|Home | About | Journals | Submit | Contact Us | Français|
Similar to the assembly of other dsDNA viruses, bacterial virus phi29 uses a motor to translocate its DNA into a procapsid, with the aid of protein gp16 that binds to pRNA 5’/3’ helical region. To investigate the mechanism of the motor action, the kinetics of the ATPase activity of gp16 was evaluated as a function of DNA structure (ss-or ds-stranded) or chemistry (purine or pyrimidine). The kcat and Km in the absence of DNA was 0.016 s−1 and 351.0 µM, respectively, suggesting that gp16 itself is a slow-ATPase with a low affinity for substrate. The affinity of gp16 for ATP was greatly boosted by the presence of DNA or pRNA, but the ATPase rate was strongly affected by DNA structure and chemistry. The order of ATPase stimulation is poly d(pyrimidine) > dsDNA > poly d(purine), which agreed with the order of the DNA binding to gp16, as revealed by single-molecule fluorescence microscopy. Interestingly, the stimulation degree by phi29 pRNA was similar to that of poly d(pyrimidine). The results suggest that pRNA accelerates gp16 ATPase activity more significantly than genomic dsDNA, albeit both pRNA and genomic DNA are involving in contact with gp16 during DNA packaging.
Most linear double-stranded (ds) DNA bacteriophages translocate their genomic DNA into a preassembled protein shell, known as a procapsid, to near-crystalline density. Due to the limitation of the spatial capacity of procapsids, these viruses have developed unique ATP-driven viral DNA packaging motors to complete such an entropically unfavorable DNA packaging process (Black, 1989).
The DNA packaging motor of many dsDNA bacteriophages consists of a dodecameric protein portal, also called the connector (Simpson et al., 2000;Ibarra et al., 2000), and two nonstructural components that comprise the packaging enzyme complex with certain characteristics typical of an ATPase (Guo et al., 1987). Mostly, the large subunit is responsible for procapsid binding, genome maturation (cleaving concatemeric DNA) and DNA translocation, while the small subunit recognizes the unique sequences on its own genomic DNA to employ the large subunit to assemble into an active terminase complex. The packaging motor of phi29 involves four essential components: terminal protein gp3, the 12-subunit connector (Guasch et al., 2002;Simpson et al., 2001;Simpson et al., 2000;Ibarra et al., 2000), the DNA-packaging protein gp16 (Guo et al., 1987), and the packaging RNA (pRNA) hexamer (Guo et al., 1998;Zhang et al., 1998).
Previously, it was found that gp16 binds dsDNA without discrimination of phi29 genomic DNA versus non-phi29 DNA (Lee & Guo, 2006). It was also revealed that gp16 specifically docks 5’/3’ paired ends of pRNA on the packaging motor. This procapsid/pRNA complex then recruits the gp16 to assemble into an active packaging motor for the further interaction with genomic DNA or for DNA translocation. Thus, gp16 serves as both a linkage between pRNA and DNA and as an essential DNA-contacting component during DNA translocation (Lee & Guo, 2006).
