Bacterioplankton Growth and Nutrient Use Efficiencies Under Variable Organic Carbon and Inorganic Phosphorus Ratios
- First Online:
- Cite this article as:
- Jansson, M., Bergström, A., Lymer, D. et al. Microb Ecol (2006) 52: 358. doi:10.1007/s00248-006-9013-4
- 199 Views
We carried out enclosure experiments in an unproductive lake in northern Sweden and studied the effects of enrichment with different dissolved organic carbon (glucose)/inorganic phosphorous (DOC/Pi) ratios on bacterioplankton production (BP), growth efficiency (BGE), nutrient use efficiency (BNUE), growth rate, and specific respiration. We found considerable variation in BP, BGE, and BNUE along the tested DOC/Pi gradient. BGE varied between 0.87 and 0.24, with the highest values at low DOC/Pi ratios. BNUE varied between 40 and 9 g C g P−1, with high values at high DOC/Pi ratios. More DOC was thus allocated to growth when bacteria tended to be C-limited, and to respiration when bacteria were P-limited. Specific respiration was positively correlated with bacterial growth rate throughout the gradient. It is therefore possible that respiration was used to support growth in P-limited bacteria. The results indicated that BP can be limited by Pi when BNUE is at its maximum, by organic C when BGE is at its maximum, and by dual organic C and Pi limitation when BNUE and BGE have suboptimal values.
Heterotrophic bacteria use organic carbon as carbon (C) and energy source . Bacterial production (BP) in aquatic systems is dependent on the concentration of available organic C and to what extent this organic C can be used for BP and for respiration (BR). The use of assimilated organic C for BP is characterized by bacterial growth efficiency (BGE), which is the share of the total bacterial assimilation of organic C used for BP (BGE = BP/BP + BR). An often used mean value for lakes is 0.26 ; however, BGE varies enormously in aquatic systems, and values between 0.01 and 0.8 have been reported . In addition, BP is also regulated by inorganic nutrient limitation, and inorganic phosphorus (Pi) is the most frequently reported limiting nutrient for BP . The P content of dissolved organic matter has also been shown to regulate bacterial use of DOC for growth . BP is thus controlled by at least three important factors: (1) availability of organic C for bacterial assimilation, (2) BGE, and (3) availability of Pi. In general, BGE is positively correlated to P concentrations in lakes [1, 3, 22], and inorganic nutrients availability was reported to be a key factor for BGE in the review by del Giorgio and Cole . Pi limitation of BP, therefore, can be the result of the fact that Pi controls BGE.
Experiments show variable responses of BP to enrichment of lake water with dissolved organic carbon (DOC) or Pi. DOC additions only have been demonstrated to stimulate BP in productive lakes [27, 30]. Stimulation by Pi addition only is the most reported response found in a large variety of lakes ( and references therein). In addition, although seldom tested for, dual limitation (i.e., stimulation of BP by addition of either DOC or Pi) has been shown in moderately productive lakes . Single DOC or Pi limitation can be understood if C and P function as classical limiting factors of biomass production (sensu Liebig). However, this conservative view of alternative limiting factors cannot account for dual limitation of BP that must be assumed to be a result of interactive forces of organic C and inorganic nutrients on bacterial metabolism. For further discussion on bacterial resource exploitation and the regulation of BP, a second efficiency measure—bacterial nutrient use efficiency (BNUE)—can be used as an analogue to BGE. BGE denotes the BP in relation to assimilated C, whereas BNUE denotes BP per unit of assimilated inorganic nutrients [26, 23, 12].
Equation (3) means that if, e.g., uptake of C increases and uptake of Pi is constant, the BNUE/BGE ratio will increase, either as a consequence of an increase in BNUE or/and a decrease in BGE. Because uptake of C and Pi depends on their availability, we hypothesize a negative relationship between BNUE and BGE along gradients with different DOC/Pi ratio of the bacterial growth medium. As BGE decreases when the DOC/Pi ratio increases [1, 22], it can be assumed that BNUE increases when the DOC/Pi ratio increases. Consequently, low BGE and high BNUE can be characteristics of Pi limitation (high DOC/Pi ratio), whereas C limitation of BP can occur simultaneously with high BGE and low BNUE. BGE and BNUE variation could therefore offer a useful tool in studies of the combined effects of DOC and inorganic nutrients on bacterial growth.
