Johannes Rousk, Environment Centre Wales, Bangor University, 2nd floor, Bangor, Gwynedd LL57 2UW, UK. Tel.: +44 756 390 68 29; fax: +44 1248 354997; e-mail: firstname.lastname@example.org
Bacterial and fungal growth rate measurements are sensitive variables to detect changes in environmental conditions. However, while considerable progress has been made in methods to assess the species composition and biomass of fungi and bacteria, information about growth rates remains surprisingly rudimentary. We review the recent history of approaches to assess bacterial and fungal growth rates, leading up to current methods, especially focusing on leucine/thymidine incorporation to estimate bacterial growth and acetate incorporation into ergosterol to estimate fungal growth. We present the underlying assumptions for these methods, compare estimates of turnover times for fungi and bacteria based on them, and discuss issues, including for example elusive conversion factors. We review what the application of fungal and bacterial growth rate methods has revealed regarding the influence of the environmental factors of temperature, moisture (including drying/rewetting), pH, as well as the influence of substrate additions, the presence of plants and toxins. We highlight experiments exploring the competitive and facilitative interaction between bacteria and fungi enabled using growth rate methods. Finally, we predict that growth methods will be an important complement to molecular approaches to elucidate fungal and bacterial ecology, and we identify methodological concerns and how they should be addressed.
Species composition, biomass, and growth rate (production) are important characteristics for any community, including soil fungi and bacteria. Of these variables, the growth rate will be most sensitive to changes in environmental conditions, thus being the prime choice to enable the detection of rapid and subtle changes in microbial communities. Bearing this in mind, the rudimentary state of our knowledge on microbial growth and production in soil is striking. Different studies that have attempted to estimate growth have yielded very different and often conflicting results, and, more importantly, there are remarkably few studies that report actual growth or turnover rates of the soil microbial community at all. This is in contrast to the state of the knowledge on aquatic environments, where thousands of bacterial growth rate measurements can be compiled (Cole et al., 1988; Fouilland & Mostajir, 2010) and where several independent methods have been used to estimate growth rates, yielding corroborative results, including for example thymidine (Fuhrman & Azam, 1980; Moriarty & Pollard, 1982) and leucine incorporation (Kirchman et al., 1985), frequency of dividing cells (Hagström et al., 1979), and biomass increase after filtering out predators or diluting with bacteria-free water (Fuhrman & Azam, 1980). One reason for the different states of knowledge on microbial growth in aquatic and soil systems is, of course, that in soil, we have two major groups of decomposers, bacteria and fungi, where the latter group also comprises both saprotrophs and mycorrhizal fungi with different feeding strategies. Moreover, the soil matrix is a complex structure interfering with the measurements. For example, while aquatic microorganisms can be easily separated from their predators without seriously compromising the system using the dilution and filtration technique, there is no easy way of separating soil microorganisms from their predators without disturbing the system. Under normal soil conditions, the growth of the microorganisms will be balanced by predation or other types of cell death, thus resulting in a quasi-steady-state system, where biomass production will not always translate into a net increase in biomass, as indicated by the usually small variations in biomass over time (Joergensen et al., 2011). Another example of the difficulties in assessing microorganisms in a soil system is the use of the frequency of dividing cells approach. This approach appears to overestimate the bacterial growth rates in soil (Bloem et al., 1992), probably because in soil, the presence of surfaces will result in cells being close together long after a division, unlike the situation in water, where dividing cells will eventually split in two.
However, progress has been made by applying methods originally used for estimating the growth rates of bacteria and fungi in aquatic environments, namely the thymidine (TdR) and leucine (Leu) incorporation techniques to estimate bacterial growth and production and acetate incorporation into ergosterol (Ac-in-erg) to estimate fungal growth. In this review, we will compare estimates of bacterial and fungal growth and turnover of biomass with earlier estimates. We will also review how different environmental factors, including both abiotic and biotic ones, affect bacterial and fungal growth in soil. Finally, we will address some of the main problems in determining fungal and bacterial growth in soil and identify measures to resolve them, and discuss the future role of growth rate measurements in environmental microbiology.
Some of the first attempts to determine the turnover rates of soil microorganisms involved modeling using microbial biomass and annual substrate input from plants as the input parameters. This initially resulted in estimates of turnover rates of the soil microbial biomass varying on the order of days to years (Hunt, 1977; Jenkinson & Ladd, 1981; Chapman & Gray, 1986). With the inclusion of recycling and maintenance in the models, even larger variations in calculated turnover times of the soil microorganisms were obtained (Chapman & Gray, 1986). Bearing in mind that both maintenance and the extent of recycling, which depend on growth yields, are virtually unknown in soil, these calculations are not very precise. Furthermore, the estimates are dependent on input data on long time scales (annual), and thus, short-term effects cannot be addressed.
