Fiber Depolymerization: Fracture, Fragments, Vanishing Times, and Stochastics in Sickle Hemoglobin

The well-characterized rates, mechanisms, and stochastics of nucleation-dependent polymerization of deoxyhemoglobin S (HbS) are important in governing whether or not vaso-occlusive sickle cell crises will occur. The less well studied kinetics of depolymerization may also be important, for example in achieving full dissolution of polymers in the lungs, in resolution of crises and/or in minimizing gelation-induced cellular damage. We examine depolymerization by microscopic observations on depolymerizing HbS fibers, by Monte Carlo simulations and by analytical characterization of the mechanisms. We show that fibers fracture. Experimental scatter of rates is consistent with stochastic features of the analytical model and Monte Carlo results. We derive a model for the distribution of vanishing times and also show the distribution of fracture-dependent fiber fragment lengths and its time dependence. We describe differences between depolymerization of single fibers and bundles and propose models for bundle dissolution. Our basic model can be extended to dissolution of gels containing many fibers and is also applicable to other reversible linear polymers that dissolve by random fracture and end-depolymerization. Under the model, conditions in which residual HbS polymers exist and facilitate repolymerization and thus pathology can be defined; whereas for normal polymers requiring cyclic polymerization and depolymerization for function, conditions for rapid cycling due to residual aggregates can be identified.     

Fiber depolymerization

Depolymerization is, by definition, a crucial process in the reversible assembly of various biopolymers. It may also be an important factor in the pathology of sickle cell disease. If sickle hemoglobin fibers fail to depolymerize fully during passage through the lungs then they will reintroduce aggregates into the systemic circulation and eliminate or shorten the protective delay (nucleation) time for the subsequent growth of fibers. We study how depolymerization depends on the rates of end- and side-depolymerization, k(end) and k(side), which are, respectively, the rates at which fiber length is lost at each end and the rate at which new breaks appear per unit fiber length. We present both an analytic mean field theory and supporting simulations showing that the characteristic fiber depolymerization time tau= square root 1/k(end)k(side) depends on both rates, but not on the fiber length L, in a large intermediate regime 1 < k(side)L(2)/k(end) < (L/d)(2), with d the fiber diameter. We present new experimental data which confirms that both mechanisms are important and shows how the rate of side depolymerization depends strongly on the concentration of CO, acting as a proxy for oxygen. Our theory remains rather general and could be applied to the depolymerization of an entire class of linear aggregates, not just sickle hemoglobin fibers.     

Poleward kinetochore fiber movement occurs during both metaphase and anaphase-A in newt lung cell mitosis

Microtubules in the mitotic spindles of newt lung cells were marked using local photoactivation of fluorescence. The movement of marked segments on kinetochore fibers was tracked by digital fluorescence microscopy in metaphase and anaphase and compared to the rate of chromosome movement. In metaphase, kinetochore oscillations toward and away from the poles were coupled to kinetochore fiber shortening and growth. Marked zones on the kinetochore microtubules, meanwhile, moved slowly polewards at a rate of approximately 0.5 micron/min, which identifies a slow polewards movement, or "flux," of kinetochore microtubules accompanied by depolymerization at the pole, as previously found in PtK2 cells (Mitchison, 1989b). Marks were never seen moving away from the pole, indicating that growth of the kinetochore microtubules occurs only at their kinetochore ends. In anaphase, marked zones on kinetochore microtubules also moved polewards, though at a rate slower than overall kinetochore-to-pole movement. Early in anaphase-A, microtubule depolymerization at kinetochores accounted on average for 75% of the rate of chromosome-to-pole movement, and depolymerization at the pole accounted for 25%. When chromosome-to-pole movement slowed in late anaphase, the contribution of depolymerization at the kinetochores lessened, and flux became the dominant component in some cells. Over the whole course of anaphase-A, depolymerization at kinetochores accounted on average for 63% of kinetochore fiber shortening, and flux for 37%. In some anaphase cells up to 45% of shortening resulted from the action of flux. We conclude that kinetochore microtubules change length predominantly through polymerization and depolymerization at the kinetochores during both metaphase and anaphase as the kinetochores move away from and towards the poles. Depolymerization, though not polymerization, also occurs at the pole during metaphase and anaphase, so that flux contributes to polewards chromosome movements throughout mitosis. Poleward force production for chromosome movements is thus likely to be generated by at least two distinct molecular mechanisms.     

