Biphenyl Dioxygenase from an Arctic Isolate Is Not Cold Adapted

Abstract

Biphenyl dioxygenase from the psychrotolerant bacterium Pseudomonas sp. strain Cam-1 (BPDO_Cam-1) was purified and found to have an apparent for biphenyl of 1.1 ± 0.1 s⁻¹ (mean ± standard deviation) at 4°C. In contrast, BPDO_LB400 from the mesophile Burkholderia xenovorans LB400 had no detectable activity at this temperature. At 57°C, the half-life of the BPDO_Cam-1 oxygenase was less than half that of the BPDO_LB400 oxygenase. Nevertheless, BPDO_Cam-1 appears to be a typical Pseudomonas pseudoalcaligenes KF707-type dioxygenase.

The cold-tolerant bacterium Pseudomonas sp. strain Cam-1, previously isolated from Arctic soil, transforms polychlorinated biphenyls (PCBs) with up to four chloro substituents at 7°C at higher rates than Burkholderia xenovorans LB400, a well-characterized PCB-degrading mesophilic bacterium (2, 14). At 50°C, PCB removal by Cam-1 was diminished, but that by LB400 was not (14). The degradation of PCBs by these two strains is further distinguished by their respective congener preferences. LB400 preferentially degrades ortho-substituted PCBs with up to six chloro substituents (6). In contrast, the congener preference of Cam-1 is similar to that of Pseudomonas pseudoalcaligenes KF707, preferentially degrading para-substituted PCBs with up to four chloro substituents (6). A third well-characterized mesophilic PCB-degrading bacterium, Pandoraea pnomenusa B-356 (previously Comamonas testosteroni B-356), preferentially transforms meta-substituted congeners with up to four chloro substituents (1).

The first enzyme of the biphenyl/PCB-degrading pathway, biphenyl dioxygenase (BPDO), is one of the major determinants of the aerobic PCB-degrading capabilities of bacteria (17). Accordingly, the PCB-degrading capabilities of strains LB400 and B-356 largely reflect those of BPDO_LB400 and BPDO_B356, respectively (7). BPDO is a three-component dioxygenase comprising a reductase, a ferredoxin, and an oxygenase component (iron-sulfur protein [ISP]) of α₃β₃ constitution. The α-subunit contains a Rieske FeS cluster and, within the active site, a mononuclear iron center. BPDO utilizes NADH to activate O₂ and catalyze the 2,3-dihydroxylation of biphenyl. A crystal structure of ISP_B356, the oxygenase component of BPDO_B356, in complex with 2,6-dichlorobiphenyl has provided insight into the determinants of congener preference in BPDOs (7). For instance, Phe336 of ISP_LB400 decreases the preference of this enzyme for di-para-substituted congeners (7). Moreover, BPDOs have been engineered to improve their PCB-degrading capabilities (3, 5, 18).

The oxygenase component of BPDO_Cam-1 shares 99%, 95%, and 70% amino acid sequence identities with those of BPDO_KF707, BPDO_LB400, and BPDO_B356, respectively. Indeed, the α-subunits of ISP_Cam-1 and ISP_KF707 differ by a single residue at position 178 (Ala in ISP_Cam-1 and Val in ISP_KF707) (13). To better understand the differences between the various BPDOs and their influence on the temperature dependence of PCB degradation by the parent strains (14), BPDO_Cam-1 was heterologously produced and purified. The biphenyl- and PCB-transforming activities of BPDO_Cam-1 were evaluated and compared with the data for BPDO_LB400, BPDO_B356, and BPDO_KF707. The impact of temperature on the activities and stabilities of BPDO_Cam-1 and BPDO_LB400 were also compared to evaluate whether the former is cold adapted.

Reconstitution of BPDO_Cam-1.

