Microsensors were introduced in microbial ecology by Niels-Peter Revsbech who ...... Nielsen M, Revsbech NP, Larsen LH, Lynggaard-Jensen A (2002) On-line ...
Use of microelectrodes to measure in situ microbial activities in biofilms, sediments, and microbial mats
ARMIN GIESEKE, DIRK DE BEER
Max Planck Institute for Marine Microbiology, Celsiusstraße 1, D-28359 Bremen, Germany
Introduction
In the last hundred years, microbiologists have elucidated an enormous variety in microbial species and metabolisms. Groundbreaking achievements have been made in understanding the regulation of metabolic activities and the relationships between different microbial processes. However, detailed studies of the activity of microbes in their natural environments are still beyond the limits of the classical approach of microbiology: the isolation of strains and their cultivation under defined conditions. This fundamental approach only provides information on the potential preferences and activities of the cultivated microorganisms. In most aquatic systems, however, the bulk microbial conversions occur by microorganisms immobilised (at least their mobility is limited) in multispecies communities as found in sediments, microbial mats, suspended flocs or bio-films adhered to solid surfaces (further referred to as immobilised microbial biomass, IMB). There, supply with substrates and disposal of products for the individual cell is hindered by the reduction or absence of free convection within these structures, subsequently limiting conversion rates. To understand activities in situ two inherent aspects have to be considered: (i) the variety of microenvironments in situ, which can occur along a very small scale, and (ii) the importance of interactions between various populations. Examples for the relevance of both can be found in many, if not all, microbial life forms, like microbial mats [66], biofilms [14], planktonic [65] and sedimentary [10, 62] consortia, and even microbial endosymbionts and their hosts [29]. Besides this, these aspects might explain why still a large number of microorganisms is not yet cultivated. (For in-depth considerations of bacterial viability and culturability the reader is referred to a review by Barer and Harwood [6]). Predicting the behaviour of IMB based on the cultivation of its cultivable members is further biased by the unknown species distribution within the IMB. The microenvironments in the IMB differ from the bulk medium because of mass transfer resistance (e.g., [38, 71]). Extrapolation of the system behaviour to that of individual populations or even microcolonies is impossible
2
without knowledge about the microenvironment of these populations. Therefore, detection techniques with high spatial resolution are needed for the analysis of both microbial community structure (phylogenetic composition, abundance and distribution of bacterial species) as well as the activities of the community (amount and distribution). Promising recent tools are (i) molecular analyses of the genetic material, which avoid the bias of cultivation methods, and (ii) the use of microsensors to determine the chemistry of microenvironments inside intact IMB. Microsensors were introduced in microbial ecology by Niels-Peter Revsbech who constructed the first reliable O2 microsensors for profiling sediments and biofilms [70]. Numerous microsensors relevant for microbial ecology have subsequently been developed and used, such as for N2O [73], pH [39], NH4+ [27], NO3- [25, 54], S2- [75], H2S [45], NO2[21], CH4 [17] and CO2 [19]. The development of new microsensors remains an active and dynamic research field. This chapter will (i) give an overview of measurement principles and types of microsensors, (ii) illustrate the preparation of ion-selective microelectrodes, (iii) address the determination of substrate and product profiles by microsensors, and (iv) explain how these profiles can be used to calculate the distribution of microbial activity. Microsensors Microsensors used in microbial ecology are needle-shaped devices with a tip size of 1-20 µm that measure the concentration of a specific compound. Due to the small sensing area at the tip highly localised measurements are possible in biofilms, flocs, aggregates, microbial mats and sediments. Although the technique is invasive, the tiny tips have only a small influence on structures and processes. Currently, four principally different microsensor types are available, based on