Photoreactivation and Dark Repair in UV-Treated Microorganisms: Effect of Temperature

Because of the lack of readily available information about theinfluence of temperature on microorganism reactivation processessubsequent to inactivation with UV radiation, a series of batchreactivation studies were performed at 5, 10, 15, 20, 25, and30°C. A special effort was made to model the reactivationprocess to enable the effect of the temperature variable tobe quantified. Because an earlier-proposed kinetic model (K.Kashimada, N. Kamiko, K. Yamamoto, and S. Ohgaki, Water Sci.Technol. 33:261-269, 1996), a first-order saturation type, doesnot adequately fit the data obtained in experiments of reactivationin conditions of light and darkness, a modification of thatmodel is proposed. The new model, which actually coincides withthe classical logistic equation, incorporates two kinetic parameters:the maximum survival ratio (S_m) and the second-order reactivationrate constant (k₂). In order to interpret correctly the reactivationoccurring in conditions of darkness, a new term for the decayis added to the logistic equation. The new model accuratelyfits the data obtained in reactivation experiments, permittingthe interpretation of the kinetic parameters S_m, k₂, and M (foronly repair in darkness), where M is mortality, a zero-orderdecay rate constant, and their relationship with various environmentalconditions, such as microbial type, light, and temperature.The parameters S_m and k₂ (and M for reactivation in conditionsof darkness) show exponential dependence on the reactivatingtemperature, and it is possible to predict their values andhence the reactivation curve from the equations proposed inthis work.

INTRODUCTION

Top

Introduction

UV light is one of the most practical methods used for disinfectionin wastewater treatment systems because it can inactivate bacteria,viruses, bacterial spores, and oocysts of protozoa (6, 9). Unlikethose for other technologies, such as chlorination and ozonation,one of the technological advantages often put forward for UVdisinfection is its lack of sensitivity to temperature variations.This independence from temperature would be expected if UV werea simple photochemical reaction (11). Severin et al. (12) performeda series of batch inactivation studies with bacteria, yeast,and viruses and demonstrated that the effect of temperatureon UV inactivation was very small. Abu-ghararah (1) found thatthe effect of water temperature on the kinetics of the UV disinfectionprocess in the normal operating range of most treatment plants(20 to 40°C) is not statistically significant.

However, the physiological effect of inactivation is not wellunderstood and is complicated by the ability of many organismsto repair UV damage inflicted to their nucleic acids. Whilethe infliction of damage is a purely photochemical reactionthat occurs in the few seconds while water is exposed to UVradiation in the disinfection channels, the interplay of reactionswhich eventually lead to inactivation includes some biochemicalreactions which can be affected by temperature. The latter phenomenacan take place during the irradiation period, but it is mostlyduring the postirradiation time that reactivation processesoccur if adequate conditions exist.

The principal inactivating effect of UV irradiation is the formationof photoproducts in DNA. Of these photoproducts, the most importantis the pyrimidine dimer formed between adjacent pyrimidine moleculeson the same strand of DNA, which can interrupt both the transcriptionand the replication of the DNA. The formation of a dimer canbe reversed by two repair mechanisms: photoreactivation anddark repair. Systematic quantitative study of photoreactivation,the more important of the two mechanisms, has suggested a two-stepreaction scheme (4).

Step 1 is the formation of a complex between a photoreactivationenzyme (PRE) and the dimer to be repaired. This step does notrequire light, but it is dependent on temperature, pH, and ionicstrength (8).

Step 2 is the release of PRE and repaired DNA. The restorationof the dimer to its original monomerized form is absolutelydependent upon light energy intensity. The reaction occurs inless than a millisecond; consequently, the limiting step ofthe whole reactivation process is the formation of the PRE-dimercomplex. An extended period of exposure to photoreactivatinglight would enable the release of PRE, which would then be availableto form new complexes (step 1).

The effect of temperature on the reactivation phase is stilllittle studied. Chan and Killick (3) investigated the effectof salinity and temperature on the reactivation of Escherichiacoli in a marine environment. They found that the effect ofsalinity is greater than that of temperature, although an Arrheniustendency was confirmed for the rate constants calculated.

The prediction of temperature effects is complicated by thenature of DNA itself and the complexity of reactions, enzymes,and molecules implicated in the inactivation and repair processes.

Because information about the influence of temperature on thereactivation process subsequent to inactivation with UV radiationis not readily available, batch reactivation studies were performedat 5, 10, 15, 20, 25, and 30°C. A special effort was madeto model the reactivation process to enable the effect of thetemperature variable to be quantified.

