Introduction

Tissue cultures of cells taken from higher, differentiated organisms normally need complex growth factors for suc­cessful cell division. These factors are nowadays usually referred to as cytokines. Their role is generally under­stood to involve binding at the cell membrane and the pro­duction of second messengers such as cGMP. These serve to activate various signaling pathways and segments of primary metabolism, which may of course include those responsible for their own synthesis (Alberts et al. 1989). By contrast, it is usually assumed in prokaryotic microbi­ology that each bacterial cell in an axenic culture can mul­tiply independently of other bacteria, provided that appro­priate concentrations of substrates, vitamins and trace ele­ments are present in the culture medium. Current labora­tory experience seems to be consistent with this in that the development of bacterial colonies from single cells on agar plates is commonplace, and the most-probable-num­ber method is based on the apparently correct assumption that a test tube containing but one viable cell will in due time display visible growth or turbidity. While it is be­coming clear that axenic bacterial cultures do not re­motely represent a statistically homogeneous population (Koch 1987; Kell et al. 1991; Davey and Kell 1996), these observations are most easily interpreted as being in favour of “autonomous” growth.

However, an increasing body of evidence has high­lighted the widespread importance of chemically medi­ated intercellular communication between bacteria in cul­ture for specific events such as sporulation, conjugation, virulence and bioluminescence. Thus, it is now clear that a variety of different chemical compounds (pheromones) (Stephens 1986) produced as secondary metabolites are responsible for such social behaviour of prokaryotes as

exhibited under conditions of obvious cellular differentia­tion [for review, see Kaiser and Losick (1993), Swift et al. (1994, 1996), Kell et al. (1995), Greenberg et al. (1996) Dunny and Leonard (1997), Kleerebezem et al. (1997)]. The question then arises as to whether similar types of signaling may be of general significance for cell multipli­cation in growing bacterial cultures (Kaprelyants and Kell 1996).

Recently we have found that viable cells of Micrococ­cus luteus secrete a factor that promotes the resuscitation of dormant, nongrowing cells to yield normal, colony­forming bacteria (Kaprelyants and Kell 1993; Kaprelyants et al. 1994; Votyakova et al. 1994). The resuscitation-pro­moting factor (Rpf) is a protein that has been purified to homogeneity (Mukamolova et al. 1998). Analysis of the nucleotide sequence of the M. luteus gene encoding Rpf suggests that the form initially secreted is an approxi­mately 19-kDa protein. Five similar genes that can encode apparently secreted proteins have been uncovered in Myco­bacterium tuberculosis by genome-sequencing projects (Cole et al. 1998), and two are currently known in Myco­bacterium leprae. Rpf has been expressed in, and purified to homogeneity from, Escherichia coli. Picomolar con­centrations of recombinant Rpf increase the viable cell count of dormant M. luteus cultures at least 100-fold and can also stimulate the growth of viable cells. Rpf also stimulates the growth of several other high G+C, gram­positive organisms including Mycobacterium avium, My­cobacterium bovis (BCG), Mycobacterium kansasii, My­cobacterium smegmatis and M. tuberculosis. Thus, Rpf is in fact the first example of a bacterial growth factor or cy­tokine (Mukamolova et al. 1998).

In the present study we describe the conditions under which the dependence of cell multiplication on the secre­tion of bacterial growth factor(s) may be observed in some bacteria.

Materials and methods Organisms and media

M. luteus NCIMB 13267 (previously described as “Fleming strain 2665”) was grown aerobically at 30°C in shake flasks in lactate minimal medium (LMM) containing 0.5% L-lactate as described previously (Kaprelyants and Kell 1992, 1993). For some experi­ments, lactate in LMM was replaced by 1% succinate (succinate medium; SMM).

M. smegmatis (“fast” strain; All-Russia State Institute for Con­trol of Veterinary Preparations, Moscow, Russia) was grown in nu­trient broth E or in Sauton medium (Connell 1994).

M. tuberculosis (avirulent, “Academia” strain) was obtained from the Physiopulmonology Center (Moscow, Rusia). Bacteria were maintained on Lowenstein-Jensen agar slopes at 37°C and were grown at 37°C in liquid Sauton medium supplemented with albumin, glucose, NaCl, 0.6% (w/v) glycerol and 0.05% Tween 80.

