BIOTECHNOLOGY PROCESSES
`Scale-up and Mixing
`
`Edited by
`
`CHESTER S. HO
`State University of New York at Buffalo
`Buffalo, N.Y.
`
`JAMES Y. OLDSHUE
`Mixing Equipment Company
`Rochester, New York
`
`American Institute of Chemical Engineers
`New York, New York
`
`PETITIONERS
`
`EXHIBIT NO. 1012 Page 1 of 5
`
`

`

`<24 %. J?
`& J?
`
`-a# ti
`
`t
`
`Copyright 1987
`American Institute of Chemical Engineers
`345 East 47th Street
`New York, New York 10017
`
`AlChE shall not be responsible for statements or opinions advanced in papers
`or printed in its publications.
`
`Library of Congress Cataloging-in-Publication Data
`
`Biotechnology processes.
`Includes index.
`1. Biochemical engineering—Congresses. 2. Bio­
`technology—Methodology—Congresses.
`I. Ho, Chester S.,
`. II. Oldshue, James Y. III. American
`1950-
`Institute of Chemical Engineers.
`660'.6
`1987
`TP248.3.B62
`ISBN 0-8169-0410-3
`
`87-14393
`
`MANUFACTURED IN THE UNITED STATES OF AMERICA
`
`PETITIONERS
`
`EXHIBIT NO. 1012 Page 2 of 5
`
`

`

`Modeling the Dynamic Behavior of Immobilized Cell/Enzyme Bioreactors
`
`TABLE 1
`Experimentally Determined Diffusivities of Some Fermentation Substrates and Products (X106 cm2/s).
`
`Type of experiment
`
`Glucose
`
`Lactose
`
`Nil^ CI
`
`Patulin
`
`6-MSA
`
`Cell-free beads
`
`With dead Cells
`
`With live Cells
`
`6.68
`6.68
`
`7.58
`
`6.4
`6.0
`
`7.2
`
`55
`
`45
`
`6.28
`
`5.03
`
`2.92
`
`1.95
`
`XVni' l t ) = Zv-C*) +
`
`"o
`
`1
`
`1
`
`n0
`
`"o
`= ZV-('/f) +
`
`1
`
`1
`ZDR n2 diR
`
`y(—)2
`„o n
`
`1
`
`1
`
`00
`
`i
`
`no 1
`
`.1 " 1 n
`
`and finally
`no
`ZVnC'/f) = ZVn(Of) +
`) Z(-)2 - Z(-)2
`2 ("7^"
`tDR K
`dtR
`00 1
`K
`Using the property Z(—)2 = T- and defining:
`6
`! n
`I\2
`
`"O
`
`^ =
`
`and
`
`s, = E(-):
`
`1 n
`
`6
`
`1 n
`
`dCb
`
`2 ^
`= ( j ) Z V n C )
`
`(A-13)
`
`Substitution in Eqn. A-l yields:
`
`D
`e — = £)(€.„ - CL) - 6(1 - e)(^-) 5>,(0(A-14)
`at
`V
`1
`A
`
`Evaluation of the infinite summation^ V«(0
`1
`
`f »
`
`. 1 x ^ 6
`
`(A-15)
`
`Equation A-12 can be written as
`d\if
`1
`n
`(—) —
`kn
`dt
`depend of the
`In Eqn. A-15, the dynamics of the functions
`value of the coefficient x = (!/£„) ; small values of x indicate
`that
`the
`functions V„(0
`follow
`the
`forcing
`function
`{\/kn)(d$ldt). It is therefore reasonable to assume that for
`x < x0, the v/s can be evaluated using the quasi-steady state
`approximation
`
`, . , 1 , d<|)
`
`or, using dimensionless quantities
`d§
`1
`("T^)
`E Dr n2 n2 ' dlR
`
`V«(fR) =
`
`(A-16)
`
`(A-17)
`
`The value of n0 is determined as the one corresponding to XQ:
`1
`ZDr K2 (no2)
`
`— Xo
`
`or
`
`1
`)a5
`1
`"0 = (—) (
`t DR XQ
`n
`It was found by numerical experimentation, that for values of Xo
`less than 0.01, the resulting response curves, for given DR were
`identical to five significant figures. Subsequently, a value of
`x0 = 0.01 was used in all the runs.
