutylcarbonic acid or benzylcarbonic acid.

8. The process as claimed in Claim 1, 2 or 3 wherein the polysilylated kanamycin A or B contains a carbobenzyloxy group of the 6'-amino moiety.

9. The process as claimed in Claim 1, 2 or 3 wherein the polysilylated kanamycin A or B starting material contains an average number of silyl groups per molecule of from 4 to 8.

10. The process as claimed in Claim 1, 2 or 3 wherein the polysilylated kanamycin A or B starting material containing a blocking group other than silyl on the 6'-amino moiety also contains an average number of silyl groups per molecule of from 3 to 7.

11. The process as claimed in Claim 1, 2 or 3 wherein the silyl groups are trimethylsilyl.

12. The process for the preparation of a l-N[L-(-)-.omega.-amino-?-hydroxyalkanoyl]kanamycin A according to Claim 1, 2 or 3, wherein the acylating derivative of the acid is of the formula

13. The process for the preparation of a 1-N[L-(-)-.omega.-amino-?-hydroxyalkanoyl]kanamycin A according to Claim 4, wherein the acylating derivative of the acid is of the formula