The optimization of poly(vinyl)-alcohol-alginate beads with a slow-release compound for the aerobic cometabolism of chlorinated aliphatic hydrocarbons

Chlorinated aliphatic hydrocarbons (CAHs), such as cis-1,2-dichloroethylene (cDCE), are prevalent in groundwater at many locations throughout the United States. When immobilized in hydrogel beads with slow-release compounds, the bacteria strain Rhodococcus rhodochrous ATCC 21198 can be used for the in situ bioremediation of cDCE. These hydrogel beads must exhibit high mechanical strength and resist degradation to extend the lifetime of slow-release compounds and bioremediation. We engineered poly(vinyl)-alcohol – alginate (PVA-AG) beads to immobilize ATCC 21198 with the slow-release compound, tetrabutoxysilane (TBOS) that produces 1-butanol as a growth substrate, for high mechanical strength. We optimized three inputs (concentration of PVA, concentration of AG, and the crosslinking time) on two responses (compressive modulus and rate of oxygen utilization) for batch incubation experiments between 1 and 30 days using a design of experiments approach. The predictive models generated from design of experiments were then tested by measuring the compressive strength, oxygen utilization, and abiotic rates of hydrolysis for a predicted optimal bead formulation. The result of this study generated a hydrogel bead with immobilized R. rhodochrous ATCC 21198 and TBOS that exhibited a high compressive modulus on day 1 and day 30, which was accurately predicted by models. These hydrogel beads exhibited low metabolic activity based on oxygen rates on day 1 and day 30 but were not accurately predicted by the models. In addition, the ratio between oxygen utilization and abiotic rates of hydrolysis were observed to be roughly half of what was expected stoichiometrically. Lastly, we demonstrated the capability to use these beads as a bioremediation technology for cDCE as we found that, for all bead formulations, cDCE was significantly reduced after 30 days. Altogether, this work demonstrates the capability to capture and enhance the material properties of the complex hydrogel beads with predictive models yet signals the need for more robust methods to understand the metabolic activity that occurs in the hydrogel beads.


SI3. ATCC 21198 liquid culturing
Liquid culture growth reactors were used to grow ATCC 21198 as we previously described in Rolsten et al. (1) The growth reactors consisted of 720 mL Wheaton bottles (DWK Life Sciences Wheaton, Stokeon-Trent, United Kingdom) sealed with screw-on caps fitted with gray butyl rubber septa (DWK Life Sciences Wheaton) filled with 300 mL of phosphate-buffered 1X MSM at a pH of 7.0 and 420 mL of air headspace.ATCC 21198 were inoculated by using an inoculating loop (VWR International) to scrape minimal medium plates (see above) and place an inoculum of ATCC 21198 into our growth reactor MSM.
The bottles were then sealed, injected with 50mL of isobutane

𝑚 𝑖
To remove the remaining liquid, we subjected the filter paper to a VWR Oven F Air 6.3CF oven (VWR International) at a temperature of 105 ˚C for 20 min to provide the necessary energy to dry the sample.A measurement of the total suspended solids (TSS) provided the concentration of suspended cells as we previously described in Murnane et al (2).The weight of the dried solids ( ) was calculated  as: (Eqn.S1) Where is the weight of the dried cells, is the weight of the initial mass of suspended cells, and is the volume of suspended cells taken from growth reactors.

SI4. Gas chromatography details
Gas chromatography was used to quantify concentrations of cDCE and oxygen in the batch bottles.
Specifically, a 6890 Series gas chromatograph (Hewlett Packard, Corvallis, Oregon, United States) equipped with a micro-electron capture device (ECD) was used to quantify the cDCE.We separated cDCE from other compounds with an Agilent DB-624 UI capillary column (30 m x 0.53 mm) (Agilent Technologies, Santa Clara, California, United States) with ultra-high purity-pure helium (Airgas, Radnor, Pennsylvania, United States) as the carrier gas (15 mL/min) at 50 ˚C to achieve retention time (RT) of 2.4 min.A 5890 Series GC II (Hewlett Packard, Corvallis, Oregon, United States) equipped with a microelectron capture device (ECD) was used to measure oxygen levels.We separated oxygen from other compounds with a Supelco 60/80 Carboxen-10000 packed stainless-steel column (15 ft x 1/8 in) (Supelco, Inc., Bellefonte, Pennsylvania, United States) with ultra-high purity-pure helium (Airgas, Radnor, Pennsylvania, United States) as the carrier gas (30 mL min -1 ), at 40 ˚C, to achieve a retention time (RT) of 4 min.The use of external standards supplied the calibration for all GC methods.

