94 Experimental Cell Research, 6, 94-116 (1954) A NEW TRANSFORMING AGENT DETERMINING PATTERN OF METABOLISM OF GLUCOSE AND LACTIC ACID IN PNEUMOCOCCUS' HARRIETT EPHRUSSI-TAYLOR Laboratoire de Génétique de la Faculté des Sciences et du C.N.R.S., Paris, France Received May 22, 1953 Tue starting point of the present investigation was the observation that the Type II encapsulated pneumococci arising from transformations induced in certain unencapsulated strains are of two sorts: the majority give rise to colonies having diameters approximately 3 times larger than rough colonies have, while a few form colonies which are 4 times larger. Clones set up from the two sorts maintain indefinitely, under appropriate conditions, their characteristic difference in colony size. The experiments to be described below are concerned with the origin and nature of the pheumococci which give rise to large colonies (LC pneumococci). The interest of the observation emerges from the following considerations. In the past it has been noted that the size of a pneumococcal colony is often correlated with the amount of capsular polysaccharide secreted (8, 9, 14). Thus, the finding of the LC clones raised the problem, already considered in an earlier publication (15), of whether the Type III capsular agent acts as a discrete unit, or whether, on the contrary, it induces the formation of a spectrum of transformed bacteria exhibiting varying degrees of polysac- charide secretion. If the latter were true, LC clones might be pnéumococci which secrete very large amounts of capsular polysaccharide, and belong thus in the upper limits of such a spectrum. In the earlier study it was con- cluded that the capsular transforming agent acts in an all-or-none fashion. The present study confirms this conclusion, for it was found that the colonies of the transformed Type III pneumococci form two disercte distributions with respect to diameter. No evidence for a spectrum could be obtained. The experiments to be described show that LC pneumococci are produced by a double transformation, in the course of which a bacterium acquires not only the capsular agent, but also a hitherto unrecognised agent, present in extracts prepared from the Type HI strain usually employed in capsular transformation studies. The new agent (LC agent) is apparently independent 1 A grant to the Faculté des Sciences from the Damon Runyon lund for Cancer Research is gratefully acknowledged. ‘ Experimental Cell Research 6 | Ae ee eet ae nm ee Pneumococcal transformations determining cell metabolism 95 of the capsular agent, both as a genetic entity and with respect to its physio- logical activities, for it can be acquired singly by unencapsulated bacteria, producing in them a change which results in their forming colonies which are slightly larger and distinctly more opaque than those of ordinary unen- capsulated clones. A study of the metabolic differences between normal and LC strains will be described. The results of this study demonstrate that LC strains have a distinctive pattern of glucose metabolism, and that they fail to oxidize lactic acid. The present report is subdivided into two chapters, one dealing with piometric and transformation studies, the other describing metabolic studies. I, BIOMETRIC AND TRANSFORMATION STUDIES Two questions were investigated: 1) Are capsular transformations induced in all-or-none fashion? and 2) What is the origin of the LC pneumococci? These problems were investigated by the measurement of many colonies of encapsulated pneumococci, obtained by spreading on petri dishes pop- ulations of unencapsulated pneumococci in which transformations had been induced by treatment with various specific transforming extracts. MATERIAL AND METHODS The methods of culture, preparation of transforming agents, and induction of transformations were those described in a previous publication (4). The only modifica- tion introduced was the more careful standardization of the blood-agar plates on which the colonies to be measured were grown. The important factors to control are glucose content and surface humidity. With respect to the former, 2.0 cc of a sterile 2.5 per cent glucose solution were added for every 150 cc of agar medium. Surface humidity was reduced by drying the plates overnight at 37°C. In any single experiment the same volume of culture, or dilution thereof, was spread on all of the plates. In the early part of the work, sterile defibrinated rabbit blood was employed in the agar medium, while later horse blood was used. Mr. Lemetayer, Director of the Institut Pasteur at Garches, and Mr. Girard, Chef des Services de Saignées, very kindly supplied horseblood in the large quantities which proved necessary for completing the present study. Dilutions of cultures fer plating were made in 2 per cent Difeo Neopeptone dissolved in 0.85 per cent NaCl. Even with above precautions, some variation was observed in the mean diameter of colonies of a given strain when plated on different preparations of medium. Consequently, direct