MATERIALS AND METHODS

Top

Materials and Methods

UV-C irradiation.
UV-C irradiation treatment was performed with a 5.0 m³/h, horizontal-lamp,open-channel UV disinfection system (Trojan Technologies, S.L.,Spain) and six low-pressure, high-intensity mercury UV lamps(Philips; 30W UV-C at 254 nm). The UV channel received the waterfrom the unfiltered secondary effluent of the Municipal WastewaterTreatment Plant of Jerez de la Frontera (Spain). The unfilteredsecondary effluent characteristics are summarized in Table 1.

View this table:
[in this window]
[in a new window]

TABLE 1. Wastewater quality characteristics during study period

Bacteria were exposed to a UV-C fluence of 100 mW·s/cm².The UV-C fluence (mW·s/cm²) supplied was calculated asa product of the average UV fluence rate into the reactor (mW/cm²)and the irradiation time (s). The average UV fluence rate wascalculated by the point source summation method (2, 5, 10, 13,14). The exposure time was calculated from the channel volumeand the influent flow rate after first ensuring that the plugflow condition existed in the channel through the study of theresidence time distribution curves. Irradiation was performedat room temperature, between 15 and 20°C.

Repair conditions.
After UV-C irradiation, the water sample was divided and transferredinto two 500-ml glass Erlenmeyer flasks (95% transparent for360 nm light). One of the two Erlenmeyer flasks was thermostatedin a controlled-environment incubator (refrigerated incubator,model no. FOC 225E; VELP Scientifica), which was equipped withone fluorescent lamp (3.7 W; PHILIPS TLD) at six different temperatures:5, 10, 15, 20, 25, and 30°C (photoreactivation). The rangeof the lamp wavelengths was 310 to 420 nm, with a broad peakat 360 nm. Irradiation periods were in the range of 30 to 240min. The UV-A fluence rate of the fluorescent lamp was 0.1 mW/cm²at 360 nm at the sample surface, as estimated by the point sourcesummation method and the distance between the samples and thelamp. The other Erlenmeyer flask was covered immediately withaluminum foil and incubated simultaneously at the same temperaturefor 240 min (dark repair). The concentration of bacteria wasmeasured every 15 or 30 min by using a pipette to take samplesfrom each Erlenmeyer flask.

Enumeration of microorganisms.
Three bacterial indicators of microbiological contaminationhave been analyzed: total number of coliforms (TC), number offecal coliforms (FC), and number of Streptococcus faecalis bacteria(SFB).

All of the microorganisms were analyzed according to the EPA'sstandard methods for the examination of water and wastewater(14) by using the membrane filter technique. Total coliformswere cultured on MF-Endo agar and incubated at 35°C for24 h. Fecal coliforms were determined on m-FC agar and with24 h of incubation at 44.5°C. Finally, S. faecalis isolateswere grown on Kenner fecal agar at 37°C for 48 h. Afterthe incubation period, bacterial colonies were counted, andthe results were calculated as CFU per 100 ml of sample. Foreach microorganism and each experimental condition used, testswere repeated at least three times and mean values were obtainedfrom these repeated experiments.

Experiments were repeated three times independently for eachmicrobial indicator and experimental condition used. Standarddeviations of triplicates are not presented on the graphs inthe interest of clarity. When standard deviation was disproportionate(coefficient of variance, >20%), data were rejected.

The rates of reactivation were assessed by determining microorganismsurvival from microbial numbers before disinfection and afterreactivation phenomenon.

RESULTS

Top

Results

Modeling the reactivation kinetics.
Reactivation is frequently expressed as a function of the survivalratio with respect to the initial microorganism concentrationexisting before the inactivating treatment. Therefore, survivalvalues were calculated using the following equation:

Formula 1 (1)
where S is the survival ratioat time t, N_o is the concentration of microorganisms beforedisinfection (CFU/100 ml), and N_r is the concentration at timet after the beginning of the reactivation phase (CFU/100 ml).

A typical inactivation-reactivation curve as a function of timeis shown in Fig. 1. In that figure, it is possible to differentiatethe various phases of the process: exponential UV inactivationand the reactivation process, which includes an induction period,the growth phase, the stabilization phase, and the decay period.As stated, reactivation can occur by two mechanisms accordingto the exposure of the samples to light or darkness.

View larger version (15K):
[in this window]
[in a new window]

FIG. 1. Typical inactivation-reactivation curve as a function of time, where N_o is the concentration of microorganisms before disinfection, N_d is that after disinfection but before reactivation, and N_r is that at time t after the beginning of the reactivation phase. N_m is the maximum concentration of microorganisms reached by reactivation.

