4.1. Ejection of DNA
High pressure inside of capsids is thought to be used for the ejection of DNA from capsids [Liu et al 2014,Leforestier A, Livolant F 2010]. The stress inside the phage capsid decreases by polyvalent cations [Evilevitch et al. 2008,Fuller et al. 2007]. The full capsid has a slightly smaller radius than the empty capsid [Effantin et al. 2006]. The latter two results imply that the internal pressure of the mature head is not higher than the external pressure. The T4 DNA produced by degradation of capsid proteins with proteinase K show compact forms (Fig. 2), not self-repulsive strings like the ejected DNA in 30 mM Pi (Fig. 1C, G). The DNA molecule itself takes a compact conformation within and without a capsid under the regular extracellular ambient concentrations of ions, pressure, and temperature without addition of any multivalent cations. The globule-to-coil phase transition of phage DNA, which is necessary for the transportation of phage DNA into the host bacteria [Chow et al. 1971,Kuo et al. 1971,Liu et al 2014,Leforestier A, Livolant F 2010,Li et al. 2015], can be induced solely by increment of ambient concentration of Pi to the concentration corresponding to cell electrolyte in T4 DNA (Fig 1). It is implicated that this process of DNA ejection may have contribution to the process of DNA ejection of T4 to the host cell of E. coli.
The step-by-step mechanism of the ejection of DNA was not investigated in this study. However, the completed head and tail are joined spontaneously after DNA packaging is completed in the assembly of bacteriophage [Aksyuk, Rossmann 2011, Arisaka et al. 2016], we presume that in the ejection of DNA from a capsid the tail may detach from the connector first and the ejection of DNA from the connector follows according to the increment of the ambient phosphate concentration.
4.2. Packaging of DNA
Prior to discussing packaging, it would be appropriate to examine again that there were no or few, if any, intact virions in the 30 mM Pi suspension. The FLM observations clearly indicated nearly complete ejection of DNA and almost no intact virion in 30 mM Pi (Figs. 1C, 4D). After 25 mM phosphate treatment virions separate into DNA, empty heads, and tails [Chow et al. 1971]. Ultracentrifugation of 3 mM sodium-citrate treated virions indicates almost no intact virion remains after chelate treatment [Kuo et al. 1971]. It becomes appropriate to estimate few, if any, intact virions remain in 30 mM Pi suspension of T4.
However, pfu values of 30 mM Pi suspensions were equivalent to the original specimens (Figs. 4A, 4D). This apparent contradiction can be attributed to the change of the concentration of Pi during plating. The high concentrations of Pi of the T4 inocula, where virions ejected their DNA, decreases when the suspensions were inoculated into the peptone agar of low Pi concentration for plating, which triggers the regeneration of virions in the agar plate. To prove this process, T4 suspensions in T-buffer were inoculated into 30 mM Pi, 60 mM NaCl and Milli-Q water and the pfu values were monitored (Fig. 5). The solution of 60 mM Na+ was chosen as the equivalent cation concentration with 30 mM Pi. The infective ability was immediately lost when 0.01 T-buffer suspension was diluted 100-fold into Milli-Q water (Fig. 5, 0.0001Tb). The monovalent cation, 60 mM Na+, preserves but no more pfu than the infectious ability of the 0.01 T-buffer inocula throughout the time course (Figs. 5, 60Na, 0.01Tb). The infectivity recovery of the inoculation from 30 mM Pi, in contrast, was higher than the 0.01 T-buffer inocula, and moreover, it was equivalent or higher than the original 1 T-buffer suspension (Figs. 5, 30 mM Pi, Tb). Plating of virions ejected their DNA in 30 mM Pi on the normal peptone agar triggers DNA packaging and produce more infective virions than inocula, which indicates not only recovering of the infective ability but also reset some virions lost their infectivity to infective virions.
