APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Apr. 1976, p. 569-575

Vol. 31, No. 4

Printed in U.S A.

Copyright ©) 1976 American Society for Microbiology

Lipolytic Fermentations of Stickwater by Geotrichum candidum and Candida lipolytical JOHN H. GREEN,* STEFAN L. PASKELL,2 AND DANIEL GOLDMINTZ National Oceanic and Atmospheric Administration, National Marine Fisheries Service, Southeast Utilization Research Center, College Park, Maryland 20740 Received for publication 1 December 1975

Stickwater, a by-product of the fish meal and oil industry, is an aqueous suspension of fish proteins, lipids, and other materials, and also contains soluble nonprotein nitrogen but no carbohydrate. It is usually partially evaporated by heat to a marketable form called "fish solubles," which is sold with an acid preservative as an animal feed supplement. However, fish solubles are only used to a limited extent in feeds, because the lipids of solubles (averages 11%) are relatively prone to oxidative rancidity development. An investigation was undertaken to digest and/or stabilize lipids in stickwater by lipolytic fermentations and, at the same time, to attempt to increase the protein content as single cell protein. Strains of the yeast Candida lipolytica and the yeast like mold Geotrichum candidum were employed for these investigations. Stickwater fermentations were performed in a laboratory bench top fermentor. Respirometric studies of lipid metabolic activity and microbial observations were periodically performed during these fermentations. Rapid microbial growth and metabolic activity were observed in well aerated cultures. Fermented products were evaluated for chemical composition. Lipid residues were characterized by thin-layer chromatographic procedures. There was evidence of abundant microbial growth, increased lipolytic activity, and decreased lipid content. However, evidence was lacking to show that the protein content of stickwater was actually increased.

Stickwater, a by-product of the fish meal and oil industry, is the aqueous portion remaining after free oil has been removed by centrifugation of the press liquor and contains fish proteins, nonprotein nitrogen, minerals, residual lipid, and other materials. Stickwater is usually acidified with H2SO4 to a pH of approximately 5 or less and evaporated to a 1:1 (watersolids) slurry called "fish solubles." The characteristics of fish solubles, often used as a feed supplement, have been described by Soares et al. (9). The high content of unsaturated lipids in fish solubles makes them undesirable for extensive storage and limits their use in animal feeds due to oxidative rancidity development. The background of this project is associated with the National Marine Fisheries Service's interest in developing fish protein concentrate from fatty fish by biological methods. The lipolytic fermentation of Atlantic menhaden (Brevoortia tyrannus) was described by Burkholder et al. (5). A. Chu (Ph.D. thesis, Columbia Univ., New York, 1969) described his research on the physiology and lipid metabolism of Geo-

trichum candidum and other lipolytic yeasts used in this fish fermentation project. The research work performed by the Columbia University group suggested to us that lipolytic yeasts might be used to develop single cell proteins from fish lipids. Lipolytic yeasts could also be used to reduce the amount of undesired residual lipid in fishery by-products, such as stickwater, and to produce single cell proteins at the same time. We proceeded to investigate these possibilities at the National Marine Fisheries Service Laboratory (J. H. Green, S. L. Paskell, and D. Goldmintz, Intl. Congr. Microbiol., 10th, Mexico City D. F., Mexico, 1970, abstr. Ga-8). Our preliminary investigations into the growth of G. candidum and Candida lipolytica on a refined menhaden fish oil-watermineral-ammonium salts medium showed that menhaden oil could serve as the sole energy and carbon source and that ammonia could serve as a nitrogen source for microbial growth. Stickwater contains virtually no carbohydrates. Li et al. (8) reported on lipolytic fermentation by C. lipolytica and G. candidum on Great Lakes chub (Leuicichthys hoyi). H. H. Hottin' Southeast Utilization Research Center report CP-322. 2 Present address: Puget Sound Blood Center, Seattle, ger (M.S. thesis, Univ. of Wisconsin, Madison, 1972) and Hottinger et al. (7) have described the Wash. 98104. 569

