Multiple origins of replication contribute to a discontinuous pattern of DNA synthesis across the T4 genome during infection

Summary

Chromosomes provide a template for a number of DNA transactions, including replication and transcription, but the dynamic interplay between these activities is poorly understood at the genomic level. The bacteriophage T4 has long served as a model for the study of DNA replication, transcription, and recombination, and should be an excellent model organism in which to integrate in vitro biochemistry into a chromosomal context. As a first step in characterizing the dynamics of chromosomal transactions during T4 infection, we have employed a unique set of macro array strategies to identify the origins of viral DNA synthesis and monitor the actual accumulation of nascent DNA across the genome in real time. We show that T4 DNA synthesis originates from at least five discrete loci within a single population of infected cells, near oriA, oriC, oriE, oriF, and oriG, the first direct evidence of multiple, active origins within a single population of infected cells. Although early T4 DNA replication is initiated at defined origins, continued synthesis requires viral recombination. The relationship between these two modes of replication during infection has not been well understood, but we observe that the switch between origin and recombination-mediated replication is dependent on the number of infecting viruses. Finally, we demonstrate that the nascent DNAs produced from origin loci are spatially and temporally regulated, leading to the accumulation of multiple, short DNAs near the origins, which are presumably used to prime subsequent recombination-mediated replication. These results provide the foundation for the future characterization of the molecular dynamics that contribute to T4 genome function and evolution and may provide insights into the replication of other multi origin chromosomes.

Introduction

All organisms share a common need to replicate their genomes. This process is required to distribute genetic material to progeny during reproduction and to create daughter cells during development of multi-cellular organisms. Often replication occurs concurrently with transcription and other DNA transactions on a genomic template cluttered with bound proteins, presumably creating an obstacle course of molecular machines, performing the various tasks involved in the maintenance of chromosomes. Such a scenario would imply that DNA replication, transcription, and other genomic events are orchestrated to prevent one transaction from impeding another, creating a complex molecular system, which provides a context for chromosome utility and arbitrates genome evolution.

The dynamics of DNA replication and other chromosomal transactions remain poorly understood on a genomic level, and, even in the well-defined Escherichia coli system, where the mechanics of replication initiation and termination have been determined (reviewed in ^{1; 2}), there is still some controversy surrounding the progression of DNA synthesis across the genome ^{3; 4}. The situation in eukaryotic systems is even less clear, as many origins contribute to chromosomal replication, raising a number of questions regarding the regulation of origin activity and the resolution of the multiple nascent DNAs arising from the various origins into a single, genomic polymer. Several groups have employed DNA micro arrays to address these questions ^{5; 6}, and these studies have helped to identify the origins and timing of replication across entire genomes.

It is not a simple task to map the dynamic accumulation of DNA synthesis products in most systems. Typically, the chromosome is only replicated a single time, highlighting background and noise issues, and it can be difficult to maintain synchrony within a population of replicating cells throughout the entire cycle. We sought to avoid these problems by using the bacteriophage T4 as a model system in which to study the mechanics and dynamics of chromosomal replication. Like the chromosomes of eukaryotic cells, the linear, 172 kb, T4 chromosome contains multiple origins of replication, yet, in contrast, T4 is replicated several hundred times during a single, 30 to 60 minute infection. This rapid genome synthesis is accomplished by a viral replisome that polymerizes DNA at a rate of 30–45 kb per minute ⁷, a spectacular feat given that transcription is concurrent with replication during infection.

The T4 chromosome is terminally redundant and about 3 kb of DNA sequence is randomly repeated at the ends of the chromosome, creating a circularly permutated genetic map. The lack of defined chromosomal telomeres and the rapid production of progeny virions presumably contribute to a unique T4 replication strategy that includes both origin dependent and recombination dependent replication. Initial T4 DNA synthesis is thought to originate de novo from one of several loci, and it has been assumed that nascent DNAs initiated at these origins are extended bi-directionally until the resultant duplex polymer reaches the ends of the parental genome ⁸. Although this origin-mediated replication is sufficient to produce some full-length, infectious virus, the vast majority of T4 DNA synthesis is dependent on the viral recombination machinery ⁸. The transition from origin- to recombination mediated replication is not well understood, but once the parental viral genome has been copied, there should be two possible ways in which homologous recombination can be initiated: Nascent daughter chromosomes could recombine with the homologous parental sequences, or the terminally redundant, circularly permutated viral chromosome could recombine with itself, in either case producing a primer for continued DNA synthesis.