The kinetics of nucleation and growth of sickle cell hemoglobin fibers

Polymerization of sickle cell hemoglobin (HbS) in deoxy state is one of the basic events in the pathophysiology of sickle cell anemia. For insight into the polymerization process, we monitor the kinetics of nucleation and growth of the HbS polymer fibers. We define a technique for the determination of the rates J and delay times theta of nucleation and the fiber growth rates R of deoxy-HbS fibers, based on photolysis of CO-HbS by laser illumination. We solve numerically time-dependent equations of heat conductance and CO transport, coupled with respective photo-chemical processes, during kinetics experiments under continuous illumination. After calibration with experimentally determined values, we define a regime of illumination ensuring uniform temperature and deoxy-HbS concentration, and fast (within <1 s) egress to steady conditions. With these procedures, data on the nucleation and growth kinetics have relative errors of <5% and are reproducible within 10% in independent experiments. The nucleation rates and delay times have steep, exponential dependencies on temperature. In contrast, the average fiber growth rates only weakly depend on temperature. The individual growth rates vary by up to 40% under identical conditions. These variations are attributed to instability of the coupled kinetics and diffusion towards the growing end of a fiber. The activation energy for incorporation of HbS molecules into a polymer is E(A)=50 kJ mol(-1), a low value indicating the significance of the hydrophobic contacts in the HbS polymer. More importantly, the contrast between the strong theta(T) and weak R(T) dependencies suggests that the homogenous nucleation of HbS polymers occurs within clusters of a precursor phase. This conclusion may have significant consequences for the understanding of the pathophysiology of sickle cell anemia and should be tested in further work.     

Effects of retinoid ligands on RIP140: molecular interaction with retinoid receptors and biological activity

Receptor interacting protein 140 (RIP140) interacts with retinoic acid receptor (RAR) and retinoid X receptor (RXR) constitutively, but hormone binding enhances this interaction. The ligand-independent interaction is mediated by the amino and central regions of RIP140 which contain a total of nine copies of the LXXLL motif, whereas the agonist-induced interaction is mediated by its carboxyl terminus which contains a novel motif (1063-1076, LTKTNPILYYMLQK). The ligand-independent interaction could be enhanced slightly by agonists, whereas the ligand-dependent interaction was strictly agonist dependent for both RAR and RXR. In the context of heterodimers, ligand occupancy of RXR played a more dominant role for both molecular interaction and biological activity of RIP140. Competition and mutation studies demonstrated an essential role for (1067)Asn and (1073)Met for a ligand-dependent interaction. A model was proposed to address the constitutive and agonist-dependent interaction of RIP140 with RAR/RXR.     

Pax7 reveals a greater frequency and concentration of satellite cells at the ends of growing skeletal muscle fibers.

The main sites of longitudinal growth in skeletal muscle are the ends of the fibers. This study tests the hypothesis that satellite cells (SCs) are at a greater frequency (#SC nuclei/all nuclei within basal laminae) and concentration (closer together) within growing fiber ends of posthatch chicken pectoralis. SCs were localized by their Pax7 expression, and fiber ends were identified by their retention of neonatal myosin heavy chains and small cross-sectional profiles. Whereas SC frequency decreased from about 20% at 9 days posthatch to <5% at 115 days, fiber ends retained a frequency of approximately 16%. Calculated mean area of sarcolemma per SC revealed higher concentrations of SCs at fiber ends. There was also a strong inverse correlation between SC frequency and fiber profile cross-sectional size throughout development. This study suggests that SCs at fiber ends play a key role in the longitudinal growth of muscle fibers, and that fiber profile size may impact SC distribution.     

Functional Analysis of the Accessory Protein TapA in Bacillus subtilis Amyloid Fiber Assembly

Bacillus subtilis biofilm formation relies on the assembly of a fibrous scaffold formed by the protein TasA. TasA polymerizes into highly stable fibers with biochemical and morphological features of functional amyloids. Previously, we showed that assembly of TasA fibers requires the auxiliary protein TapA. In this study, we investigated the roles of TapA sequences from the C-terminal and N-terminal ends and TapA cysteine residues in its ability to promote the assembly of TasA amyloid-like fibers. We found that the cysteine residues are not essential for the formation of TasA fibers, as their replacement by alanine residues resulted in only minor defects in biofilm formation. Mutating sequences in the C-terminal half had no effect on biofilm formation. However, we identified a sequence of 8 amino acids in the N terminus that is key for TasA fiber formation. Strains expressing TapA lacking these 8 residues were completely defective in biofilm formation. In addition, this TapA mutant protein exhibited a dominant negative effect on TasA fiber formation. Even in the presence of wild-type TapA, the mutant protein inhibited fiber assembly in vitro and delayed biofilm formation in vivo. We propose that this 8-residue sequence is crucial for the formation of amyloid-like fibers on the cell surface, perhaps by mediating the interaction between TapA or TapA and TasA molecules.     