The ATPase activity of gp16 can be stimulated either by pRNA or by DNA. Similarly, procapsids with pRNA stimulate the ATPase activity of gp16 ten-fold, as compared with the lack of stimulation by procapsids alone. The maximum stimulation has been observed only when all packaging motor components, including pRNA, procapsid, gp16, and DNA-gp3, were present (Guo et al., 1987;Shu & Guo, 2003a;Grimes & Anderson, 1990;Grimes & Anderson, 1997;Ibarra et al., 2001). These stimulation effects were also found in other viral packaging enzymes. For instance, in phage lambda, each subunit of terminase possesses distinct ATPase activities (Hwang et al., 1996). The small subunit, gpNu1, is responsible for DNA recognition and DNA cutting, and shows DNA-dependent low affinity ATPase activity (Km=469 µM with DNA). Pac-ATPase subunit of T3 terminases also shows viral DNA-dependent ATPase activity, while the other subunit, termed Non-pac ATPase, was stimulated by non-packageable DNA (i.e., single-stranded or circular) or RNA (nonspecific) (Morita et al., 1993). These individual ATPase activities also can be stimulated by the interaction with other components. Formation of holoenzyme increases their ATPase activity. In SPP1, interactions between two or three G2P and one G1P, assembled into holoenzymes, showed the enhanced ATPase activity (Camacho et al., 2003). The ATP consumption in DNA packaging, in the defined in vitro DNA-packaging system of phi29, was estimated to be that every ATP hydrolysis corresponds to 2 base pairs movement of DNA (Guo et al., 1987). In addition, the recent single molecule studies revealed that the phi29 packaging motor has shown an initial rate of 100 base pairs per second with an extra load of 2.2 µm polystyrene beads on the DNA. It also showed that the force corresponds to 57 pN (Smith et al., 2001;Fuller et al., 2007), which makes it one of the most powerful bionanomotors constructed to date. Although it seems obvious that gp16 possesses a consensus ATP-binding sequence (Guo et al., 1987) and contributes as part of the ATPase complex to provide the packaging energy, no biochemical kinetic study has been reported yet on the ATPase activity of gp16 in phi29 and on the correlation to the stimulation effect of DNA binding by gp16. It is critical to understand how phi29 packaging motor gains the driving force from ATP hydrolysis by investigating ATPase activity of gp16 and its stimulation through the interaction with DNA substrate.
Conventional thin layer chromatography (TLC) has been used to separate the inorganic phosphate (Pi) released from the hydrolysis of [γ-32P]ATP by gp16 (Huang & Guo, 2003a;Shu & Guo, 2003b). However, due to the low reproducibility in autoradiography analysis and the poor resolution in early stage events of ATP hydrolysis, it is not an ideal method for kinetic studies. It is essential to observe the hydrolysis events during the early stages of reaction to accurately calculate the initial velocity. For those reasons, an MDCC-PBP (N-[2-(1-maleimidyl)ethyl]-7-(diethylamino)coumarin-3-carboxamide-phosphate binding protein) system has been employed. The phosphate binding protein (PBP) of Escherichia coli is a periplasmic protein, which is produced under low inorganic phosphate (Pi) concentrations and transports Pi into the cytoplasm (Brune et al., 1998). To be used as a Pi probe, a cysteine was introduced at position 197 on the PBP, and was selectively labeled with a fluorophore (MDCC). The resultant labeled protein (MDCC-PBP) binds Pi tightly (Kd ~ 0.1 µM) and gives an increase in fluorescence at 460 nm. MDCC-PBP has been used to investigate the kinetics and role of Pi release in a number of systems that involve ATP hydrolysis (He et al., 1997;Brune et al., 1994). In this report, we determined the kinetic parameters for the ATPase activity of gp16 by using the fluorescent phosphate sensor system, MDCC-PBP. We also investigated the effect of pRNA and various types of DNA, including single-stranded (ss)-poly d(purine) or poly d(pyrimidine) or their duplex, or different sizes of randomized ds DNA, on the ATPase activity of gp16.
To determine the kinetic parameters for ATP hydrolysis of gp16, the reaction was initiated by adding ATP and gp16 into a buffer containing MDCC-PBP as an indicator for the release of Pi (Fig. 1A). The results revealed a relationship of initial velocity Vinit (µM s−1) to concentration of gp16 (µM) (Fig. 1B). ATP hydrolysis was not detectable when the concentration of gp16 in the reaction mixture was under the 0.6 µM, while it was a linear relationship in the range of 0.6 and 2 µM of gp16. From this observation, 1 µM of gp16 was used in the rest of studies and the concentration of MDCC-PBP was optimized as 2 µM enough to cover the released Pi from ATPase hydrolysis. To examine the steady-state parameters of ATP hydrolysis by gp16, the assay was performed over a range of ATP concentrations. The ATP hydrolysis assay showed a linear relationship of increasing Vinit to increased concentration of ATP (Fig. 2), which implies that gp16 is a Michaelis-Menten-type enzyme. The catalytic rate of ATP hydrolyzing reaction by gp16 (kcat) was determined by linear regression plotting based on the Michaelis-Menten equation. The kcat of gp16 was 0.016 ± 0.002 s−1 with a Km of 351.0 ± 34.0 µM and kcat/Km of 45.0 M−1s−1, suggesting that gp16 itself is a slow-ATPase with low affinity for substrate (Table 1). This is the first kinetic measurement of the ATPase activity of gp16.