M. luteus spent medium preparation

Supernatant was obtained after the centrifugation of late-expo­nential-phase M. luteus cultures (200–1000 ml) grown in LMM.

The inoculum consisted of 2% of cells grown in rich medium (Broth E, Amersham) and then washed in LMM lacking lactate. The supernatants were passed through a 0.22- m filter (Gelman) before use.

Preparation of M. luteus Rpf

Rpf was prepared from supernatant by ion-exchange chromatogra­phy as described in Mukamolova et al. (1998) to a final concentra­tion of 1–60 μg ml^–1.

Determination of M. luteus cell viability

Cell viability was determined by plating or by a most-probable­number (MPN) assay. For plating, plates consisting of 1.3% Nutri­ent Broth E (LabM), or LMM or SMM were used. Dilutions were made in quadruplicate with LMM lacking lactate. Plates were in­cubated at 30°C for 3–5 days.

The MPN assay was performed using serial dilutions in 2-ml test tubes or in a Bioscreen C optical growth analyzer (Labsys­tems, Finland) with 100-well plates, each well containing 0.2 ml medium. Various media were used for the MPN assay: LMM, LMM supplemented with 0.05% yeast extract, SMM, and Sauton medium. Cell suspensions were diluted as described (Kaprelyants et al. 1996) in appropriate medium with or without supernatant or Rpf. Growth (optical density) was monitored using a 600-nm filter. Plates or test tubes were incubated at 30°C with intensive contin­uous shaking. The calculation of the MPN was based on published tables (Meynell and Meynell 1965)

Chemicals

Nutrient Broth E, yeast extract and agar were obtained from Lab M (Difco), whilst L-lactate (Li salt) and succinate were obtained from Sigma. Other chemicals were of analytical grade and were ob­tained from Sigma or Fisher.

Fig.1 Effect of inoculum size and the addition of culture super­natant on the regrowth of Micrococcus luteus. Cells were grown to stationary phase in lactate minimal medium and reinoculated into lactate minimal medium at the concentrations shown. The length of the lag phase was determined optically ( ● “apparent” lag) or from CFU count ( ∆ “true” lag). In some cases (o), supernatants were taken from a batch culture of the organism grown to an OD of 2, slightly before the beginning of stationary phase, and mixed 1:1 with the lactate minimal medium (Kaprelyants et al. 1994)

Table 1 Apparent and true lag phase of Micrococcus luteus growth. The resuscitation-pro­moting factor (Rpf) concentra­tion was 1–10 ng ml^–1 (LMM lactate minimal medium, SMM succinate minimal medium)

^aLimited number of divisions only

Results

It is well-known that the duration of the lag phase in batch cultures often depends more-or-less inversely on the size of the inoculum, which, in turn, could reflect the accumu­lation of some growth inducer(s) secreted by cells during the lag phase. To check this possibility, the “true” inocu­lum-dependent lag phase (when bacterial growth is moni­tored by counting viable cells) and the “apparent” lag (when the lag phase is estimated directly from uncor­rected optical density traces) need to be studied sepa­rately. The latter must, of course, necessarily be “inocu­lum-size-dependent”. Figure 1 shows that the apparent lag phase for an M. luteus culture grown on LMM is indeed dependent on the inoculum size, especially in the range of less than one cell ml^–1. The true lag phase was almost con­stant (18–20 h) over a broad range of inoculum size. Ad­dition of supernatant or of isolated Rpf resulted in a sig­nificant decrease of the apparent lag phase (Fig.1), whilst it had almost no effect on the true lag phase . Evidently, supernatant increases the growth rate of bacteria signifi­cantly. A similar effect was observed for cells grown in rich medium (Broth E) or in LMM supplemented by yeast extract (0.05%). However the difference between the ap­parent lag (± Rpf) was visible only when a small inocu­lum had been used (not shown).