Reorganization of actin filament bundles in living fibroblasts

We investigated how actin bundles assemble, disassemble, and reorganize during cell movement. Living chick embryonic fibroblasts were microinjected with actin molecules that had been fluorescently labeled with tetramethylrhodamine. We found that the fluorescent analogue of actin can be used successfully by both existing and newly formed cellular structures. Using time-lapse photography coupled to image- intensified fluorescence microscopy, we were able to detect various patterns of reorganization in motile cells. Assembly of stress fibers occurred near both the leading and the trailing ends of the cell. The initial structure appeared as discrete spots that subsequently extended into stress fibers. The extension occurred unidirectionally. The site of initiation near the leading edge remained stationary relative to the substrate during subsequent cell advancement. However, the orientation of the fiber could change according to the direction of cell movement. In addition, existing stress fibers could merge or fragment. The shortening of stress fibers can occur from either end of the fiber. Shortening from the proximal end (centrifugal shortening) was accompanied by a decrease in fluorescence intensity, as if the bundle were disassembling, and usually led to the total disappearance of the bundle. Shortening from the distal end (centripetal shortening), on the other hand, is usually accompanied by an increase in fluorescence intensity at the distal end of the bundle, as if this end had pulled loose from its attachment and retracted toward the center of the cell. Besides stress fibers, arc-like actin bundles have also been detected in spreading cells. These observations can explain how the organization of actin bundles coordinates with cell movement, and how stress fibers reach a highly regular pattern in static cells.     

Ganglioside GM3 blocks the activation of epidermal growth factor receptor induced by integrin at specific tyrosine sites

The epidermal growth factor receptor (EGFR) can be activated by both direct ligand binding and cross-talk with other molecules, such as integrins. This integrin-mediated cross-talk with growth factor receptors participates in regulating cell proliferation, survival, migration, and invasion. Previous studies have shown that ligand-dependent EGFR activation is inhibited by GM3, the predominant ganglioside of epithelial cells, but the effect of GM3 on ligand-independent, integrin-EGFR cross-talk is unknown. Using a squamous carcinoma cell line we show that endogenous accumulation of GM3 disrupts the ligand-independent association of the integrin beta1 subunit with EGFR and results in inhibition of cell proliferation. Consistently, endogenous depletion of GM3 markedly increases the association of EGFR with tyrosine-phosphorylated integrin beta1 and promotes cell proliferation. The ligand-independent stimulation of EGFR does not require focal adhesion kinase phosphorylation or cytoskeletal rearrangement. Stimulation of EGFR and mitogen-activated protein kinase signaling by GM3 depletion involves the phosphorylation of EGFR at tyrosine residues 845, 1068, and 1148 but not 1086 or 1173. The specific blockade of phosphorylation at Tyr-845 with Src family kinase inhibition and at Tyr-1148 with phosphatidylinositol 3-kinase inhibition suggests that GM3 inhibits integrin-induced, ligand-independent EGFR phosphorylation (cross-talk) through suppression of Src family kinase and phosphatidylinositol 3-kinase signaling.     

Quantitative evaluation of threshold fiber strain that induces reorganization of cytoskeletal actin fiber structure in osteoblastic cells

The cytoskeletal stress fiber structure plays essential roles in various kinds of cellular functions such as shape maintenance, active motility and mechanosensing, and its structure is dynamically reorganized under each functional process. In known reorganization mechanisms of the stress fibers, a change in its mechanical condition has been suggested as one of the key mediators that affect the reorganization process. Some experimental studies have clarified that tension release in the stress fibers induces fiber depolymerization that is considered to be the initial phase of the reorganization process. However, quantitative mechanical values such as strain or stress that induce depolymerization have still not been evaluated. This study is aimed at the quantitative evaluation of the mechanical value that induces stress fiber depolymerization, to gain a basic understanding of the reorganization phenomenon from a mechanical viewpoint. Osteoblastic cells (MC3T3-E1) were cultured on prestretched silicone rubber substrate. Compressive deformation was applied to the cells by uniaxially releasing the prestretched substrate strain and change in the stress fiber structure was observed. The results indicated that the compressive strain magnitude, not in the whole cell body but in the stress fiber itself, is important to induce disassembly of the stress fiber structure. The existence of a threshold strain magnitude for initiating fiber disassembly was also suggested; the threshold strain magnitude was evaluated as approximately -0.20.     