570

GREEN, PASKELL, AND GOLDMINTZ

utilization of alewife (Alosa pseudoharengus), fish oil, and stickwater by C. lipolytica and G. candidum. The objectives of this investigation were to reduce the lipid content, produce single cell proteins, and improve Atlantic menhaden stickwater as a feed supplement. MATERIALS AND METHODS Microorganisms used. G. candidum strains LGO C-4 and LGO-PSM-179 and C. lipolytica strain NRRL-Y-57, obtained from 0. A. Roels of the Marine Biology Laboratory, Lamont-Doherty Geological Observatory of Columbia University, New York, were described by A. Chu (Ph.D. thesis, Columbia Univ., New York, 1969). C. lipolytica, strain no. 6026, was obtained from H. J. Pfaff, Univ. of Calif., Davis. All cultures were maintained on slants of potato dextrose agar (Difco Laboratories, Detroit, Mich.) adjusted to pH 5.6. Stickwater media preparation. The stickwater medium used for all work described below was prepared from the same lot of Atlantic menhaden solubles (Brunswick Navigation, Inc., Beaufort, N.C.) by reconstituting with distilled water (1:5), adding 1 g of KH2PO4 per liter, and neutralizing with 6 N KOH or 6 NH3PO4 to pH 4.5. A 200-ml portion was placed into a 1.5-liter flask, sterilized, and used to grow the final inoculum. A 9-liter portion was placed into three or four 4-liter aspirator bottles or 3liter flasks and sterilized (121 C, 45 min). This latter portion was aseptically added to a sterile 14-liter fermentor tank of a lab bench-scale fermentor system (Fermentation Design, Inc., Allentown, Pa.). Inoculum. The inoculum for each fermentation was started by transferring the desired strain from a fresh agar slant culture into a 250-ml flask containing 100 ml of a medium consisting of 1% peptone (Difco) in 0.85 M phosphate buffer (pH 7.0) plus 10 ml of a 5% oleic acid in a 5% gum arabic emulsion. The flask was placed on a shaker at room temperature for 24 h. From this starter culture 10 ml was transferred to 200 ml of stickwater medium and placed on the shaker for 24 h. The entire 200 ml was used to inoculate the 9 liters of stickwater medium in the fermentor jar. Shake-flask culture methods. The fermentation of stickwater medium in shake-flask cultures used methods similar to those described by A. Chu (Ph.D. thesis, Columbia Univ., New York, 1969). We used a reciprocal shaker (box carrier type, Arthur H. Thomas Co.) adjusted to 120 rpm and set into a BOD (Hotpack, Philadelphia, Pa.) at 25 C. Fermentation and controls. Fermentation was maintained for 48 h with an air flow rate of 1,000 cm:'/min, coupled with maximum obtainable agitation (800 rpm.). The temperature was controlled to 25 C. It was necessary to add a silicone-based antifoam (F.G.-10, Dow Corning Corp., Midland, Mich.). At the end of 48 h the fermentation was stopped, and the fermented stickwater was pasteurized in situ at 80 to 85 C for 2 h. The source, preparation, and treatment of stickwater and samples were maintained constant throughout these experiments so that observed dif-

APPL. ENVIRON. MICROBIOL.