amperometric, voltammetric, potentiometric and optical working principles, respectively. In Fig. 1 the three most used sensor types and measuring principles are depicted. Amperometric microsensors have been described in detail previously (see references below). Very recently, voltammetric microsensors have been introduced to microbial ecology, but this type has only been applied in particular cases so far. The optical microsensors are restricted to the measurement of a few solutes and temperature. Thus, these sensor types will be reviewed shortly. The making and use of the most versatile type, the potentiometric microsensors, will be described in more detail. Amperometric microsensors are based on current measurements induced by the electrochemical reduction or oxidation of the analyte in the tip, with a rate proportional to its concentration (Fig. 1A). Useful sensors based on this principle have been established for O2 [72], N2O [3, 73], H2S [45] and HClO [22]. The O2 microsensor is the most used type and has been applied by many research groups to study photosynthesis and respiration in sediments [26, 31, 71, 88], microbial mats [75, 96], biofilms [24, 52], activated sludge flocs [82, 85], marine snow [67, 68], corals [1, 20, 51] and foraminifera [49, 74, 77]. The N2O sensor is used for denitrification studies in biofilms, sediments and microbial mats [73]. In freshwater environments or in systems where the acetylene inhibition technique cannot be used (e.g., due to co-occurrence of nitrification) the potentiometric nitrate microsensor [89] is preferred (e.g., [43, 44]). In marine or purely
3
denitrifying systems, however, the N2O sensor is still a valuable tool. With the H2S microsensor, sulphate reduction and sulphide oxidation have been studied in biofilms [21, 78, 79], activated sludge [80, 85] and microbial mats [97]. The HClO microsensor has been used in biofilm disinfection studies [22, 87]. The O2 microelectrode is the oldest microsensor used in microbial ecology, and in its present form (with internal reference) probably the most reliable one. The principle and practice of this microsensor is described in detail elsewhere [72, 74]. Oxygen diffuses into the sensor through a silicon membrane in the 2-10 µm diameter tip, and is reduced at a cathode close to the tip (see also scheme in Fig. 1A). The resulting electrical current is proportional to the oxygen concentration near the tip. A small silicone membrane also covers the tip of the H2S sensor, allowing only the diffusion of gaseous H2S into the tip, where the electrons are transferred to the electrode via a redox mediator (i.e., a ferri-/ferrocyanide redox system; see also Fig. 1A). As the membrane is dense and, thus, permeable only for small uncharged molecules (not for ions with their large hydrate shell), the sensor exclusively detects H2S, but not S2-or HS-. Subsequently, it is most sensitive for total S(-II) at low pH. To calculate the total S(II) concentrations from the H2S measurements also the pH has to be measured. Precise alignment of pH and H2S profiles, however, is difficult, and minor displacements can result in large errors if steep gradients are encountered. If the experimental conditions allow it, a strong buffer (e.g., 10 mM phosphate) should be used to dissipate the pH gradients in the sample, significantly simplifying interpretation of sensor readings [52]. The selectivity of amperometric microsensors is determined by three factors: (i) the applied polarisation voltage, (ii) the specific permeability of the membrane at the tip and, (iii) where required, by specificity of redox mediators (see also Fig. 1A). The polarisation voltage of -0.8 V applied to an O2 sensor is insufficient for reduction of nitrous oxide that needs -1.2 V. On the other hand, N2O sensors are sensitive for O2. They can only be used in the absence of oxygen. In oxic environments, a combined sensor in which O2 cannot reach the N2O cathode must be used [73]. Alternatively, an oxygen-insensitive N2O sensor has been described for the use in oxic environments. The tip of this sensor is surrounded by a capillary filled with ascorbate, where O2 is consumed in front of the inner N2O sensor tip [3] (see also Fig. 1A). The HClO sensor operates at -0.2 V, which is too low for O2 reduction, and HClO is not sensed by an O2 microsensor because it does not permeate through the silicon membrane of the latter sensor. The selectivity of the H2S microsensor is further based on