`
`After the above considerations, the infinite summation Zv* (') is
`i
`estimated as follows:
`
`Dispersal of Insoluble Fatty Acid Precursors
`in Stirred Reactors as a Mechanism
`to Control Antibiotic Factor Distribution
`
`FLOYD M. HUBER
`RICHARD L. PIEPER
`ANTHONY J. TIETZ
`Fermentation Technology Department
`Eli Lilly and Company
`Indianapolis, Indiana 46285
`
`Biosynthesis of factors in the A-21978C antibiotic complex was controlled by
`addition of appropriate fatty acid precursors. Toxicity associated with higher fatty
`acids was avoided by continuous addition of the fatty acid at a rate nearly equal to
`the uptake by the producing organism. Inability of the producing organism to assimilate
`a solid, insoluble long-chain fatty acid was overcome by dissolution of the acid in
`another substrate.
`
`Substance A21978C is a complex of antibiotics
`produced by Streptomyces roseosporus, having
`a common cyclic polypeptide nucleus and dif­
`ferent fatty acid side chains (JO (2)
`(Figure 1). Separation of the major factors
`CI, C2, and C3 revealed differences in both
`in vitro antibiotic activity and toxicology
`for the different naturally occurring com­
`pounds. Deacylation of the alkanoyl side
`chain, followed by reacylation with a series
`of fatty acids indicated the n-decanoyl sub­
`stitution at position R, resulted in the best
`therapeutic potential (3) (40 C5).
`The natural occurrence of the n-decanoyl
`factor, designated LY146032, was too low to
`
`permit isolation in sufficient quantity
`directly from fermentation broth (Table 1).
`
`In an attempt to produce LY146032 by
`fermentation, the addition of decanoic acid
`during the antibiotic production phase of
`the fermentation was proposed. Initial
`efforts in shaken cultures were unsuccessful
`due to either the toxicity or insolubility
`of the fatty acid. In this paper we will
`describe efforts to direct the biosynthesis
`of A21978C factors in continuously stirred
`reactors operating in a fed-batch mode.
`MATERIALS AND METHODS
`A mutant strain of Streptomyces
`roseosporus NRRL 11379 was used to inoculate
`50 ml of vegetative medium of the following
`composition: Trypticase soy broth
`(Baltimore Biological Laboratories,
`Baltimore, Maryland), 30 mg/ml; potato
`dextrin, 25 mg/ml. The inoculated medium was
`incubated for 48 hours at 30oC in a 250 ml
`Erlenmeyer flask on a shaker rotating through
`an arc of two inches in diameter at 250 RPM.
`One-half ml of the mature vegetative culture
`was dispensed into multiple containers and
`stored in the vapor phase of liquid nitrogen.
`One ml of the stored culture was used to
`inoculate 800 ml of the vegetative medium
`described above. The inoculated vegetative
`medium was incubated in a 2000 ml Erlenmeyer
`flask at 320C for 120 hours on a shaker
`
`249
`
`we obtain
`"o
`+ ^57 <^> <•" T> <A-18>
`»•<")"
`i
`i
`The concentration at the surface, <}i(0. can be expressed in terms
`of the bulk concentration CL{t) using the partition coefficient
`(p), namely (KO?) = CL )/p
`
`Hence Eqn. A-18 becomes:
`"o
`Zv* ('*) = !>„('«) +
`i
`i
`Subsequent substitution Eqn. A-19 into A-14 yields:
`*0
`
`l
`6 (3 e
`
`dtR
`
`(1 - -rf )(A-19)
`S,
`
`dCL
`dtR
`
`1 +
`
`i
`
`Cin - CL - 6ep(l - e ) DR
`EpO - £)
`SN
`0 -TT-)
`p £
`S,
`
`(A-20)
`
`TABLE 1
`Distribution of Naturally Occurring Factors in A21978C
`Fermentation
`
`The normalized form of Eqn. A-12 is:
`
`=- e (n n )2 Dr
`
`dtR
`
`+ (-^) {~T
`P
`dtR
`
`).
