Effects of Elevated Dissolved CO2 Levels on Batch and Continuous ...

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Nov. 1997, p. 4171–4177 0099-2240/97/$04.0010 Copyright © 1997, American Society for Microbiology

Vol. 63, No. 11

Effects of Elevated Dissolved CO2 Levels on Batch and Continuous Cultures of Aspergillus niger A60: an Evaluation of Experimental Methods MHAIRI MCINTYRE

AND

BRIAN MCNEIL*

Fermentation Centre, University of Strathclyde, Glasgow G1 1XW, United Kingdom Received 18 March 1997/Accepted 16 July 1997

The effects of elevated levels of dissolved carbon dioxide (dCO2), produced by gassing with CO2-enriched gas mixtures, upon an industrial strain of Aspergillus niger (strain A60) producing citrate and gluconate were quantitatively assessed. Particular attention was paid to the reliability and accuracy of the steam-sterilizable dCO2 probe, especially in the presence of high concentrations of potentially interfering acidic species. The response of the organism to elevated dCO2 levels was assessed by using both batch and chemostat cultures, and the sensitivity of the organism in different growth phases (lag, exponential, and stationary) was examined. Chemostat cultures showed markedly less inhibition (in terms of biomass and organic acid synthesis) than did batch cultures. Studies in batch culture indicated that lag-phase cultures were especially sensitive to elevated dCO2 levels. Overall, the results of this study indicate that previous experimental methods used to examine dCO2 effects in submerged cultures (continuous CO2-enriched gassing of batch cultures from time zero) have been inappropriate and have led to systematic overestimation of the inhibitory effects of dCO2 on mycelial organisms.

num was the fact there was no quantification of the morphological effect of CO2. In the absence of such information, it is difficult to assess the precise effect of CO2 in terms of morphological alterations. Given the profound impact of mycelial morphology on culture rheology and thus on mass momentum and heat transfer, this is unfortunate. Recent technological advances in tools for morphological analysis and in the application of these tools have emphasized the importance of such data in bioprocess monitoring (1, 4, 12, 31). More recent studies on CO2 effects (20, 21) on mycelial microorganisms in submerged culture have addressed some of the problems detailed above and have provided a greater understanding of the effects caused, particularly with reference to morphology. With some of the difficulties in techniques and applications of technologies already dealt with, in the current study we attempted to advance this area of work further by addressing some of the remaining difficulties. One piece of technology which has become available relatively recently is the steam-sterilizable dCO2 probe (27), which is of great value in providing direct measurements of the concentration of dCO2 (the inhibitory species) in a fermentation vessel. However, it has been reported that the performance of the dCO2 probe may be affected by the presence of organic acids (27). As this could present a problem in investigations of the effects of elevated dCO2 levels, a study of the probe characteristics is necessary. Another difficulty with previous studies on the effects of elevated CO2 levels on fungi is the fact that these studies were usually carried out with batch cultures. Normally, the effects of elevated dCO2 levels in late-exponential-phase industrial cultures have been simulated by gassing with CO2-enriched air from the start of the laboratory scale batch process. This is almost certainly not an appropriate method, in view of the problems relating to the complex relationship between dCO2 and gas phase CO2 (see above) and also because lag-phase or early-exponential-phase cultures are exposed (inappropriately) to elevated dCO2 levels. In a largescale industrial process, high levels of CO2 are present only