comparison of various experiments was not always possible. Duplicate plates gave, however, very consistent results. AH measurements of colonies were performed after 17 hours of incubation at 37° C. Longer incubation produces larger colonies, but autolysis becomes very pronounced when incubation is continued Experimental Cell Research 6 6 H. Ephrussi-Taylor yeyond the chosen time interval. Measurements were carried out with a Zeiss ocular crew-micrometer mounted on a binocular dissection microscope The overall magni- ication was 9 fold. The petri dishes were held on an inclined stage, and their surface Numinated obliquely, With an even illumination, the colony margins were very harp. The diameter of a colony was measured at right-angles to the direction of the nclination. The pneumococcal strains which were transformed were R36A, the rough strain ised in the classical experiments of Avery, MacLeod and McCarty (2), and SHI-1-TS0, in intermediate smooth strain produced by transformation of the above with an xtract obtained from the mutant Type HI strain SIII-1 (15). At the outset, transforming extracts were prepared from the Type ILI strain AG6. ‘his strain contains the normal, or SIII-N capsular agent, and forms very large mounts of polysaccharide. As the present study evolved, transforming extracts vere prepared from various clones isolated from transformation expcriments. These rill be described in the course of presenting the results. EXPERIMENTAL RESULTS 1. The diameters of SUI-N colonies of independent origin. If the capsular gent induces transformations in an all-or-none fashion, independently ‘ansformed bacteria should be identical with respect to capsule. A de- nitive study of this point is not possible, owing to lack of quantitative and ualitative methods of studying polysaccharide secretion in a large sample f clones set up from such bacteria. However, a procedure was developed hich, with certain reserves, can be considered to serve this purposeé: Differ- nces in the amounts of polysaccharide secreted are often reflected in differ- nees in the diameters of colonies: the more polysaccharide, the larger the slony. Thus, measurements were made on many colonies derived from ype HI pneumococci which had arisen independently of cach other by ansformation of an unencapsulated strain. The characteristics of the dis- ibution obtained were compared with those of a second distribution, ob- ined from measurements of colonies derived from the progeny of a single ype III pneumococcus, also made by transformation. This second distri- ution is a control, describing the variability of the diameters attained by lonies grown from sister cells. If the capsular agent acts in an all-or-none shion, the characteristics of these two sorts of distributions should be entical, for the variability in both instances should be due only to normal iriation in the growth on plates. The first, or experimental distribution, was obtained in the following way. he strain undergoing transformation was treated with the Type III agent ader optimal transformation conditions. After incubation, dilutions of the eated cultures were spread on blood-agar plates. All colonies which were perimental Cell Research 6 Pneumococcal transformations determining cell metabolism 97 more mucoid than the inoculated strain were measured, provided they did not fall in a crowded area. At least four treated cultures were employed for establishing a single distribution. Thus, while some of the mucoid colonies on a plate made from a given culture might be the progeny of a single trans- formed bacterium, the colonies measured for a single distribution cannot be derived from less than four independent transformation events. It is indeed probable that they are derived from very many more. The addition of as little as 0.005 micrograms of DNA, isolated from an SULN strain, to an appropriate medium containing transformable pneumococci suffices to induce the transformation of a few bacteria. It has been found that the addition of 1000 times this amount of nucleic acid induces transformation of about 0.5 per cent of the unencapsulated bacteria during a time interval of 30 minutes, provided the pacteria are in the state which permits them to react with the nucleic acid (10). Under the conditions of the present experiments, the transforming agent is present in about 1000 fold excess of the minimal amount required to obtain transformation. At the time when the bacteria become capable of reacting with the agent, their population density is 16 000 per ce. The reactive state persists for two hours. Assum- ing an incidence of transformation of 0.5 per cent which is a minimum estimate, the encapsulated bacteria appearing at the end of the incubation period normally given to transformation cultures (17-20 hours) would on the average derive from 165, or 80, transformations in cach of the treated cultures. In the earlier experiments, treated and control cultures were incubated - 18-20 hours prior to spreading on plates. However, it was