Another classical way of studying the growth rates in soil is by following increases in the biomass of fungi or bacteria, usually during a situation of rapid growth after adding a substrate, increasing the water content, or during recolonization by microorganisms of partially sterilized soil. For example, frequently repeated counting of bacterial cells in forest humus (summing periods of increasing biomass as growth) resulted in an estimated annual turnover of 12 times, that is, a bacterial turnover time of 30 days (Clarholm & Rosswall, 1980). A problem with the rationale for this approach is that bacteria are certainly also growing under conditions of no net biomass increase, when growth is balanced by predator consumption and death due to other causes. Both increasing (Clarholm, 1981; Christensen et al., 1992b, 1995) and decreasing (Christensen et al., 2007) bacterial growth has been shown to result in increased and decreased bacterial predator biomass, respectively, emphasizing that bacterial growth may not always be resolved from the measurement of net biomass change. Consequently, the estimates of growth rates using changes in biomass must be considered minimum values. We can easily study a biomass increase after adding a substrate (e.g. Nannipieri et al., 1983), but it is often unclear how this situation can be translated to normal soil conditions with low substrate availability. Furthermore, even after adding a substrate, predators will eventually start to grow, corrupting the information provided by a biomass increase regarding the growth rates during longer experiments.
The problems listed above can be largely circumvented using techniques based on the addition of tracer amounts of radioactively labeled precursors of macromolecules synthesized during growth and measurement of tracer incorporation over short time periods. During the brief incubation periods, the potentially altered growth conditions following the tracer additions do not affect the initial growth rates. The estimated macromolecular synthesis can then be used as a relative estimate of in situ microbial growth, because macromolecular synthesis will be approximately proportional to increased biomass. Alternatively, conversion factors can be used to calculate actual growth in terms of for example number of cells produced or the production of microbial carbon (C), or by dividing the concentration of biomass by the rate of biomass production to estimate the turnover time.
Bacterial growth rates measured using TdR and Leu incorporation
Initially, TdR and Leu incorporation in soil were estimated using protocols similar to those used for sediments in aquatic habitats. The precursor was added in trace amounts to a slurry, incubated briefly, after which the bacteria were killed. This was followed by the extraction of radioactively labeled macromolecules from the soil slurry using different acid/base extraction steps and finally quantification using liquid scintillation to measure the amount of precursor incorporated into the macromolecules (Christensen et al., 1989; Bååth, 1990; Michel & Bloem, 1993; Uhlířová & Šantrùčková, 2003; Amalfitano et al., 2008).
A problem with measuring bacterial growth in a soil slurry is the requirement for a high concentration of precursor during incubation, as well as a large and variable background noise derived from, for example, the precursor binding to soil particles. To overcome this, Bååth (1992) combined the TdR incorporation technique (and later the Leu incorporation technique, Bååth, 1994a, b) with a method to extract bacteria from soil using homogenization, followed by low-speed centrifugation (Bakken, 1985). The tracer is then added to the suspension of cells recovered from the soil. Although the homogenization/centrifugation technique introduced yet another step in the procedure, the addition of the precursor and subsequent washing steps were less laborious than using the slurry technique. Also, these steps could be performed exactly as in the application of Leu and TdR incorporation techniques in aquatic samples. In the modified method, tracer amounts of the precursor are added to the extracted bacterial suspension. After a short incubation period (usually between 0.5 and 4 h at room temperature), the bacteria are killed, followed by several washing steps using filtration to remove nonincorporated precursor molecules. Later, the filtration step was replaced by centrifugation using microcentrifuge vials (Smith & Azam, 1992; Bååth et al., 2001), which made the method even more rapid. By counting bacteria in the extracted bacterial suspension, bacterial turnover rates could also be estimated. Perhaps more importantly, this simplified technique made it possible to process hundreds of samples in a day, allowing for high-resolution studies of environmental effects, including temperature, moisture, pH, toxic substances, substrate additions, etc. It is important to consider that the homogenization/centrifugation step can affect the results by the selective recovery of some cell types; thus, the cells in the suspension may not fully represent the community present originally in the soil (Uhlířová & Šantrùčková, 2003). However, the effect on the measured growth rates appeared to be minor (Bååth, 1996), as indicated by the lack of difference between bacterial growth estimates from soil slurries and samples processed by the homogenization/centrifugation method.
