The Determination of Carbon Dioxide in Fermenting Mixtures. Co-Zymase: a Study of Purification Methods Citation Winegarden, Howard Merlin (1927) The Determination of Carbon Dioxide in Fermenting Mixtures. Co-Zymase: a Study of Purification Methods. Master's thesis, California Institute of Technology. doi:10.7907/R7YV-A670. https://resolver.caltech.edu/CaltechETD:etd-12272004-144412 Abstract No abstract. Item Type: Thesis (Master's thesis) Subject Keywords: Chemistry Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Unknown, Unknown Thesis Committee: Unknown, Unknown Defense Date: 1 January 1927 Record Number: CaltechETD:etd-12272004-144412 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-12272004-144412 DOI: 10.7907/R7YV-A670 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 5139 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 29 Dec 2004 Last Modified: 16 Aug 2023 23:00 Thesis Files Preview PDF (winegarden_hm_1927.pdf) - Final Version See Usage Policy. 1MB Repository Staff Only: item control page

THE DETERMINATION OF CARSON DIOXIDE IN FERMENTING MIXTURES Thesis by Howard Merlin Winegerden In Partial Fulfillment of the Requirements for the Degree of Master of Science California Institute of Technology, Pasadena, California, 1927 THE DETERMINATION OF CARBON DIOXIDE IN FERMENTING MIXTURES By Albert L. Raymond* end Howard M. Winegarden * National Research Fellow in Chemistry Introduction In connection with a research on enzymic behavior it was necessary to develop a simple and expeditious method for the determination of the carbon dioxide formed during fermentations. The most suitable method of carbon di- “oxide analysie appeared to be the procedure evolved by Cain and Maxwell? for the determination of carbon in steel. They absorbed the carbon dioxide, formed by combustion,in known volumes of barium hydroxide solution,and follow- ed the precipitation of barium carbonate by measurements of the electrical conductivity of the solution. The idea has been applied by Spoehr and MeGee*, 5 to the determination of carbon dioxide in their studies on plant respiration. This paper describes the development of thie general method so as to render it ap~ plicable to studies on fermentation. The. technique whichwas finally adopted consisted in liberating the carbon dioxide from the fermentation mixtures by shaking,carrying it by means of a stream of air,free from carbon dioxide,to the absorption vessels con- taining barium hydroxide,and measuring the change in resistance of the barium hydroxide solutions. We wish to thank Professor A.A.Noyes for grants which made possible this investigation. Conductivity of Barium Hydroxide Solutions Before undertaking the carbon dioxide determinations,it was necessary to determine the conductivities of pure barium hydroxide solutions. The conductivities found in the literature are almost exclusively at 25°C. and are not in very good agreement. As we wished to employ both 30°C, and 37°C.,we determined the conductivity at these temperatures over the concentration range of 0,12 to 0.065 N.,and in addition repeated the observations at 25°C. for 6om- “Parison with previously published data. In making these determinations the usual precautions were observed. Con- ductivity water of specific conductance of from 0.8 to 1.2 x 1076 mhos was pre= pared by redistilling the laboratory supply of distilled water first from acid permanganate and then from barium hydroxide,all im a current of air tree from carbon dioxide. The potassium chloride employed as a conductivity stendard was Baker's cep. twice recrystallized. The barium hydroxide was Merk's c.p. similarly treated, Sodium carbonate was used as an acidimetric standard and was prepared by gentle ignition of pure sodium bicarbonate,which in turn was secured by sat= urating with carbon dioxide a cooled solution of Baker's c.p. sodium carbonete, Various samples of the final carbonate prepared as above,gave reproducible analyses. Baker's c.p. hydrochloric acid was diluted and standardized against the sodium carbonate using methyl orange as indicator. In all titrations the neutral solution was boiled to expel carbon dioxide, cooled,and the end-point -redetermined, The final value chosen for the normality of the hydrochloric acid was the average of five determinations in which the greatest difference was 0.15%. The barium hydroxide solutions were standardized against this hydro= - ¢hloric acid,using the average of three or four determinations,which in general showed differences of not more than 0.25%. The temperature of the thermostat was read to # 0.09" by means of a stand~ ardized thermometer, Barium hydroxide solutions of various concentrations were immersed in the thermostat,and after allowing them to come to temperature, their resistances were determined with a dip electrode. The thermostat temperature was then changed and the determination repeated, Measurements were made on a portion of each sample at 25°C.,30°C.,and 37°C. At each temperature the cell- constant of theelectrodes was determined by immersion in an accurately prepared potassium chloride solution both before and after measuring the conductivities. In no case did the eell-constant undergo appreciable change during the measure- ments. After several preliminary experiments,the data given in Table I were ' secured in three runs,as indicated in the first colum. Each sample was used at tne three temperatures,and the normalities given in colum two are those which the solutions would have at 20°C. For convenience in our later work, the ' eoncentrations were recalculated and expressed in milligrams of carbon dioxide per 100 cc. of solution,by which is meant the milligrams of carbon dioxide etoichiometrically equivalent to the barium hydroxide in 100 cc. of solution, Thus 0.1 N. solution is expressed as 220 mg. 0/100 cc, These values appear in colum three. The last three colums in the table give the specific resistances as calculated from the observed conductivities and cell constants, Zable 1 Norm. at MgCOg per Specific Resistance Run 20 100 cc. 25° 350° 37° I 0.1151 248.9 43.43 ho 13 36.16 9. 1000 220.0 48.78 45.04 40.60 0.0907 199 66 53.21 4o 12 44.30 0.0817 | 179.9 58.63 54.12 48.85 0.0723 15961 65.40 60.33 54.46 0.0633 139.3 73.68 68.13 61.48 II 0. 1939 228.6 47.05 43,45 59.25 0.9954 209.8 51.00 47.07 42,40 0.0862 189.6 55.01 51.64 46.61 0.9771 169.7 61.90 57.15 51.5 0.0680 149.5 69.49 64.15 57.88 Ilr 0.1172 257.9 42.42 58.94 35.15 0. 