ISP_Cam-1 and ISP_LB400, the oxygenase components of BPDO_Cam-1 and BPDO_LB400, respectively, were produced in Escherichia coli strain C41(DE3) (15) by using derivatives of pT7-6a and pT7-7a, respectively (8, 13). The ISPs were purified and handled anaerobically in a glovebox, concentrated to 20 to 30 mg/ml, and flash frozen as beads in liquid nitrogen essentially as described previously (7, 11). The yields of ISP_LB400 and ISP_Cam-1 were approximately 15 and 23 mg/liter of cell culture, respectively, and the R values (A₂₈₀/A₃₂₃) and iron and sulfur content of these preparations indicated that the ISPs contained essentially their full complements of Rieske-type [2Fe-2S] clusters and mononuclear iron centers (11). BPDO_Cam-1 and BPDO_LB400 were reconstituted by using His-tagged ferredoxin from LB400 (Ht-BphF_LB400) and His-tagged reductase from B-356 (Ht-BphG_B356) (4, 10). The specific activities of BPDO_Cam-1 and BPDO_LB400 were 3.1 ± 0.4 (mean ± standard deviation) and 0.27 ± 0.01 U/mg, respectively, at 25°C. The sequences of BphF_LB400 and BphF_Cam-1 are identical, and the results of a previous study indicate that exchanging electron transfer components between closely related BPDOs, including using Ht-BphG_B356, does not significantly affect enzyme activity (10). Thus, differences in the activities and steady-state parameters of BPDO_Cam-1 and BPDO_LB400 reflect differences in the respective ISPs of these systems.

A comparative analysis of BPDO_Cam-1 reactivity.

The steady-state parameters of BPDO for biphenyl were evaluated by following the consumption of O₂ as previously described (7, 11). All experiments were performed by using air-saturated 50 mM morpholineethanesulfonic acid (MES), pH 6.0, and the assay mixtures contained 0.6 μM ISP, 3.6 μM Ht-BphF_LB400, 1.8 μM Ht-BphG_B356, 320 μM NADH, and 0.5 to 160 μM biphenyl. Under all conditions studied, the steady-state consumption of O₂ by BPDO as a function of biphenyl concentration obeyed Michaelis-Menten kinetics with the fitted parameters shown in Table 1. Moreover, the hydroxylation of biphenyl was well coupled to O₂ activation. That is, in the presence of 160 μM biphenyl and excess NADH, the amount of O₂ consumed by BPDO_Cam-1 corresponded to the amount of biphenyl utilized (Table 2) and no H₂O₂ was detected upon the addition of catalase to these reaction mixtures. At 25°C, the steady-state utilization of biphenyl by BPDO_Cam-1 was remarkably similar to that of BPDO_B356 (apparent k_cat [] = 4.1 ± 0.2 s⁻¹ and = 20 ± 4 μM [7]) and different from that of BPDO_LB400 previously measured under similar conditions. More particularly, the values of BPDO_Cam-1 and BPDO_B356 for biphenyl are 10 times higher than that of BPDO_LB400, while the apparent specificities (/) of the former are approximately 1/10 that of BPDO_LB400. This is true of BPDO_Cam-1 and BPDO_LB400 at 10°C as well. However, BPDO_LB400 did not detectably transform biphenyl at 4°C.

TABLE 1.

Effect of temperature on the reactivities of purified BPDOs with biphenyl^a

Temp (°C)	BPDOCam-1		BPDOLB400
	(s−1)	(μM)	(s−1)	(μM)
4	1.1 (0.1)	18 (3)	ND	NA
10	2.2 (0.1)	11 (1)	0.26 (0.01)	1.6 (0.3)
25	4.6 (0.2)	29 (3)	0.4 (0.1)b	0.18 (0.03)b

^aValues in parentheses represent standard errors (n = 3 or 4). ND, not detected; NA, data not available.

^bData taken from Gomez-Gil et al., in which assay mixtures contained 0.6 μM ISP, 3.6 μM Ht-BphF_LB400, 1.8 μM Ht-BphG_B356, 350 μM NADH, and 0.1 to 150 μM biphenyl (7).

TABLE 2.