Repair of laser-severed stress fibers in myocardial non-muscle cells

Phase-dense stress fibers in cultured non-muscle cells from neonatal rat ventricles were severed using the 532 nm wavelength of a Q-switched Nd Yag laser microbeam. The breaks were confirmed using anti-actin antibodies and Coomassie blue staining. SEM showed that no visible membrane damage resulted from the laser. Following irradiation, severed stress fiber ends quickly retracted 3–5 μm apart and repaired, averaging 12.2 min, in 84% of the control cells. Most fibers not repairing had much longer, > 10 μm, retraction distances. Disruption of microfilaments by cytochalasin B (CB) or chlorpromazine (CPZ) resulted in increased retraction distances and a dose-dependent decrease in the ability of stress fibers to repair. Fibers not repairing in CB or CPZ consistently displayed directional depolymerizations of fiber segments on the proximal side of the cut relative to the cell center and, at the extreme, condensations of stress fiber material into ‘knob-like’ structures. It appears to us that increased retraction distances might reflect CB or CPZ disruption of stress fiber-membrane attachments. Directional depolymerization suggests that stress fibers are unipolar structures, yet we failed to see any directional repair. Microtubule removal by colcemid, vinblastine, or podophyllotoxin resulted in a doubling of stress fiber repair rates. This in vivo evidence suggests that a relationship does exist between stress fibers and microtubules. Finally, inhibition of protein synthesis by 95% had little effect upon fiber repair, therefore indicating that protein synthesis is not necessary for stress fiber repair.     

Gelsolin as a calcium-regulated actin filament-capping protein.

Various concentrations of gelsolin (25-100 nM) were added to 2 microM polymerized actin. The concentrations of free calcium were adjusted to 0.05-1.5 microM by EGTA/Ca2+ buffer. Following addition of gelsolin actin depolymerization was observed that was caused by dissociation of actin subunits from the pointed ends of treadmilling actin filaments and inhibition by gelsolin of polymerization at barbed ends. The time course of depolymerization revealed an initial lag phase that was followed by slow decrease of the concentration of polymeric actin to reach the final steady state polymer and monomer concentration. The initial lag phase was pronounced at low free calcium and low gelsolin concentrations. On the basis of quantitative analysis the kinetics of depolymerization could be interpreted as capping, i.e. binding of gelsolin to the barbed ends of actin filaments and subsequent inhibition of polymerization, rather than severing. The main argument for this conclusion was that even gelsolin concentrations (100 nM) that exceed the concentration of filament ends ( approximately 2 nM), cause the filaments to depolymerize at a rate that is similar to the rate of depolymerization of the concentration of pointed ends existing before addition of gelsolin. The rate of capping is directly proportional to the free calcium concentration. These experiments demonstrate that at micromolar and submicromolar free calcium concentrations gelsolin acts as a calcium-regulated capping protein but not as an actin filament severing protein, and that the calcium binding sites of gelsolin which regulate the various functions of gelsolin (capping, severing and monomer binding), differ in their calcium affinity.     

Sphingosylphosphorylcholine induces stress fiber formation via activation of Fyn-RhoA-ROCK signaling pathway in fibroblasts

Sphingosylphosphorylcholine (SPC), a bioactive sphingolipid, has recently been reported to modulate actin cytoskeleton rearrangement. We have previously demonstrated Fyn tyrosine kinase is involved in SPC-induced actin stress fiber formation in fibroblasts. However, Fyn-dependent signaling pathway remains to be elucidated. The present study demonstrates that RhoA-ROCK signaling downstream of Fyn controls stress fiber formation in SPC-treated fibroblasts. Here, we found that SPC-induced stress fiber formation was inhibited by C3 transferase, dominant negative RhoA or ROCK. SPC activated RhoA, which was blocked by pharmacological inhibition of Fyn activity or dominant negative Fyn. Constitutively active Fyn (ca-Fyn) stimulated stress fiber formation and localized with F-actin at the both ends of stress fibers, both of which were prevented by Fyn translocation inhibitor eicosapentaenoic acid (EPA). In contrast, inhibition of ROCK abolished only the formation of stress fibers, without affecting the localization of ca-Fyn. These results allow the identification of the molecular events downstream SPC in stress fiber formation for a better understanding of stress fiber formation involving Fyn.     