Both origin-mediated and recombination-mediated DNA replication use an overlapping set of T4 gene products, which include the gene 43 DNA polymerase, 45 polymerase clamp, and 44/62 clamp loader, as well as the 41 helicase and the 59 helicase loader ^{8; 9; 1}. The 59 protein binds several DNA structures that mimic recombination intermediates used to prime recombination-mediated replication and targets 41 helicase loading to these substrates, allowing formation of a processive DNA replisome ¹¹. The role of 59 protein during origin mediated replication is less clear. It stimulates replication on R-loop substrates designed to mimic RNA primers presumed to initiate origin-mediated replication in vitro ^{12; 13} but is not absolutely required for viral replication during infection ^{8; 14; 15}.

Like other T4 recombination genes, including the recA homologue uvsX, mutation of the helicase loader gene 59 causes viral DNA synthesis to halt, prematurely, giving rise to the so-called DNA arrest phenotype (reviewed in ^{8; 15; 16}). Although the timing of this arrest is thought to mark the transition from origin to recombination-mediated replication, little is known about the switch between the two modes of replication, and the null recombination deficient phenotype is poorly defined. Many of the T4 recombination genes contribute to multiple biological activities, making it difficult to ascribe their replication phenotypes to a particular defect, and the actual amount of origin-mediated DNA synthesis occurring in either mutant or normal infections remains unclear.

Historically T4 DNA synthesis in vivo has been measured by supplementing infected cultures with radioactive thymidine and measuring the incorporation of this pulsed nucleotide into growing DNA polymers. Unfortunately, using this methodology, the actual amount of DNA synthesized is not readily apparent, as there is no way to correlate the incorporation of label with the number of genomes synthesized. Rather, genome replication can only be estimated by infectious virion production, a method that may not account for all DNA synthesis. Moreover, early during T4 infection, exogenous label is not efficiently incorporated into nascent viral DNA, as nucleotides are scavenged from the degraded host chromosome (reviewed in ¹⁷), so pulsed nucleotides may not provide an accurate measurement of initial DNA synthesis.

The experimental difficulties presented by the bipartite T4 replication strategy are not limited to shared protein requirements, as is aptly demonstrated by earlier attempts to identify the viral origins of replication ¹⁸. In one of the most exhaustive studies, Kreuzer and Alberts restricted and shotgun cloned the T4 genome and simply asked if any of the resultant clones could be maintained in trans and packaged as concatemers during an otherwise wild type infection. After several rounds of selection, a number of putative origins were recovered. Several of these mapped to regions of the T4 genome previously shown to be hotspots for recombination ¹⁸, and at least one putative origin supports recombination-mediated replication ¹⁹, implying that origin and recombination mediated replication are not only biochemically linked, but also physically linked to the same regions of the T4 genome.

In total, at least 7 putative origins have been identified by a number of groups, oriA, oriB, oriC, oriD, oriE, oriF, and oriG (see figure 1, reviewed in ^{8; 20; 21}). Several of these loci were identified as restriction fragments enriched for nascent T4 DNA early during infection, including oriA ²², oriB ²³, oriC ²³, oriE ^{22; 24; 25}, and oriF ²². Some, including oriD ²⁶, were observed using electron microscopy, while others, oriF and oriG, were recovered by Kreuzer and Alberts as autonomously replicating sequences ^{18; 27}. With the exception of oriF and oriG, most of these putative origins remain uncharacterized, and neither the cis elements necessary for origin function nor the actual start sites of DNA synthesis have been defined.

Figure 1.

The T4 genomic macroarray. Although the terminally redundant, 172 kb T4 chromosome is linear, the ends are is circularly permutated, and there are no fixed telomeres. This results in a circular genetic map as depicted here. The location and position of PCR fragments included in the genomic array are indicated with yellow tabs with identifying numbers that correspond to the PCR loci (LP numbers) listed in Table 1. Putative T4 origins of DNA replication have been placed on the map as reported previously ⁸, ²¹ and are indicated with green lollipops and identifiers (a, b, ect.). Major open reading frames (>100 amino acids) are indicated with arrows. These were placed on the genetic map using pDRAW32 and were color coded to indicate the timing of transcription, green, early; yellow, middle; and red, late transcripts based on Luke et al ⁴⁴. Additionally, the position of two smaller T4 late genes soc and rI.-1 are indicated with red arrows near oriA and oriC, respectively.