Reactivities of purified BPDOs with selected biphenyls at 25°C^a

Congener	BPDOCam-1		BPDOLB400b		BPDOB356b
	Rate of congener removal (nmol·min−1)	H2O2/O2	Rate of congener removal (nmol·min−1)	H2O2/O2	Rate of congener removal (nmol·min−1)	H2O2/O2
Biphenyl	95 (9)	ND	12 (4)	ND	55.3 (0.6)	ND
2,2′-Dichlorobiphenyl	3.2 (0.3)	0.36 (0.01)	15 (4)	ND	15 (1)	0.6 (0.2)
3,3′-Dichlorobiphenyl	8 (2)	0.40 (0.05)	6 (2)	0.3 (0.1)	32.2 (0.2)	0.20 (0.04)
4,4′-Dichlorobiphenyl	2.6 (0.1)	0.26 (0.05)	3.9 (0.3)	0.4 (0.2)	9 (5)	0.4 (0.2)

^aO₂ consumption (H₂O₂/O₂ ratios) values are based on the assumption that the amount of O₂ detected upon catalase addition corresponds to 50% of H₂O₂ produced. The rates of congener removal and H₂O₂/O₂ ratios are based on data obtained after 1.5 min for biphenyl and after 3 min for dichlorobiphenyls. Values in parentheses represent standard errors (n = 3 or 4). ND, not detected.

^bData taken from Gomez-Gil et al., in which assay mixtures contained 50 μM of each congener (7).

In studying the congener preference of BPDO_Cam-1, it was found that the enzyme depleted biphenyls in the following order of apparent maximal rates at 25°C: biphenyl ≫ 3,3′-dichlorobiphenyl > 4,4′-dichlorobiphenyl ≈ 2,2′-dichlorobiphenyl (Table 2). Thus, as with the steady-state kinetic parameters of biphenyl utilization, the congener preference of BPDO_Cam-1 is similar to that of BPDO_B356 (biphenyl > 3,3′-dichlorobiphenyl > 2,2′-dichlorobiphenyl ≈ 4,4′-dichlorobiphenyl) and distinct from that of BPDO_LB400 (biphenyl > 2,2′-dichlorobiphenyl > 3,3′-dichlorobiphenyl > 4,4′-dichlorobiphenyl) (7, 11). Consistent with what has been reported for BPDO_B356 and BPDO_LB400, the utilization of O₂ by BPDO_Cam-1 in the presence of 68 μM 2,2′-dichlorobiphenyl, 80 μM 3,3′-dichlorobiphenyl, or 40 μM 4,4′-dichlorobiphenyl was not fully coupled (Table 2). The extent of uncoupling between O₂ and dichlorobiphenyl utilization was dependent on both the isozyme and the PCB congener. For example, BPDO_Cam-1 was significantly more uncoupled in the presence of 2,2′-dichlorobiphenyl than BPDO_LB400 (Table 2). Moreover, the uncoupling was not necessarily lowest for dichlorobiphenyls that were depleted at the highest rates.

In general, the congener preference data reported for the purified BPDOs are consistent with those reported for the parent strains in whole-cell assays, except that whole cells of strains Cam-1 and B-356 depleted 2,2′-dichlorobiphenyl very poorly, if at all. Thus, in contrast to the purified BPDOs, cells of strain Cam-1 did not detectably transform 2,2′-dichlorobiphenyl (14) and B-356 cells depleted 4,4′-dichlorobiphenyl better than 2,2′-dichlorobiphenyl (1, 11). While these data might reflect the low sensitivity of the whole-cell assay, it is possible that PCB transformation by whole cells is determined by other factors, such as PCB transport across cell membranes and the presence of other PCB-transforming enzymes in the cell. Finally, differences in apparent congener preference between whole cells and purified BPDO might reflect the enzyme's relative protection against oxidative damage in whole cells. Notably, whole-cell assays are typically conducted over much-longer periods of time (hours) than enzyme assays (minutes).