the specific redox chemistry of [Fe(CN)6]3-/[Fe(CN)6]4- as the redox mediator [45]. The variety of measurable analytes has been expanded by applying either purified enzymes or whole bacterial cells as catalysts for the formation or consumption of redox compounds in the sensor. A scheme of such a biosensor is given in Fig. 1A (right panel). Enzyme-based microscale biosensors are the glucose sensor [16] that is based on immobilised glucose oxidase, and the CO2 microsensor that uses carbonic anhydrase [19]. Cultures of methane oxidisers are used in methane microsensors [17]. The decrease in oxygen during consumption of methane in the tip is sensed by an internal O2 microsensor. The signal can then be converted into a methane partial pressure by calibration. Due to its measuring principle, the methane sensor is sensitive to oxygen. A modified sensor system with an oxygen "trap" (a chamber with heterotrophic bacteria in
4
front of the tip) can be used for measurements of methane in oxic environments. Pure cultures of incomplete denitrifiers (lacking the enzyme nitrous oxide reductase) have been applied to construct nitrate/nitrite sensors [54, 55]. Both nitrate and nitrite are converted to N2O, that is then detected by an internal N2O microsensor. This type of sensor has been applied in freshwater [58], marine sediments [57, 60], and in activated sludge [53]. Further modifications of this type of sensor have recently led to the development of a microscale nitrite biosensor (by replacement of the denitrifying strain by another one that only utilises nitrite, but not nitrate, [63]), and a microsensor for volatile fatty acids (by use of a medium containing nitrate but no electron donor in the front chamber; [61]). We can expect the list of biosensors to be extended in the future. The large advantage of whole cells over enzymes for the use in biosensors is their stability: The life time of cell-based biosensors is in the range of weeks (nitrate/nitrite biosensor, [54]) to months (methane biosensor, [17]), whereas that of enzyme-based sensors is a few days (CO2 biosensor, [19]) to one week (glucose biosensor, [16]). To increase sensitivity of the sensors, a charge can be applied across the tip membrane, that enhances transport of ions into the tip (electrophoretic sensitivity control, ESC [47, 61]). A serious disadvantage of all amperometric microsensors, especially the microscale biosensors, is their complexity. For example, the combined O2-N2O microsensor, and the microscale biosensors for nitrate/nitrite and glucose need to be made by a few highly skilled and well-trained personnel. Furthermore, a low success rate and short life expectancy have contributed to the limited popularity of this sensor principle. Also the making of the O2 microelectrode and the H2S microsensor, both with internal references, needs considerable experience, but it is worth the investment. After some practice, several sensors can be produced in one day, with a life time of up to several months. The complexity of the biosensors is balanced by their relatively long life time, and the absence of alternatives. In certain cases a more simple design might be sufficient for oxygen sensors: a simple cathode coated with DePeX, an oxygen permeable resin. The catalytic site of the sensor is in close contact with the sample, that should be low in calcium and magnesium ions, and have a pH above 6 [71, 72]. If these conditions are not fulfilled (sea water, acidic springs) the calibration will change gradually during the measurements. The HClO microsensor is also of this single cathode type and simple to make and use. For amperometric microsensors, a picoampere meter (ranging down to 1 pA, but preferably lower) and a voltage source for polarising the sensor are needed; both can be integrated most conveniently. Amplifiers can be either bought, e.g., from Keithley (28775 Aurora Road, Cleveland, OH 44139, USA) or WPI (175 Sarasota Center Blvd, Sarasota, FL 34240, USA), or home-built for a fraction of the price (about € 500,- in parts). UniSense A/S (Science Park Aarhus, Gustav Wieds Vej 10, DK-8000 Aarhus C, Denmark) sells complete microsensor measuring systems, including microsensors and electronics. Voltammetric microsensors for in situ measurements in sediments have been described by Luther et al. [59]. For this sensor type, which has been applied only recently in the field, the reader is referred to this and related publications for further information.