The toxicity of carbon dioxide to a wide range of microorganisms has been demonstrated (8, 16). Initial studies of this process focused on the potential use of CO2 as an agent to control the growth of pathogenic organisms (25) or food spoilage organisms (10, 22, 32). Later, it became apparent that organisms utilized in submerged liquid culture systems might also be subject to CO2 inhibition. In these cases, CO2 produced by the metabolism of the microorganism itself is the cause of the inhibition (24). CO2 inhibition is likely to be most apparent in large fermentors (i.e., industrial-scale fermentors), in which increased hydrostatic pressure near the bases of the vessels results in increased dissolved CO2 (dCO2) concentrations (15). Inhibitory effects of CO2 in submerged bioprocesses have been well-documented, particularly in studies involving bacteria (11, 18, 23) and yeasts (5, 16, 17). Decreased growth (5, 8, 11, 16–18) and decreased product formation (8, 16, 23) are the most important effects observed. Recently, there has been considerable interest in the effect of elevated dCO2 levels in submerged liquid cultures of filamentous fungi, including Penicillium chrysogenum (13, 14, 26, 29, 30) and Aspergillus niger (20, 21). The general trend reported was that, as the CO2 level in the influent gas phase was increased, growth and product formation decreased. There may, however, be several fundamental problems with the techniques used in these studies. First, there is the question of which species of dCO2 is responsible for the effects observed. It is only at pH levels below 4.0 that dCO2 (i.e., CO2 (aq)) is the major species of CO2 present (6). In the studies on P. chrysogenum (4, 13, 26, 29, 30) pH was controlled at 6.50, a value at which the inhibitory effects cannot be clearly ascribed to a single inhibitory species unambiguously. An additional limitation of previous studies on P. chrysoge-

* Corresponding author. Mailing address: Fermentation Centre, Royal College Building, University of Strathclyde, 204 George St., Glasgow, G1 1XW, United Kingdom. Phone: 0141 552 4400. Fax: 0141 553 1161. 4171

4172

MCINTYRE AND MCNEIL

APPL. ENVIRON. MICROBIOL.

toward the ends of periods of rapid growth or when biomass levels are high (15), not from the start of a batch process. In view of the difficulties with previous experimental methods, in the current investigation we attempted to determine the sensitivity of each of the growth phases of A. niger to CO2 inhibition in batch fermentations. Our findings were then related to previous studies to determine their usefulness in simulating the response of actual cultures at the industrial scale. More realistic strategies for experimental design are also discussed below, particularly the use of chemostat continuous cultures, which has been seriously neglected in studies of CO2 inhibition of microorganisms. MATERIALS AND METHODS dCO2 probe. The response of a steam-sterilizable dCO2 probe (Ingold) to 6 and 12% (vol/vol) CO2 in air in the influent gas stream was monitored (Fig. 1). Calibration of the probe has been described previously (27). The starting point was a steady probe response to gassing with air (at a flow rate of 1 vol/vol/min [vvm]), which returned to a steady response following sparging with a CO2-air mixture. The probe output (in millibars) was recorded at 15-s intervals, and the CO2-air mixture was replaced with air once a steady readout had been achieved with the gas mixture for 15 min. The effect of citric acid on the probe response was monitored. The bioreactor was supplied with a gas mixture containing 5% (vol/vol) CO2, and, when a steady readout had been achieved, citrate additions were begun. Aliquots of a stock solution of citrate were added to give final concentrations of 10, 20, 40, 60, 80, and 100 kg m23 in the bioprocess broth. Successive additions of citrate were made following a steady probe response for 15 min. Probe response was again recorded at 15-s intervals. No pH change occurred upon citrate addition. The conditions used for the probe experiments were the same as those used for the batch fermentations, as detailed below; the medium used was the medium used for the batch experiments, and the pH was 1.8. Submerged liquid processes. (i) Microorganism. An industrial strain of A. niger (strain A60) was used in this study. Spore slants for inoculum purposes were prepared on potato dextrose agar (Unipath, Ltd.) in 0.02-dm3 universal bottles. Cultures for spore production were grown at 30°C for 7 days. The medium used for inoculum development, batch processes, and the batch phase of continuous processes contained (per cubic meter) 140 kg of sucrose, 2.0 kg of KH2PO4, 2.0 kg of (NH4)2SO4 z 7H2O, 0.5 kg of MgSO4 z 7H2O, 0.1 3 1023 kg of ZnSO4 z 7H2O, 0.1 3 1023 kg of (NH4)2SO4Fe2(SO4)2 z 24H2O, and 0.06 3 1023 kg of CuSO4 z 5H2O. The pH was adjusted to 3.5. The medium used for the continuous phase of chemostat studies was the same as the medium described above except that it contained (per cubic meter) 50 kg of sucrose, 1.0 kg of KH2PO4, 1.0 kg of (NH4)2SO4, and 0.1 kg of MgSO4 z 7H2O. The pH was adjusted to 3.5. The bioreactor inoculum was 0.8 dm3 of a 48-h-old shake flask culture grown at 30°C and 200 rpm. (ii) Batch processes. The bioreactor used in this study was a Biostat ED ESIO bioreactor (B. Braun Biotech International) with a working volume of 8 dm3. Fermentations were performed at 30°C, and the stirrer speed was 300 rpm. The culture pH fell steadily to 1.80, reaching this value after between 43 and 74 h depending on the influent CO2 concentration used. The culture was aerated at a rate of 1 vvm. Gas flow rates and pressures were monitored and independently controlled by using a gas-mixing unit (model 881611; B. Braun Biotech International) and a thermal mass flow meter (B. Braun Biotech International). Processes were carried out with a pulse or pulses of CO2 at a defined stage in the batch process. In each case the pulse consisted of 7.5% (vol/vol) CO2 in air. The dCO2 in the fermentation broth was measured with a steam-sterilizable carbon dioxide probe (Ingold). In addition to a standard process with air as the input gas, processes were carried out with 2, 4, 7.5, 10, and 18% (vol/vol) CO2 in air in the influent gas stream. The dynamic phase response of the culture to step changes in the influent CO2 level was closely monitored. (iii) Analyses. Dry weight was estimated by filtering 0.005 dm3 of culture broth through Whatman GF/C filter paper (diameter, 4.25 cm). Filter cakes were washed with 2 volumes of sterile distilled water, dried for 20 min in a microwave oven (650 W) on low power, and cooled before weighing. The amount of sucrose in culture filtrates was determined by the method of Dubois et al. (9). The citric acid concentration was determined by the method of Marier and Boulet (19). The concentration of NH41 in culture filtrates was determined by using a kit (catalog no. 640) provided by Sigma, U.K. Ltd. The concentration of carbon dioxide in the exit gas was monitored with a model ADC 7000 infrared gas analyzer (The Analytical Development Co.), and the concentration of oxygen in the fermentor offgas was monitored with a series 500 Servomex Paramagnetic oxygen analyzer (Sybron Taylor).