It has been reported that gp16 interacts with DNA in a sequence independent manner (Lee & Guo, 2006;Guo & Lee, 2007). To investigate the effect of dsDNA chemistry and structure on the ATPase activity of gp16, dsDNA fragments with size of 25, 54 and 160 bp, were tested in the defined ATPase assay system (Table 1). The presence of dsDNA affected both Km and kcat of the ATPase activity of gp16 so that kcat/Km was increased at least 10 times greater than the absence condition (Table 1). Although their sizes were different the rate of ATP hydrolysis activity (kcat) of gp16 for these different size dsDNA fragments were similar. In addition, the Km for all DNAs was not much different. Furthermore, the stimulation extent by random sequence dsDNA of 54 bp was very similar to that by a homopolymeric dsDNA (poly dA:poly dT) with similar size of 50 bp (Table 1). The results confirm the previous findings that gp16 randomly binds dsDNA in a non-sequence specific manner and indicate that any specific sequence is not required for the stimulation of ATPase activity of gp16, as also expected from the previous report (Lee & Guo, 2006).
The stimulation effect of 54 bp dsDNA on the ATPase activity of gp16 was further compared to that of 50 bases homopolymeric ssDNA, such as poly dT and poly dA. The use of these homopolymers avoids possible interference from partial double-stranded regions or other secondary or tertiary structures. The results showed that the slope for dsDNA lies near that for poly dA, with the interesting finding that poly dT stimulates the ATPase activity of gp16 most efficiently (Fig. 2). The kcat for poly dT was higher than that for dsDNA. However their kcat/Km were close, since in its interaction with gp16, dsDNA had a lower Km and kcat than poly dT (Table 1). To the contrary, kcat/Km for poly dA was lower than that of dsDNA or poly dT (Table 1), which means that the dsDNA or poly dT binding to gp16 increase their affinity to ATP. In addition, it suggests that the stimulation of the ATPase activity of gp16 caused by dsDNA or poly dT binding was due to an increase in the catalytic rate of the ATPase activity of gp16 together with increase of the affinity for ATP.
It has been reported that the ATPase activity of gp16 was stimulated by pRNA (Grimes & Anderson, 1990;Ibarra et al., 2001;Shu & Guo, 2003a). It has also been reported that gp16 serves as a bridge between pRNA and genomic DNA by binding to both pRNA and dsDNA (Lee & Guo, 2006). However, the strength of stimulation effect has not been quantified. Kinetic analysis revealed that the slope of pRNA-dependent ATPase activity was close to that of poly dT (Fig. 2). kcat and Km in the presence of the pRNA were similar to that of poly dT, and far apart from that of dsDNA (Table 1). This indicates that pRNA is a more preferred structure than dsDNA in stimulating the ATPase activity of gp16.