To support this, we tried to digest the naked DNA in 30 mM Pi suspension with DNase I. However, abundant coil and globular DNA molecules were observed after the DNase I treatments and no decrease of pfu counts was observed (Figs. 1A, 4E). Consequently, we compared the DNase I degradation in situ and on-filter vs. the suspensions of T4 virions in EL, 30 mM Pi and TE (Fig. 4). As described at Materials and Methods section, in “in situ treatment”, the DNA molecules in the original solvents were degraded, i.e. in EL, 30 mM Pi and TE, and “on-filter-dry treatment”, the suspensions were filtered through once and DNA molecules collected on 0.02 μm Anodisc to remove the original solvents. By this filtration, coil DNA molecules and virion particles are exposed to air and adsorbed on the filter surface. Afterwards, the collected DNA molecules and virion particles were degraded on-filter with DNase I in EL or 10 - 30 mM Pi, according to the original suspensions. In “on-filter-wet treatment”, the suspensions with DNase I were mounted on 0.02 μm Anodisc and filtrate slowly. When the sample suspension decreased, additional DNase solution was supplied. By this filtration, coil DNA molecules and virion particles are filtrated on the filter surface but not exposed to air during DNase degradation. In the T4 suspensions in EL, DNase I treatment did not change the FLM abundances and the pfu counts among the original specimens, in situ and on-filter-dry DNase I treated specimens (Fig. 4A, B, C). On-filter-wet DNase I treatment also did not change the FLM abundance (data not shown). Intact virions are not digested with DNase I. The ejected DNA in 30 mM Pi was not digested with in situ DNase I treatment (Fig. 4E). During the DNase I treatment parts of DNA molecules became globular (Fig. 4E) which were bigger than intact virions (Fig. 4A, B, C), and it is known that the globular conformation is more stable than coil conformation and DNA in the latter conformation naturally transformed to the former conformation [Yoshikawa, Matsuzawa 1995]. On-filter-dry DNase I in 10 mM and 30 mM Pi treatments digested almost all DNA molecules (Fig. 4F). While on-filter-wet DNase I in 10 mM and 30 mM Pi treatments did not digest coil or globular DNA molecules (data not shown). This indicates following facts; first, the DNase I in 10 mM and 30 mM Pi maintained the ability of the degradation of DNA, and second, almost no intact virion survived in 30 mM Pi. The abundances of DNase I resistant DNA particles in 30 mM Pi, determined by FLM, were ca. 1% or less of the original viral abundances (Figs. 1, 4). Contrary, few globules of DNA remained after DNase I treatment in TE (Fig. 4H), while both the initial population in TE and DNase I treated suspensions showed no infectious ability. All the DNA of T4 in TE was degraded with the on-filter-dry DNase I treatment (Fig. 4I). It becomes clear, DNase I, an endonuclease from bovine pancreas that digests single- and double-stranded DNA, digests T4 DNA in TE, but not the coil DNA in 30 mM Pi, the intracell mimic concentration of Pi. After two days of the above preparations, the on-filter-dry preparation of T4 virions in EL decreased their size dramatically, while the on-filter-wet preparations of T4 virions in EL and coil or globular DNA molecules in 30 mM Pi did not change the FLM condition (data not shown). The compact DNA molecules in the capsids were gradually digested by DNase I after the virions were exposed to air. Probably the virion heads were deformed by the exposure to air and a trace of DNase I might seeped into the heads to digest packaged DNA molecules. Modification of DNA protects T4 DNA from the nuclease digestion [Bryson et al. 2015]. The protection of T4 DNA against DNase I works when the DNA is in coil form in intracell mimic concentration of Pi, but not when it is attached on a filter or in the compact conformation. The inactivity of DNase I in 30 mM Pi also can be explained by the lack of enough concentration of metal ions, e.g. Mg2+, Ca2+ and Mn2+ which are necessary for DNase activity [Price 1972, Price 1975], because of the chelation of metal ions by Pi. While DNase I in 10 – 30 mM Pi digests coil DNA attached on a filter, the concentrations of metal ions in 10 – 30 mM Pi are high enough for the activity of DNase I. The behaviors of ejected DNA in 30 mM Pi and in TE were also somewhat different. The coil DNA molecules in 30 mM Pi showed active Brownian motion during the microscopic observation, while the ejected DNA molecules in TE were more likely to stick on the surface of the filter.
The high pfu abundances of DNase I treated 30 mM Pi suspensions are not derived from the intact virions in 30 mM Pi suspensions which are resistant to DNase I, but originated from DNA-ejected virions in 30 mM Pi which convert into infective virions during the plating processes which dilutes the concentration of phosphate and re-produce infective virions. The ejected DNA molecules from virions in situ are resistant from the degradation by DNase I (Figs. 4 D, E) and maintain their intact physiological activity (Fig. 5). The re-production of infective virions during the plating of DNA-ejected T4 suspension on peptone agar plate not only recovers the infectivity but also converts some non-infective virions into infective virions.
The currently accepted theory of DNA packaging of most dsDNA bacteriophages initiates when a terminase creates an end of the concatemeric DNA, which attaches to the portal vertex of a capsid and packages the self-repulsive DNA string into the empty capsid with the molecular motor, fueled by the energy of ATP hydrolysis [Casjens 2011,Aksyuk, Rossmann 2011,Vafabakhsh et al. 2014,Zinchenko 2016]. This type of ejection-packaging system is hereinafter called ‘motor-ATP system’. Accordingly, the motor-ATP system cannot work without ATP [Kottadiel et al. 2012].