`
`n = 1,2,3,....no.
`(A-21)
`Solution by numerical integration of Equations A-20 and A-21
`yields the dynamic response of the reactor.
`
`% of Total
`A21978C
`Complex
`27.3
`40.1
`25.5
`
`A21978C
`Factor
`
`CI
`C2
`C3
`C5
`LY146032
`
`Concentration
`yg/ml
`77
`113
`72
`trace
`trace
`282
`Eli Lilly and Company, Indianapolis,
`Indiana.
`
`248
`
`BIOTECHNOLOGY PROCESSES
`
`BIOTECHNOLOGY PROCESSES
`
`PETITIONERS
`
`EXHIBIT NO. 1012 Page 3 of 5
`
`

`

`Dispersal of Insoluble Fatty Acid Precursors
`
`Floyd M. Huber, Richard L. Pieper, and Anthony |. Tietz
`
`rotating through an arc of two inches in
`diameter at 250 RPM. The entire contents of
`the two flasks (approximately 1400 ml after
`incubation) were used to inoculate 1900 liters
`of a secondary vegetative stage having the
`following composition (mg/ml): soybean flour,
`5.0; yeast extract (Difco Laboratories,
`Detroit, Michigan), 5.0; calcium gluconate.
`10.0; KC1, 0.2; MgSO^yHzO, 0.2; FeSOWHaO,
`0.004; Sag 471 antifoam (Union Carbide,
`Danbury, Connecticut). The potassium,
`magnesium, and ferrous salts were prepared
`separately as follows: 7.6 g FeSGi,*71120 was
`dissolved in 76 ml of concentrated HC1. 380 g
`of MgSOi, *71120 and 380 g of KC1 and deionized
`water were added to bring the total volume to
`3800 ml. The inoculated medium was incubated
`24 hours in a stainless steel vessel at 30oC.
`The vessel was aerated at 0.85 v/v/m and
`stirred with conventional agitators.
`The mature secondary seed (8.33% v/v)
`was used to inoculate a production medium of
`the following composition (mg/ml): soybean
`flour, 22.0; Fe(NHi4)2S0it*6H20, 0.66; glucose
`monohydrate, 8.25; Sag 471, 0.22; potato
`dextrin, 33.0; and molasses (blackstrap).
`2.75.
`Two types of stirred reactors were used.
`The smaller vessel, operated at 120 liters,
`was agitated with two conventional flat
`Rushton type impellers at relatively high
`power input. The larger vessel, operated at
`4550 liters, was equipped with impellers
`having curved paddles and was operated at
`relatively low power input. Air flow in both
`reactors was supplied at 0.5 v/v/m by large
`open tubes which were estimated to contribute
`very little to the overall mixing. Respira­
`tion rates were estimated by difference in
`inlet and exhaust gas concentration via a
`Perkin-Elmer mass spectrometer. Distribution
`of A21978C factors was estimated by high
`performance liquid chromatography as
`described previously C2).
`Examination of the batch fermentation
`medium suggested that the growth limiting
`nutrient was carbon in the form of carbohy­
`drate. It was then hypothesized that in a
`fed-batch operation a moderately toxic sub­
`strate, such as decanoic acid, could be fed
`continuously to the fermentation if the
`metabolic consumption rate exceeded the
`addition rate.
`Delivery of decanoic acid to the
`culture presented a problem. With a melting
`point of 340C the compound is a solid at the
`fermentation temperature of 30oC, and the
`
`compound has very low solubility in water.
`In order to avoid the obvious problems aris­
`ing in supplying a limiting nutrient as a
`solid phase, the substrate was dispensed to
`the stirred reactor as a five percent solu­
`tion dissolved in a fifty percent ethanol/
`water mixture. There was an immediate
`response in oxygen uptake to the onset of
`the decanoic acid feed, as illustrated in
`Figure 2. Also, a significant improvement
`in LY146032 concentration was immediately
`realized (Table 2).