1092 240.2 44,80 44.50 37.41 0.0916 200.3 53.15 40.11 4h 32 0.9711 156 .4 66.70 61.52 55 058 4 L For comparison with the data of Cain and Maxwell ,and Spoehr and MeGee*> they are all plotted in Fig,I. A single point falling in this region determined by A. A.Noyes* is included. Figur oO 70 AA. Noyes. Spoege v McGee. Coin we Maxwell Spoghr v MeGee. Raypond w Winegarden (0) fe) Specific Resistance ~y® @ (OJ df. L 0e@06 + 3 4 = h Ht 30 140 160 180 200 229 240 260 Mg. per 100 cc. Apparatus and Method The electrical equipment for the determination of conductivities differed but slightly from that employed by Gain and Maxwell', A Weibel. type A.C. galvano- meter operating on the tto v. 56 cycle light circuit,supplied 6 v, A.C. for the bridge. The galvanometer wae of low-res istance, and we therefore substituted a low-resistance Kohlrausch bridge and a four-dial Curtis box for their self-con- tained bridge. An advantage of this change will be discussed later, The instru- ments were manufactured by Leeds and Northrup and gave complete satisfaction, Resistances up to 10°90 ohms could be measured to within 0.95%,which was greater accuracy than we required. Carbon dioxide determinations were made with an accuracy of + 0.0006 mg./cce of absorbing solution,or + 0.15 mg. for a 250cc. cell, For convenience as well as for reproducibility and accuracy in the carbon dioxide determinations,it was decided at the outset to immerse both the fermenta— tion vessels and barium hydroxide solutions in a single large thermostat. One equipped with the usual stirrer and regulator and capable of maintaining a desired temperature to within + 2 degrees was employed throughout, For transferring the carbon dioxide to the absorption vessels,a stream of air free from carbon dioxide was used. The air was purified by passing it through a calcium chloride tower and then through three soda-lime towers in , teries. By refilling the most used one of these every three to four weeks and moving the towers in a counter-current fashion,no trouble was experienced from carbon dioxide in the gas stream. A pressure regulating device,consisting of an eight inch U tube about one fourth filled with mercury,was connected to the air line by means of a large T tube. If tne line pressure exceeded tne desired value,the gas stream escaped through the mercury. This was found to give as close regulation as was necessary. To maintain constant rates of gas flow,a piece of glass capillary tube was placed in the line following the pressure regulator. By suitably choosing the size of the capillary and the amount of mercury in the U tube,the gas flow was easily adjusted to any desired value. For our work we found 200 ec. of air per minute to be satisfactory. An exactly constant rate of gas flow is,however,probably unnecessary,as will be indicated later. After trying several types of gas absorbers,we decided that the design of Weaver and Edwards® was best suited to aur purpose. It is a compact and very efficient absorber,even at high rates of gas flow. Their original design was simplified and the lower stopcock omitted,as illustrated in figure ITI, The absorber is for 250 cc. of solution but may be altered to permit the use of any quantity. The gas stream enters through the tube A,carries the absorb- ing liquid up the spiral,and finally escapes through the outlet B. Fig. IT. ]> ip) Scale in cms. ‘The spirals were built of tubing which was & mm. inside diameter,and although - considerable precipitation of barium carbonate occurred in them, the absorbers could be used for several runs before cleaning and refilling. A diffioulty which has been experienced, however,is that the injector may become clogged during a run. To avoid this the constriction could be omitted as the bubbles are still small enough to produce efficient absorption at low rates of gas ‘flow. A soda-lime tube was attached to the small upper outlet of the absorber to prevent carbon dioxide entering from the air. A further advantage of this type of absorber is that is admirably suited to the use of pipette conductivity cells,which were inserted through the rubber _ stopper © and left in place. The cell which we employed is illustrated in figure III. IF Ill. J > }@ 0 s pe ee | Scale in ems, The upper electrode is sealed into the tube A,which also serves to insulate the lead wiree The connecting wire to the lower electrode is stretched along the side of the cell and held in place by the two-hole stopper into which the two tubes are inserted. The open tube B serves to admit air free from carbon dioxide, ‘in order to blow out the contents of the electrode as well as to mix the solution in the absorber. The cell is immersed so0 that the upper electrode is below the liquid level,and the pipette need therefore be only blown out,as it refills of itself. It was found by blowing out the contents three or four times that the solution was uniform,and further mixing produced no changes in resistance. A three-way stopcock was connected to a source of air,free from carbon dioxide in such a way that the electrode might be alternately blown out and allowed to refill by merely turning the stopcock. In order to liberate the carbon dioxide from the fermenting mixtures it was desirable to expose as large a liquid surface as possible to the gas stream. For this reason the Permentation tube illustrated in figure IV was employed, the horizontal part of the tube being half filled by 20-25 ec, of solution, Scala in cms. Solutions could be added or withdrawn through the straight tube with a pipette without disconnecting the current of air free from carbon dioxide which entered through the branch tube. The straight tube was capped,except during such removals or additions. In order to accelerate the evolution of carbon dioxide, the tube was fastened with small clamps to a rocking table immersed in the thermo- stat. The table was oscillated by means of a small electric motor fitted with a Worm drive and so arranged that both the frequency and the amplitude of the oscillation could be changed. After some preliminary experiments an amplitude of 4 6 degrees from the horizontal and a frequency of about 120/min. were adopted, In our final apparatus we mounted six identical sets side by side in the thermostat, the shaking table being built to accomodate six tubes.,,The air lines for transferring the carbon dioxide and for blowing out the pipette electrodes were manifolded. Gas-washing towers filled with water were placed in the air lines and immersed in the thermostat to saturate the incoming air -and prevent evaporation. For convenience an automatic pipette delivering exactly 259 cc. Was used to fill the absorption vessels. A further time-saving device was made possible by our electrical hook-up. The cell constants of our different electrodes were in general not the same. However,by setting the four dial resistance box to correspond to the constant of the cell employed( the setting takes but a few seconds) it was possible to compensate for these differences. The strength of the barium hydroxide solutions were then calculated from the data of Table I and plotted as a function of the bridge reading. The graph was found to be very useful,as the calculations were otherwise quite tedious. Reliability of the Analytical Method The observation of Spoehr+ that the constant of an immersed cell is subject to change due to precipitation of barium carbonate, led us to examine this b Fe) question. We found that removing the cell from solution and allowing carbon dioxide from the air to form barium carbonete,as suggested by Spoehr ,did preduce a change in the cell constant, However,as long as the electrodes were kept completely immersed no appreciable change in cell