FIG. 1. (a) dCO2 probe response to 5% (vol/vol) CO2 in air in the influent gas. (b) dCO2 probe response to 12% (vol/vol) CO2 in air in the influent gas. In panels a and b the arrow pointing up indicates a switch from air to a CO2 mixture, and the arrow pointing down indicates a switch back to air. (c) Effect of citrate concentration on probe response when the influent gas contained 6% CO2. The arrows indicate points of citrate addition.

dCO2 AND A. NIGER

VOL. 63, 1997

4173

FIG. 2. Levels of dCO2, exit CO2, and oxygen consumption and biomass, sucrose, citrate, and NH41 concentrations over the time course of a batch process exposed to 7.5% (vol/vol) CO2 in air between 15.5 and 18 h. (iv) Morphological analysis. Morphological measurements were made by using a PC-based system and image analysis software (Aequitas Image Analysis; Dynamic Data Links). Images were captured with a monochrome video camera (model XC-77CE; DDC video camera module; Sony) connected to a microscope (model CH; Olympus) with a total magnification of 3400. Measurements were made manually by using frozen images and a mouse to skeletonize individual hyphal trees (“organisms”) and to measure hyphal dimensions. By using individual mycelial particles (hyphal trees), the mean main hyphal length, the mean branch length, and the mean hyphal growth unit were determined for each sample. The mean values were estimated from measurements of 50 mycelial particles for each fermentation sample.

RESULTS The response of the dCO2 probe was monitored when the influent gas contained a defined concentration of CO2 in air. Figures 1a and b show the responses to 6 and 12% (vol/vol) CO2 in air, respectively. In both cases the CO2-containing gas mixture was introduced into the fermentor once a steady readout with air had been achieved, and the CO2 mixture was replaced with air once a steady response had been achieved for 15 min.