`
`TABLE 2
`Distribution of A21978C Factors with Decanoic Acid Feed3
`
`% of Total
`A21978C
`Complex
`19.8
`29.9
`11.5
`5.2
`33.5
`
`A21978C
`Factor
`CI
`C2
`C3
`C5
`LY146032
`
`Concentration
`yg/ml
`72
`109
`42
`19
`122
`364
`(a) N-decanoic acid/ethanol/water 1:2:2 fed
`50 ml per hour to 120 L operating volume.
`Material balances suggested that only
`a small portion of the decanoic acid that
`was fed could be accounted for by incorpora­
`tion into the product. Thus, most of the
`fatty acid was apparently catabolized,
`presumably by the beta-oxidation pathway.
`In an attempt to increase the amount of
`decanoic acid available for the incorpora­
`tion, the concentration of fatty acid in the
`
`TABLE 3
`Distribution of A21978C Factors with Increased Decanoic Acid
`Feed3
`
`% of Total
`A21978C
`Complex
`10.4
`15.0
`8.5
`4.1
`62.1
`
`A21978C
`Factor
`CI
`C2
`C3
`C5
`LY146032
`
`Concentration
`yg/ml
`131
`189
`107
`52
`784
`1263
`(a) N-decanoic acid/ethanol/water 1:2:2 fed
`50 ml per hour to 120 L operating volume.
`
`feed was increased to twenty percent—the
`solubility limit of decanoic acid in aqueous
`ethanol. A significant increase in both the
`LY146032 and total yield was observed
`(Table 3).
`
`Although the fed-batch fermentation
`employing the fatty acid/ethanol/water addi­
`tion proved an effective method of directing
`the synthesis of the A21978C complex, the
`presence of the volatile alcohol presented
`both safety problems and uncertainties in
`quantitating the carbon balance. It was not
`known to what extent the producing organism
`could metabolize ethanol, since it was likely
`that much of the alcohol was escaping in the
`exit gases. Methyl oleate was identified in
`batch shaken cultures as a metabolizable,
`non-toxic and low volatile solvent. A mix­
`ture of equal volumes of methyl oleate and
`decanoic acid remained liquid at 30oC. The
`results of feeding the mixture of decanoic
`acid dissolved in methyl oleate on an equiva­
`lent basis to the previously used aqueous
`ethanol feed resulted in a slightly higher
`total yield with a similar concentration of
`the desired LY146032 (Table 4).
`
`TABLE 4
`Distribution of A21978C Factors in a Decanoic Acid/Methyl
`Oleate Fed Fermentation3
`
`% of Total
`A21978C
`Complex
`
`10.6
`13.1
`8.7
`4.7
`62.8
`
`A21978C
`Factor
`CI
`C2
`C3
`C5
`LY146032
`
`Concentration
`yg/mi
`154
`191
`127
`68
`913
`1453
`(a) N-decanoic acid/methyl 1:1 oleate fed
`13 ml per hour to 120 L operating volume
`
`SCALE-UP CONSIDERATIONS
`The initial scale-up of the process to
`a larger 6000 L pilot-scale stirred reactor
`did not produce equivalent LY146032 factor
`distribution (Table 5).
`
`The percentage of LY146032 was approxi­
`mately one-half of that obtained in the
`smaller equipment. It was hypothesized that
`the microbial population was oxidizing the
`decanoic acid preferentially over the methyl
`oleate, and due to the poorer mixing in the
`
`250
`
`BIOTECHNOLOGY PROCESSES
`
`BIOTECHNOLOGY PROCESSES
`
`TABLE 5
`Distribution of A21978C Factors with Decanoic Acid/Methyl
`Oleate Fed to a Larger Reactor3
`
`% of Total
`A21978C
`Complex
`17.0
`22.1
`17.7
`11.1
`32.0
`
`A21978C
`Factor
`CI
`C2
`C3
`C5
`LY146032
`
`Concentration
`yg/mi
`260
`337
`270
`170
`489
`1527
`(a) N-decanoic acid/methyl oleate 1:1 fed
`490 ml per hour to 4550 L operating
`volume
`
`large vessel, the uptake of the fatty acid
`was not equivalent throughout the population
`in the reactor (Table 6).