To further verify the effect of ssDNA on the ATPase activity of gp16, poly dG and poly dC were compared with poly dA and poly dT. The poly d(pyrimidine), poly dT and poly dC, was a more effective stimulator than poly d(purine), poly dA and poly dG (Fig. 2). kcat for poly dT and poly dA were 0.141 ± 0.017 and 0.044 ± 0.001 s−1, respectively (Table 1), indicating that poly dT stimulated the ATPase rate of gp16 at least 3 times higher than poly dA and almost 9 times higher than gp16 itself. In addition, kcat/Km for poly d(purine) was at least 2 times lower when it was compared to that for poly d(pyrimidine). However, Km was almost not affected by the presence of ssDNA, while ssDNA mostly changed kcat. The data indicates that the stimulating effect of ssDNA on the ATPase activity of gp16 depended on its chemistry (type of nucleotide) (Table 1).
The sensitivity of gp16 to the type of DNA leads to a more important question of whether the structure of DNA plays a significant role in interacting with other motor components during translocation. If this is the case, how the helical nature of dsDNA is involved in the mechanism of motor motion would be of interest. Recently, it was proposed that one of the motor components, the connector, does not rotate during DNA translocation (Baumann et al., 2006;Hugel et al., 2007). Nevertheless, most DNA packaging models of phi29 favor a rotary motor mechanism. According to this model, the helical nature of dsDNA plays an essential role so that the different pitch and depth in different DNA types might affect the interaction with gp16. In alternative packaging models, which are predominantly based on the alignment of the packaging enzymes with recognition sites on the DNA, higher tolerance toward structural alterations would be expected (Simpson et al., 2000;Guasch et al., 2002). In these models, the binding sites on the DNA need to be sufficiently intact for the interaction with the packaging enzyme. Studies on the packaging activity for structurally modified DNA substrates might help in the understanding of the specific interaction between gp16 and genomic DNA during packaging, which can be useful to turn the motor off (and potentially back on) at specific sites on the DNA. Indeed, changing of the DNA from helix to single strand may serve as a mechanism to completely stop the translocation of phi29 DNA (Moll & Guo, 2005). This speculation was supported by the finding that the stimulating effect of homopolymeric dsDNA (poly dA:poly dT) containing 50% purine and 50% pyrimidine lies between those of poly d(purine) and poly d(pyrimidine), and that the poly dA:poly dT made no difference in the stimulating effect from random dsDNA also containing 50% purine and 50% pyrimidine (Fig. 2). Whether the observation that only 50% (pyrimidine, but not purine) of the nucleotides had a strong effect on stimulating the ATPase activity of gp16 can be correlated with the previous findings that, on the average, one ATP consumption packages two base pairs of substrate DNA (Guo et al., 1987;Morita et al., 1993;Chemla et al., 2005) remains to be investigated. In addition, the finding that the DNA frequently slipped out of the procapsids, as the DNA was packaged during the packaging process (Smith et al., 2001), might have a certain relationship with the sequence on the DNA-gp3. Since our single molecule photoimaging data shows that poly dT binding to gp16 was more stronger than poly dA, it can be inferred that a purine-rich region on the DNA might decrease the binding affinity of the packaging motor against DNA. Then, the loose interaction between the DNA and gp16 in the packaging motor gives rise to a slipping phenomenon due to the highly built up pressure inside the capsids during packaging. This theory is based on the assumption that gp16 interacts with one strand out of the duplex DNA during DNA translocation.