Figure S1
A Typical O 2 uptake data for hydrogel beads obtained from gas chromatography.B Typical cDCE uptake data for hydrogel beads obtained from gas chromatography.The data was fitted with zeroorder rates as data was taken over narrow ranges of time for uptake and transformation and remained linear over the entirety of the time period.
Metabolic activity tests were performed as described in the main text.The beads used in these experiments were formulated with the same formula as the center point of the central composite orthogonal design (Experiment No 15, Table 2), following the methods described in the main text (Methods: Immobilizing ATCC 21198 and TBOS with PVA -AG beads) and excluding the addition of PVA.Thus, the formula is comprised of 0% (w/v) PVA, 1.5% (w/v) AG, 10% (v/v) TBOS, 0.1% (v/v) Span 80, and 0.5 mg/mL ATCC 21198.Negative control consists of poisoned cells.Gas chromatography (GC) data was used to evaluate the cometabolic and metabolic transformation of cDCE and O 2 , respectively, with ATCC 21198 immobilized in all bead formulations.See Section 2.4 in the main text for sample preparation and injections.The total mass measured from the GC was taken over time, and zero-order rate laws were applied to the data (Figure S1A-B).Note that zero-order rate laws were used in this study as the change in mass over time for both O 2 and cDCE was linear for the range of time used to determine the rates.Uniaxial compression measurements were performed on hydrogel beads to obtain the compressive modulus, E (Figure S3).The Hertz equation (Eq. 3) can be rearranged and plotted with against .The compressive modulus, E, is taken as the slope of the linear portion 3/4   -0.5 Δ = ( - ' ) 1.5   of the curve between a small range of deformation.After the small deformation ranges, the data transitions to an exponential growth.The linear portion was selected by taking the first point as the start and where the absolute value of the second derivative of the data exceeded 0.01.We approximated the second derivative using the diff() function in MATLAB.

SI8. Design of experiments statistical method
The software MODDE-Pro 12. values for interactions between the independent variables; and is the random error of the response.

𝜖
Factor effect plots were used to compare the individual contributions of each factor on each of the responses measured.Each input factor was varied over its specific range while all the other input factors were held constant at their averages.In contrast, response surface maps were used to evaluate the individual and combinatorial effects of multiple input variables on each of the responses measured.Both factor effect plots and response surface maps were generated from the predictive equations for each response as a function of the three input factors.

SI9. cis-1,2-dichloroethylene (cDCE) rate data design of experiments models
We fit cDCE rate data on 1 and 30 days ( and described in the main text under Section 3.1 with second order multivariate regression models and eliminated terms that were not statistically significant.Experimental and predicted data are shown for all bead formulae tested with the central composite design described in the main text (Table S2).This data was not used to optimize the bead parameters and therefore was excluded from the main text.The models were evaluated based on the criteria provided in the main text under Section 3.3.That is, the statistical significance of each model was then evaluated with ANOVA tests, with a p-value of < 0.05 indicating a significant model and a p-value > 0.10 for lack of fit indicating a model with a negligible pure error.We provide the ANOVA assessment for the and regression models to demonstrate that the models were statistically significant and  ,1  ,30 had negligible pure error (Table S3).The unscaled models obtained for and were: negative quadratic influence on , whereas had a positively quadratic influence (Figure S4A).

𝑘 𝑐𝐷𝐶𝐸,1 𝑡 𝑥𝑙𝑖𝑛𝑘
The maximum point of was predicted to occur below the midpoint for ( % w/v) and near the midpoint for ( % w/v), when was held constant at min (Figure S4B).
For the factor effect plot for , was shown to have no effect, whereas both and had a  ,30       negative linear effect on the response (Figure S4C).Thus, the maximum of point of was predicted to occur at the low value of where was held constant at min (Figure S4D).

𝐶 𝐴𝐺 𝑡 𝑥𝑙𝑖𝑛𝑘 𝑡 𝑥𝑙𝑖𝑛𝑘 = 75
While these data could be used to maximize the bioremediation capability, the objective of this research was to maintain strong beads and to ensure that transformation of cDCE was not inhibited by the immobilization process.Thus, these models were excluded from the main text.All data are expressed as average  SD, n = 3.Rates are similar between the first 5 batches (1-3,4-6,7-9,10-12,13-15), but the optimal batch was observed at a much higher rate.
Metabolic activity tests were performed as described in the main text.The beads used in these experiments were formulated with the same formula as the center point of the central composite orthogonal design (Experiment No 15, Table 2), following the methods described in the main text (Methods: Immobilizing ATCC 21198 and TBOS with PVA -AG beads) and excluding the addition of PVA.Thus, the formula is comprised of 0% (w/v) PVA, 1.5% (w/v) AG, 10% (v/v) TBOS, 0.1% (v/v) Span 80, and 0.5 mg/mL ATCC 21198.2g of beads were added to 150 mL batch reactors and placed on a shaker table and oxygen was measured on a GC, as described in the main text.The control was made with each batch of beads, indicated by the batch id (1-3, 4-6, 7-9, 10-12, 13-15, optimal).Six controls were completed throughout the testing with three replicates, each.Between each batch, the oxygen rates for the control bottles were variable (Figure S5).Batches 4-6 and 13-15, and batches 7-9 and 10-12 were observed to have similar rates.The optimal batch was measured to have the lowest rate of oxygen utilization rate between all batches, which was likely due to the change in a 90% purity TBOS to a 98% purity TBOS described in the main text (Section 2.2).This data suggests that there could be differences between live/dead cells immobilized in the beads.2) at day 1 and day 30.For beads under abiotic conditions, data are expressed as average  SD, n = 3.For experiment 15, data is expressed as average  SD, n = 9. * p <0.05 between bead formulations, # p <0.05 between day 1 and day 30 obtained using a two-way ANOVA test with Tukey's honestly significant difference post-test.B Elastic loss ΔE for beads under abiotic conditions and center point formulation (Experiment No 15, Table 2) at day 1 and day 30.For beads under abiotic conditions, data are expressed as average  SD, n = 3. experiment 15, data is expressed as average  SD, n = 9. * p <0.05 between bead formulas obtained using a one-way ANOVA test with Tukey's honestly significant difference post-test.