Here we assess the variation of BP, BGE, and BNUE in enclosure experiments, in which water in an unproductive lake was enriched with inorganic nutrients (P and N) and DOC (glucose). The N/P supply ratio was high to ensure that Pi was the limiting inorganic nutrient. We measured BP, and calculated and compared BGE and BNUE at different DOC/Pi stoichiometry levels in the enclosures. We use the results to test the hypothesis that the relative availability of DOC and Pi (DOC/Pi ratio) determines BP by its influences on BGE and BNUE. We also use our experimental results to discuss how BP can be regulated by limiting organic C and inorganic nutrient resources.
An experiment was conducted in a subarctic lake near Abisko in northern Sweden (Lake Diktar Erik, 68°27′N, 18°36′E) during 9 days (June 24–July 3) in 2003. Eight open enclosures (300-L plastic bags) were filled with lake water and anchored in the lake. Analysis of the lake water at the start of the experiment revealed nutrient-poor conditions [DOC: 4.2 mg C L−1, total nitrogen (Tot-N): 270 μg N L−1, total phosphorous (Tot-P): 4 μg P L−1]. The midday water temperature in the enclosures varied between 16.6 and 18.7°C during the course of the experiment. Organic C in the form of glucose was added daily to six of the enclosures forming a gradient (12.5, 25, 50, 100, 200, and 400 μg of added C L−1 day−1). All the bags received daily additions of N (30 μg N L−1 day−1 as NH4NO3) and P (2 μg P L−1 day−1 as Na3PO4). One bag (Ctrl) served as a control, and one bag (T1) received only N and P to control for the effect of nutrients on phytoplankton primary production (PP) and bacterioplankton production (BP).
The growth rate (μ) of bacteria was calculated as ln(BPglucose + BB)/BB for each sampling occasion and enclosure. We used the mean of these values for comparison with other variables. Specific bacterial respiration (spec. R) was expressed as the estimated BR on glucose-C for the whole experimental period divided by the mean bacterial biomass during the period.
Enclosure experiment in Lake Diktar Erik conducted during 9 days in June and July 2003
C addition (mg C L−1)
BP (μg C L−1 day−1)
PP (μg C L−1 day−1)
BB (μg C L−1)
DOC (mg L−1)
SRP (μg L−1)
DIN (μg L−1)
BNUE (g C gP−1)
Spec. R (h−1)
We introduced two sources of uncertainty with this calculation. First, our calculation of Pi uptake by bacterioplankton, and consequently the BNUE, was dependent on the phytoplankton C/P ratio used to calculate Pi uptake in phytoplankton. However, the phytoplankton production was low (Table 1), and even when we tested the end members of the C/P ratios (0.6–12 μg P C−1) for phytoplankton reported by Vadstein , we only obtained small changes in BNUE (Fig. 4). Second, we assumed that the share of the total Pi uptake by bacterioplankton that was used for BPglucose in each treatment was the same as the share of BPglucose in relation to BP on other substrates (i.e., the BP difference between Ctrl and T1). Because BPglucose was dominant in most enclosures, this potential error source was also small when tested assuming the possibility that all bacterial uptake of Pi was used for BPglucose. The sum of these potential errors is shown in Fig. 4. We consider that the uncertainties introduced by our calculations caused small errors in our BNUE values, which did not interfere with our interpretations of the results.
Our estimates of bacterial C and Pi uptake assume that insignificant amounts of C and Pi were lost to biofilms on the walls of the bags. Previous experience from this type of short-term enclosure experiments with addition of small amounts of nutrients shows that losses of added nutrients to walls are minimal . We also examined strips cut out from the bags after termination of the experiment, and found that the Chl-a content on the walls was less than 4% of that in the water. Wall effects should thus have had only minor influences on our uptake estimates.