Figures 1a and b show that after an initial lag the probe level reached 80% of its steady-state value within 5 min of being exposed to elevated CO2 levels. The effect of citrate on the probe response was also monitored (Fig. 1c) when the fermentor was sparged with a gas mixture containing 6% (vol/vol) CO2 in air. Prior to the addition of citrate, a steady readout with 6% CO2 had been achieved for 15 min. With additions up to the 40-kg m23 level, minor fluctuations in the probe response were observed, (66% of the steady-state value). From around 100 min after citrate addition, however, the response was approximately steady, with only minor fluctuations in the recorded dCO2 values. A series of batch fermentations was carried out with a pulse of elevated CO2 introduced into the fermentor at defined stages in the fermentation process (Fig. 2 through 4). In all pulse experiments CO2 was introduced into the fermentor in the influent gas stream at a level of 7.5% (vol/vol) in air. This set of experiments was designed to determine which stage of the fermentation was most sensitive to elevated levels

4174

MCINTYRE AND MCNEIL

APPL. ENVIRON. MICROBIOL.

FIG. 3. Levels of dCO2, exit CO2, and oxygen consumption and biomass, sucrose, citrate, and NH41 concentrations over the time course of a batch process exposed to 7.5% (vol/vol) CO2 in air between 16 and 18.5 h and between 183.5 and 186 h.

of dCO2. Levels of dCO2 equivalent to those produced by gassing with a gas mixture containing a CO2 concentration of 7.5% may be expected in large-scale fermentors containing rapidly growing cultures or high levels of biomass (16). For this reason, pulses of 7.5% CO2 lasting 2.5 h were introduced into batch cultures of A. niger in the early exponential phase (Fig. 2) and the stationary phase (Fig. 3). A third batch process experiment was carried out with A. niger which was exposed to an influent gas mixture containing 7.5% CO2 for the first 24 h of cultivation to examine the effect of moderately elevated CO2 levels on subsequent culture behavior. The time course for this process is shown in Fig. 4. DISCUSSION As can be seen from Fig. 1, the dCO2 probe was shown to be a reliable means of measuring dCO2 in fermentation broth. The probe response to changes in influent CO2 level was rapid, with a good correlation between influent CO2 level and the level of CO2 dissolved in the liquid phase of the fermentor. Recently, there has been much discussion concerning the

concentrations of carbon dioxide in a culture medium containing an active microorganism (2, 15, 28, 33). It is usually assumed that the concentration of CO2 in the liquid phase in a bioreactor is in equilibrium with the concentration measured in the exit gas phase, that is, that Henry’s law is applicable (2, 3). This assumption underpins many previous studies on effects of elevated CO2 levels in submerged bioprocesses. However, it has been reported that under normal fermentation conditions, the level of CO2 dissolved in the culture broth may be significantly higher than the equilibrium concentration, particularly when culture viscosity is high (7). The use of a dCO2 probe and exit gas analysis in the present study, however, made it possible to measure the level of CO2 dissolved in the liquid phase and the CO2 concentration in the gas phase. Results from this study show good correlation between gas and liquid phase CO2 concentrations, with a linear relationship between the measured values for dCO2 levels and exit CO2 levels under both test conditions used (Fig. 1a and b) and also in the elevated CO2 processes discussed in previous studies (20, 21). Similarly, the results presented here (Fig. 1c) do not confirm

VOL. 63, 1997

dCO2 AND A. NIGER

4175

FIG. 4. Levels of dCO2, exit CO2, and oxygen consumption and biomass, sucrose, citrate, and NH41 concentrations over the time course of a batch process exposed to 7.5% (vol/vol) CO2 in air between 0 and 24 h.