ferences could be ascribed to the growth rates and metabolism of the various lipolytic microorganisms employed. Comparisons of fermented stickwater were made to a sterile stickwater control that had undergone the same preparation and treatment but without the inoculation of microorganisms. Sampling procedures. Samples were periodically removed by aseptic technique for microscopic and respirometric examination. At the end of each fermentation experiment, including the control, the pasteurized slurry was removed, placed in shallow pans, and freeze dried, and portions of the freezedried material were removed for chemical analysis. Viable plate counts using potato dextrose agar (pH 5.6) and microscopic observations of wet mounts were made periodically on all fermentation experiments. For G. candidum strains, an initial 1:10 dilution in a blank test tube was vigorously agitated on a Vortex mixer for 30 s to break up multiple-cell mycelia prior to subsequent dilution and plating. Respirometric methods. A differential respirometer (Gilson Medical Electronics Inc., Middleton, Wis.) was used, and the following simple manometric procedure was employed. The center well of a Warburg reaction vessel was filled with 0.2 ml of 6 N KOH plus a thin-folded filter paper strip (approximately 6 by 50 mm) as a wick to remove CO2 from the cup's atmosphere. Fermenting stickwater was placed in the cup of the vessel. To the vessel sac was added either 0.3 ml of 5% oleic acid in a 5% gum arabic aqueous emulsion for exogenous or 0.3 ml of distilled water for endogenous reactions, respectively. Paired vessels were used and averaged for each reading; reactions were run at 30 C with maximum possible shaking. Volume changes were read directly from the respirometer's volumometers and then corrected later for atmospheric conditions according to the manufacturer's directions. Chemical analysis. Unless otherwise stated, current procedures described by the Association of Official Agricultural Chemists (AOAC) (3) were used. (i)Protein. Kjeldahl N x 6.25 (AOAC Section 2.044). (ii) Ash. Samples (1 or 2 g) were preheated at a lower temperature to drive off lipid material and then ashed overnight in a muffle furnace at 550 C (AOAC Section 18.008). (iii) Nonprotein nitrogen. A cold trichloroacetic acid (10%) precipitation method was adapted from Bell (4). (iv) Ammoniacal nitrogen. Kjeldahl N determination was performed without prior digestion (AOAC Section 2.050). (v)

Nitrate/nitrite. Xylenol method (AOAC Section 23.010). (vi) Moisture. Oven-dried samples (AOAC Section 23.005). (vii) Total fat. Chloroform-methanol extraction (2). (viii) Ether fat. Ethyl ether 16-h extraction (AOAC Section 22.033). (ix) Amino acids. Determined by Beckman/Spinco amino acid analyzer method. Glyceride analysis by thin-layer chromatography. To obtain the stickwater glycerides from each fermented or control freeze-dried sample, a 5-g portion was extracted twice with 25-ml volumes of CHClI:, dried over anhydrous Na2SO4, and filtered, and the extract was stored in glass-stoppered flasks at 4 C. The resulting 10:1 (vol/w) extract was

LIPOLYTIC FERMENTATIONS OF STICKWATER

VOL. 31, 1976

used in volumes of 2 to 6 ,u for all thin-layer chromatography work. In general, procedures recommended by Dieckert and Reiser (6) were used to separate mono-, di-, and triglycerides. Prepared thin-layer chromatographic sheets impregnated with silicic acid and silica gel (Gelman Instrument Co., Ann Arbor, Mich.) were used. Reference lipids (Sigma Chemical Co., St. Louis, Mo.) were employed. Chromatograms shown in this report were sprayed with 90% sulfuric acid and charred for 10 min at 230 C. Densitometer readings of each sample chromatogram were made with an automatic scanning, recording, and integrating instrument (Gelscan, Gelman Instruments Co.), and relative concentrations were interpreted from integrator data. RESULTS