Despite previous efforts, it has been unclear to what extent any of the putative origins are used during the course of a normal T4 infection, as incomplete representation of the T4 genome in the restriction fragment libraries used to map nascent DNA, disparate experimental conditions between studies, and incongruent experimental results have prevented a consensus. Without such basic information, it is impossible to determine how exactly DNA synthesis proceeds across the T4 genome, not to mention how the interplay between origin- and recombination-mediated replication and other DNA transactions contribute to the dynamics of replication and the fruition of full length genomes. Hence, in this study, we have mapped the origins of replication used during T4 infections and monitored the DNA synthesis originating from these loci, in real time, across the viral chromosome.

Results

The switch between origin and recombination mediated replication during infection

To begin our characterization T4 DNA synthesis during infection, we empirically defined the switch between origin and recombination-mediated synthesis during infection using viral recombination mutants. Needing an accurate way to measure viral DNA synthesis early during infection, we developed a simple dot blot assay, wherein DNA was harvested at various times post infection, blotted to nylon, and hybridized to a random-primed, ³²P labeled probe derived from full-length virus. The amount of T4 DNA associated with each time point was then measured with a phosophoimager, and accumulation over time was calculated as the fold-increase in viral DNA over that present at 2 minutes post infection, prior to the onset of replication. This allowed us to then estimate the average number of T4 genome equivalents in each infected cell at various times post infection.

We used a straightforward infection protocol, and E. coli was grown in LB at 37°C and directly infected with T4 at an MOI (multiplicity of infection) of 5 viruses to each cell with no additional manipulation of cultures. One advantage of the T4 system is that a population of viruses can be effectively synchronized during infection, and, under these conditions, host cells absorb more than 90% of the virus within 4 minutes of infection (data not shown). To control for experimental variances, multiple, independent infections were repeated on several different days for all viruses, and two replicates of each independent infection were blotted, probed, and averaged.

We chose to monitor the onset of recombination-mediated DNA synthesis using a virus harboring a mutation in uvsX, a recombinase gene. Many of the previously characterized uvsX mutants contain amber stop codons, which can be leaky, though the small amount of full-length protein expressed seems to have little affect on the much-reduced virus production ²⁸. In our hands, the uvsX mutant used in this study produced an average of 15 infectious virions per cell after a 1 hour infection, compared to 260 for wt (data not shown), essentially the same amount of virus produced by verified uvsX null mutants ²⁸.

As expected, the uvsX mutation caused an arrest in DNA synthesis, and mutant DNA accumulation was much reduced compared to wt (Figure 2A). However, this reduction began much earlier than anticipated and was noticeable as early as 7 minutes post infection, as is evident in the expanded graph (Figure 2B). This observation was confirmed with a second recombination mutant, containing a nearly complete deletion within gene 59, constructed as described in Materials and Methods. At an MOI of 5, gene 59 mutant infections showed similar amounts of DNA synthesis as uvsX mutant infections: The accumulation of viral DNA was clearly reduced compared to wt, and this reduction in DNA synthesis was apparent from at least 7 minutes post infection (Figures 2A and B).

Figure 2.

T4 DNA synthesis during wt and recombination deficient infections. The increase in T4 genomes over the course of infection was monitored by hybridization to viral T4 DNA as described in Materials and Methods. (A) E. coli BL21(DE3) cells were infected at a multiplicity of five viruses per cell with either wt T4D (n=6), uvsX mutant (recA homologue, n=2), or gene59 mutant (helicase loader, n=5), where n is the number of independent experiments. (B) The first 15 minutes of the infection in (A) are plotted on an expanded scale. (C) E. coli BL21(DE3) cells were infected at a multiplicity of 0.5 viruses per cell with either wt T4 (n=6), uvsX mutant (n=5), or gene59 mutant (n=3). Thus, on average, most infected cells contain a single virus. (D) The first 15 minutes of the infection in (C) are plotted. The symbols used in graphs are as follows: wt, open squares; uvsX mutant, open circles; gene59 mutant, open triangles. Error bars indicate standard error.

To determine whether our experimental conditions contributed to the early onset of DNA arrest observed with the recombination mutants, we repeated the experiment with the MOI reduced to 0.5 viruses per cell, so that each infected cell contained on average a single virus. As can be seen in Figures 2C and D, this change in the MOI had a profound effect on viral DNA synthesis. Under otherwise identical conditions, at an MOI of 0.5, the uvsX mutation has little affect on synthesis, compared to wt, until 15 minutes post infection. Thus the number of infecting viruses appears to influence the timing at which recombination begins to contribute to T4 DNA synthesis. Notably, replication of the gene 59 mutant at the lower MOI of 0.5 arrested earlier than the uvsX mutant, implying that 59 protein is required prior to the onset of recombination-mediated replication.