`
`TABLE 6
`Mixing Time and Power Input in Pilot-Scale Equipment Used in
`A21978C Fermentation
`(a)
`Power Input
`Vessel Size Mixing Time
`(Hp/100 Gal.)
`(Seconds)
`(1)
`1.36
`150
`7
`6000
`0.82
`43
`(a) Time required for pH to reach equilib­
`rium after addition of a sufficient
`quantity of 6M NaOH to increase pH by
`0.3
`The first approach to achieve better
`incorporation of the fatty acid precursor in
`the poorly mixed larger vessel was to
`increase the feed rate of the decanoic acid/
`methyl oleate mixture. The percentage of
`LY146032 increased to nearly that achieved
`in the smaller equipment, but the total yield
`of A21978C was reduced to about two-thirds of
`the amount achieved at the lower feed rate
`(Table 7).
`
`The reduction in overall yield associ­
`ated with the increased feed of the decanoic
`acid precursor was believed to be a result of
`the feed rate of the fatty acid approaching
`the metabolic rate of consumption. The con­
`sequences of introducing the toxic fatty acid
`
`251
`
`PETITIONERS
`
`EXHIBIT NO. 1012 Page 4 of 5
`
`

`

`D-Ala
`
`/
`1
`\
`
`L-Asp
`
`L-Orn
`
`Gly
`
`3-MeGlu (L-threo)
`
`Gly
`\
`D-Ser
`\
`/
`L-Kyn
`oy" 0 O
`
`.50
`
`| .40-
`IJ
`o
`E .30
`5
`a
`iS
`5. .20
`
`D s
`g .10
`x
`o
`
`Start
`Decanoic Acid
`Feed
`
`(
`
`15
`
`30
`
`45
`
`Hours
`
`60
`
`75
`
`FIGURE 3. The toxic effect of overfeeding decanoic acid.
`
`L-Asp
`
`L-Thr
`
`L-Asp
`
`L-Asn
`
`I
`t
`t
`t
`
`L-Trp
`
`NH
`/ w
`R 0
`
`Factor
`
`A21978Co
`A21978C,
`A21978C2
`A21978C3
`A21978C4
`A21978C5
`
`* mixture of two isomers
`
`R
`
`CJO alkanoyl*
`8-methyl decanoyl
`10-methyl undecanoyl
`10-methyl dodecanoyl
`0,2 alkanoyl t
`0,3 alkanoyl t
`t structure uncertain
`
`FIGURE 1. Naturally occurring A21978C factors.
`
`.50-
`
`X
`
`.40-
`
`Decanoic Acid
`Overfed
`
`-30-
`
`E
`«
`9
`je
`a
`| .20-
`>
`X o
`
`.10-
`
`Feeder #1
`Feeder #2
`
`D
`D
`
`>JJ
`
`O
`O
`
`0.00
`0.00
`
`T
`15.00
`
`T
`T
`30.00
`45.00
`Hours
`
`T
`60.00
`
`75.00
`
`FIGURE 2. The effect of n-decanolc acid feed upon oxygen uptake.
`
`FIGURE 4. Feeding of n-decanolc acid/methyl oleate to a large reactor
`via two entry points.