Two critical prerequisites for the Leu and TdR incorporation methods to measure bacterial growth are specificity (incorporation only into bacteria and not fungi) and to what extent the growth rate in the slurry or the extracted bacterial suspension reflects the original conditions in the soil. The first will be especially problematic in the slurry system. However, it has been suggested that the low concentrations of the added tracer molecules select for organisms with a maximal surface area to volume ratio, which should benefit bacterial uptake (Bloem & Bolhuis, 2006). Specificity is less problematic when extracting bacteria by homogenization/centrifugation, because very few hyphae will be present in the bacterial suspension, and thus this step selects for bacteria. In agreement with these predictions, in soils manipulated to be bacterial and fungal free using soil sterilization and inoculation, Bååth (1990) showed that TdR incorporation only occurred in soils with bacteria, even when using the soil slurry technique. It has also been shown that antibacterial antibiotics decreased bacterial growth (Leu incorporation), but increased fungal growth (as Ac-in-erg incorporation) (Rousk et al., 2008), also suggesting that the Leu and TdR incorporation specifically reflects soil bacterial growth.
Both the slurry and the homogenization/centrifugation technique will alter the nutrient status and other conditions for the bacteria due to the addition of the potential substrates TdR and Leu and by releasing substrate from the soil matrix into the aqueous phase, with the potential to affect the growth rates. However, there are several lines of evidence that indicate that the altered growth conditions will not affect bacterial growth rates during several hours at normal laboratory temperatures. Both TdR and Leu incorporation rates were stable up to at least 1 h after preparing the slurry (Bååth, 1990) and over a period of 4 h for the homogenization/centrifugation method (Bååth, 1992, 1994b) at room temperature, with measurements made already 15 min after tracer addition. Later studies have shown that this incubation period with stable incorporation rates can be extended, especially at low temperatures. For example, at 5 °C, constant incorporation rates (linear increase in cumulative Leu incorporation) were found for a period of at least 24 h (Fig. 1). Also, addition of more substrates such as glucose does not alter the bacterial growth rate for a period of up to 10 h (Bååth, 1990; Iovieno & Bååth, 2008; Rousk et al., 2009a; S. Reischke, J. Rousk, & E. Bååth, unpublished data), showing that altered nutrient conditions do not result in immediate changes in bacterial growth rates. This is consistent with the substrate-induced-respiration response after adding glucose that remains at a stable level usually for up to 10 h or more, indicating no altered microbial growth during that time (Anderson & Domsch, 1985, as also discussed by Bååth, 1990).
Fungal growth rates measured using Ac-in-erg incorporation
The method is based on the incorporation of radioactively labeled acetate into the fungal-specific lipid ergosterol. The method was originally used on plant litter in aquatic habitats (Newell & Fallon, 1991), but was later adapted to soil (Pennanen et al., 1998; Bååth, 2001). Briefly, 14C-acetate is added to a soil slurry and incubated to allow the fungi to incorporate the substrate into newly synthesized ergosterol, where the rate of ergosterol produced will be proportional to fungal growth. Originally, an incubation time of 16 h was used (Bååth, 2001), but this has lately been reduced to 4–5 h at 22 °C (Bapiri et al., 2010; Rousk et al., 2010) and even down to only 2 h (S. Reischke, unpublished data). Ergosterol is then extracted and separated using HPLC using a UV detector and a fraction collector. The ergosterol peak is collected and the amount of acetate incorporated is determined using liquid scintillation.
In the same way as for the TdR and Leu incorporation technique, the lack of specificity and altered substrate concentrations during the incubation of the soil slurry might confound the results. The former problem is, however, less problematic because ergosterol is specific to fungi. Additionally, the lack of Ac-in-erg incorporation into fungal-free soil (Bååth, 2001), and the lack of negative effects of antibacterial antibiotics (Bååth, 2001), which can totally inhibit bacterial growth (Rousk et al., 2008), represent corroborative evidence for the specificity of the method. Bååth (2001) also found a constant incorporation rate up to 18 h after adding acetate to the soil slurry, and we have later verified this up to around 12 h (Fig. 2). This suggests that the present use of a 2–4-h incubation time would not alter the fungal growth rate to any large extent compared with the original growth before adding the acetate, and that the method can therefore be used to estimate in situ fungal growth. It should be noted, however, that although only fungi will incorporate Ac into ergosterol, making the method specific, bacteria are also able to use the labeled Ac, affecting the supply to fungi, which could potentially influence the result. A short incubation period should also minimize this problem, however.