Melt electrospinning of poly-(ethylene glycol-block-epsilon-caprolactone)

Various block copolymers of poly(ethylene glycol) and poly(epsilon-caprolactone) (PEG-b-PCL) with molecular weights between 7000 and 26,900 g/mol were synthesized, and melt electrospun at temperatures between 60 degrees C and 90 degrees C. Two types of fibers were collected, including excellent quality fibers - highly coiled and continuous, with a constant diameter and relatively defect free. Such fibers, termed "solid fibers", were sufficiently cooled during their path between the spinneret and the collector that the symmetric fiber shape is maintained after landing on the collector. The second type of melt electrospun fiber were poor quality, large diameter fibers, flattened on the collector - termed "molten fibers". The solid and molten fibers were morphologically distinct from each other as determined from scanning electron microscopy (SEM). Using an SEM imaging method to assess the regional variations of collected electrospun material, we found the spinneret pump rate largely influenced the fiber quality. The polymer flow rate to the spinneret and the molecular weight of PEG-b-PCL had the greatest effect on the electrospun fibers collected, with an optimum rate of 0.05-0.1 mL/h for the highest molecular weight copolymers. The lowest molecular weight PEG-b-PCL tended to electrospray, while the material collected from higher molecular weight copolymers were conducive to fiber formation. The highest quality fibers were PEG-b-PCL block copolymers (22,000 and 26,900 g/mol) melt electrospun at temperatures of 85 degrees C and 90 degrees C, corresponding to shear viscosities of the polymer of between 28.1 and 39.4 Pa.S.     

Ligand-independent activation of the androgen receptor by interleukin-6 and the role of steroid receptor coactivator-1 in prostate cancer cells

The androgen receptor (AR) can be activated in the absence of androgens by interleukin-6 (IL-6) in human prostate cancer cells. The events involved in ligand-independent activation of the AR are unknown, but have been suggested to involve phosphorylation of the AR itself or a receptor-associated protein. Steroid receptor coactivator-1 (SRC-1) has been shown to interact with the human AR and to modulate ligand-dependent AR transactivation and is regulated by phosphorylation by MAPK. To date, no one has examined the role of SRC-1 in ligand-independent activation of the AR by IL-6 or other signaling pathways known to activate the full-length receptor. This study addressed this and has revealed the following. 1) SRC-1 similarly enhanced ligand-independent activation of the AR by IL-6 to the same magnitude as that obtained via ligand-dependent activation. 2) Androgen and IL-6 stimulated the MAPK pathway. 3) MAPK was required for both ligand-dependent and ligand-independent activation of the AR. 4) Phosphorylation of SRC-1 by MAPK was required for optimal ligand-independent activation of the AR by IL-6. 5) Protein-protein interaction between endogenous AR and SRC-1 was dependent upon treatment of LNCaP cells with IL-6 or R1881. 6) Protein-protein interaction between the AR N-terminal domain and SRC-1 was independent of MAPK. 7) Ligand-independent activation of the AR did not occur by a mechanism of overexpression of either solely wild-type SRC-1 or mutant SRC-1 that mimics its phosphorylated form.     

Oxygen and nitric oxide rebinding kinetics in nonsymbiotic hemoglobin AHb1 from Arabidopsis thaliana

Type 1 nonsymbiotic hemoglobin from Arabidopsis thaliana (AHb1) shows a partial bis-histidyl hexacoordination but can reversibly bind diatomic ligands. The physiological function is still unclear, but the high oxygen affinity rules out a function related to O2 sensing, carrying, or storing. To gain insight into its possible functional roles, we have investigated its O2 and NO rebinding kinetics after laser flash photolysis. The rate constants of the rebinding from the primary docking site for both O2 and NO are higher than CO, with lower photolysis yields. Moreover, the amplitude of the geminate phase increases and, as for CO, the numerical analysis of the experimental curves suggests the existence of an internal pathway leading, with high rate, to an additional docking site. However, the accessibility to this site seems to be strongly ligand-dependent, being lower for O2 and higher for NO.     

Bicarbonate concentration and osmolality are key determinants in the inhibition of CHO cell polysialylation under elevated pCO(2) or pH.