Based on the result that the poly d (pyrimidine) stimulates the ATPase activity of gp16 stronger than the poly d(purine), it is central to investigate the binding ability of gp16 to explain the cause for such a distinctive effect. Single molecule fluorescence microscopy was employed to compare their binding affinity for gp16. Gp16 was first immobilized on the quartz surface pre-coated with anti-gp16 IgG., biding of the fluorescent ssDNA, or pRNA directly imaged using SMDV-TIRF single molecule dual viewing system (Shu et al., 2007;Zhang et al., 2007). Each visualized spot represents one single DNA molecule that binds to gp16, as confirmed by single molecule photobleaching technique (Shu et al., 2007;Zhang et al., 2007). It was found that Cy5-poly dT was bound to gp16 even at 10 pM, while Cy5-poly dA bound to gp16 was not observed at the lower concentration of 10 nM. However, dsDNA (Cy5-poly dT:Cy3-poly dA) showed stronger binding than poly dA, but weaker than poly dT (Fig. 3A). In addition, Cy3-pRNA binding to gp16 was greater than both Cy3-poly dA and dsDNA (Cy5-poly dT:Cy3-poly dA). The order of the binding strength was in agreement with the order of the ATPase activity stimulated by the respective DNA or pRNA. Mean fluorescence intensity analysis showed that the difference of gp16 binding intensity between poly dT and poly dA is far greater than their difference in stimulating the ATPase of gp16 (Fig. 3B). Nevertheless, the results suggested that gp16 binds poly dT, representing poly d(pyrimidine), or pRNA with higher affinity in comparison to poly dA, representing poly d(purine), or dsDNA. The agreement between the order of the binding to gp16 and the stimulation effect on the ATPase activity of gp16 implies that the enhancement of ATPase stimulating activity was a result of their higher binding affinity for gp16.
Our data reveals that DNA and pRNA enhanced both the binding affinity and the catalytic activity of gp16 on ATP. It also reveals that the enhancement was correlated to DNA or pRNA binding to gp16 (Fig. 3A and Table 1). The following model might explain why DNA or pRNA binding to gp16 affects the ATPase function of gp16. DNA or pRNA binding to gp16 could trigger a conformational change of gp16. Such structural alteration activate a catalytic site on gp16 to enhance its affinity to ATP, thus elevates ATPase activity. Our data also shows that both the ATP binding and hydrolysis by gp16 was correlated with DNA structure and chemistry, however, which did not affect the Km in gp16 ATPase kinetic analysis. These observations raise some challenging questions: Does gp16 or the phi29 packaging motor involve single stranded DNA during the motor assembly or packaging process? Does the phi29 double-stranded genome partially separate into a single-stranded form in some contact region of the packaging motor due to the involvement of any unknown functions such as gyrase, helicase or topoisomerase during DNA translocation? The previous finding that nicked phi29 DNA can still be packaged (Moll & Guo, 2005), and similar phenomena in other phages including T3 (Fujisawa et al., 1987), T5 (Hayward & Smith, 1972), T7 (Khan et al., 1995) and Rhodopseudomonas sphaeroides bacteriophage RSI (Donohue et al., 1985), do not support this argument. Any relationship between ATPase activity of gp16 and DNA packaging process depending upon the structure and chemistry of DNA still remains to be answered.
Gp16 was prepared as previously described (Huang & Guo, 2003a;Huang & Guo, 2003b;Lee & Guo, 2006). Gp16 was purified to homogeneity and tested for its full DNA packaging activity in the in vitro phi29 assembly system (Lee & Guo, 1995) before the ATPase assay. Poly d(N) was synthesized to be 50 bases (except 25 bases of poly dG) (IDT Inc., USA). For fluorescence single molecule microscopy, Cy5 or Cy3 was conjugated on the 5’ end of poly d(A or T) (IDT Inc., USA). MDCC-PBP was purchased from Invitrogen, USA. All dsDNAs used in our experiments were expressed with the prefix “ds-”, while the prefix “ss-” was used to refer to the single stranded DNA. Poly d(N) refers to homopolymeric single stranded DNA and poly dA:poly dT referes to the hybridized double stranded heteroduplex DNA.
Purified gp16 was mixed with MDCC-PBP in HMS (20 mM HEPES, pH 8.0, 100 mM NaCl and 10 mM MgCl2) containing 0.2 U/mL of purine nucleoside phosphorylase (PNPase) (Sigma), 0.04 mM 7-methyl guanosine (Sigma) and 6% glycerol, then placed in a quartz microcuvette positioned in a Cary Eclipse spectrophotometer (Varian, Inc.). After the background fluorescence was checked for 10 seconds, ATP solution was added to initiate the ATP hydrolysis reaction. Pi was scavenged by MDCC-PBP when released from the ATPase reaction. The Pi-bound MDCC-PBP was excited at 425 nm wavelength and the emitted fluorescence was monitored at 464 nm wavelength with 0.25 second time interval under control by the Cary Eclipse analysis software (Varian, Inc.). The reaction was terminated by the addition of 250 µM of KH2PO4 (pH 6.5).