Photoreactivation kinetic.
Kashimada et al. (7) proposed an asymptotic model, assumingthat the photoreactivation phenomenon follows a saturation-typefirst-order reaction, as shown by the following equation:

Formula 2 (2)
where S_m is the maximum survivalratio (N_m/N_o · 100 [N_m is the maximum concentration ofmicroorganisms {CFU/100 ml}]) and k₁ is the first-order reactivationrate constant.

In the model, the term S_m – S acts as the driving forcefor the reactivation. As the survival ratio, S, is reachingits maximum value (S_m), the process decelerates, showing anasymptotic tendency.

Reactivation curves have been obtained by means of the experimentsdescribed below. In Fig. 2a, the typical asymptotic sigmoidalshape of these curves can be observed. In the induction period,the curve suffers imperceptible reactivation, then a rapid exponentialgrowth can be seen, and finally a stabilization period is reachedwhen growth ceases. After applying the model of Kashimada etal. (7), we observed that it did not fit the data correctly,mainly at the beginning of the curve, when an induction periodis observed (Fig. 2b). Hence we decided to modify the model,but without increasing the number of parameters. The new modelis represented by equation 3:

Formula 3 (3)
wherek₂ is the new growth, second-order reactivation rate constant.This relationship is simply a combination of the second-orderequation and the driving force concept employed by Kashimadaet al. (7). The equation is really not new because it coincidesin its mathematical form with the logistic equation proposedby Verhulst in 1838 (15) to interpret biological populationgrowth. Nevertheless, the originality of our work lies in theinnovative application of the equation to microorganism reactivationprediction. The model has the advantage that both kinetic parameters,S_m and k₂, have clear physical significance. On one hand, S_mis the maximum limit of the survival of the microorganisms byreactivation and, on the other hand, k₂ represents the rateat which that value is reached. It can be seen in Fig. 2b thatthis proposed model correctly fits the experimental data.

View larger version (17K):
[in this window]
[in a new window]

FIG. 2. (a) Typical photoreactivation curve. (b) Curves from the model proposed by Kashimada et al. and the logistic model.

By the integration of equation 3, the following equation isobtained:

Formula 4 (4)
whereS_o is the survival immediately after UV disinfection (N_d/N_o· 100 [N_d is the concentration of microorganisms afterdisinfection {before reactivation} {CFU/100 ml}]).

From equation 4, it is possible to express the variable S asa function of the kinetic parameters k₂, S_m, S_o and time (equation5) and, as a result, we can easily obtain both parameters S_mand k₂ by nonlinear regression:

Formula 5 (5)
Equation 5 allows the photoreactivation curveover time to be simulated (Fig. 3).

View larger version (20K):
[in this window]
[in a new window]

FIG. 3. Survival ratio (logarithmic scale) versus time of exposure to photoreactivating light for different reactivation temperatures (5 to 30°C). Experimental data and prediction by the logistic model are shown.

Photoreactivation experiments.
As stated, samples were exposed to an inactivating UV-C doseof 100 mW·s/cm² and then to six reactivation temperatures(5, 10, 15, 20, 25, and 30°C). Figure 3 represents the survivalratio (on a logarithmic scale) versus time for TC, FC, and SFB,and the asymptotic shape of the curves can be seen, includingan induction period, an exponential growth phase, and finallya stabilization phase. The model described in equation 5 wasapplied to experimental data by nonlinear regression. Table2 gives the values of the estimated kinetic parameters S_m andk₂, together with the r² statistics and the optimization function(observed – predicted)². The good fit of the model tothe experimental data can be observed.

View this table:
[in this window]
[in a new window]

TABLE 2. Kinetic parameters of the logistic model applied to photoreactivation experiments^a