found that this long incubation led to a wide spread in the measured diameters, owing, no doubt, to the onset of adverse conditions after the cessation of growth in the transformation medium. In later experiments, platings were made after only 12-14 hours of incubation. Nonetheless, the distributions of colony diameters obtained from transformation experiments are somewhat wider than those observed when the same bacteria are taken from blood-broth medium. The clone utilized for preparing the control distribution was derived from a single Type III colony isolated from a previous transformation of the same strain upon which the experimental distribution was to be obtained. Control clones were grown under the same conditions as those employed for inducing transformations, in order to eliminate differences in the distributions which might be due to or result from differences in cultural conditions prior to plating. The results of a typical expcriment are shown in Fig. 1. In this particular instance, the strain undergoing transformation was SILI-1-T50. As has been already stated, the control clone came from a single SIII-N colony recovered Experimental Cell Research 6 98 I. Ephrussi- Taylor 1S) MEAN 1.615 4 0.47 Fig. 1. Histograms constructed from the meas- urements of diameters of colonies of encap- sulated Type III pneumococci. On the left, the bacteria came from four cultures in which multiple transformations had been induced. On the right, the bacteria came from a clone iso- lated a few days previously from a transformation experiment. The strain undergoing transforma- tion was SIIIJ-I-T50. Both the means and the Standard deviations are essentially the same. 5 NUMBER OF COLONES a 14° 16 18 20 14° 16 * MICROMETER UNITS 20 2.2 from a transformation of strain SHI-1-T50 a few days earlier. It can be seen that the distributions of the diameters of the SIII-N colonies are essentially the same whether the bacteria come from the control clone, or directly from a transformation culture. It is clear, therefore, that there is no obvious differ- ence between the various SIII-N colonies which arise from the control clone, and those which arise from different transformed SHI-1 pneumococci. Thus, the SILI-N agent does not appear to give rise to a spectrum of pneumococci having graded polysaccharide secretion, but on the contrary seems to behave as a discrete entity. If now new clones are set up, seleeting the largest and smallest of the colonies of a particular distribution, usually, all of these clones give rise to essentially identical distributions, when the colonies formed by them are measured, and these distributions do not differ from the original one. However, the following exception to this rule is observed: from time to time, when a clone is set up from a very large colony, it gives rise to colonies which are on the average about 25 per cent larger in diameter than the majority of SIII-N colonies measured. These constitute the LC clones. The appearance of these clones in the transformation cultures is very rare. Often, none are found. However, in one out of ten distributions obtained from transformed cultures of strain SII-1-T50, a definite bimodality was found. This aberrant distribution is shown in Fig. 2. The second mode was found to be due to the presence of an unusually large number of LC clones in this particular experiment. Attempts were made to determine the average number of cells per colony in the two kinds of clones. Several well isolated colonies from cach were homogenized in a measured volume of medium, and serial dilutions of the suspensions plated. It was found that whereas the results were concordant when the colony came from an SIII-N- Experimental Cell Research 6 we one Pneumococcal transformations determining cell metabolism 99 vig. 2. Histogram constructed from _Measure- ments of encapsulated Type HI colonies grown m bacteria of four cultures in which multiple Hensformations had been induced. Out ten sets of such measurements, only this one s owe l pimodality. The peak on the right is caused by the presence of an exceptionally large number of large colonies in this particular experimen’. The peak on the left corresponds to the single i 6 28 30 ak in Figure 1. 12°14 16 18 2.0 22 24 2 peak in Fig MICROMETER UNITS 3 NUMBER OF COLONIES i isted LC clone, they were aberrant when the colony was SIII-N. The aberration const ee 7 n excess of colonies in the higher dilutions. An Pere was UP to see wy ether concordant colony counts would not be obtained if the initia : P yon whe incubated a short time before diluting and plating. This was tore ailating ine ve. ions 15 minutes at 37°C, and one hour in the refrigerator before di i e a sureading on plates. The counts on serial dilutions were then in nani ettnn pes ile th erage number of cells per colony 0 -N- , { clones. Furthermore, while the av oN Pe comained unchanged by this treatment, the number of certs pe vheexperiments ‘ a nt. ' io of these averages was 3:1. > increased by 30 per cent. The ratio o , ; was