constant was observed during a rume This is presumably due to the fact that the barium carbonate is largely precipitated in the spiral before coming into contact with the electrodes,changes in cell constant being due to the precipitation of bariun carbonate upon the platinized surface and not to adsorption of previously precipitated barium carbonate, It was observed, however,that the resistance of the barium hydroxide solutions increased slightly when kept 24 to 48 hours. in our cella. This occurred,nevertheless,even in the absence of carbonate, but was not of sufficient magnitude to introduce appreciable error, The cause of this phenomenom wae not investigated. An additional difficulty which we had anticipated wae that the precipi- tation would pe*S low as to require a correcting factor, To examine this possi- bility,resistence determinations were made one-half,one,and five minutes after introducing carbon dioxide and again after ten to twenty-four hours. Changes . _ after one minute were found to be negligible as long as the solutions always ° contained a relatively large excess of barium hydroxide,as was the case in the range of concentrations which we employed. In order to show that the method actually gave correct results on carbon dioxide,we prepared samples of gas by adding measured volumes of standard sodium carbonate solution to an excess of dilute sulphuric acid, A stream of air,free from carbon dioxide,was bubbled thru the mixture and then through the abeorber. After %5 to 20 minutes the amount of carbon dioxide as determined by the change in resistence was compared with that from the amount of carbonate tt used. The differences were always less than one percent,and were systematically in the direction which would indicate that not all of the carbon dioxide had been evolved from the acid solution. Rate of Evolution of Carbon Dioxide from Solution Using the apparatus described above the rate of removal of carbon dioxide from solutions was investigated. Figure V illustrates the results obtained, The graphs represent the carbon dioxide removed in different periods of time, the removal being expressed in percent of the total amount originally present, ‘ai act Eel = 4 Ax IN / Zs=s=e —- = CO,removed in percent of total Pe 10 20 30 40 50 Time in min. t2 In tne first experiment 20 ce. of mercury were placed in the fermentation tube,and the carbon dioxide introduced in the space above. The shaker was started and the air stream turned on. In one half minute 95% of the carbon dioxide was precipitated as carbonate,as shown by Curve I. The total amount present was ‘taken as being the value after 10 minutes,as this was identical with the results. at 3 and 5 minutes. The experiment was then repeated with the exception that 20 cc. of water saturated with carbon dioxide were substituted for the mercury, In this way Curve II was obtained. For comparison,this second run was repeated,except that the shaker was not operated. The readings were continued and the value at 24 hours was taken as representing the total amount present. The data for the first 50 minutes are given in Curve III. From curves I and II may be seen that removal of the gas from the space above the liquid is,for all practical purposes complete in one minute, but that the rate at which carbon dioxide escapes from the liquid into the gas space is in comparison very slow. From the data for @urve III the rate of escape from an. unagitatéd liquid was found to be 4% per minute at the start,dropping rapidly to 2% per ‘minute. From the agitated liquid, the rate of escape was. about 60% of the re- sidual carbon dioxide per minute. This illustrates the extreme importance of the shaking. In,creasing the shaking,placing beads in the tube,and using a corrugated tube that broke up the liquid surface more thoroughly, each produced slightly better results,but not enough to compensate for. certain mechanical dis- ‘advantages. With different mechanical arrangement,which would allow very violent shaking,it shouldbe possible to still further increase the rate of removal» The rate of gas flow was found,as mentioned above,to be of but slight con- sequencee By shaking the mixture with the air stream off, stopping the shaking, and turning on the air,a curve almost identical with Curve III was obtained. The average rate in this case was about 55% of the residual carbon dioxide removed per: minute. If we assume 50% per minute to be certainly obtainable,it is possible to 3 calculate the limitations of the method, If carbon dioxide is being produced at @ constant rate,then the error in observation the first minute after the beginning of such production will amount to 50% of the production. The second minute the error will be only 25%,and the third minute only 12%,etc. Similarly,if some acceleration (either positive or negative) in the rate of production suddenly occurs, then after 5 minutes this will be reported to within about 3% of the incre- ment. For small uniform acceleration, the error is not appreciable. It may there- fore be stated that the method may be used with negligible errors for uniformly accelerated changes,while for instantaneous changes the error is negligible after five minutes. Observations may, therefore,not be made lees than five minutes after some sudden change in rate. As an example,there may be included the results (Fig. VI) on a fermentation employing live yeast. Fig. Vi pon. Mg. COg NN 20 0 1 2 3 Time in hrs. 14 At the point P saturated mercuric chloride solution Was introduced. Thie reagent causes an almost immediate cessation of fermentation, but the observations do not report this fact for several minutes,as may be seen from the curve, A further factor which must be conesidéred in this connection is the pH of the solutions. In the experiments just reported,since the solutions were unbuffered, the pi was between 4 and 7. In this region the carbon dioxide is largely present as C05 (or HgC0z). If,however,the solution is made more alkaline,the proportion of carbon dioxide as 003" and HOOs” increasés,as shown ‘in Table II. Since 50% of the total carbon dioxide is removed per minute in Table il Percent of total _ pl present as. 4 5 6 1 8 9 H005~ 05 | 29 3.1 | 75.0 | 96.8 | 95.9 005" - | | 0.001 | 6.03 | 0.6 + 5.6 G0 + 1200; 99.7 | p71 76.9 | 25.0 | 26 | 045 the acid solutions,it might be expected that 50% of the free carbon dioxide (COg not present as HCOz~ or 005" ) would be in general removed per minute. This was investigated experimentally. Buffered solutions were substituted for the water and the shaker was operated throughout. To obtain the total amount originally present,the solutions were acidified after 50 to 50 minutes, and shaken until equilibrium was reached. Curves IV,V,and VI in Fig, V represent the results at pH 7,8,and 9,respectively, It may be seen that the rate of evolution is much less in these more alkaline solutions. The rates which would be predicted from the fraction present as COQ + Hp00z at tne different hydrogen ion concentrations were compared with rates actually 4 observed at 16, 20, and 30 minutes. In the