Growth rates in soil
Most studies on bacterial growth rates using either the TdR or the Leu incorporation method suggest that the turnover times of the soil bacterial community are in the order of days to weeks at a temperature of around 20 °C (Table 1), irrespective of whether the slurry or the centrifugation/homogenization method is used. However, some earlier studies have suggested much longer turnover times using TdR incorporation, with values varying between 107 and 160 days at 25 °C (Harris & Paul, 1994). As discussed by the authors, differences in the methods used could be the reason for this large discrepancy. Longer turnover times were also found by Tibbles & Harris (1996) for soils from Antarctica. However, these soil samples were incubated at 10 °C. The incubation temperature will of course be of utmost importance in determining growth rates. Assuming a Q10 for bacterial growth between 10 and 20 °C of 2.5 (Rinnan et al., 2009), the estimates by Tibbles & Harris (1996) can be recalculated to estimate growth rates at 20 °C. This recalculation results in estimates of 4.4–86 days at 20 °C, which overlap with the range found in the other studies (Table 1).
Table 1. Estimated bacterial turnover times in soil
Turnover time (days)
Incubation temperature (°C)
TdR, thymidine incorporation; Leu, leucine incorporation; slurry, incubation using a soil slurry; susp., incubation in a bacterial suspension extracted from soil using homogenization/centrifugation; FDC, frequency of dividing cells; NA, not applicable.
Bacterial growth in the rhizosphere represents a special situation. The rhizosphere will be a place of rapid proliferation of bacteria compared with the surrounding soil due to the input of root exudates into the soil. This will result in an initial increase in microbial biomass, making it possible to estimate growth rates by repeated counting of bacterial cells. This has resulted in estimates of bacterial turnover times of 12–19 h (Olsson et al., 1987), with doubling times of 24 h on young roots increasing to over 100 h on old roots (Barber & Lynch, 1977), and sometimes with generation times as short as 7.5–9.1 h (Bowen & Rovira, 1976). Although some early attempts to use the TdR method to compare the growth rates in the rhizosphere and in the bulk soil could not detect differences (Christensen et al., 1992a; Christensen, 1993; Christensen & Christensen, 1994), several studies have later indicated that the bacterial growth rates in the rhizosphere are higher than bulk soil, with the actual growth rates similar to those estimated by repeated counting. Turnover times of 9–60 h for 6-day-old roots were found by Bååth & Johansson (1990) and around 2–3 days by Söderberg & Bååth (1998), and the growth rate measured in rhizosphere soil was consistently higher than that in bulk soil estimated by both the TdR and the Leu incorporation techniques (Christensen et al., 1995; Olsson et al., 1996a; Söderberg & Bååth, 1998, 2004; Söderberg et al., 2002).
Fewer attempts have been made to determine fungal growth rates in soil (Table 2), and all of these suggest longer turnover times of the fungal community compared with the bacterial community, with turnover times in the range of tens of days to several hundred days (Table 2). Because measurements are scarce in soil, we have also included estimates obtained from aquatic studies of fungi growing on leaf litter in our comparison (Table 2). Although this will be a situation of no moisture limitations and initially a surplus of fresh substrate, turnover rates comparable to those reported for soil fungi were found (Table 2). Thus, it appears that the turnover times of soil fungi are up to one order of magnitude greater than those of soil bacteria. This pattern is consistent with assessments in aquatic systems, with longer turnover times for the fungal compared with the bacterial community (Buesing & Gessner, 2006).
Table 2. Estimated fungal turnover times in soil and on decomposing leaf litter in water using the Ac-in-erg incorporation method
There are several problems and uncertainties in the calculation of growth rates using the incorporation of radioactive precursors. These include the extent of isotope dilution, the extent of incorporation into nontarget macromolecules, and the choice of conversion factors used to calculate the actual production of biomass from the amount of TdR incorporated into DNA and Leu incorporated into proteins for bacteria, and Ac incorporated into ergosterol for fungi. The extent of isotope dilution, that is, the extent to which the labeled precursor added is diluted by exogenous substrate already present in soil or de novo synthesized during incubation, can be determined using the isotope dilution approach (see Bååth (1998) for examples from soil), while the degree of nonspecific labeling of other macromolecules can be estimated using a combination of acid–base extraction steps (Riemann & Söndergaard, 1984). Isotope dilution was more varied for TdR than for Leu incorporation, but could be avoided by adding large amounts of TdR (Bååth, 1998).