However, when washed cells were used as an inocu­lum, there was a pronounced influence of Rpf on cell growth. In the absence of Rpf, no increase in CFU up to 10 days was observed (Table 1). Thus, the absence of any Rpf effect on the true lag phase in Fig.1 may be due to the carryover of a small amount of the putative cytokine by exponentially growing cells, either on their cell sur­face or in spent medium (notwithstanding the significant dilution). It is worth mentioning that cell washing did not influence the number of CFU on either rich plates or plates with LMM. Similar to that of M. luteus, growth of M. smegmatis cells also revealed an Rpf-dependence when washed cells had been grown in Sauton medium (Fig.2).

As an extreme, we found that Rpf supports bacterial growth even in a medium in which M. luteus had to date never been cultivated. We used succinate instead of lac­tate in minimal medium and obtained the results shown in Table 1 and Fig.3. SMM does not normally support macroscopically observable M. luteus growth when the inoculum is less than 10⁵ cells ml^–1. Cells underwent only a few divisions, after which growth stopped; this might be due to carryover of some Rpf with the unwashed inoculum. However, the addition of purified Rpf (4 ng ml^–1) re­sulted in cell growth sufficient to form a turbid suspension (Fig.3). It should be stressed that washed cells did not lose the ability to form colonies on LMM agar, and nor­mal cells can also grow on agar prepared with SMM with­out added Rpf. It is interesting that the number of CFU per unit volume of culture medium as assessed on plates with succinate medium depended on the total number of cells that were plated out on the agar, changing from 5% to 70% (of the CFU that could be observed on rich agar plates) (Fig.4).

Fig.2 Effect of Micrococcus luteus resuscitation-promoting fac­tor (Rpf) on the growth of Mycobacterium smegmatis in batch cul­ture. M. smegmatis cells washed five times in Sauton medium be­fore inoculation were grown in 25-ml flasks on Sauton medium at 37°C with agitation. In some cases, Rpf was added in dilutions of 1:10,000 or 1:100,000 (initial concentration of Rpf was 60 µg ml^–1) Growth was monitored by sampling aliquots and plating them out on agar plates supplemented with Broth E and incubated at 37°C

Earlier we noticed that M. luteus cells taken from a long stationary phase as an inoculum grew very poorly on LMM. We checked the culturability of such cells by two methods: by plating out on agar plates, and by an MPN assay. Figure 5 shows that the incubation of cells for 100 h in the medium in which they had grown to stationary phase resulted in significant differences between the MPN count and CFU (the latter was almost identical for both LMM and rich-medium plates). Underestimation of viable cells by the MPN method means that at least 1000 viable cells (as judged by CFU) should be present in a test tube if they are to go on to produce visible growth. When Rpf was added to the MPN dilutions, the estimation of the

number of viable cells was similar to that determined by plate counts, which demonstrates the ability of one “old” cell to grow in a test tube in the presence of Rpf.

Fig.3 Effect of Micrococcus luteus resuscitation-promoting fac­tor (Rpf) on the growth of M. luteus in batch culture in succinate minimal medium (SMM). M. luteus was grown in Broth E until the end of the exponential growth phase, washed once and resus­pended in SMM. Flasks (20 ml) with SMM were inoculated with approximately 1,000 cells per milliliter, and growth was monitored by sampling aliquots and plating them out on agar plates supple­mented by Broth E and incubated at 30°C. In some cases (), Rpf was added at a dilution of 1:1,000 (initial concentration, 1 μg ml^–1)

Fig.4 Dependence of CFU number on solid medium supple­mented by rich (Broth E) medium or by succinate minimal me­dium (SMM) on the initial concentration of cells per plate. Before plating, cells were grown in Broth E until the end of the exponen­tial phase. Cells were diluted 10⁶-fold before plating. Plates were incubated for up to 20 days at 30°C

The results of our study of the culturability of M. lu­teus cells obtained by the two methods are summarised in Table 2. The underestimation of viable cells by the MPN method (in comparison with viable counts judged on plates) after cell washing or using old cells or cells grown in SMM indicates that more than one cell must initially be present per well if visible growth is to be produced. Again, when Rpf was added to the MPN dilutions, the es­timation of the number of viable cells was similar to that determined by plate counts (Table 2). Such a comparison demonstrates that at least 10⁵ cells grown on succinate me­dium, 10³ of cells from old stationary phase, and approxi­mately 10–500 washed cells are required per millilitre of inoculum for visible growth. At the same time, perfor­mance of the MPN assay for “washed” and “old” cells in rich liquid medium (Broth E) resulted in a viable count similar to the CFU number.