Mapping early nascent T4 DNA synthesis

Having defined the time course of origin dependent replication under our conditions, we next wanted to map the nascent DNA produced during T4 infection at an MOI of 0.5. The hybridization strategy used in Figure 1 proved to be quite useful in quantifying the amount of nascent DNA made over time, and we wanted to apply a similar strategy to the mapping of these synthesis products to discrete regions within the T4 genome. Hence we created an array of PCR products that provided equal representation of the entire T4 genome and could be used for mapping studies. This array is comprised of discrete 2.5 to 3.0 kb PCR products amplified from 30 loci, spaced evenly across the phage genome, and includes products positioned near seven putative origins of DNA replication, oriA, oriB, oriC, oriD, oriE, oriF, and oriG, identified in previous studies (see Figure 1).

Although the T4 replisome progresses at a rapid rate of 30–45 kb per minute ⁷, at the onset of replication in a population of infected cells, there should be some number of origin initiated, leading strands less than 10 kb, but longer than the 2 kb average Okazaki fragments ²⁹. To map these nascent DNAs, we harvested DNA from infections and fragmented contaminating E. coli chromosomal DNA with the restriction enzyme HaeIII and resolved the undigested T4 DNA on alkaline agarose gels. HaeIII cleavage is blocked by methylation of the internal cytosines within the recognition sequence 5′-GGCC-3′ ³⁰, and it does not appear to cut the hydroxymethylated cytosine containing T4 DNA produced early during our infections (data not shown).

The shorter, nascent viral DNAs were then excised from the gels, purified, random primed labeled, and used to probe the T4 genomic macro array. The enrichment of these size selected DNA’s at specific loci, and hence their origin, was determined using a relative abundance score. This score allows the abundance of DNA at a locus during infection to be compared to the abundance of DNA at that same locus in the packaged virus, before infection and subsequent replication, and is calculated as the ratio between the percentage signal at a given locus compared to all loci on the array and the same percentage obtained with a control probe derived from packaged viral DNA. Thus a relative abundance score at a particular locus greater than one indicates that this locus is enriched in nascent DNA as compared to the packaged, full length T4 genome.

A large enrichment at all three time points was observed in the nascent 3 to 6 kb wt DNA at a PCR locus near the putative T4 origin, oriE^{22; 24; 25} (Figure 3A). A similar enrichment pattern was also seen in the 6 to 10 kb size selected wt DNA, with the greatest relatives abundances observed near oriE (Figure 3B). There was also a less pronounced enrichment at another locus which overlaps a second previously identified T4 origin, oriF ^{18; 31}. Although the sequence elements required for oriE activity have not been delineated, oriE and oriF are thought to be mechanistically distinct ³². Thus, it appears that two functional classes of origins are active within a single population ofT4 infected cells.

Figure 3.

Relative abundance of size-fractionated nascent T4 DNA synthesized early during infection. Wild type T4D and the recombination uvsX mutant were used to infect E. coli BL21(DE3) cells at a multiplicity of 0.5 viruses per cell. Total viral DNA harvested from infections at 6, 7, and 8 minutes post infection was size fractionated on alkaline agarose gels and used to generate random primed probes as described in Materials and Methods. These labeled DNAs were then used to probe the T4 genomic macroarray, immobilized on nylon membranes, and the relative abundance of viral DNA at each locus along the array was calculated as described in Materials and Methods. (A) Wild type nascent T4D DNA, 3–6 kb. (B) Wild type nascent T4D DNA, 6–10 kb. The filled circles used in (A) and (B) are 6 minutes, black: 7 minutes, orange: 8 minutes, blue. (C) Mutant uvsX nascent DNA, 3–6 kb. (D) Mutant uvsX nascent DNA, 6–10 kb. The open circles used in (C) and (D) are 6 minutes, black: 7 minutes, orange: 8 minutes, blue. The position along the 168 kb T4 genome is identified in kilobases along the x-axis.

To determine if the synthesis observed in Figures 3A and B was true origin-dependent rather than recombination-dependent synthesis, size selected nascent DNA from uvsX mutant infections was prepared and hybridized to the array. Similar to wt, the greatest relative abundance in both size selected populations of mutant DNA was observed near oriE and oriF, indicating that DNA synthesis originating from these loci during the time monitored is not dependent on uvsX protein (Figures 3C and D). Thus it appears that there are at least two origins of de novo synthesis active in a population of infected cells under our conditions. Of these, oriE appears to be the most active.