Obtaining appropriate conversion factors to use for converting incorporation of TdR and Leu for soil bacteria and Ac-in-erg for fungi to the actual production of biomass is more uncertain. In aquatic habitats, conversion factors for bacterial growth can be estimated by comparing different methods used for estimating microbial growth. However, in soil, no such comparisons for bacteria have been made, although growth in an extracted bacterial suspension from soil over 24 h suggested that a similar conversion factor for TdR incorporation as commonly used in water, 2 × 1018 cells produced mol−1 TdR incorporated, is valid for soil (Bååth, 1992). Conversion factors for Ac-in-erg to fungal biomass C can been established in soil by comparing the uptake of radiotracer to the increase in ergosterol concentrations during periods of fungal biomass accumulation (Rousk & Bååth, 2007b). These conversion factors rely on subsequent conversion from the estimated ergosterol concentration to fungal biomass, which can only be determined easily in pure cultures grown under conditions rather different from those in the natural soil habitat for fungi, and it is likely that these conversion factors overestimate fungal biomass (as discussed in Rousk & Bååth, 2007b).
Another uncertainty is to what extent a single conversion factor can be used for soils under different environmental conditions, i.e. how does the conversion factor vary depending on the growth conditions of the microorganisms? In a recent study of coastal bacterioplankton (Franco-Vidal & Morán, 2011), the conversion factor for both Leu and TdR incorporation varied by a factor of around 10 over the course of the year, although the mean values were close to the theoretical ones: 3.1 kg C mol−1 Leu incorporated and 2 × 1018 cells mol−1 TdR incorporated. This suggests that although a rough estimate of growth rate can be obtained using a common conversion factor, more data are needed, not only on conversion factors for soil organism per se but also on the extent of variation in different soils and under different growth conditions. Furthermore, it may be necessary to establish specific conversion factors for each soil to enable correct estimates of absolute growth.
No less important than the calculation of growth and production of the microorganisms is actual biomass determination. The biomass concentration is used in the numerator of the calculation of turnover rates (biomass :microbial production). Estimating bacterial and fungal biomass in soil is not trivial, and different methods can yield very different results. This was extensively discussed by Rousk & Bååth (2007b) in connection with fungal growth rate determinations.
To avoid the biomass problem, one can calculate the actual biomass production from the incorporation data, although the problems pertaining to the conversion factor used will remain. Such estimates are scarce in soil, but using TdR incorporation, bacterial production between 30 and 60 g C m−2 year−1 was estimated in a forest humus, which can be compared with an estimated C input from plants of 200 g C m−2 year−1 (Bååth, 1990). Using Leu incorporation, a bacterial production of 7–14 × 10−5 g C h−1 g−1 soil C at room temperature was estimated in an agricultural soil (Bååth, 1994a), which was comparable to soil respiration rates. A lower fungal production rate of around 5–6 × 10−6 g C h−1 g−1 soil C was estimated for a garden soil (Rousk & Bååth, 2007b).
Environmental effects on fungal and bacterial growth in soil
The optimum temperature for growth usually is much higher than the mean annual temperature. For instance, in Antarctic soils, the Topt can be above 20 °C (Rinnan et al., 2009). At first glance, it may seem counter-intuitive that the temperature for the optimum growth rate for most of the year will be at a temperature far above the in situ temperature. However, it is the highest temperature during the year that is probably the most important in determining Topt. If the temperature is higher than Topt, even for a short period, this will result in the death of the community, being replaced with one having a higher Topt (Bárcenas-Moreno et al., 2009). This process, death of strains that cannot tolerate high temperatures and growth of more high-temperature-adapted strains, is rapid and will rapidly alter the microbial community's temperature relations. When the temperature decreases, this will not kill the new community, because low temperatures only slow growth. The recovery of the initial community with lower Topt will consequently be a slow process, especially at very low temperatures when the growth and turnover of the soil community, even of better adapted strains, is very slow (Ranneklev & Bååth, 2001).
That Tmin for bacterial growth, around −4 to −8 °C in temperate soils, is substantially below 0 °C may seem unrealistic, because there should be no growth in frozen soil where there will not be any free water. Thus, Tmin is often stated as an apparent minimum temperature for growth. However, for normal soil bacteria, with a Topt around 25–30 °C in a pure culture, estimates of Tmin are also lower than 0 °C (Rosso et al., 1993). Thus, the growth of bacteria in soil will be possible even below 0 °C if free water is present. Indeed, microbial biomass synthesis in soil at −4 °C was demonstrated recently (Harrysson Drotz et al., 2010).
Pietikäinen et al. (2005) reported that fungal growth was less negatively affected by low temperatures compared with bacteria, as indicated by a lower Tmin for fungal growth than for bacteria. However, subsequent studies (Bárcenas-Moreno et al., 2009) did not find this difference. A more systematic comparison between the temperature sensitivities of fungal and bacterial growth still remains before general conclusions can be established.