previously stated reservations regarding the probe response being affected by the presence of organic acids (27). As Fig. 1c shows, minor fluctuations in probe response were observed when citrate was introduced into the fermentor at levels of 10 to 40 kg m23. However, the response returned to the steadystate value within 10 min. In an organic acid-producing fermentation system, such as a system in which citric acid is produced by A. niger, the rate at which citrate appears in the broth is slower and more gradual than the rate observed in the test. We concluded, therefore, that the probe was a reliable means of determining dCO2 levels in cultures, even in the presence of citrate. Once the reliability of the dCO2 probe had been established, bioprocess experiments were performed with elevated CO2 levels in the influent gas stream. The effects of exposure to a pulse of CO2 on the growth and metabolism of A. niger A60 are shown in Fig. 2 through 4. By comparison with a standard batch process for this organism (20), no significant effect was observed on the culture exposed to 7.5% CO2 for 2.5 h in the early exponential phase. However, both growth (as measured

by biomass concentration) and citrate production were reduced in the process when 7.5% CO2 was introduced into the fermentor for 2.5 h in both the early exponential phase and the stationary phase, and a significantly greater inhibitory effect was observed in the process exposed to 7.5% CO2 for the first 24 h (the lag phase). However, in none of the pulse fermentations was the inhibitory effect as marked as the inhibitory effect in the process continuously exposed to 7.5% CO2 (Table 1). From the results presented here, it appears that a culture was most vulnerable to CO2 inhibition in the lag phase. When a culture is inoculated into a fermentor, it is subject to many pressures due to the differences between the shake flask environment and the bioreactor environment, such as differences in pH and substrate concentration. When elevated dCO2 levels in the fermentation fluid are also considered, it is likely that there are marked inhibitory effects on the culture. In the process exposed to elevated CO2 levels during the lag phase (up to 24 h) (Fig. 4), a certain amount of recovery appears to have occurred once the stress of the early elevated CO2 level was

4176

MCINTYRE AND MCNEIL

APPL. ENVIRON. MICROBIOL.

TABLE 1. Reduction (relative to the control process value) in maximum biomass concentration (batch processes) or steady-state biomass concentration when the influent gas stream contained an elevated level of CO2a Concn of CO2 (%, vol/vol) in influent gas stream

% Reduction in biomass concn relative to control process Batch processes

1 2 3 4 5 7.5 10 15 18 a

Steady-state processes

16

0

26

6

32 39

17 20

51

31

Air was the sparge gas in the control.

removed. This can be seen from the data in Table 2, which show that biomass and citrate production were not reduced by as much as in the batch process continuously exposed to 7.5% CO2. The culture was, however, inhibited to a degree, and this inhibition may be attributed solely to the elevated CO2 level entering the fermentor in the first 24 h of the process. Relating the results described above to the situation in industrial-scale cultures, it is unlikely that a culture growing on an industrial scale would be subjected to elevated CO2 concentrations at the early stage of a process. It is only when biomass levels or growth rates are high that levels of metabolically produced CO2 are high in the fermentation broth (24). In view of this, our results call into question the methods employed in previous studies to determine the effects of elevated carbon dioxide levels in submerged liquid bioprocesses. Most previous investigations involved the use of batch cultures which were continuously gassed (from time zero to the end of the process) with a set level of CO2 in the influent gas stream (4, 5, 13, 17, 20, 26, 29, 30) to simulate elevated dCO2 levels in the process fluid. As the results presented in Table 1 demonstrate, gassing with an elevated CO2 level from time zero represents an unrealistic simulation method. In such studies, observed degrees of inhibition are likely to be of a magnitude greater than that expected in an industrial-scale fermentor when metabolically produced CO2 causes increases in dCO2 levels much later in the bioprocess. In industry, high levels of dCO2 are not likely to be present during the most sensitive phase of growth (i.e., the lag phase). A disproportionate level of inhibition is TABLE 2. Maximum biomass concentration and maximum citrate concentration obtained when air (standard) or 7.5% CO2 was the sparge gas for the entire process or when the process was exposed to a pulse of CO2 Fermentation process

Standard 7.5% CO2 continuous gassing Pulse 1 (exponential phase)a Pulse 2 (exponential and stationary phases)b Pulse 3 (first 24 h of lag phase)c a b c

See Fig. 2. See Fig. 3. See Fig. 4.