To investigate the poor transformation of 2,2′-dichlorobiphenyl by whole cells and to facilitate comparisons of results for BPDO_Cam-1 with whole-cell data for BPDO_KF707 (18), PCB removal by E. coli C41(DE3) cells freshly transformed with pT7-7a (bphAEFG_Cam-1) was tested (Table 3) as described previously (7). These data are consistent with the results of previous studies of Pseudomonas sp. strain Cam-1 (14) but differ from the current data obtained with purified BPDO_Cam-1, which removed 2,2′-dichlorobiphenyl and 4,4′-dichlorobiphenyls at similar rates. Importantly, these experimental results establish that in whole cells of E. coli, BPDO_KF707 and BPDO_Cam-1 have the same apparent preference for 3,3′-dichlorobiphenyl and 4,4′-dichlorobiphenyls and that both enzymes transform these congeners at significantly higher rates than biphenyl (18). These data further indicate that the congener preferences of BPDO_Cam-1, BPDO_KF707, and BPDO_B356 are quite similar.

TABLE 3.

Depletion of biphenyls by E. coli cells containing BPDO_Cam-1

Congener	% Depletion ata:
	3 h	6 h
Biphenyl	59 (10)	100 (0)
2,2′-Dichlorobiphenyl	38 (1)	64 (3)
3,3′-Dichlorobiphenyl	95 (1)	100 (0)
4,4′-Dichlorobiphenyl	99 (1)	98 (1)

^aValues in parentheses represent standard errors (n = 3).

Temperature dependence of BPDO activity.

The dependence of BPDO activity on temperature was investigated by performing the standard activity assay at temperatures from 4 to 60°C (Fig. 1). In the results of the standard assay, the concentration of biphenyl exceeded the BPDOs' for this substrate by approximately an order of magnitude. Studies of BPDO_B356 indicate that this is also the case for O₂ ( = 28 μM [11]). The relative activities of the enzymes should therefore be essentially unaffected by temperature-dependent changes in the solubility of the two substrates. At any given temperature between 4 and 45°C, the activity of BPDO_Cam-1 was higher than that of BPDO_LB400. As noted above, at 4°C, the activity of BPDO_LB400 did not differ significantly from the background rate, whereas BPDO_Cam-1 transformed biphenyl at a significant rate (Table 1). Further, the highest observed activity of BPDO_Cam-1 occurred at a lower temperature than that of BPDO_LB400 (47°C versus 55°C). The thermodynamic parameters were derived from Arrhenius plots (9, 12) by using the data up to 42.5°C and 35°C for BPDO_Cam-1 and BPDO_LB400, respectively (Table 4). The free energy of activation (ΔG^‡) of BPDO_Cam-1 is lower than that of BPDO_LB400, consistent with the former's higher . However, determination of the apparent steady-state parameters of the two BPDOs at 10°C revealed that their respective catalytic efficiencies (/) were not significantly different (see Table 1).

FIG. 1.

Effects of increasing temperature on the biphenyl-degrading activities of ISP_Cam-1 (♦) and ISP_LB400 (□) (n = 3; error bars indicate standard deviations). k_rel, relative k_cat.

TABLE 4.

Thermodynamic parameters of ISP_Cam-1 and ISP_LB400 with biphenyl at 25°C^a

Source of ISP	ΔG‡ (kJ mol−1)	ΔH‡ (kJ mol−1)	TΔS‡ (kJ mol−1 K−1)
Pseudomonas sp. strain Cam-1	69.9 (0.05)	38 (2)	−32 (2)
B. xenovorans LB400	74.2 (0.07)	48 (5)	−26 (5)

^aΔG^‡, free energy of activation; ΔH^‡, enthalpy of activation; TΔS^‡, product of temperature and entropy of activation. Values in parentheses represent standard errors (n = 3 or 4).