calculations it was assumed that 55% per minute is evolwed at pi 5, (Table III). Table III Calc. rate ____ Observed rate pH % per min. 10 min. 20 min. 30 min. 7 14 : 12 1 | 10 8 1.5 : 6 | 4 3 9 243 2 2.2 |e 9 We may therefore conclude that without employing correction factors, the metnad is not easily applicable in its present form to investigations in alkaline media. Applications A particular advantage of the method is that it is ideally suited to investigations on aerobic and anaerobic fermentations,as a change from com- pressed air to compressed nitrogen is all that is required, A disadvantage involved in the application of the procedure to biological studies, however, lies in the difficulty of securing. sterility. The tubes may be easily sterilized, but the air stream and connections may present difficulties. Our own work has been confined to enzyme systems,and we have therefofe used an inhibiting agent in the mixtures Toluene,which is most often used for this purpose,is too volatile,and we have employed instead a 125090 concentration of gentian violet, At this concentration there is no appreciable inhibition of zymase,but no live yeasts have been observed even after 3 or 4 days. 16 Summary The specific resistances of barium hydroxide solutions from 0.12 to 0.065 N. have been determined at 25, 30, and 37°. The rate of the evolution of carbon dioxide from aqueous solutions under various conditions has been examined. A method has been developed for the rapid and accurate determination of the carbon dioxide formed in fermenting mixtures. The method is applicable to either eerobic or anaerobic investigations. The advantages and limitations of the method are pointed out. 4 . Cain and Maxwell, Jour. Ind. Eng. Chem,,11, 852 (1019). 2s poehr and McGee, Ibid, 16, 128 (1924). . *8poehr and McGee, Carnegie Inst. Publ,325 (1923). +A AsNoyes, Garnegie Inst.Publ.,63,254 (1907). 7n.E.Weibel, Bur. of Stand. Sci. Paper, 297 (1917). Sweaver and Edwards, Jour. Ind. Eng. Chem.,7, 534 (1915). GO-ZYMASE - A STUDY OF PURIFICATION wEYRODS whesis by Howard Merlin Winegarden in Partial fulfillment of the Kequirements for the Degree of master of science California institute of wvechnology, Pasadena, California 1927 vO-ZYMASE - a STUDY OF eURIAISATIGN METHODS L oy Albert L. Raymond and noward ::. winegarden. 1 Hational xesearch sellow in vhemistry. introduction Uo-ezymase is one of the essential components of the complex enzyme mixture which effects alcoholic fermentation in the ab- sence of living cells. ‘the separation of the mixture into "zy- masé’ and "co-zymase" was first accomplished by uarden and Young by means of ultrafiltration through a gelatin impregnated Uhamber- land filter candle. iwhe residue and filtrate as thus prepared possessed, separately, no fermentative action, but when mixed were found to produce a rapid fermentation. ‘the active constituent of the residue was named zymase, while that constituent of the fil- trate responsible for the reactivation of the residue was named co-zymase. whe mechanism of the activation of the zymase was investigated by tuler and yrbach”. who came to the conelugion that the co- zymase was involved in that stage of the process in which inorganic phosphates are converted to carbohydrate esters. ihe same authors have described a comprehensive and successful series of experiments on the purification of co-zymase. A complete knowledge of the properties and function of co- zymase would be of extreme importance, as it should throw consider- able lizht upon the mechanism of enzyme action as a whole, and should in addition have important bearing upon the entire question of carbohydrate utilization. it has been shown, for example, by Meyerhort? that a substance, either identical with, or very similar to co-zynase exists in most animal tissue, and the demonstration of a@ complete identity of the two would have important metaoolic con- notations. | In view of the importance of the field, as well as of the Ge interest of one of us° in the problem or carbohydrate utilization, we decided to attempt an extension of the work of wuler ani Myr- bach on the surification of co-zymase. at the beginnin--, however, we found that our yeast produced much less active enzyme prepara- tions than those of suler and »yrbach, as regards both the zymase and co-zymase content, and also that the technic of the above authors with lead precipitation, wnaich enabled them to secure an initial purification of from ten to thirty fold, was in our case practically useless. ve therefore abandoned our original intention and have instead Studied the purification produced in our material by a variety of reagents. in the investisation we have repeated much of the work done by sulsr and uyroach, and Ssev=ral differences have been foniid, Wnhich aprvear difficult to explain solely upon the basis of the lower initial purity of our material. as certain of the experi- ments show distinct promise, we hope to be adle to extend the work upon a material of considerably hizner oris:neal purity, such as was employed by suler and iyrbach. v@ desire to thank 2rofessor 4.4. Woyes for grants and facil- ities whicn allowed us to oursue these studies, and vrofessor aeV. VOlman Yor his constant interest and encouragement. inaterials ana wechnic of assay. “he yeast which we employed throughout was a bottom yeast Supplied by the sastside orewery! in wos angeles. it was obd- tained fresh, transported imuediately to rasadena, and used the Same day. wor analyzing extracts for co-zymase activity, the analytical Ge procedure adopted was similar to that employed by wuler and Nyr- bach. iihen zymin (acetone-dried yeast) is washed with water, the co~zymase is largely removed. he washed material is therefore a source of gymase practically free from co-zymase, and on addition of the latter is again capable of producing fermentation. wsuler and iiyrbach found on adding co-zymase extracts to such washed zymin that within reasonable limits of concentration the maximum rate of carbon dioxide production was proportional to the amount of extract added. Uur experience in general confirmed this, as well as the fact that it is necessary to keep the phosphate concentration con~ Stant throughout. All solutions were adjusted to pu 6.3 - 6.5 before testing, but even in those cases in which ali precautions were observed, we found variations of 5 to 10 per cent in certain analyses and oui assays are theretore only reilable within these limits. In order +o prepare our zymin the prodecure of Albert® was used. #resh yeast was filtered through canvas in a small plate and frame press under five to fifteen pounds pressure. ‘ihe press Was then blown out with compressed air to secure partial drying, and the yeast was pressed in a hand press to remove as much water as possible. ‘nis drying was facilitated by making cakes of the washed yeast, wrapping them in sheets of fiiter paper, and placing them between pads of blotting paper. ‘he bundle was wrapped secure- ly in canves, and pressed at about 300 los. per sq. in., producing a friacle, light-colored product. whis