ie in that they suggest that SIII-N colonies may be smaller pecause these ae teria have a lowered viability under aerobic conditions, undergoing autoly ac @ more readily. . | From these experiments it may be concluded that the SII-N preamoe produced by the action of the SJH-N transforming agent on ne, se and SIII-1-T50 fall into one of two categories of wet the ea frequent: clones which form colonies of the diameter nn n : oe and clones characterized by og Le a cee tate fen this fave ive diameters of normal anc clones. voce obtained by measuring the colonics produced by one ne Fone and one normal clone, grown under identical conditions and measure the Stes ne Fig. 3 also includes measurements made the same day on strain sr oean intermediate smooth Type HI clone which has been mentione< vee quently in previous publications (4, 13, 15). The three clones q distinct. In the course of this study, a surprising mutability was oper in ihe clones derived from the transformed bacteria. In one experiment, an Sens once clones, two contained mutant forms which were less meet a ie ” i eco Growth of the tet tear fare. ‘This may be ‘due toa teal mutability of the capsular oe 7 i these tran stor . the other hand, it may be that unencap- sulted strains ‘lend to nnvotee toward a state in which capsular synthesis is disad- van W. f Ww ues V ransformation upsets the tageous, so that endo ing them ith capsu ] b t et pat te oO lee der 1 1ces, Mutations of capsular agent mn f sele ction. Un er these circumstance ’ the a ’ Experimental Cell Research 6 100 . Ephrussi- Taylor SII-N 30} wn w z 920 Fig. 3. Histograms constructed from measurements of 8 colonies grown from three different clones of pneumococci, . Clones SIII-N and SHI-N-LC were isolated from single ° colonies, the measurements of which are included in Fig. 2. 2 10 Clone STII-N came from a colony in the left hand distribu- a tion, while clone SIII-N-LC from a colony in the right z hand one. For comparison, colonies of mutant Type If 2 strain SIUI-2 were measured simultancously and included in the diagram. Os 1.0 1.5 MICROMETER UNITS leading to diminished capsule synthesis, would be favored, as would also be favored mutations reversing the trend established by prolonged existence in the unencapsu- lated state. 2. The homogeneity of type III strain A66. Since the transforming extract of strain A66 obviously induces the formation of two kinds of Type HI pneumo- cocci, it becomes essential to know whether strain AGG is heterogeneous, being composed of these two cell-types, or whether, on the contrary, both activities reside in the nucleic acid fraction of one cell-type. Therefore, meas- urements were made upon colonies of the stock strain AGG. On the same day, and on the same media, measurements were also made on colonics of a clone of SIII-N pneumococci, isolated from a previous transformation experi- ment. The results of these measurements are shown in Figure 4. It is clear from these measurements that the colonies of strain A66 have a mean dia- meter about 20 per cent larger than that of the SILI-N colonies of the control clone. Thus, strain A66 is best described as an SIJI-N-LC clone. There is no indication that the smaller colony type is present in the population AG6 in any appreciable numbers. This being so, the simplest working hypothesis consists of supposing that both colony types arise from the activities of the nucleic acid isolated from an essentially homogeneous population, and that the two activities are due to two factors which reside in each pneumococcus in the AG66 population. One of these factors is the capsular agent, while the other is presumably a new agent, responsible for the large colony size. The rareness with which SIN-N-LC pneumococci are formed by transformation would thus be the consequence of the rareness of two different transformations taking place in the same bacterium. Experimental Cell Research 6 ae er et ee ee ~ or Pneumococcal transformations determining cell metabolism 101 1$L | ” w Zz. Stok 4 a ° V9 wh 6 4 . at 4 Fig. 4. Histograms constructed from measurements of colonies & derived from strain AGG, on the right, and an STEN clone on 2 the left. The SILI-N clone was made by transformation of an 2 SUF-1 bacterium by an extract of strain AGG. 14 16 is 20° 22° 2% MICROMETER UNITS 3. The capsular agents present in SIII-N and SILU-N-LC strains. lf the above hypothesis is correct, the capsular agents in both types of clones should be the same. To test this point, nucleic acid extracts were prepared from two clones, one SIH-N-LC and one SHI-N (TPs 36E and 30B, respectively). With these extracts, transformations were induced in strains R36A and SIII-1-T50, the treated populations were spread on plates, and the colonies obtained measured in the usual fashion. In a typical experiment, 63 smooth colonies from plates of cultures treated with TP 36E and 96 from plates of cultures treated with TP 30B were measured, vielding the mean values of 2.02 + 0.173 and 1.954-0.117 micrometer units respectively. The