Maximum amt of biomass (kg m23)

Maximum citrate concn (kg m23)

18.2 11.8 17.5 16.3

78 27 69 61

13.0

40

caused during this early stage possibly due to the added pressures on the organism, as discussed above. A more realistic approach for quantifying the effects of carbon dioxide inhibition in growing cultures is the use of continuous cultures. Few such studies exist in this area, but those which have been carried out (21) have shown the limitations of the earlier approach. Table 1 summarizes the degree of inhibition of biomass formation in both batch (20) and continuous steady-state (21) processes with various levels of CO2 added to the influent gas stream. It is apparent that the level of inhibition is far lower in a continuous steady-state culture than in a batch culture. In a continuous culture at steady state, growth occurs in a constant environment, and a steady state is established. Therefore, such an environment provides a realistic setting for studies in which only one variable is altered. It can be assumed that in such studies the change in the nature of the influent gas was the only variable, with no changes in nutrient or product levels influencing the observed effects. The much reduced level of CO2 inhibition in chemostat cultures compared with the level of CO2 inhibition noted in batch cultures continuously gassed with elevated CO2 levels may be due to the impact of CO2 inhibition in the early batch phase (lag and early exponential phases), whereas, as Fig. 2 shows, exponentially growing cultures may be less sensitive to CO2 inhibition. Conclusions. This study confirms the value of the dCO2 probe as a direct means of measuring the concentration of dCO2 in submerged fungal cultures and indicates that probe response is not interfered with by the presence of significant concentrations of organic acids. The results of this study call into question the relevance of the most common previous experimental methods used to simulate elevated dCO2 levels. This is especially so because of the sensitivity of the lag-phase culture to dCO2 inhibition. Combined with the much lower levels of inhibition noted in chemostat cultures, our findings indicate that inadequate or inappropriate experimental methods have led to a systematic overestimation of the inhibitory effects of elevated dCO2 levels on microbial cultures in previous studies. ACKNOWLEDGMENT This work was kindly supported by the Chemical and Pharmaceutical Directorate of the Biotechnology and Biological Sciences Research Council, United Kingdom. REFERENCES 1. Adams, H., and C. R. Thomas. 1988. The use of image analysis for morphological measurements on filamentous micro-organisms. Biotechnol. Bioeng. 32:707–712. 2. Aiba, S., and H. Furuse. 1990. Some comments on respiratory quotient (RQ) determination from the analysis of exit gas from a fermenter. Biotechnol. Bioeng. 36:534–538. 3. Beuksom, C. N. J., H. E. De Kok, and J. A. Roels. 1981. A method for the determination of dissolved carbon dioxide. Biotechnol. Bioeng. 23:2379– 2401. 4. Braun, S., and S. E. Vecht-Lifshitz. 1991. Mycelial morphology and metabolite production. Trends Biotechnol. 9:63–68. 5. Chen, S. L., and F. Gutmanis. 1976. Carbon dioxide inhibition of yeast growth and biomass production. Biotechnol. Bioeng. 28:1455–1462. 6. Covington, A. K. 1985. Potentiometric titrations of aqueous carbonate solutions. Chem. Soc. Rev. 14:265–281. 7. Dahod, S. K. 1993. Dissolved carbon dioxide measurement and its correlation with operating parameters in fermentation processes. Biotechnol. Prog. 9:655–660. 8. Dixon, N. M., and D. B. Kell. 1989. The inhibition of CO2 of the growth and metabolism of micro-organisms. J. Appl. Bacteriol. 67:109–156. 9. Dubois, M., K. A. Gilles, J. K. Hamilton, P. A. Revers, and F. Smith. 1956. A colourimetric method for the determination of sugars and related substances. Anal. Chem. 28:350–356. 10. El-Goorani, M. A., and N. F. Sommer. 1979. Suppression of post harvest plant pathogenic fungi by carbon monoxide. Phytopathology 69:834–838. 11. Gill, C. O., and K. H. Tan. 1979. Effect of carbon dioxide on growth of