was converted to zymin in 500 gram lots by passing it through a ten mesh sieve into three liters oy c.p. acetone. after stirring for ten minutes, the acetone Was removed by suction filtration. whe partiy dried yeast was next introduced into one liter of fresh acetone for two minutes, filtered, ground with 500 ce. or ether for three minutes, and again filtered. ‘the almost white powder which resuited was air-aried. Althouzh the procedure appears aguite simple, we found it ditiicult to prepare an actively fermenting zymin from our yeast. Jui gsuostituting technical acetone for the c.p. material, a completely inactive preparaivion resuited, aS Was the case when the acetone was added with stirring to the yeast insvead oi the reverse as described aoove. uven toose lots which were prepared by apparently identical procedures showed considerable differences in activity. whe experi- mentai wor. to be reported was, therefore, pertormed with two very Similar lots wnich were, incidentally, the vest wnich we secured. ror assay purposes, the zymin was freed of co-zymase by mixing with eight parts of water ana centrifuging. whis wesning was re- peated twice more, and the zjmin then produced almost no fernenta- tion in the absence of added co-zymase. whe zymin was freshly washed as needed. ithe solution used for making the assays was prepared as follows. isleven grams ot zymin were washed as above, and dilutea to a total weight of 43.4 ams. (44 cc.) oy adding distilled water and a small amount of “gymophospnate" solution. where were then added 18.9 cc. ot a solution of the following composition: WapHPO, .onpO 13.83 grams KH 204 15.6 =" Glucose 583.0 " Gentian violet 0.75 # Distiilec water to 1000 ec. the mixture was made homogeneous by stirring and shaking, and 10.0 cc. 5. were removea and added to 9.0 cc. of the co-zymase solution to be tested (or a lesser amoant diluted to 9.0 cc.). The finai concen- tration in this fermenting mixture were: Na.HPO, 0.013 moial a KHePOg . 0.020 " Glucose 10. O% Gentian violet 0.125% “ymin 10.0% The “gymophosphate solution" was added because of its known property of reducing the incubation period of fermentation. ithe "“gymophosphate* was prepared from the first zymin washes by precipi- tation first as a barium and finally as a lead salt. Lead was pre- Gipitated with hydrogen sulfide, the latter removed in vacuo, and the resulting solution adjusted to pH 6.4. it had no co-zymase activity, and enough was added to diminish the incubation period to from o to # hours. the gentisan violet was added as an antiseptic, and at the above concentration we ooserved no growths for the periods over which we worked (two to twelve hours), and there Was no inhibition of the enzyme action. ‘he pH of the mixture was 6.4, and the total phosphate concen- tration was optimum for our yeast. for determining the carbon dioxide formed an electrometric® method devised for the purpose was employed. keadings were made at 15 minute intervais, and the maximum rate was taken as a measure of the co-zymase activity of the sample. vtypical fermentation curves are given in #is. i. in these the rates of carbon dioxide production are plotted as a function of the elapsed time. tsurves 1 and 11 were obtained with different quanti- ties of co-zymese, using the glucose-phosphate mixture described above, while Uurves ilI and iV are the corresponding ones for runs employing twice the concentration of phos:hnate, other factors be- ing unchanged. she longer periods before attainment of the maximum rave in the latter case are characteristic of high phospnate con- centrations, as are the larger maximum rates for tne same co-zymase concentration. 8 Rate of C0, production in mg. per hr. Time in hrs. Fig. 1. Te it is of interest that at no time did we find co-zymase pre- parations which gave curves markedly different from those of vig. I. With many of the purified preparations, however, an incubation period of from one to three hours was observed, and as “zymophos- phates" have been Stated to remove these lsg peéericdas, we added such preparations, aS mentioned before, to our fermentation mixtures. nydrogen ion concentrations were determined throughout by means of indicators, the colors being checked with electrometrically standardized buffers. in addition to Ulark and Luo's siandard set of indicators, we employed shenolphthalein, thymophthalein, and in a few cases, methyl blue. . #or convenience in reporting our results, we have defined a "unit" as being that quantity of co-zymese required to produce a faximum rate of fermentation of one mg. of carbon dioxide per hour. We have also used wuler's activity coefficient, which is defined as the maximum rate of carbon dioxide production in ec. per hour, divided by tne solid content of the sample expressed in grams. ihe Symbol Aco was suggested by suler to designate this unit and has been employed throughout this paper. ‘the Aco, except where otherwise noted, refers to the ash-free material. solids were determined by drying to constant weight at 105-110" and ash by igniting at dull red until the residue was gray or white in color. uxtraction of vo-zymuse from Yeast Our co-zymase solutions were prepared by extracting yeast with hot water, a procedure which has been shown by others to destroy the zymase activity without appreciably impairing the co-zymase. zresh and washed yeasts were both investigated, but as the only difference between the two extracts lay in the fact that those from the washed yeast conteined slightly less solids, the latter were not extensively employed. in order to obtain an efficient extraction, various factors were examined with a view to determining the optimum yield of co- zymase, Without regard to the purity of the resultant product. sateches of 10 to 20 liters were prepared by adding yeast to water at. 90-92° | stirring two to three minutes, and cooling as rapidly as possible by immersing the metal container in running water. the effect of concentration of the yeast upon the yield was studied, and it was found that the efficiency of the extraction in- Greased largely with decreasing concentration down to 0.1 gm. fresh yeast per cc. Solution. More dilute extracts were not examined, as solutions of much lower activity required too extensive evaporation before they could be used. Acidity was also examined with regard to its effect on the yield, glacial acetic acid and sodium carbonate being added to the hot water to vary the acidity. whe pH was determined and the solu- tion neutralized to pH 6.4, and tested. whe results are ineluded in table i. table I. neagent HOAG Control Naovdg voncentration (g/L) 5 1 1 5 pH of extract 3.8 4.6 6.0 7.0 10.0 Activity (Units/cc.) 3.6 3.6 4.7 2.9 2.5 Klectro Dialysis No attempts appear to have been made toward purifying co-zymase by means of electro-dialysis. as the method offers great possibili- ties with those substances which migrate under the influence of an 9. electric field, we decided to investigate the results with our mater- ial. A cell was built consisting of seven 50 cc. compartments, in @ row, and separated by