5
Opt(r)odes are based on the change of fluorescence (fluorescence intensity, life time of fluorescence state or absorption) of an indicator dye covering the tip of an optical fibre. Optical microsensors have been developed for O2, pH and temperature. The presence of the substrate induces quenching of the fluorescence intensity or decrease of the fluorescence life time. The response is not linear as with amperometric sensors. Since the quenching follows the well-described Stern-Volmer kinetics, a two-point calibration is sufficient. A description of the theory and practice is given by Klimant et al. [48]. Optodes typically have a tip diameter of ≥ 20 µm and a response time in the order of seconds. Their attractions are their stability, mechanical strength and most of all: ease of manufacturing. The sensor part consists of an optical fibre, tapered in a flame and dipcoated with the desired compound. An O2 optode can be prepared within a few minutes. The ease of making optodes will probably lead to more widespread use of this type of microsensor among microbial ecologists. The indicator dye can also be spread on a larger foil, that can be brought into contact with samples such as sediments and microbial mats [36], and biofilms [37], allowing 2-D measurements. New measurement systems working with transparent planar optodes allow a large versatility [40]. The optical and electronical components needed for an optode setup include a LED light source, excitation and emission filters, beam splitters, and a photomultiplier system (Fig. 1C). Integrated opto-electronics and sensors can be purchased by PreSens Precision Sensing GmbH (Josef-Engert-Str. 9, 93053 Regensburg, Germany). Details of the microoptode technique are given in a review by Holst et al. [41]. Potentiometric microsensors are based on the measurement of an electrical potential at the tip generated by charge separation at an interface. Usually, (the electrode part of) this interface is formed by an ion-selective membrane. Three different types of membranes are used: 1. Full glass. The membrane consists of a pH-sensitive glass similar to typical pH electrodes (e.g., pH-sensitive glass tube, Sensortechnik Meinsberg, Meinsberg, Germany). The most used design is the Hinke type [39, 91]. The glass pH microsensor can be applied in many environments [74], but due to its rather long tip (50-100 µm) it has a limited spatial resolution. The response time of this sensor is comparatively long (90% of final signal change from one concentration to another reached in about 30 s). 2. Metal oxide. A few designs for an iridium oxide pH sensor have been described [8, 93]. These sensors, however, have yet to be applied in environmental samples. In contrast, the Ag2S membrane sulphide electrode [75] has been widely used to study the sulphur cycle in microbial mats [75] and biofilms [42, 52]. The peculiarities of this sensor have been the subject of hot debates. Most of the problems could be attributed to imperfect sealing of the Ag2S membrane and the presence of oxygen. The use of this sensor is restricted to completely anoxic environments, because the tip potential of 600-700 mV is sufficient to reduce oxygen. The induced current and concomitant pH effects lead to unreliable signals and changing calibrations. The newly developed (amperometric) H2S sensor, mentioned above, has no such problems and can replace
6 2-
the S sensor in most research. The H2S sensor is most suitable for environments with low pH ( 30s), but reading approaches a defined value
Solder joint between coax and Ag/AgCl wire in contact with filling electrolyte A. Wrong tip geometry
Move the cable upwards
Make new capillary with a more conical tip (exception: pH microsensor) B. Membrane and/or electrolyte Prepare new membrane cocktail chemicals or composition inade- / filling electrolyte quate C. Sensor too old Fill new sensors Readings show drift, i.e., no A. Leakage at glass surface / If drift is small, recalibrate defined value is reached, membrane interface causing regularly and correct for drift in reading over time looks "log-like" dissipation of potential calculations difference Fill new sensors; (wrong/insufficient silanisation, If older capillaries used, predry humidity at glass surface, etc.) in the oven before filling; If problem persists, prepare new capillaries and take special care for silanisation step B. Reference electrode defective Use ref. electrode properly, or inadequate refresh if needed; refer to manual C. Ref. electrode (type) Use other (more reliable) type of unreliable ref. electrode Slope is too low, i.e.