Initial fluorescence represents the intensity of free MDCC-PBP (If). The ATP hydrolysis reaction was initiated by addition of ATP solution (pH 7.0), and the released Pi was proportional to the burst of fluorescence as an observed intensity at time t (Iobs(t)). The amount of MDCC-PBP bound to Pi (Xb(t)) was calculated by subtracting the intensity of free MDCC-PBP (If) from the observed intensities at time t (Iobs(t)), and the difference was divided by the difference between the saturated Pi-bound MDCC-PBP (Ib) in the presence of excess Pi and free MDCC-PBP (If). The equation can be expressed as Xb(t)= [Iobs(t) − If]/[Ib − If]. The initial velocities (Vinit) were calculated as the linear slope of released Pi (µM) from ATPase activity of gp16 over time (s). Hanes-Woolf plot was used to determine kinetic parameters during steady-state ATP hydrolysis of gp16 by rearranging Michaelis-Menten equation (Vinit = Vmax [ATP]/(Km+[ATP]) to be [ATP]/Vinit = ([ATP]+ Km)/Vmax. In the equation, the ratio of ATP concentration ([ATP], µM) to Vinit (µM s−1) is plotted over [ATP], in which data is expressed as a straight line of slope 1/Vmax and a y-intercept of Km/Vmax. The linear regression analysis of data was performed using the OriginPro version 7.5 software (Origin Lab Corp.).
The quartz surface of the perfusion chamber (15 µL internal volume) was coated with 100 µg/mL of anti-gp16 rabbit IgG (Proteintech Group, Inc) in HMS incubated for overnight at 4°C. The gp16 (100 nM) diluted in HMS was infused into the microchamber and incubated for 30 min at RT for Ab-Ag binding. The unbound gp16 was washed using HMS. Each Cy5-or Cy3-labeled DNA in HMS was then incubated in the gp16-immobilized microchamber for 20 min at RT. After washing with HMS, an oxygen depletion solution containing glucose oxidase, catalase, β-D-glucose, and 2-mercaptoethanol (Ha et al., 2002) was infused inside the chamber before observation to reduce the effect of photobleaching of fluorescent materials. Observation of bound fluorescent-DNA to gp16 was conducted by the total internal reflection fluorescence (TIRF) imaging technique as described previously (Shu et al., 2007;Zhang et al., 2007). An Olympus IX71 inverted microscope with 60x oil immersion objective (PlanoApo, 60 ×, NA=1.4) was used, and an extra 1.6 × magnification lens was inserted for the further magnification of the obtained image. A laser beam with a 638 nm wavelength was used to excite Cy5-DNA, and 532 nm wavelength for Cy3-DNA or Cy3-pRNA. The beam size is about 80 µm × 180 µm at the quartz/solution interface. Fluorescence signals were collected by an EMCCD camera (iXon 877V, Andor Technology). The fluorescence images were recorded with an exposure time of 0.2 second. For each fluorescence image, the mean fluorescence intensity over a 37 µm × 77 µm region was calculated using Andor iQ version 8 software (Andor Technology, USA), then the mean intensities were plotted versus the molar concentrations of fluorescent DNA.
We wish to thank Drs. Christopher Williams, Donald Bergstrom, Carlo Montemagno, David Wendell and Dan Liang for their technical assistance in setting fluorometry. We also give our thanks to Dr. Andrew Herr for helpful discussion and review of the manuscript. This work was supported mainly by NIH grant GM59944, and partially by EB03730 as well as by NIH through the NIH Roadmap for Medical Research (PN2 EY018230).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.