parchment paper partitions. vthere were placed in the compartments in order: 1, copper sulfate solution (1 molal) and a copper electrode; 11, potassium chloride (0.1 m.); 1II, potas- Sium chloride (0.Im.) adjusted to the desired pH; 1V, co-zymase | solution similarly adjusted; Vv, VI, and VII were duplicates of ili, iI, and Ll respectively. ‘he potassium chloride solution in il and VI prevented contemination of that in iIl and V. Yo see if the px of the solution affected either the magnitude or direction of migration of the activity, a set of four experiments was performed covering pH range of 2 to 10. A potential drop of 4 volts/cm. was maintained in the co-zymese solution and 1000 to 1800 coulombs passed in 40 to 60 minutes. in the case of experiment TIi, the apparatus was kept in the ice-box to decrease the destruction due to the alkalinity. ‘the pn did not change appreciably during the rune ‘tne activities of the solutions are given in tsehle il. Yable if. Activity (Units/cc.) Expt. pro Anode Center Vathode LV 2.0 0.2 1.4 0.3 il 4.4 0) 2.7 6) I 6.4 0.2 3.2 0.2 LII 9.7 6) 2.3 fe) in no case, therefore, was an appreciable amount of co-zymase trans-~ ferred to the potassiim chloride compartments, although tests showed the parchment paper was permeable to the co-zymase over the entire LO. ph range employed. in view of the fact that the paper was permeable to the ¢o- gymase and yet practically no transference occurred in an nour, it seems justifiable to conclude that the co-zymase carries no appreciable fraction of the current over the po range examined. Whether this is | due to its being present at low concentration, or to its having a low degree of ionization, or to both, can not be determined from the above data. However, the possibility of purifying erude solutions by this means seems to be definitely excluded. Alcohol vrecipitation A8 alcohol has been suggested for purifying co-zymase, we per- tormed a number of experiments with this reagent. we found that an alcohol concentration of 40% produced a fine flocculent precipitate and left the majority of the activity in the solution, but that 80% alcohol, on the other hand, produced a less satisfactory separation and trequently occasioned considerable losses of activity. it seemed well to learn, therefore, how much puritication was being produced, and a set of experiments was performed with this in mind. 4A co-zymase solution was concentrated tive fold in vacuo and divided into two portions. ‘the first was brought up to 80% by the addition of 95% alcohol, and the tractions separated. (Table i111) PYable if Yield (%) ACO ury ash-free riltrate 59 . 28 38 rrecipitate 48 | 30 41 Uriginsl 26 56 il. vo the remaining portion of the concentrated solution alcohol was added to 40% and the filtrate was then brought up to 80% as before. (Yable 1V) whe Aco was determined on each fraction and the yields calculated. whe yields are expressed as the percentage of the original co-zymase, present in the different tractions. ‘he failure to add up to 100% is presumably due either to varying phosphate con- centrations in the difterent tractions, or else to the removal of some inhibiting material. this phenomenon was frequently observed. wable iV rield (%) ACO LLy Ash-free 40% precipitate 11 15 29 80% " 51 36 42 — BO% filivale 61 31 44 vriginal B6 36 it will be seen that even though there were apparently no losses of activity, the purification wes not appreciable in any case. it must be noted that these experiments were performed only upon ernde solutions and the possibility, therefore, remains that different results might be obtained on a purer material. Lead vrecipitation bue to the great success ot wuler and Myrbach using lead pre- cipitation as a means of purification, we attempted to repeat their experiments with this reagent. ithe procedure evolved by them con- sisted in a precipitation at pu 6, which left the active material in the filtrate, followed by precipitation of the activity at pH 10. Lg. We employed both washed and unwashed yeast and prepared an ex- tract exactly as they described. Precipitation at pH 6 guve a mixture which was very difficult to filter and which was turbid after centrifugation. Practically none of the activity passed into the precipitate. Solid determinations indicated only very slight increase in the Aco. Buler and Myroach, on the other hand, obtained three fold purification and reported no mecnanical difficulties. At pH 10 we found the activity to be again almost entirely in the filtrate. ‘his is contrary to the results of the above autnors, who found precipitation to begin at ph 8.5 and be complete at pH 10.. We &lso duplicated our own results Witn pu 6 filtrates to which additional lead acetate had been added, indicating that the differ- ence in the alkaline precipitation was not due to insufticient lead. 9 ascertain whether we were dealing solely with a concentration eftect, we evaporated a sample of our material to one sixth volume in vacuo and repeated the experiment. Again the activity remained ainost completely in the filtrates, both at pH 6 and 10. In view of these facts, as well as the one mentioned above that toe precipitation is not at satistactory at pH 6, we decided to modity Euler's first procedure as follows. satches of co-zymase solution were prepared by adding 2% kg. of fresh yeast to 15 L of water at 90-92°. ‘ne solutions were filtered immediately and cooled. Centrifugation probably would have been preferable but no continuous centrituge was available. une cooled, filtered solutions were placed in a large glass container capable of holding 25-30 L and a hot concentrated lead acetate solution was added to the ex- tent of 7.5 g. of lead acetate per liter. ‘The pH was then adjusted to 9.0 by the addition of 6 N. NaOH with constant stirring, and the mixture was allowed to settle for one or two minutes. ‘the super- natant portion was occasionally perfectly clear and filtered rapidly 13. giving a light yellow, crystal clear solution. When this was not the case, additional portions of 0.5 g/L of lead acetate were added, and the pH readjusted after each addition to 9.0. in no case did we need to add more than 8s g/L total lead acetate to produce perfect 10 of the solution dropped precipitation. The phosphate content from an average of 0.35 mg/cec. t2 about 0.007 mg/cc. which accounts for a large portion of the added lead. these final solutions, after neutralization and removal of lead, contained from 90-100% of the activity of the original ex- tract, out solid determinations showed only slight changes in Aco. & few experiments chosen at random sre given to illustrate these points. (Table v) Table V Expk.. remarks Yield (%) Aco Dry Corr. 19 Orig. unwashed 36 Precipitated 109 39 75 40 Orig. unwashed 50 Precipitated 96 45 73 41 Orig. washed 73 Precipitated 97 62 130 66 Orig. unwashed 21 Precipitated 90 20 32 The solid content of the final solution may be corrected for those ions added during the precipitation which would be carried through into the filtrate. These include Na* and 0, from the base and acid used and Ac