571

9

a

7

6 2

u15 .2

e

,, E

4

za J 3

In preliminary studies on the stickwater fermentation process using a 14-liter fermentor, it 2 was observed that for the same air flow rate more cells were produced by rapid rather than by slow agitation. It was also observed that more viable cells were produced and at a faster 0 30 40 50 60 20 10 growth rate than in our previous experience Hrs of Culture with shake-flask culture methods; a typical FIG. 1. Growth of G. candidum LGO-C-4 at 23 to comparison appears in Fig. 1. It is evident that aeration is important in growth of these lipo- 25 C. Comparison of the rate and amount ofgrowth lytic microorganisms. Hence, the high agita- between the shake culture and 14-liter fermentor cultion rate, coupled with the high air input, was ture under conditions described in the text. used for all fermentations. TABLE 1. C. lipolytica fermentation of stickwater: Microscopic observations of both shake and respirometric oxygen uptake measurement ofaliquots aerated fermentor cultures were made periodiof cultural medium at increasing hours of culture cally. In G. candidum cultures, transformation Oxygen uptake (,jl/ml/h) from multiple-cell mycelia to free arthrospores started at the time when maximum growth was Culture time (h) Endoge- Relative reached, which was approximately 20 to 24 h nous exogenous True exog(strain no. 60-26) (stickwa- (0.5% oleic enougs after inoculation. By 48 h observations of microter acid lipscopic fields revealed nearly all arthrospores. added) ids) In C. lipolytica spore stages were not discerned, whereas budding yeast cells were observed at 0 (uninoculated) 26 1.5 24.5b 37 44.5 7.5 0.5 all times. 7.0 198 191 The rapid agitation and aeration caused a 7.5 120 835 955 20.5 severe foaming of the stickwater medium. The 820 1080 260 foam formation greatly increased, occasionally 27 888 452 336 beyond control, as the maximum number of 48 viable cells (all strains) was approached. This Effect of added oleic acid; column 3 minus colcould be controlled by using both a silicone umn 2. b In uninoculated (control) cultures of stickwater antifoam and only 9 liters of medium out of 14liter maximum capacity. On uninoculated con- that are agitated, aerated in the 14-liter fermentor, endogenous and exogenous 02-uptake meastrol fermentations, using the same agitation both and aeration, this increase in foam formation urement remain nil throughout the 48-h period. was not observed. Typical results of respirometric observations the endogenous activity decreased, presumably because of depletion of metabolizable lipids. At on fermenting stickwater are shown in Table 1. The fermenting stickwater had enough lipids to the same time the true exogenous activity ingive endogenous activities representing the creased, indicating that the microbial cells are amount of growth and cellular activity. The capable of utilizing more free fatty acids such as addition of oleic acid (exogenous) caused only a oleic acid, if they were available. The results of chemical analysis are shown in moderate increase in total activity as the maxiTables 2 and 3. The apparent increase in both mum growth was approached. After 20 to 24 h, 0

"

572

GREEN, PASKELL, AND GOLDMINTZ

APPL. ENVIRON. MICROBIOL.

TABLE 2. Chemical analysis of fermented, freeze-dried stickwatera

Control

Protein (%)

Ash (%)

59.2

20.2

C. candidum 61.2 22.4 Strain no. 179 63.2 Strain no. CA4 24.9 C. lipolytica Strain no. Y-57 63.0 26.4 Strain no. 60-26 64.8 25.9 11 Adjusted to dry weight basis. b NPN, r

NPN (S oN))(%

NH,/NH o N4

6.72

MofN

+

NO2! NO:,

Total fat (%)

Ether fat (%)

1.18

(mg/g) 2.26

24.2 (1001'

17.30 (100)

5.76 8.50

1.30 1.15

2.17 2.17

12.7 (52) 7.08 (28.3)

3.86 (21.8) 3.53 (20.5)

6.51

1.56

2.51

6.20

1.39

2.50

8.0 (30.8) 5.43 (22.4)

2.63 (14.1) 2.77 (15.9)

Non-protein nitrogen.

Percentage of control lipids.

TABLE 3. Amino acid analysis of fermented stickwater Amino acid

Lysine

Histidine Ammonia Arginine Taurine Aspartic acid Threonine Serine Glutamic acid Proline

Glycine

Alanine Valine Methionine Isoleucine Leucine

Control

3.65 1.62 2.75 3.63 4.03 4.22 1.66 1.84 7.20 4.36 8.66 5.49 2.12 1.31 1.45 3.15 0.81 1.76

a

G. candidum 179

G. candidum C-4

3.70 1.48 3.32 3.75

3.70 1.36 2.59 3.71 3.93 5.02 2.14 2.28 8.39 4.29 7.59 4.85 2.44 1.11 1.85 3.32 1.09 2.01

strain no.

4.59 1.93 1.94 8.09 4.37 8.19 4.88 2.28 1.13 1.67 3.18 0.99 1.94

Tyrosine Phenylalanine Total 55.68 57.43 61.67 Expressed as percentage of Kjeldahl N reported (Table 2) for the sample.