This study indicated similar behavior with respect to the reactivationtemperature for the three microbial indicators. Thus, TC, FC,and SFB all showed an expected result: the higher the temperature,the greater the maximum reactivation observed. By comparingupper reactivation limits (S_m), we found the following sequence:TC > FC > SFB. The maximum survival ratio (TC exposedto 30°C) did not exceed 1%. Despite this apparently lowpercentage, for an N_o of 10⁶ CFU/ml, a reactivation of 10⁴ colonies/mlwould be produced; this could cause serious health and environmentalproblems. On the other hand, if the temperature dependence ofk₂ is considered (Table 2 and Fig. 4), the following relationshipis observed: the higher the temperature, the smaller the valuefor the kinetic parameter k₂. Normally, chemical and biochemicalrates are increased by temperature, according to the Arrheniusrelationship; however, in this case the opposite behavior wasobserved. In fact, k₂ is not a pure reaction rate constant,but rather is a model parameter that is adjusted to predictthe experimental data. Its physical meaning is related to thetime required to reach the maximum survival ratio and then thestabilization phase: high k₂ values signify short inductionand growth phases. The trend of k₂ versus temperature foundin these experiments could be explained as follows: since anelevated temperature provides a higher maximum of reactivation(S_m) then to reach that maximum, more time is required, andthus, the k₂ constant is lower. By comparing the k₂ values foreach different microbial indicator (Table 2 and Fig. 4), thefollowing sequence can be seen: TC ${approx}$ FC < SFB. This meansthat SFB reaches the maximum survival ratio sooner than do TCand FC (Fig. 3), although this maximum is lower than those reachedby TC and FC. Therefore k₂ values may depend on particular biochemicalrepair mechanisms that vary among different microorganisms.

View larger version (23K):
[in this window]
[in a new window]

FIG. 4. Kinetic parameters S_m, k₂, and M of the logistic model for the three microorganism groups assayed: TC, FC, and SFB. A comparison between dark repair and photoreactivation is shown.

In Table 3 and Fig. 4, the consistent behavior of S_m and k₂with temperature can be seen. On one hand, S_m (the asymptoticlimit of the microorganism survival due to photoreactivation)shows a positive exponential trend with the reactivation temperature.This finding indicates that low temperatures represent impedimentsfor the bacteria, making their reactivation less easy, whilehigh temperatures encourage a more extended photoreactivationprocess. On the other hand, k₂ (the kinetics rate constant)shows an inverse tendency: it decreases with increasing temperature.This finding means that the reactivation phenomenon finishessooner when the temperature is low because a lower survivallevel is reached than that with higher temperatures.

View this table:
[in this window]
[in a new window]

TABLE 3. Exponential dependence of the reactivation parameters S_m, k₂, and M on the temperature of the reactivation phase

Dark repair kinetic.
After carrying out the dark repair experiments, it was observedthat the curve of the survival of microorganisms versus timeshowed, after a low and brief reactivation period, a decay phasenot detected in photoreactivation experiments. We concludedthat in darkness, the reactivation did occur but to a lesserextent than that in illuminated conditions, and then the survivalcommenced a decreasing tendency (Fig. 5). Therefore the modelproposed (equation 5) did not fit these data, and it was necessaryto include a term corresponding to the decay process. Observingexperimental curves, we decided to probe a zero-order kineticfor this decay process, following equation 6:

Formula 6 (6)
where M, mortality, is a zero-order decayrate constant.

View larger version (23K):
[in this window]
[in a new window]

FIG. 5. Survival ratio (logarithmic scale) versus time of exposure to darkness for different reactivation temperatures (5 to 30°C). Experimental data and prediction by the logistic model are shown.

Equation 7 is obtained by integrating equation 6, and it indicatesthat the decay term is linear over time

Formula 7 (7)
By combining equations 5 and 7, the modifiedversion of the integrated model is obtained, which is also validfor dark repair:

Formula 8 (8)

Dark repair experiments.
Figure 5 shows that in conditions of darkness, reactivationoccurs to a considerably lower degree than that in the caseof photoreactivation. After the maximum survival is reached,a decay process is observed, with survival diminishing accordingto a linear trend over time. For this reason, a modificationhas been made to the model (equation 8). In Table 4, the resultsof applying the model are listed, including the values of thethree kinetic parameters estimated: S_m, k₂, and M (mortalityrate constant). The last parameter, M, is calculated experimentallyfrom the slope of the linear end of the survival curve (betweenthe 120 min point and the last point at 240 min). The good fitof the model to the experimental data from dark repair experimentscan also be seen. The parameters S_m and k₂ have exponentialrelationships with temperature, similar to those found in thephotoreactivation experiments, whereas M seems to be dependenton temperature and the microorganism, having a slight tendencyto increase when temperature increases. The value of M rangesbetween 1 x 10^–5 and 6 x 10^–5 (percent survivalper min). Those values are relatively small, since during the240 min of reactivation, a maximum reduction of survival ofonly approximately 0.015% will be produced. TC presented lowervalues of M than did FC and SFB, and the last two reached figuressimilar to each other (Fig. 4).