Soil moisture and drying/rewetting
Moisture is also a very important regulating factor for microbial activity in soil. The influence of soil moisture on the overall microbial activity in soil, as indicated by C mineralization, is well understood. In accordance with C mineralization, bacterial growth also increases with a higher moisture content of the soil (Iovieno & Bååth, 2008), a relationship that has also been corroborated for bacterial communities in drying Mediterranean river beds (Amalfitano et al., 2008). The relationship between fungal growth and soil moisture has not been studied directly, although it is reasonable to expect that low moisture inhibits and higher moisture is also conducive for fungal growth in soil. To what extent soil moisture affects the balance of fungal and bacterial growth remains to be investigated systematically. A special situation with regard to moisture is the sudden increase in the water content during a rewetting event. Although the effects of these perturbations on microbial biomass, biomass composition, and fluxes of nutrients and greenhouse gases have been rather extensively studied (see discussion in Bapiri et al., 2010), the effects on direct measurements of microbial growth are more scarce. Following the addition of water to a dry soil, there is an abrupt (<1 h) increase in respiration to a rate that considerably supersedes the basal respiration rate of the moist soil. The pulse quickly passes and the respiration rate asymptotically converges with that of a continually moist soil within days. Bacterial growth does not follow this pattern (Iovieno & Bååth, 2008). Instead, the bacterial growth rate initially is very low, but an immediate and steady increase results in a recovery of bacterial growth that converges with that of the moist control soil after around 10 h. This indicates that the bacterial growth response cannot be inferred from respiration measurements. Furthermore, the bacterial response is very different from that after adding a substrate like glucose. A glucose addition initially does not alter the growth rate, but after a lag phase, results in an exponential increase in the growth rate. Thus, although altering the nutrient conditions during incubation with a radioactive precursor in a slurry or a bacterial suspension does not appear to alter the initial growth rate for several hours, precaution should be exercised when working with soil that is much drier than the optimum moisture conditions. In such situations, the duration of the incubation times should be minimized to avoid the potentially confounding influence of the altered growth conditions used in the method. For example, Iovieno & Bååth (2008) used 15 min as their shortest incubation time.
Bapiri et al. (2010) compared bacterial and fungal growth responses, using Leu and Ac-in-erg incorporation, following repeated drying/rewetting events of soil in a microcosm system. Drying–rewetting decreased bacterial growth while fungal growth remained unaffected, resulting in an elevated fungal : bacterial growth ratio. This effect was found irrespective of the initial fungal : bacterial biomass ratio (three soils of various fungal : bacterial ratios were investigated). Many drying–rewetting cycles did not, however, affect the fungal : bacterial growth ratio compared with a few cycles. In contrast, Williams (2007) found that microbial growth, determined as 13C-glucose incorporation into phospholipid fatty acid (PLFA) markers, was unaffected by the drying–rewetting treatment, suggesting no difference between the fungal and the bacterial growth responses. Williams (2007) measured microbial growth 0–6 and 0–24 h following rewetting (with identical results), while Bapiri et al. (2010) measured the microbial growth responses 3 days after rewetting. It remains to be seen whether the discrepancy between the results from Bapiri et al. (2010) and Williams (2007) was due to the different techniques used or due to the differences in the timing of measurements after rewetting.
An association between fungal dominance in acid soils, and bacterial dominance in neutral or slightly alkaline soils, is well known (Brady & Weil, 2008). However, clear evidence for the differential growth of fungi and bacteria in soils with different pH has been lacking until recently. Bååth & Arnebrant (1995) studied the effects of lime- and ash-treated forest soils on bacterial growth. The treatments resulted in a range of pH values from about pH 4 to 7, which resulted in an approximately fivefold increase in the bacterial growth rate, as measured by TdR incorporation. Higher TdR incorporation in ash-treated soils was also found by Fritze et al. (2000) and Perkiömöki & Fritze (2002). Similarly, in a study including 19 different soils under various land uses, spanning pH 4–8, there was a positive correlation between bacterial growth and higher pH as measured with Leu incorporation (Bååth, 1998). Bacterial growth increased fourfold between pH 4 and 8. The influence of soil pH on both fungal and bacterial growth in soils with a more restricted pH range (pH 3.6–4.1) in forest humus (Pennanen et al., 1998) and the acute effects of artificially increasing pH using 13C-Ac incorporation into PLFAs (Arao, 1999) have also been determined. Both studies found increased bacterial and lower fungal growth at higher pH, and vice versa. However, in all these cases, other factors also covaried with pH, possibly confounding the pH effect, making causality difficult to assign.