In our experiments, the initial concentration of ATP molecules in dialysed virion suspension prior to packaging was ca. 20 pM (Fig. 3), which is more than 1000 times lower than the minimum ATP concentration for packaging DNA in motor-ATP system, 25 μM ATP [Kottadiel et al. 2012]. The packaging of DNA molecules into capsids occurred at the condition up to 107-fold dilution from this concentration of ATP; 102-fold dilution for ejection of DNA in 30 mM Pi solution, following 103-fold dilution with 30 mM solution plus 102-fold dilution with EL at the compaction process (column IV of Table 2). Practically there was no ATP at the compaction/packaging process of DNA. This indicates the packaging system of DNA into a capsid introduced here is not ATP dependent nor a motor-ATP system. The sole agent causing the ejection and packaging of DNA in our protocol was the change of the ambient Pi concentrations. The DNA molecules in the capsids of T4 were first ejected by increasing the ambient concentration of Pi. Without addition of multivalent cations of the valences higher than three [Zinchenko 2016, Todd et al. 2008], simply decreasing of the ambient concentration of Pi induced the fluid-to-solid conformational change of DNA. This compaction of DNA occurs inside of a capsid and the DNA is packaged into capsids, which is confirmed by the resistance of the compacted DNA particles against DNase I and the ability of infection (Fig. 1). The efficiency of packaging to form the infective virion is nearly 100% (Fig. 1). This process does not require additional ATP for packaging of DNA. The concentration of the ambient ATP indicates the process is an ATP-free system. However, the process itself includes the change of the ambient concentrations of Pi, which implies the differences of the concentrations of ambient ions or the density gradient energy can be the source of the energy for the ejection and packaging processes [Vranjes, Kono 2015]. This type of packaging system is hereinafter called ‘conformational change system’.
Characteristic differences between the motor-ATP system and the conformational change system are: 1) ATP dependent in the former [Casjens 2011,Aksyuk, Rossmann 2011], while independent from the extra ATP in the latter, 2) the packaging of DNA in the former is one-by-one mode and takes several minutes to package one DNA into a capsid [Black, Rao 2012]. In the latter, a group of packaging happens in parallel and in the ensemble average, ca. 80% of packaging might have been done within three minutes (data not shown), 3) the ratios of the regenerated virions were ca. 10% in the former [Black, Peng 2006], while the efficiencies in the latter were >90% in both FLM and pfu counts (Fig. 1). Besides these differences, the motor-ATP system has a limitation for working as the sole packaging system of DNA in a cell. As the globular DNA cannot be packaged into capsids, packaging of DNA should be initiated when they are in a coiled state under the intracellular high Pi concentration. On the other hand, packaged DNA flows out of a capsid in the high Pi condition [Chow et al. 1971,Kuo et al. 1971,Shafia, Thompson 1964]. Indeed, if the packaging process is interrupted by the addition of ATPγS during the transportation process, the packaged DNA runs out from the capsid [Morita, Fujisawa 1997]. This indicates that during the packaging process, once the pumping activity of the motor stops, the DNA molecule will automatically run out from the capsid, and complete virions cannot stay in a cell. Stabilization of the complete virions needs a lower concentration of Pi, which induces DNA coils into the globular conformational change (Figs. 1, 2, 4) [Gosule, Schellman 1976,Yoshikawa, Matsuzawa 1995], and disables the packaging by the ATP-motor system. This can be an obstacle if the ATP-motor system is the only packaging process. On the other hand, in the conformational change system, once the free ends of DNA molecules are connected, or partially packaged into capsids spontaneously [Vafabakhsh et al. 2014] or by the ATP-motor system [Casjens 2011,Aksyuk, Rossmann 2011,Vafabakhsh et al. 2014], the decrease of the chelate (Pi) concentration induces the packaging of DNA (this study), which is followed by the spontaneous joining of head and tail to form infectious virions [Aksyuk, Rossmann 2011,Arisaka et al. 2016]. Accordingly, the sole motor-ATP system or sole conformational change system may not be able to accomplish the packaging of DNA in a cell, but they need each other to accomplish DNA packaging. When one end of DNA attached or packaged into capsids by the motor-ATP system and the ambient Pi concentration is reduced at the end of viral reproduction [Black, Peng 2006] or even after the burst of the host cell (pfu of Figs. 4D, E), the packaging of DNA will be completed by the conformational change system. Because the process of the conformational change system does not require extra ATP and proceeds automatically, it has been ‘invisible’ and might be overlooked in previous observations.