Table 2 Viable count of Micrococcus luteus estimated by two methods. M. luteus cells were grown until stationary phase in lac­tate minimal medium (LMM), and their viability was estimated by MPN in well plates [using succinate minimal medium (SMM) as dilution medium] and by CFU (“untreated” cells) The same cells were either washed several times with LMM and resuspended in LMM (“washed cells”) before MPN estimation (using LMM as di­lution medium) and CFU estimation or they were held in station­ary phase for more than 100 h before estimation of their viability [as in the case of (“stationary”). In some cases, resuscitation-pro­moting factor (Rpf; 1–10 ng ml^–1) was added to each well or tube for the MPN assay

Discussion

The results of the present experiments show that Rpf has a significant influence on the growth of producer cells or cells that secrete Rpf homologues (and, in addition, Rpf increases the final yield of bacterial biomass, as shown in Table 3). However, the observability and character of its action depends (as with the results from any viability esti­mation; Barer et al. 1998; Kell et al. 1998) on the condi­tions used in any particular experiment. For example, if untreated cells from the late exponential phase were used to test the activity of Rpf on the lag phase, we found that Rpf mainly influenced the growth rate rather than the du­ration of the true lag phase, changing only the apparent lag phase (Fig.1). At the same time, Rpf decreased the true lag phase if washed cells or cells grown in a very in­appropriate medium (succinate medium) were used (Table 1). Indeed, it has been known that an inoculum-dependent lag phenomenon may be observable only under a re­stricted range of conditions: in the case of Achromobacter delmarvae, an inoculum-dependent lag has been detected only in a poor medium, but not in a rich one (Shida et al. 1977). Similarly, the study of an inoculum-dependent lag for various Bacillus spp. has been performed by using poor medium (Lankford et al. 1966). Dagley et al. (1950) have found that the effect of supernatant on the inoculum­dependent lag itself depends on the size of the inoculum: the largest supernatant effect was observed with the small­est inoculum. We may suggest that a small amount of Rpf that is carried over by unwashed M. luteus cells used as an inoculum can be sufficient to stimulate the initial multi-

plication of cells, while added Rpf further stimulates ex­ponential growth. In the extreme situation, when growth of cells is either absent (washed cells) or occurs only un­til a limited number of divisions have been made (SMM), Rpf allows cells to grow with a division time of approxi­mately 10 h (for SMM) and 5 h (for LMM). Normal M. luteus cells grow on LMM with a division time of 4 h.

We have to stress that washed cells have not lost the ability to form colonies on LMM agar medium, and cells can grow on agar prepared with SMM without Rpf. This may be due to the fact that neighbouring cells are in inti­mate contact during colony development, which should facilitate cell-cell communication by locally accumulated Rpf as well as by juxtacrine signaling by cell-surface­associated proteins. The dependence of the CFU number on the number of cells spread on SMM agar is also in favour of the importance of cell-cell interactions for their growth (Fig.4).

The application of two independent methods for the es­timation of bacterial viability gives one the possibility of estimating quantitatively how many cells in a liquid me­dium can support growth in the absence of any externally added growth factor. For various treatments and media the minimal number of M. luteus cells required per millilitre for visible growth was from 10² to 10⁵ (Fig.5, Table 2), despite the fact that almost all of the cells in these popula­tions were viable and able to form colonies on agar plates. To our knowledge, this is the first detailed demonstration [but see also Votyakova et al. (1994) for resuscitation] of the existence of a “threshold” initial number of bacterial cells that must be present for their further growth, a phe­nomenon known for nucleated cells (Wheatley et al. 1993; Christensen et al. 1995, 1998). Hence, we may sug­gest that the basic principle of microbiology – “one cell­one culture” – may not work in some circumstances. The involvement of bacterial growth factors in this phenome­non is likely since the addition of Rpf resulted in very similar viable counts as judged by the two methods.

Table 3 Effect of resuscitation-promoting factor (Rpf) on the maximum optical density of broth cultures of various high G+C, gram-positive bacteria. For conditions for growing the bacteria, see Material and methods. Rpf concentration was 1–10 ng ml^–1(LMM lactate minimal medium)