dCO2 AND A. NIGER

VOL. 63, 1997 Pseudomonas fluorescens. Appl. Environ. Microbiol. 38:237–240. 12. Glasbey, C. A., and G. W. Horgan. 1994. Image analysis for the biological sciences. Statistics in practice series. John Wiley and Sons, Inc., New York, N.Y. 13. Ho, C. S., and M. D. Smith. 1986. Effect of dissolved carbon dioxide on penicillin fermentations: mycelial growth and penicillin production. Biotechnol. Bioeng. 28:668–677. 14. Ho, C. S., and M. D. Smith. 1986. Morphological alterations of Penicillium chrysogenum caused by carbon dioxide. J. Gen. Microbiol. 132:3479–3484. 15. Ho, C. S., M. D. Smith, and J. F. Shanahan. 1987. Carbon dioxide transfer in biochemical reactors. Adv. Biochem. Eng. 35:83–125. 16. Jones, R. P., and P. F. Greenfield. 1982. Effect of carbon dioxide on yeast growth and fermentation. Enzyme Microb. Technol. 4:210–223. 17. Kuriyama, H., W. Mahakarnchanakul, and S. Matsui. 1993. The effects of pCO2 on yeast growth and metabolism under continuous fermentation. Biotechnol. Lett. 15:189–194. 18. Lowenadler, J., and U. Ronner. 1994. Determination of dissolved carbon dioxide by colourimetric titration in modified atmosphere systems. Lett. Appl. Microbiol. 18:285–288. 19. Marier, J. R., and M. Boulet. 1956. Direct determination of citric acid in milk by an improved pyridine, acetic anhydride method. J. Dairy Sci. 41:1683– 1692. 20. McIntyre, M., and B. McNeil. 1997. Dissolved carbon dioxide effects on morphology, growth and citrate production in Aspergillus niger A60. Enzyme Microb. Technol. 20:135–142. 21. McIntyre, M., and B. McNeil. Effect of carbon dioxide on morphology and product synthesis in chemostat cultures of Aspergillus niger A60. Enzyme Microb. Technol., in press. 22. Molin, G., L.-M. Stenstrom, and A. Termstrom. 1983. The microbiological

23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

4177

flora of herring fillets after storage in carbon dioxide, nitrogen or air at 2°C. J. Appl. Bacteriol. 55:49–56. Mollah, A. H., and D. C. Stuckey. 1992. The influence of H2, CO2 and dilution rate on the continuous fermentation of acetone-butanol. Appl. Microbiol. Biotechnol. 37:533–538. Onken, U., and E. Liefke. 1989. Effects of total and partial pressure (oxygen and carbon dioxide) on aerobic microbial processes. Adv. Biochem. Eng. Biotechnol. 40:137–169. Pasteur, L., and P. Joubert. 1877. Etudes sur la maladie charbonneuse. C. R. Acad. Sci. 84:900–906. Pirt, S. J., and B. Mancini. 1975. Inhibition of penicillin production by carbon dioxide. J. Appl. Chem. Biotechnol. 25:781–783. Puhar, E., A. Einsele, H. Buhler, and W. Ingold. 1980. Steam sterilisable pCO2 electrode. Biotechnol. Bioeng. 22:2411–2416. Royce, P. N. 1992. Effect of changes in the pH and carbon dioxide evolution rate on the measured respiratory quotient of fermentations. Biotechnol. Bioeng. 40:1129–1138. Smith, M. D., and C. S. Ho. 1985. The effect of dissolved carbon dioxide on penicillin production: mycelial morphology. J. Biotechnol. 2:347–363. Smith, M. D., and C. S. Ho. 1985. On dissolved carbon dioxide in penicillin fermentations. Chem. Eng. Commun. 37:21–27. Thomas, C. R. 1992. Image analysis: putting filamentous organisms in the picture. Trends Biotechnol. 10:343–348. Wolfe, S. K. 1980. Use of CO- and CO2-enriched atmospheres for meats, fish and produce. Food Technol. 34:55–58. Zeng, A. P. 1995. Effect of CO2 absorption on the measurement of CO2 evolution rate in aerobic and anaerobic continuous cultures. Appl. Microbiol. Biotechnol. 42:688–691.