SI11. Compression tests under abiotic conditions
Beads with autoclaved cells were used to evaluate the abiotic degradation of beads.Compression tests were performed as described in the main text.Beads under abiotic conditions were formulated with the same formula as the center point of the central composite orthogonal design (Experiment No 15, Table

Figure S2 A
Figure S2 A Hydrogel Bead before compression with top of plate in contact.B Hydrogel bead irreversibly deformed after compression test Hydrogel beads were irreversibly deformed during compression tests.Before compression, beads were

Figure S4
Figure S4 The factor effect plots (left column) include all input variables, , , and denoted       with red (solid), green (dashed), and yellow (dashed-dotted) lines, respectively.3-D response surface plots (right column) consist of the output response on the z-axis predicted for a range (factor value / [coded units]) of and on the x and y-axes, respectively and constant ∈ [ -1, 1]       = 75 [min].The color bar represents the magnitude of the response from low (purple) to high (yellow).A Factor effect plot of .B 3-D response surface map of .C Factor effect plot of .D  ,1  ,1  ,30 3-D response surface map of . ,30

Figure S6 A
Figure S6 A Compressive modulus E measured for beads under abiotic conditions and center point formulation (Experiment No 15, Table2) at day 1 and day 30.For beads under abiotic conditions, data are expressed as average  SD, n = 3.For experiment 15, data is expressed as average  SD, n = 9. * p <0.05 between bead formulations, # p <0.05 between day 1 and day 30 obtained using a two-way ANOVA test with Tukey's honestly significant difference post-test.B Elastic loss ΔE for beads under abiotic conditions and center point formulation (Experiment No 15, Table2) at day 1 and day 30.For beads under abiotic conditions, data are expressed as average  SD, n = 3. experiment 15, data is expressed as average  SD, n = 9. * p <0.05 between bead formulas obtained using a one-way ANOVA test with Tukey's honestly significant difference post-test.
Solution 2 consists of 15.5% (w/v) dipotassium hydrogen phosphate (Thermo Fisher Scientific Inc, Waltham, Massachusetts, United States) and 8.5% (w/v) monosodium phosphate (Sigma-Aldrich, St.Louis, Missouri, United States) dissolved into ultra-pure water (Table1).Solution 1 and solution 2 are autoclaved separately to ensure reactions between compounds occur.Finally, Solution 1 and Solution 2 are diluted with deionized water at a 10:1:100 ratio to form a 1X MSM solution.
(1)butane (Gas Innovations, La Porte, Texas, United States) at 30 ˚C.Minimal medium plates were prepared as we previously described in Rolston et al.(1)A solution of 1.7% (w/v) Difco 247940 agar (Thermo Fisher Scientific Inc) was dissolved in ultra-pure water then autoclaved.Once cooled (~< 30 ˚C), MSM was added to the agar solution and the solution was poured into polystyrene disposable sterile petri dishes (VWR International) to form gels.The culture is validated by streaking heterotrophic growth plates made with premixed Difco 247940 agar (Thermo Fisher Scientific Inc, Waltham, Massachusetts, United States) or with 3, 10, and 15 % (w/v) of tryptic soy, glucose, and agar, respectively, in ultra-pure water(1).
Eppendorf, Hamburg, Germany) at 200 RPM at 30 ˚C to incubate.Cell harvest ensued by harvesting ATCC 21198 in the late exponential growth phase by centrifugation with a Beckman J2-MI (Gas innovations, La Porte, Texas, United States) into the headspace, and placed the growth reactors on a New Brunswick Scientific G10 Gyratory shaker table ( 1 (Sartorius, Fremont, California, United States) was used to generate the CCO experimental matrix and analyze the experimental data obtained.Multiple linear regression was used to obtain predictive models for each response variable.Experimental data was obtained first and used to fit models.Singular value decomposition was used to obtain regression coefficients.Analysis of variance (ANOVA) tests were used to determine the statistical analysis of the experimental data and regression coefficients, and to obtain interactions between the variables and the responses.A desirability function approach was applied to identify the optimized condition that produced the most desirable responses on dependent variables.The quadratic equations used in DOE for all four dependent variables were: =  0 +  1   +  2   +  3   + 11  2  +  22  2  +  33  2  +  12     +  13     +  23     +

Table S2
Bead formulae for each experiment, generated by a central composite orthogonal design, plus an additional bead formulation (optimal bead).Experimental and predicted data for