Results and Discussion
Growth and Respiration in Relation to DOC and Pi Availability
Abiotic characteristics, PP, BP, BB, BGE, BNUE, growth rates, and specific respiration rates are given in Table 1. It shall be stressed that the bacterial responses in our experiment did not reflect the response of bacteria to the abiotic conditions as in a pure bacterial culture, because of competition with phytoplankton for inorganic nutrients, and because elimination of bacteria was significant in the enclosures. Grazing on bacteria by heterotrophic and mixotrophic flagellates corresponded to between 25 and 35% of BP over the experimental period (Bergström, unpublished data). Cladoceran grazing on bacteria was responsible for the elimination of ca. 5% of BP (Karlsson, unpublished data), and viral lysis of cells was estimated to correspond to between 30 and 50% of BP (Lymer, unpublished data). Total elimination of bacteria by these processes was thus of the same magnitude as bacterial production, and explains the minor increase in bacterial biomass in all treatments. Bacterial production in the enclosures, therefore, represented the actual bacterial exploitation of added DOC and inorganic nutrients as determined by the prerequisites of the experiment. Because the only variable that differed between the enclosures at the start of the experiment was the DOC dose, we can nevertheless attribute differences in bacterial growth and respiration between the enclosures to differences in DOC dose and DOC/Pi supply ratios.
That low BGE was not synonymous with low growth was indicated by comparing specific growth rate (μ), specific respiration, and BGE and BNUE. Growth rate was positively correlated with BNUE (μ = 0.071 + 0.083 log BNUE, r2 = 0.94, p = 0.001) and negatively correlated with BGE (μ = 0.074–0.00068 BGE, r2 = 0.92, p = 0.003) over the studied gradient in DOC/Pi supply. Consequently, μ was also positively correlated to specific respiration (μ = 0.021 + 0.22 spec. R, r2 = 0.93, p = 0.002). The positive correlation between μ and BNUE may seem confusing. BNUE should reflect the C/P quotient of the bacteria that has been shown to be negatively correlated to μ for P-limited bacteria , in accordance with the “growth rate hypothesis” . However, it appears as if the growth rate hypothesis is not applicable when bacteria are not P-limited. Thus, Makino and Cotner  showed that bacterial C/P ratios increased with μ in P-saturated bacteria and explained this observation with use of stored polyphosphates for growth. In our experiment, there was a gradient in the degree of Pi limitation with no, or very low, limitation in T2 to Pi limitation in T7 (see below). It is very probable that there was a variable degree in “luxury uptake” of Pi that could be stored as polyphosphates [4, 14, 16] in the enclosures. It is also probable that when the DOC supply was increased in relation to Pi, the result was an increased growth that reduced the buildup of polyphosphates. In that situation, a positive relation between μ and BNUE is logical. The negative relation between μ and BGE may also seem confusing. However, it is implicit in this relation that μ and the specific respiration were positively correlated, i.e., that high growth rates were connected with high respiration rates.
The response of bacteria in our DOC/Pi gradient experiment was in agreement with various previous reports on bacterial growth and respiration, including dependence of BGE on P supply [6, 22], high specific respiration in P-limited cultures , positive relationship between μ and specific respiration in nutrient- and carbon-limited bacteria .
Our results stressed the impact of the relative distribution of available DOC and Pi sources on BGE, which has often been difficult to separate from the effects of organic substrates in aquatic bacterioplankton assemblages . Both BGE and the comparison of μ and specific respiration showed that a greater share of DOC was used for growth at C limitation, whereas DOC was used more for respiration at Pi limitation—i.e., C-limited bacteria tended to maximize DOC use for growth, whereas Pi-limited bacteria tended to respire a large portion of assimilated C.
The positive relationship between μ and specific respiration along the DOC/Pi gradient indicates the possibility that high respiration rates of “excess C” was partly used to support growth and not only for maintenance [18, 24]. If so, high respiration may be important for growth, especially when growth is restricted by Pi limitation.
Regulation of BP by DOC and Pi
Our results can be used to assess the question of how BP is regulated by resource limitation. It should be stressed that the discussion here concerns principles of DOC and Pi regulation of BP. We do not consider that the values of BGE and BNUE obtained in our experiments with glucose and Pi enrichment can be used for definition of the divides between DOC and Pi limitation in natural systems. The critical in situ values of, e.g., BGE and BNUE are probably quite different when bacteria grow on less available organic substrates and forms of P.