Finally, we investigated the respective half-lives of ISP_Cam-1 and ISP_LB400 at 57°C by following the characteristic absorbance of the iron-sulfur cluster of these proteins: this temperature was chosen because the enzymes lost activity at a significant rate above 50°C. Briefly, a sample of each ISP was diluted to 2.2 μM in MES buffer (pH 6.0) preequilibrated to 57°C. The sample was incubated at 57°C, and its absorbance was continuously monitored at 323 nm, where a decrease in absorbance reflects loss of the Rieske-type [2Fe-2S] cluster and, thus, loss of ISP function. The half-lives of ISP_Cam-1 and ISP_LB400 were 16 ± 2 min and 38 ± 8 min, respectively. The longer half-life of ISP_LB400 at 57°C is consistent with the higher relative k_cat of BPDO_LB400 that was observed at temperatures above 50°C.

The principal distinguishing characteristic of cold-adapted enzymes is a significant decrease of their enthalpy of activation (ΔH^‡) with respect to that of their mesophilic homologues (5 to 42 kJ mol⁻¹), which contributes to a higher at lower temperatures (12). Cold-adapted enzymes also tend to be less thermostable than their mesophilic homologues. Thus, comparison of these two characteristics of BPDO_Cam-1 and BPDO_LB400 could be interpreted to indicate that BPDO_Cam-1 is cold adapted. However, several considerations argue that ISP_Cam-1 is not cold adapted. First, the difference in ΔH^‡ of the two reactions (10 kJ mol⁻¹) is at the low end of that which distinguishes mesophilic from psychrophilic enzymes. Second, the steady-state kinetic parameters of BPDO_Cam-1 for biphenyl at 25°C are similar to those reported for a mesophilic enzyme, BPDO_B356, determined under similar conditions (7, 11). Third, the specificities of BPDO_Cam-1 and BPDO_LB400 for biphenyl (/) are not significantly different at 10°C, contrary to what has been observed in other cold-adapted enzymes (9). Fourth, the α-subunit of ISP_Cam-1, which encompasses the BPDO active site, is almost identical to that of another mesophilic enzyme, ISP_KF707.

Overall, our data suggest that BPDO_Cam-1 is a typical KF707-type enzyme. Nishi et al. (16) reported that Pseudomonas putida KF715 encodes a KF707-type BPDO on a 90-kbp conjugative element (16). Thus, it is conceivable that such an element was transferred to a psychrotolerant Pseudomonas species, such as Cam-1. It is nonetheless possible that ISP_Cam-1 is less stable than ISP_KF707 and ISP_LB400. In particular, the β-subunits of ISP_LB400 and ISP_KF707 differ by a single residue and they differ from the β-subunit of ISP_Cam-1 by the same five residues (Arg16, Pro33, Gln34, Pro40, and Phe79 in ISP_Cam-1). Inspection of the crystallographic structure of ISP_B356 reveals that all five residues occur at or very close to subunit interfaces. Perhaps the most intriguing of these is Pro33 in ISP_Cam-1, which is an arginine residue in ISP_LB400, ISP_KF707, and ISP_B356. In ISP_B356, this arginine is involved in an extensive network of hydrogen bonds involving residues at the interface of adjacent β-subunits. The loss of this network in ISP_Cam-1 could influence this enzyme's thermostability.

Although the combined data reported here suggest that BPDO_Cam-1 is not cold-adapted, the differences between BPDO_Cam-1 and BPDO_LB400 nevertheless account for the differences between their respective parental strains, Cam-1 and LB400. Thus, whole cells of Cam-1 remove certain PCB congeners at higher rates than LB400 cells at 7°C (14). Furthermore, PCB removal by LB400 increased at high temperature; in contrast, high temperature inhibited PCB removal by Cam-1. Both observations are consistent with the higher of BPDO_Cam-1 for biphenyl at low temperature and the low thermal stability of BPDO_Cam-1 compared to that of BPDO_LB400. Overall, the current data indicate that a distinguishing feature of KF707- and LB400-type enzymes is the formers' superior ability to degrade PCBs at low temperatures.