from the lead acetate. the final column of the 14. table gives the activities corrected by tnese factors, from which it may be seen that an actual slight purification did result in each case. Solutions prepared as above, stiii containing a small amount of lead, were quite stable. . Unlike the original boiled yeast extracts, they exhibited little tendency to develop mold or bac- terial growths, and lost less than 50% of their activity on being kept in the ice~box for two or three weeks. As mentioned above, inorganic phosphates were almost completely removed by the lead precipitation, and therefore did not interfere with additional reagents. Hor these reasons, we adopted this precipitation as standard ,and the further work described in this paper wes largely done on such lead treated solutions, from which excess lead was not removed. io distinetion will be made between such solutions prepared from washed and unwashed yeast, as no differences in their behavior were observed. In order to determine the effect of higher pk on the preci- pitation of the active material, we added lead acetate in differ- ent concentrations and tried px values of 10.5 and ll. ‘ithe be- havior was quite variable and appeared to depend upon the parti- cular batch of material employed. sor example, lot no. 41 made 0.005 mM. with lead acetate, gave a precipitate containing 79% of the activity at pH 10.5 and 72% at px 11.0, while another, lot no. 66, using lead ion concentrations of 0.005, 0.01, and 0.015 M., at p's 10, 10.5, and 11 gave a maximum precipitation of 27% It is possible that the differences petween Euler's results and our own regarding the precipitation at pH 10 are dependent upon the same factors that must be operating in the cases cited. in any event, the lead is not a satisfactory precipitant until these 15. "apparent anomalies are understood. In those cases in which the lead did precipitate the activity, the purity was considerably increased. ‘the results obtained for lot no. 41 are given in Vable vI. Table vI Ppt. pH Yield (%) Aco Dry Ash-free Orig. © 73 10.5 79 | 358 524 11.0 72 | 291 413 Although this lot gave the most effective purification which we secured with lead, it was but seven fold increase over the ori- ginal material. in view.of the fact that similar precipitations in the hands of guler and iyrbach produced 10 to 30 fold increase over their original, the above results do not appear particularly encouraging. here is one further point to ove noted in connection with the experiments on lead, regarding the poisoning of zymase by that substance. A duplicate experiment was performed, using in one case the lead treated co-zymase solution directly, and in the other an identical quantity from which the lead had been removed with hy- drogen sulfide. ‘the two runs were identical in fermentative activity within the limits of experimental error. an analytical determination demonstrated a lead concentration of 0.00048 M. in the fermentation mixture, which wonld seem to exceed the "trace" which Euler and wMyrbach found produced poisoning. 16. Mercury Precipitation The results with mercury as a precipitant for co-zymase were particularly interesting, as they showed certain peculiarities that do not appear to have been previously noted, in biological purifications. | mercuric chloride is by far the most common meroeury salt employed in biological work. vue, however, to its low degree of ionization, it might be predicted that less precipitation would occur with this salt than with mercuric nitrate, for example, at the same molalities. this was verified experimentally. On adding HZ(CN)p, ugClp, and ug(NOz)o of the same concentration to the lead- treated solution, there were observed, respectively, no, slight, and heavy precipitation. this is in exact correspondence with the degree of ionization of these salts. kiore quantitative experiments were therefore performed with HeCle and g(NOs) g. 10 Samples of co-zymase solution, ugUls solution was added to concentrations of 0.0025, 0.005, ana 0.0075 fhe, the pH was adjusted to 7.g - 8.0, and the precipitate centrifuged off. the addition of a few drops more of HgUle solution produced no precipitation or opalescence in the filtrates, indicating "somplete precipitation" in each case. it was therefore of great interest that not only did the yield of active material carried down by the precipitate inerease with increasing concentration of Hgvle, but the Aco's likewise progressively increased. ‘rhe data are given in wable vII. 17. Table vil Hgvly conc. yield (%) aco bry Ash-free orig. 20 Bl 0.9025 25 75 139 0.005 42 125 185 0.0075 65 117 192 the above experiment was then repeated with He (NOz)o at the Same concentrations. whe clear solutions, after centrifuging off the precipitate, were tested with more of the reagent and further precipitation was found to occur with the 0.0025 and 0.005 M. concentrations, but not with the 0.0075. in this case, the yields increased as before, but the purity underwent a constant decrease. (Table VII1) table VIil Molal Yield (%) aco Dry Ash-free orig. 20 31 0.0025 45 135 214 0.005 57 106 166 0.0075 75 106 140 in order to elaborate uvon these observations, the experiments were repeated in a more acid medium. aA pH of 6.5 - 6.7 was employed, and the same concentrations of mercury as above. vrecipitation was practically complete with the uguls at 0.0025 mw. and with the dg(NUg)p at 0.005 «=. in this experiment the results were similar to those at the higher pH, but the yield was in each case lower, as illustrated in table 1X. 18. table 1X Precipitate Orig. HgClo Hg(NOz) 5 gone. molal 0.0025 0.005 0.0075 0.0025 0.005 0.0075 Yield percent — 12 37 38 42 46 59 Aco (ash-free) 31 47 83 121 93 93 the results confirm the experiment of Kuler to the extent that HgUlp to "complete precipitation" in neutral or acid solution is a poor precipitant. “hey are, however, quite different in their final interpretation, for they indicate HgUls to have considerable possibilities in the purification, if used in larger quantities than necessary for "complete precipitation". the experiments are also of importance in connection with biologiecsl purifications employins mercury, as they indicate in- portant differences resulting from different salts of this element, and serve as well to invalidate “complete precipitation" as a reliable guide. | Silver Precipitation Silver, like mercury, was found to be a fairly satisfactory precipitant for co-zymase. Our preliminary experiments indicated approximately neutral solutions to be satisfactory, and the effect of concentration wes therefore investigated at pH 7.2 - 7.3. ‘The results are listed in Table xX. fable xX Conc. (molal) 0.0 0.0025 0.005 0.0075 Yiela (%) 48 61 63 Aco (ash-free) 30 173. - 206 203 Attempts to increase the above yields by altering the acidity met with failure. Occasional batches, however, gave yields of 75% | 19. to 85% at the above pH. As the Aco was likewise fairly high, as will be observed above, it would appear that silver might perhaps be profitably employed. removal of Metals as Sulfides In the case