A study where the confounding influences by factors other than pH were minimized was conducted on a century-old continuous soil pH gradient: Hoosfield acid strip, Rothamsted Research, UK (Rousk et al., 2009b). This experiment provides a uniform pH gradient, ranging from pH 8.3 to 4.0, within 180 m in a silty-loam soil on which barley has been continuously grown for more than 100 years. Growth-based measurements revealed a fivefold decrease in bacterial growth, and a fivefold increase in fungal growth, with a lower pH. This resulted in a c. 30-fold increase in the fungal : bacterial growth ratio, from pH 8.3 to pH 4.5. The different relationships with soil pH for fungi and bacteria in this soil, when variations in other factors were minimized, have since been verified (Rousk et al., 2010). Using soils from the 150-year-old Park Grass experimental grassland experiment at Rothamsted Research, Rousk et al. (2011) also found that bacterial growth decreased with lower pH, while fungal growth increased, resulting in a 50-fold increase in the fungal : bacterial growth ratio between pH 7.5 and 3.3. A similar pattern was also found in a set of Iberian Vineyard soils, varying between pH 4 and 7 (D. Fernández-Calviño & E. Bååth, unpublished data). We conclude that 30–50-fold increases in the fungal : bacterial growth ratio can be expected between soils of neutral pH and soils of pH 4–5.
It is commonly thought that fungi are more important for the degradation of complex substrates such as wood, composed mainly of macromolecules like cellulose and lignin, while bacteria are considered more competitive in degrading easily available substrates. Specialized fungal groups, such as white- and brown-rot fungi, are mainly involved in the former. A study on the bacterial and fungal growth responses to pure substrate additions corroborated this. Meidute et al. (2008) found that bacterial growth responded relatively more than fungal growth following the addition of simple C compounds (glucose and gelatin), while there was a relatively higher fungal response following the addition of a more complex C substrate (cellulose) to soil.
More complex substrates from plant material have also been shown to have a differential effect on fungal and bacterial growth. Adding alfalfa (C/N=15) or straw (C/N=75) to soil in a time-series laboratory microcosm experiment resulted in the alfalfa addition increasing bacterial growth more than fungal growth, and the reverse was observed following straw amendment (Rousk & Bååth, 2007a). A possible explanation could be that bacteria benefitted from conditions of higher N availability compared with fungi and that the results were simply due to the stoichiometric difference in the C : N ratio between the plant materials. This was, however, not the case. When mineral N was added to compensate for the difference between the C : N ratios, the same differences in growth responses from the two plant material were still obtained. Moreover, bacterial growth was unaffected by extra mineral N addition, while fungal growth increased between days 4 and 7 when mineral N was added to the straw. This suggests that the difference between the plant materials was derived from the quality of the C compounds, rather than the elemental composition. Similar results were found by Meidute et al. (2008), where adding extra N enhanced fungal growth on cellulose even more, while adding N with glucose resulted in increased bacterial growth.
It is not only the quality and type of substrate that affects the balance of bacterial and fungal growth in soil. The concentration of substrate is also of importance. Griffiths et al. (1999) added different concentrations of a mixture of easily available carbon substrate (sugars, amino acids, organic acids) and found increasing fungal compared with bacterial growth with increasing loading rates of substrate (measured as increase in PLFAs indicative of the two groups). This was also found with the addition of different concentrations of glucose (S. Reischke, J. Rousk, & E. Bååth, unpublished data) using Leu and Ac-in-erg incorporation. At low concentrations (<1 mg glucose-C), only bacterial growth increased significantly, while at high concentrations (>4 mg glucose-C), increased fungal growth was also detected. At very high concentrations (>16 mg glucose-C), bacterial growth was inhibited, while fungi proliferated extensively.
The increased substrate flow in the rhizosphere due to root exudation has been shown to increase bacterial growth rates significantly (see references above). However, it has also been shown that there is a plant species effect on bacterial growth in the rhizosphere (Söderberg et al., 2002), presumably due to different exudation rates. Even different genotypes of the same plant can result in different bacterial growth rates (Velasco et al., 2009; Aira et al., 2010).
Measurement of growth rates is ideal for studying the interactions between fungi and bacteria not only because the methods are sensitive to subtle changes but also because it is equally easy to rapidly detect both increases and decreases in growth rates, which is not the case when using biomass measurements to infer interactions. There have been indications of a pronounced interaction between fungi and bacteria from soil systems. For instance, studies by Thiele-Bruhn & Beck (2005) and Feeney et al. (2006) obtained results suggesting bacterial inhibition of fungal growth (without reciprocity), while Meidute et al. (2008) found indications of synergistic interactions, where the initial fungal degradation of cellulose seemed to promote bacterial growth. Observations from studies of fungal and bacterial biomass production during the decomposition of litter in soil and water also indicated pronounced interactions between fungi and bacteria, including reports of reciprocal negative interactions (Mille-Lindblom & Tranvik, 2003), antagonistic interactions (Møller et al., 1999; Mille-Lindblom et al., 2006), and synergistic interactions (Bengtsson, 1992; Romani et al., 2006).