BNUE and BGE variations in Fig. 3 show that, for a given amount of Pi, we can expect a stimulation of BP by increasing organic C availability (increasing the DOC/Pi ratio) until the maximum BNUE is reached. In our experiments, the maximum BNUE was probably not reached; however, Fig. 3 indicates that it should be approximately 40–50 g C g P−1 and be obtained at DOC/Pi ratios > 200. Furthermore, Fig. 3 indicates that it should have been possible to increase BP by increasing the Pi supply in relation to DOC over most of the studied DOC/Pi gradient. Stimulation of BP by Pi is thus possible as long as the bacteria can respond with increased BGE to increased Pi availability in relation to DOC (in our experiment in all treatments, with the probable exception of T2 and T3). The difference in BGE between T2 and T3 (Table 1) was very small, indicating that the bacteria were P-saturated and limited mainly by C only. Therefore, the relationships between BGE, BNUE, and DOC/Pi supply ratios indicated that the bacteria were P-saturated and probably limited by C only at DOC/Pi ratios below ca. 10 and by P only at DOC/Pi ratios > ca. 200.
BGE and BNUE variations also offered a clue to the phenomenon of dual limitation. We use this expression for situations when BP can be stimulated by increased supply of either of two resources. In our study, it should in principle be possible to increase BP over most of the studied gradient by increasing either the Pi or the DOC supply. Thus, Pi should increase BP as long as BGE can be increased, and DOC should increase BP as long as BNUE can be increased. Dual limitation should, however, be particularly pronounced when both BGE and BNUE are clearly below suboptimal values that correspond to “intermediate” DOC/Pi ratios.
We suggest that dual limitation of BP is related to the fact that DOC serves as a source of both C and energy, that the allocation of C to these processes is dependent on P availability, and that bacteria can take up and store P, which is not immediately used for growth. In a situation corresponding to “intermediate” DOC/Pi ratios in our experiment, it is possible to increase the use of assimilated C for growth by increasing the P supply (increased BGE). In the same situation, the bacteria take up and keep an excess of P not used for growth, and an increase in the DOC supply can link assimilated P to increased growth instead of to storage. The latter scenario presumes that, as in our experiment, the positive effect on BP of increasing BNUE is greater than the negative effect of decreasing BGE.
Our model for assessment of DOC and Pi limitation is largely in line with reported responses of lake bacterioplankton to Pi and DOC enrichment. P limitation, which—according to our model—should be expected when BGE is suboptimal, has been demonstrated in a wide variety of lakes . Studies comparing the effects of Pi and organic C enrichment have shown that BP in oligotrophic lakes and ponds in the Arctic  and in Canada  were stimulated by Pi, but not by organic C. These systems have extremely low inorganic nutrients availability and may very well meet our criteria of single P limitation, namely, that BNUE is at maximum due to high DOC/Pi ratios. Dual limitation should, according to our model, be particularly pronounced at intermediate DOC/Pi ratios when both BGE and BNUE are clearly below their potential maxima. It is therefore logical that dual limitation, when tested for and reported, has been found in moderately eutrophic and humic lakes . Finally, single organic C limitation in lakes should be found only when Pi supply is so high in relation to DOC that BGE is at its potential maximum. In agreement with our model, single C limitation has been reported for eutrophic lakes during autumn circulation when surface waters were enriched with large amounts of inorganic nutrients from deeper water layers. Single C limitation in a eutrophic lake was also shown by Wang et al. .
In summary, we suggest that: (1) BP can be limited by Pi (solely limitation) at high DOC/Pi ratios, which permit a maximum BNUE under ambient conditions; (2) organic C (solely limitation) can limit BP at low DOC/Pi ratios, which permit maximum BGE under ambient conditions; and (3) BP can be dual limited by organic C and Pi at intermediate ratios where both BGE and BNUE are at suboptimal values. In this model, limitation by P is mainly carried out through repression of BGE, and limitation of C is mainly by repression of BNUE. Pi limitation should be most frequent in oligotrophic lakes, organic C limitation in eutrophic lakes, and dual C and Pi limitation should occur in a wide variety of eutrophic–oligotrophic lakes.
This study was financially supported by the Swedish Research Council. We thank Ulf Westerlund for excellent assistance in the field and in the laboratory.