difference between the means is not significant, and it can be concluded, therefore, that both TP prepara- tions induce the formation of essentially the same sort of encapsulated bacteria. Thus, strain SILI-N-LC contains the same capsular agent as strain SITI-N. However, it was noted in the course of these experiments that in the cultures treated with TP 36E there occasionally appear some SILI-N-LC clones, while none are found in the cultures treated with TP 30B. In its biological activity, TP 3615 thus resembles the transforming extracts prepared from strain A66. This finding is entirely in agreement with the hypothesis that the LC trait is determined by an agent similar to the capsular agent, which can be transferred to pneumococci by the transformation technique. 4. The isolation of unencapsulated LC clones. If the properties of the LC strains are determined by a transforming agent, and if this agent is indepen- ent of the capsular agent, then one should expect unencapsulated bacteria to be able to acquire the former without the latter. Indeed, this transformation should occur more frequently than the double transformation in which Experimental Cell Research 6 102 H. Ephrussi- Taylor both are acquired. Detection of R-LC or SIII-N-LC bacteria would depend upon whether transformation by the LC agent alone suffices to produce a recognizable morphological change. A careful search among the previously supposed “‘untransformed’’ mem- bers of treated populations of R and SIII-1 pneumococci revealed that some of the colonies produced by them were, in fact, different from ordinary R or SIII-1 colonies in that they were somewhat larger and very much more opaque. The incidence of the aberrant colonies in the treated ‘cultures was never greater than 0.7 per cent. None were ever found in cultures which had not received a nucleic acid preparation from an LC strain. Clones derived from the aberrant colonies were established, but these could be maintained free of reversions only when carried on solid medium (blood-agar). Measure- ments of colonies of normal and aberrant clones of strain SIII-1-T50 gave the average values of 0.796+0.143 and 0.983+0.192 micrometer units. The differences between the means is not significant, and, indeed, on the basis of size alone it would be difficult to ascertain that two distinct colony types exist. However, the opacity of the aberrant colonies permits an almost certain differentiation when the plates are examined in transmitted light. Since it was suspected that these aberrant colonies might be derived from pneumococci which had acquired the hypothetical LC agent but not the SHII-N agent, the next step was to transform them with the capsular agent, to see whether or not they yielded SIII-N-LC bacteria. Just prior to subjecting various clones to transformation, they were transferred from plates into liquid medium, using large inocula. These cultures served to inoculate trans- formation cultures and were discarded afterwards. In the particular experi- ments to be described, the aberrant clones were transformed with a nucleic acid extract of strain A66. At the same time, controls were made by trans- forming normal strains with the same nucleic acid preparation. The treated populations were streaked on sectors of agar medium plates, rather than being spread for measurement. From each treated population, one trans- formed ‘colony was chosen at random, grown out in blood-broth medium, and samples of the resulting culture plated for measurement: Quadruplicate transformation cultures were made of each aberrant clone and of each control strain. Thus, for each strain treated with the nucleic acid fraction of strain AG6, the progenies of four independently transformed bacteria were measured. This method of study was chosen to avoid the greater variability in colony diameter which is observed when measurements are made directly on the colonies of the transformation cultures themselves. The results of some meas- urements are shown in Fig. 5. From these observations, it is clear that when Experimental Cell Research 6 Pneumococcal transformations determining cell metabolism 103 8 Fig. 5. Histograms constructed from measurements of SILI-N colonies derived from transformation of normal and aberrant clones of strains SIII-1-T50 and R36A. For the pair of curves on the left, the left-hand distri- pution came from measurements of colonies of trans- formed normal] SIII-1 pneumococci, while the distri- pution on the right came from measurements of trans- formed aberrant SIII-1 pneumococci. The pair of gistribution curves to the right are the results of similar measurements on transformed normal and aberrant R36A. In both cases, the transformed aber- rant clones gave rise to larger colonies on the average than did the transformed normal clones. oF COLONIES 8 NUMBER & O 45 2.0 25 30 1s 290 25 MICROMETER UNITS the clone undergoing transformation is an aberrant clone, be it R or SIII-1, the SIH-N bacteria induced are ail SIII-N-LC; that is, they form colonies with a mean diameter roughly 25 per cent larger than that of colonies of the majority of the transformed bacteria arising in the normal strains of R or SIII-1 pneumococci. This result is entirely consistent with the hypothesis that the aberrant R and SIII-1 strains already contain the postulated LC agent, which they had acquired independently of the capsular agent in a previous trans- formation. The aberrant R and SIT-1 clones will thus be referred to as R-LC and SUI-1-LC clones. 