of the three metals described above, lead, mercury, and silver, it was frequently observed that the loss in total co-zymase Was very great. this Was assumed to be due to the fact that when the metals were removed as sulfides, as a prelim- inary to the tests of activity, some of the co-zymase was adsorbed on the precipitated sulfides. An experiment was undertaken to ex- amine this possibility. fo pairs of samples of co-zymase solution ,the above three ions were added to a concentration of 0.0075 ii., and the pH was adjusted in one of each pair to 10 and in the other to 5. Without removing any precipitate, the solutions were saturated with hydrogen sulfide, filtered, and the hydrogen sulfide removed in vacuo. The solutions were then tested and the losses in activity thus determined. ‘they are listed below in able XI in percent of the original amount present. Taple AI Percent Loss in Activity by the Precipitation of Sulfides pi Lead Mercury Silver 5 6 - 16 22 10 0 9 2 In purer samples the above phenomenon was of even greater mag- nitude. With silver, in particular, we had great difficulty, and in a few cases losses of as high as 60% and 80% occurred. As the precipi-~ 20. tation of the sulfides is besi accomplished in acid solutions and as it is here that the greatest losses occur, this factor must be given close attention. We avoided the difficulty as besi we could by pre- cipitating the sulfides from dilute solutions, in which ease the losses were much lower. The matter is of interest hecause of its general bearing on the problem of biological purifications by means of the heavy metals. Cadmium Turning from the heavy metals, we examined the behavior of cadmium, iron, and aluminum hydroxides. Cadmium was not particular- ly effective in carrying down the activity, as indicated by the table below (XII). ‘the experiments were performed by adding a solution of CdClp to the desired molality, adjusting the solution to pH 10.5 - 10.7, centrifuging off the precipitate, neutralizing both portions, and removing the cadmium as a sulfide. Preliminary experiments had shown practically no losses from precipitation of cadmium sulfide. fable XII Conc. (molal) 0.005 0.0075 0.01 Orig. Yield in precipitate i%) 9 24 37 Aco (ash-free) 39 42 42 30 It will be observed that not only are the yields low, but that al- most no purification results. Iron As the precipitation of iron hydroxide in the co-zymase solu- tions was found to be difficult, the nydroxide was freshly prepared, washed, and addea to the co-zymase solution to 0.05 molal. ‘he | él. Solution was made alkaline to pH 10.1 - 10.2, and shaken for ten to twenty Minutes. On centrifuging off the precipitate, neutralizing, and testing, 88% of the original activity was found to remain in the filtrate, while the precipitate was inactive. The 12% loss might well have been a result of the alkalinity of the solution. In any event, iron would not appear to be a useful reagent. Aluminum - The use of aluminum hydroxide as an adsorbent for the co~zymase activity has been discussed by Kuler. He found that the hydroxide was quite an effective adsorbent at pH 10, and deduced the Aco of the material to be rather high. He did not, however, utilize the method. On original extracts, untreated with leag, we found aluminum hydroxide to be quite ineftective. Aluminum sultate was added to give aluminum ion concentrations of 0.005, 0.01, and 0.015 molal, and then NaOH to bring the pH to 10. ‘he precipitates were centri- fuged off and both portions neutralized. ‘Yaole xIII gives the distribution of the co-zymase in per cent of the original. able XIII Gonc. (molal) 0.005 0.01 0.015 Frecipitate 3 7 25 viltrate 90 61 59 However, in repeating the experiment, with the exception that a lead precipitated solution was employed, more promising results were secured, as indicated in Table XIV. . PAA Table XIV Gone. (molel) 0.005 0.01 0.015 Precipitate 77 103 75 filtrate 33 4 0 The difference in the behavior with the two solutions may perhaps be accounted for by the absence of phosphate in the lead precipitated material, as aluminum phosphate, which is surely formed in the original material, may interfere. fie problem of recovering the activity from the hydroxide still remained. Two methods were examined and both found to be useful. whe first consisted in dissolving the precipitate in a minimal quantity of 6 N. hydrochloric acid, diluting, and neutralizing. In the second method the precipitate was suspended in water and adjust- ed to px 5.9 - 6.1 and kept in the ice-box for 12 to 24 hours. As the mixture became more alkaline on standing, it was readjusted once or twice during the period. | Both methods were found to give filtrates containing the major portion of the activity. ‘the second procedure, moreover, resultea in a fairly pure product, the Aco being increased, for example, in one experiment from 30 to 275 (ash-free) by precipitating at pH 10 and reversing the adsorption at pH 6.0. xepeating the atlsorption and reversal once more caused an increase to only 335. Miscellaneous. The remaining reagents which we employed were examined only casually and will, for this reason, be only briefly mentioned. Barium, in accordance with the statements of Kuler, was found to produce no precipitation of the activity, even in solution as 25. alkaline as pi 10.5. | Tannic acid was likewise found to precipitate none of the activity from dilute solutions, and concentrated solutions were not investigated. ‘the filtrate from the tannic acid precipitation was noted to develop only faint precicitates with either phos- photungstic or silicotungstic acids. Silicotungstic and phosphotungstic aciaus were both tried, but without particular success. considerable loss of activity in gen- eral resulted with no very great increase in the purity of the co- aymase preparation. summary Electro dialysis and a number of reagents have been examined with regard to their applicability to purifying co-zymase extracts. Lead salts were found to give variable results. ‘the experi- ments of Kuler and iiyrbach could not be duplicated on the available co-zymase extracts. mercuric chloride and nitrate were found to effect different results. the experiments appear to have considerable bearing on biological purifications with this element. Mercury, silver, and aluminum were found to be most useful. Pasadena, California. 24. HKOOTNOTES « 2. Harden, A., and young, wWed., ge rhysiol., 1905, xxxii, ETO.* bor. poc., 1906, B Iixxvii, 405. 3. Vv. suler, H., and Myrbach,K., 4eit. Physiol. Uhem., 1924, ecxxxix, 15. 4. v. woler, H., and myrbach,s& Zeit. physiol. Uhem., 1924, oxxxix, 281. 5. wtieyerhof, 0., zeit, Physiol. Ghem., 1918, Gi, 165; cii, 1. 6. xaymond, A.l., Pro. tliat. Acad. Sci., 1925, vii, 622 7. 1 wish to express my gratitude to the many persons connected with the brewery, whose continued courtesy and assistance facilitated this research. A.leik. 8. Albert, m., suchner u., and Rapp, R., Ber. chem. ves., 1902, xXxxv, 2378. 9. Accompanying paper. 10. Determined by the method of Sen. senedict and k.U. theis ds biol. vhem., 1924, lIxi, 63