View this table:
[in this window]
[in a new window]

TABLE 4. Kinetic parameters of the logistic model applied to dark repair experiments^a

In Fig. 4, it is possible to compare the photoreactivation anddark repair processes through the kinetic parameters obtainedresulting from the application of the model. It can be seenthat, at elevated temperatures, S_m is more than 1 order of magnitudehigher in the case of photoreactivation, unlike k₂, which ishigher in dark experiments where reactivations finish earlier.Table 3 lists the coefficients of the fitted exponential thatare useful for predicting the S_m, k₂, and M parameters by typeof microorganism, light or dark conditions, and reactivationtemperature. Good correlations can be seen, except in the caseof k₂ in SFB and darkness, where no variations with temperatureare observed and hence a low r² is obtained.

DISCUSSION

Top

Discussion

The kinetic model proposed by Kashimada et al. (7), a first-ordersaturation type, does not adequately fit the data obtained inreactivation experiments conducted in conditions of light anddarkness, because it does not take into account the existenceof an induction period.

A modification of that model is proposed, which consists ofchanging from a first- to a second-order reaction while keepingthe concept of saturation model suggested by Kashimada et al.The new model, which actually coincides with the classical logisticequation, incorporates two kinetic parameters, S_m (maximum survivalratio) and k₂ (reactivation rate constant). In order to interpretcorrectly the reactivation that occurs in conditions of darkness,a new term for decay, based on the parameter M, the mortalityrate constant, is added to the logistic equation. The modelaccurately fits the data obtained in photoreactivation experimentsand permits the interpretation of the estimated kinetic parametersS_m, k₂, and M (only for dark repair) and their relationshipwith various environmental conditions, such as microbial type,light, and temperature.

With respect to the type of microbial indicator exposed to photoreactivatinglight, the maximum survival ratio S_m follows the sequence TC> FC > SFB, whereas for the pseudo-second-order kineticsconstant k₂, the sequence is SFB > TC = FC. Dark repair wasless important than photoreactivation, with the differencesbetween the microbial indicator types therefore being less relevant.The maximum survival ratio (TC exposed to 30°C and to reactivatinglight) did not exceed 1%; however, despite this apparently lowpercentage, it could cause serious health and environmentalproblems.

In the experiments carried out in darkness, S_m was 1 order ofmagnitude lower than that in the experiments of photoreactivationand, to the contrary, k₂ was greater in darkness experiments.

The new kinetic parameter, M, shows a slight tendency to increasewhen temperature increases. We suppose that this mortality rateis due to the residual effect of radiation on the bacterialDNA, since the biochemical mechanism of actuation needs sometime to be manifested completely. It is possible that mortalitydoes not occur in photoreactivation experiments since, in thissituation, the repair of damaged DNA is more effective.

The parameters S_m, k₂, and M, both in photoreactivation andin conditions of darkness, have shown exponential dependenceon the reactivating temperature, and it is possible to predicttheir values and then the reactivation curve from the equationsproposed in this work.

S_m increases with the temperature, indicating that the extentof reactivation is favored by elevated temperatures.

However, k₂ diminishes with the temperature of the reactivation.This fact is due to k₂ not being a pure reaction rate constant:it is a model parameter that is adjusted to predict the experimentaldata. Its physical meaning is related to the time required toreach the maximum survival ratio and then the stabilizationphase: high k₂ values signify short induction and growth phases.

M also seems to be dependent on temperature and has a slighttendency to increase when temperature increases. The value ofM ranges between 1 x 10^–5 and 6 x 10^–5 (percentsurvival per min). These are relatively low values since duringthe 240 min of reactivation there is a maximum reduction ofsurvival of only approximately 0.015%.

Since UV irradiation occurs at room temperature (15 to 20°C),the reactivation experiences at the most extreme temperaturesassayed (5 and 30°C) could cause a temperature shock tothe bacteria and therefore alter the reactivation process. Nocontrol experiments were conducted to quantify this effect separately,but the results of our experiment permit the modeling of themicrobial survival after UV-C irradiation at several temperatures.That survival would be a consequence of the combined effectof reactivation and temperature shock.

Nevertheless, TC, FC, and SFB are each heterogeneous groupsand the composition of these indicators may change; this newmodel has the capability of fitting to experiment data fromdifferent wastewaters, resulting in new parameters that willpermit the prediction of the reactivation process.

ACKNOWLEDGMENTS

We thank M. García and E. Gracia of Trojan Technologies,S.L., Spain, for their assistance. We also express our gratitudeto AJEMSA and AQUALIA for providing technical support.

FOOTNOTES

* Corresponding author. Mailing address: Facultad de Ciencias del Mar y Ambientales, 11510 Puerto Real, Spain. Phone: 34 956 016198. Fax: 34 956 016746. E-mail: enrique.nebot{at}uca.es.

Published ahead of print on 5 January 2007.

REFERENCES

Top