5. A quantitative study demonstrating that the LC agent is a specific com- ponent of nucleic acid extract 36E. Pneumococci exhibiting the LC phenotype were found only when R or SIH-1 bacteria were treated with a nucleic acid fraction prepared from strain AG6, or from an SIIT-N-LC strain recovered from a transformation experiment. Since in general the incidence of LC clones is low in a transformed population, it was conceivable that failure to find such clones after treatment with other nucleic acids might be due to sampling error. Accordingly, an experiment was performed in which popula- tions of strain R3G6A were treated with nucleic acid extracts 36E and 30B, the former obtained from an SIII-N-LC strain and the latter from SJII-N, the incidence of R-LC colonies was scored, and the data analysed statistically. A total of 11 transformation cultures were studied, 4 receiving TP-36E and 7 receiving TP-30B. Samples of the treated cultures were spread on plates, and after incubation, the total number of colonies and the number of LC colonies per plate were estimated from counts. A large number of colonies from the seven cultures treated with TP 30B were carefully examined. After Experimental Cell Research 6 104 HW, Ephrussi-Taylor searching for LC colonies by holding the plates against a light source, they were examined under a binocular microscope. Whereas ten LC colonies were found on the plates made from the four cultures treated with TP-36E, none were found on those made from the seven cultures treated with TP-30B (see Table I). In view of the large number of colonies examined, it is very unlikely that the observed difference is due to sampling. The experiment thus confirms the hypothesis that TP-36E contains a specific agent responsible for the appearance of the LC bacteria in the treated populations and that this agent is absent from TP-30B. All of the biometric and transformation studies agree with the hypothesis that the nucleic acid fractions of strain A66 and of SIII-N-LC strains contain a hitherto unrecognized transforming agent which is responsible for the LC phenotype. The presence of this agent in strain AG6 accounts for the size difference observed between colonies of this strain and the majority of colonies arising from transformations induced with nucleic acid derived from it, for the two transformable strains most frequently used in transformation experiments differ from strain AG6 by at least two factors: the capsular agent and the LC agent. Transformations induced by nucleic acid extracts prepared either from strain A66 or from SIII-N-LC strains produced by transforma- tion thus consist of the following inductions: SII-N Single transformation R (or Sil-1) —2P-A86 R-LC (or SIII-1-LC) single transformation SITI-N-LC i double transformation Ii. METABOLIC STUDIES . The inconvenience of colony diameter as a criterion for the identification of clones is that it lacks specificity. Thus, the homogeneity of the class of strains which have been called LC remained doubtful until a more specific criterion for identifying them was found. The metabolic studies to be pre- sented below provided such a criterion. The observed differences in colony size are manifested on blood agar plates; that is, under aerobic conditions and where glucose limits the amount of growth. A metabolic difference was therefore sought in the acrobic meta- bolism of glucose by normal and LC strains. In addition, a second possible difference was sought, since it had been noticed that the two types of colonies produced different degrees of greening of the blood. Since greening is due to the liberation of H,O, by the bacteria, manometric experiments on glucose Experimental Cell Research 6 eee ean Pneumococcal transformations determining cell metabolism 105 oxidation were set up in such a way as to permit a simultaneous study of ihe amount of H,O, formed by each of the strains. No systematic study has apparently ever been made of the oxidative metabolism of pneumococci. Finkle (5) observed differences in the Qo, of three smooth races of different capsular type, as well as in rough races derived from the latter. In phosphate buffer, the optimum O, consumption was found to occur at pH 7.8. Observations of Sevag and Maiweg (13) demonstrated that without the addition of protective agents to the bacterial suspensions, respiration was rapidly inhibited py the accumulation of H,O,. Catalase and pyruvate were effective in maintaining enzymatic activity, the former by destroying peroxide, the latter by reacting chemi- cally with it. —_—