protein (Kjeldahl N x 6.25) and ash of the fermented samples, compared to the control, is probably due to loss of solids by the metabolic conversion of some of the lipid to CO2 and H20 (Table 2). Preliminary experiments indicated that G. candidum and C. lipolytica are capable of utilizing urea or ammonia as the nitrogen source for cell growth. The G. candidum strains are not proteolytic (A. Chu, Ph.D. thesis, Columbia Univ., New York, 1969), but the C. lipolytica strains are capable of proteolysis after 24 to 48 h of culture (1). The major source of nitrogen for cell growth is conceivably some form of nonprotein nitrogen. The data in Table 2 do not show significant differences in NH3/NH4+, NO3/NO2, or nonprotein nitrogen to indicate the source of nitrogen utilized. The amino acid analysis shown in Table 3

C.

lipolytica Y-56

3.55 1.17 3.40 3.31 4.50 4.99 2.15 2.22 7.69 4.58 7.85 4.94 2.36 1.09 1.88 3.37 1.17 2.13

62.35

C. sipolytica strain no. 60-

3.29 1.12 2.80 3.14 4.75 1.96 1.98

7.68 4.70 8.31

4.95 2.36 1.11 1.75

3.09 1.02 1.98 55.99

does not show significant increases in lysine or methionine, which are important essential amino acids. However, the summation of the total percentage of amino acids over Kjeldahldetermined protein indicates a trend for a higher percentage of amino acids in the fermented samples than in the control. The other purpose of our investigation, the reduction of lipids, is evident from the data in the last two columns of Table 2. It is evident from this data that total lipids are reduced by about half or more and that ether-soluble lipids are greatly reduced by comparison to the unfermented stickwater. In subtracting the ether lipids from the respective total lipids, it can be seen that some of the non-ether-soluble lipids have apparently been utilized during fermentation. Identification of residual glycerides was done

sgEi~

VOL. 31, 1976

LIPOLYTIC FERMENTATIONS OF STICKWATER

573

by thin-layer chromatography (Fig. 2 and 3). control spots (2-1l application). When the indiThe silicic acid chromatograms (Fig. 2) show vidual chromatograms were subjected to densithat the fermentation control contains a large tometer readings and corrected for dilution facamount of triglycerides (top spot on mono- and tors, quantitation of residual glycerides was diolein and trilinolein control spot), a fair made (Table 4) by comparison with oleic acid, amount of mono- and diglycerides (not clearly mono- and diolein, and trilinolein standards (1 separated into distinct zones), and free fatty ,l of a 10-mg/ml sample spotted). It is apparacids (compared to oleic acid spot). The extrac- ent that the lipolytic fermentation results in a tions of the fermentations contained only faint large reduction of mono-, di-, and triglycerides traces of mono-, di-, and triglycerides but did and, in some cases, fatty acids. It is assumed contain a moderate to large amount offree fatty that the triglycerides are transformed by the acids. The silica gel chromatogram (Fig. 3) re- lipase systems to glycerol and free fatty acids confirmed the above observations and provides plus residual mono- and diglycerides and that a clearer picture of the mono- and diglyceride the viable cells metabolize the free fatty acids separations. Because the control trilinolein is and glycerol to cellular material plus CO., and tangent to the solvent front, it would be diffi- H,O. A. Chu (Ph.D. thesis, Columbia Univ., cult to compare triglycerides in this chromato- New York, 1969) demonstrated this by use of gram. However, all of the fermented samples radioactive lipids. display only faint zones in the triglyceride (solDISCUSSION vent front) area. In the chromatograms, the fermented samThe objective of reducing the amount of lipids ples represent more concentrated spots (6-Al in stickwater (reconstituted solubles) was application of extract) than the fermentation achieved, as indicated by the data related to MEDIUM: ITLC-SA

SOLVENT SYSTEM: ISOOCTANE/ETHYL EMER 80: 20 (SULFURIC ACID CRARR)

ei

fK~~~C

,v1&.L < h2i'si - -t

KS-S-~''-' 0-, --

ojvt#Oj-y kA

' ':F rt'5*->',,.Wt, _t,',a t' i 9-i *

s

-

>-~g

CO.NT............................+ILi 0.-,-0-0S:+0 -0; ........ fRLILI LEI ACI

DIOLEIN (MONOOLEIN)

_- '

a, ttJ

'

0

;-'''

F3NT&TO

G.Candidu

FER4ETTIi UIN: CONTOL

;

FER.t,,