`
`252
`
`BIOTECHNOLOGY PROCESSES
`
`BIOTECHNOLOGY PROCESSES
`
`Dispersal of Insoluble Fatty Acid Precursors
`
`Floyd M. Huber, Richard L. Pieper, and Anthony J. Tietz
`
`TABLE 7
`Distribution of A21978C Factors with Increased Decanoic Acid/
`Methyl Oleate Feda to a Larger Reactor
`
`A21978C
`Factor
`CI
`C2
`C3
`C5
`LY146032
`
`% of Total
`A21978C
`Complex
`11.9
`11.5
`8.1
`9.1
`59.4
`
`Concentration
`yg/ml
`122
`117
`83
`93
`607
`1022
`N-decanoic acid/methyl oleate 1:1 fed 590
`ml per hour to 4550 L operating volume
`
`(a)
`
`at a rate greater than the microbial consump­
`tion rate are illustrated in Figure 3. In
`this reactor the feeding system failed,
`resulting in significant overfeeding of the
`fatty acid. The resulting accumulation of
`the fatty acid caused rapid lysis of the
`culture as evidenced by the rapid cessation
`of respiration.
`A second approach to solve the mixing
`problem in the large fermenter was to intro­
`duce the fatty acid/ester mixture at two
`widely separated entry points. One entry
`point was at the top of the vessel and a
`second was in between the two impellers
`(Figure 4). The subsequent LY146032 concen­
`tration obtained (Table 8) indicated that a
`greater incorporation of the n-decanoyl side
`
`TABLE 8
`Distribution of A21978C Factors with Decanoic Acid/Methyl
`Oleate Fed to a Larger Reactor via Two Entry Points3
`
`% of Total
`A21978C
`Complex
`5.3
`8.5
`7.7
`1.1
`77.4
`
`A21978C
`Factor
`C1
`C2
`C3
`C5
`LY146032
`
`Concentration
`yg/ml
`75
`119
`108
`16
`1090
`1408
`(a) N-decanoic acid/methyl oleate 1:1 fed 212
`ml per hour to each of two entry points.
`4550 L working volume.
`
`chain had occurred as a result of the dual
`introduction.
`
`In summary, the following scale-up prob­
`lems associated with the production of
`LY146032 have been addressed and solved:
`A sound strategy was provided to
`feed the microbial culture a very toxic
`substance.
`A mechanism was devised to supply
`the normally solid substrate to the reactor
`in a convenient liquid form.
`Preferential substrate utilization
`was identified and the necessary feeding
`hardware was constructed to promote homogen­
`eous utilization of lipoidal materials when
`the fermentation was scaled up to larger
`reactors.
`
`LITERATURE CITED
`Hamill, R.L. and M.M. Hoehn, U.S. Patent
`4,208,403 to Eli Lilly and Company
`(June 17, 1980).
`Debono, Manuel, U.S. Patent 4,399,067 to
`Eli Lilly and Company (August 16, 1983).
`Debono, Manuel, B.J. Abbott, V.M.
`Krupinski, R.M. Molloy, D.J. Berry, F.T.
`Counter, L.C. Howard, J.L. Ott and R.L.
`Hamill, "Synthesis and Structure Activity
`Relationships of New Analogs of the New
`Gram Positive Lipopeptide Antibiotic
`A21978C," Abstract 1077, Interscience
`Conference on Antimicrobial Agents and
`Chemotherapy (ICAAC) (October 1984).
`Fukuda, D.S., B.J. Abbott, D.J. Berry,
`L.D. Boeck, Manuel Debono, R.L. Hamill,
`V.M. Krupinski and R.L. Molloy, "Deacy-
`latipn and Reacylation of A21978C,
`Acidic Lipopeptide Antibiotic: Prepara­
`tion of New Analogs," Abstract 1076,
`Interscience Conference on Antimicrobial
`Agents and Chemotherapy (ICAAC)
`(October 1984).
`Counter, F.T., P.J. Baker, L.D. Boeck,
`Manuel Debono, P.W. Ensminger, R.L.
`Hamill, V.M. Krupinski, R.M. Molloy and
`J.L. Ott, "LY146032 [N-(n-decanoyl)
`A21978C Nucleus], a New Acidic Lipopep­
`tide Antibiotic: Synthesis and Biological
`Evaluation," Abstract 1078, Interscience
`Conference on Antimicrobial Agents and
`Chemotherapy (ICAAC) (October 1984).
`
`!
`
`^ j
`
`PETITIONERS
`
`EXHIBIT NO. 1012 Page 5 of 5
`
`