Accumulation of CO(2) in animal cell cultures can be a significant problem during scale-up and production of recombinant glycoprotein biopharmaceuticals. By examining the cell-surface polysialic acid (PSA) content, we show that elevated CO(2) partial pressure (pCO(2)) can alter protein glycosylation. PSA is a high-molecular-weight polymer attached to several complex N-linked oligosaccharides on the neural cell adhesion molecule (NCAM), so that small changes in either core glycosylation or in polysialylation are amplified and easily measured. Flow-cytometric analysis revealed that PSA levels on Chinese hamster ovary (CHO) cells decrease with increasing pCO(2) in a dose-dependent manner, independent of any change in NCAM content. The results are highly pH-dependent, with a greater decrease in PSA at higher pH. By manipulating medium pH and pCO(2), we showed that decreases in PSA correlate well with bicarbonate concentration ([HCO(3)(-)]). In fact, it was possible to offset a 60% decrease in PSA content at 120 mm Hg pCO(2) by decreasing the pH from 7.3 to 6.9, such that [HCO(3)(-)] was lowered to that of control (38 mm Hg pCO(2)). When the increase in osmolality associated with elevated [HCO(3)(-)] was offset by decreasing the basal medium [NaCl], elevated [HCO(3)(-)] still caused a decrease in PSA, although less extensive than without osmolality control. By increasing [NaCl], we show that hyperosmolality alone decreases PSA content, but to a lesser extent than for the same osmolality increase due to elevated [NaHCO(3)]. In conclusion, we demonstrate the importance of pH and pCO(2) interactions, and show that [HCO(3)(-)] and osmolality can account for the observed changes in PSA content over a wide range of pH and pCO(2) values.     

Cyclic stretch-induced stress fiber dynamics - Dependence on strain rate, Rho-kinase and MLCK

Stress fiber realignment is an important adaptive response to cyclic stretch for nonmuscle cells, but the mechanism by which such reorganization occurs is not known. By analyzing stress fiber dynamics using live cell microscopy, we revealed that stress fiber reorientation perpendicular to the direction of cyclic uniaxial stretching at 1 Hz did not involve disassembly of the stress fiber distal ends located at focal adhesion sites. Instead, these distal ends were often used to assemble new stress fibers oriented progressively further away from the direction of stretch. Stress fiber disassembly and reorientation were not induced when the frequency of stretch was decreased to 0.01 Hz, however. Treatment with the Rho-kinase inhibitor Y27632 reduced stress fibers to thin fibers located in the cell periphery which bundled together to form thick fibers oriented parallel to the direction of stretching at 1 Hz. In contrast, these thin fibers remained diffuse in cells subjected to stretch at 0.01 Hz. Cyclic stretch at 1 Hz also induced actin fiber formation parallel to the direction of stretch in cells treated with the myosin light chain kinase (MLCK) inhibitor ML-7, but these fibers were located centrally rather than peripherally. These results shed new light on the mechanism by which stress fibers reorient in response to cyclic stretch in different regions of the actin cytoskeleton.     

Decreased pCO(2) accumulation by eliminating bicarbonate addition to high cell-density cultures

High-density perfusion cultivation of mammalian cells can result in elevated bioreactor CO(2) partial pressure (pCO(2)), a condition that can negatively influence growth, metabolism, productivity, and protein glycosylation. For BHK cells in a perfusion culture at 20 x 10(6) cells/mL, the bioreactor pCO(2) exceeded 225 mm Hg with approximate contributions of 25% from cellular respiration, 35% from medium NaHCO(3), and 40% from NaHCO(3) added for pH control. Recognizing the limitations to the practicality of gas sparging for CO(2) removal in perfusion systems, a strategy based on CO(2) reduction at the source was investigated. The NaHCO(3) in the medium was replaced with a MOPS-Histidine buffer, while Na(2)CO(3) replaced NaHCO(3) for pH control. These changes resulted in 63-70% pCO(2) reductions in multiple 15 L perfusion bioreactors, and were reproducible at the manufacturing-scale. Bioreactor pCO(2) values after these modifications were in the 68-85 mm Hg range, pCO(2) reductions consistent with those theoretically expected. Low bioreactor pCO(2) was accompanied by both 68-123% increased growth rates and 58-92% increased specific productivity. Bioreactor pCO(2) reduction and the resulting positive implications for cell growth and productivity were brought about by process changes that were readily implemented and robust. This philosophy of pCO(2) reduction at the source through medium and base modification should be readily applicable to large-scale fed-batch cultivation of mammalian cells.