Using an approach different from the addition-type experiments previously conducted on litter (above), Rousk et al. (2008) utilized a removal-type experiment to investigate the interaction between bacteria and fungi. Using a range of soils with various fungal : bacterial growth ratios, the application of three different bacterial inhibitors resulted in bacterial growth being reduced to virtually zero. In response to the bacterial growth inhibition, fungal growth increased. Consequently, the study demonstrated that bacteria exerted an important competitive pressure on fungi, and that competition was alleviated when the influence by bacteria was terminated. Negative interactions between fungal and bacterial growth responses have been suggested repeatedly, for instance in response to substrate (Rousk & Bååth, 2007a), heavy metal (Rajapaksha et al., 2004) and salt addition (Tobor-Kapłon et al., 2005), and following land-use change (Lopez-Sangil et al., 2011) or during the recolonization after a fire event (Bárcenas-Moreno et al., 2011).
One of the clearest examples of a reproducible and very pronounced negative correlation between bacterial and fungal growth is along pH gradients (Rousk et al., 2009b), suggesting that the interaction between the groups could be influential in shaping pH relationships of the decomposer groups. The interaction between fungi and bacteria was further studied in an arable soil pH gradient (Rousk et al., 2010). In the absence of a bacterial inhibitor, the pH relationships for bacterial and fungal growth corroborated those documented previously (Rousk et al., 2009b). When a bacterial inhibitor was added, the bacterial growth was suppressed across the entire pH gradient, and the strong pH dependence of fungal growth, starting low at a high pH and increasing toward a lower pH, was offset; fungal production was high across the entire pH gradient. This result indicated that the pH dependence of fungal growth was, at least partly, mediated by competitive interaction with bacteria.
In the light of modern molecular method, which presently allows us to investigate the vast microbial diversity in soil at a high resolution, the use of methods that integrate measurements across large groups like bacteria and fungi might be considered obsolete. The use of stable isotope probing both in connection with PLFA (Boschker & Middelburg, 2002; Bengtson et al., 2009) and DNA (Dumont & Murrell, 2005; Buckley et al., 2007) may also be an alternative way of estimating microbial growth rates in soil, at a higher taxonomic resolution. However, the ease, accuracy, and speed with which bacterial and fungal growth can be estimated using the incorporation of radiolabeled tracers means that they will continue to be invaluable in soil ecological studies, both as a reference to compare with other methods and as a way of comparing environmental effects on these broadly defined important decomposer groups. Keeping in mind that bacteria and fungi drive different food webs in soil, differentiating between the growth of these two groups will also be important for understanding the ecosystems in which they exist. Last, but not least, the use of the same methods in both soil and aquatic habitats enables proper cross-ecosystem comparisons.
There are still several points of uncertainty in using TdR and Leu incorporation and Ac-in-erg incorporation to estimate bacterial and fungal growth in soil, respectively, that need to be addressed in the future. Several of these have been touched upon in this review. For example, conversion factors to calculate bacterial and fungal growth from incorporation data need to be better constrained. This will not only enable the estimation of actual production of bacteria and fungi in soil – which will ultimately be transferred into higher trophic levels – but, together with data on substrate loss or respiration, will allow us to determine the growth efficiencies of soil microorganisms. We also need to develop methods allowing for a higher resolution in time (shorter incubation periods) and space (allowing for measurements on milligrams of soil). The taxonomic resolution of the tracer incorporation methods should also be increased in the future. We need, for example, to be able to differentiate between the growth of saprotrophic and mycorrhizal fungi in soil. However, there are possibilities for even higher taxonomic resolution, for example by combining immunomagnetic separation of specific bacteria with the Leu incorporation method (Sengeløv et al., 2000). Finally, the growth rates of archaea in soil still remain to be determined and related to those of bacteria and fungi. It was shown that both archaea and bacteria incorporate Leu in sea water (Herndl et al., 2005). The combination of Leu incorporation with specific bacterial inhibitors has also suggested that the contribution by archaea to the overall heterotrophic activity was low in marine systems (Ionescu et al., 2009). Similar studies in soil are, however, lacking.
This review was written with financial aid from the Swedish Research Council to E.B. (Project No. 2009-4503) and J.R. (Project No. 623-2009-7343).