Analysis of T4 replication dynamics

Though the previous analysis gives some indication as to the origins of phage DNA synthesis, our size selection strategies preferentially enriched for nascent DNA derived from origins at which 3–10 kb nascent DNA polymers accumulate. This could be problematic if, as one would expect, once initiated, nascent DNA polymers continue to grow at 30–45 kb per minute, until the entire genome has been replicated. By size selecting nascent DNAs, most of these predicted, growing polymers would be excluded from analysis, and with them, the ability to follow T4 DNA synthesis across the genome and illuminate the dynamics replication.

A second mapping strategy was employed to address these issues. Herein a time course of total T4 DNA harvested early during infection was blotted onto nylon strips, and these strips were independently probed with each of the PCR loci from the genomic array. This allowed the amount of DNA synthesized at a particular locus, at a particular time point to be calculated as the fold increase over the amount at the same locus present at 2 minutes post infection, before significant replication of the parental genome. An advantage of this methodology is that the fold-increase in DNA at a given locus should reflect the number of times that locus is replicated, providing a direct measurement of DNA accumulation. Hence this approach allows one to plot the synthesis of DNA across the genome in real time and to visualize the dynamics of viral replication.

Similar to our total viral DNA analysis, multiple, independent infections were repeated on several different days, and the results from each were averaged. During the first 8 minutes of wt infections a two to three fold increase in the amount of DNA was observed near oriE (Figure 4A), an expected result, given that the greatest enrichment of size selected nascent DNAs was also observed at the same PCR locus. As the infection progressed, the DNA accumulation near oriE continued, and by twelve minutes a nearly 10-fold increase was observed at this locus (Figure 4B), implying that on average replication was initiated from this region nearly 10 times per infecting virus during this time.

Figure 4.

T4 DNA replication dynamics in wt and recombination deficient infections. DNA synthesis was monitored across the viral genome using labeled PCR fragments from the T4 macroarray to probe blotted DNA from infections. The amount of viral DNA present at a given time point at a given locus was plotted as a fold increase over the DNA in the same infection at the same locus at 2 minutes post infection, as described in Materials and Methods. (A) E. coli BL21(DE3) cells were infected with wt T4D (n=3) at a multiplicity of 0.5 viruses per cell, where n is the number of independent experiments. (B) The same wt T4D infections at later time points. The filled circles used in (A) and (B) are 7 minutes, orange: 8 minutes, blue: 10 minutes, red; 12 minutes, purple; 15 minutes, green. (C) E. coli BL21(DE3) cells were infected with mutant uvsX (n=4) phage at a multiplicity of 0.5 viruses per cell. (D) The same uvsX infections at later time points. The open circles used in (C) and (D) are 7 minutes, orange: 8 minutes, blue: 10 minutes, red; 12 minutes, purple; 15 minutes, green. The position along the 168 kb T4 genome is identified in kilobases along the x-axis. To maintain graphic clarity only the upper extent of the standard error at each data point is indicated with error bars.

DNA also accumulated at several other loci with the T4 genome during the first 8 minutes of infection, including those near oriA, oriC, oriF, and oriG, as well as a discrete locus located between the latter two origins (Figure 4A). Moreover, the amount of DNA continues to increase near oriA, oriC, oriF, and oriG, and by 12 minutes post infection a ~5-fold increase in accumulation is observed near all of these loci (Figure 4B). Thus multiple T4 origins are active within a single population of infected cells, and these origins continue to fire, producing multiple nascent DNAs, as the infection proceeds.

Replication dynamics in the uvsX recombination mutant

Our data indicate that at an MOI of 0.5 both wt and uvsX mutant viruses produce essentially the same amount of DNA during the first 15 minutes of infection (Figure 2C and D). After that point, DNA synthesis in the uvsX mutant plateaus, and the virus exhibits the DNA arrest phenotype associated with T4 recombination mutants. The obvious conclusion is that T4 recombination-mediated DNA synthesis does not contribute significantly to phage DNA replication at the lower MOI until 15 minutes post infection. This conclusion is supported by the fact that short nascent DNA recovered early during uvsX infections maps to the same loci as wt (Figure 3). So it is not surprising that the pattern of total DNA accumulation in the uvsX mutant is also similar to that of wt virus, at least during the first 12 minutes of infection, with most synthesis mapping near oriE, and less mapping near oriA, oriF and oriG (Figure 4C and D). Yet, as the time course progresses to 15 minutes post infection, some distinct differences are observed between the two viruses, most notably that DNA accumulation near oriA is reduced during uvsX mutant infections, as compared to wt. This could indicate that the T4 genome contains at least two functionally distinct classes of origins, one more dependent on recombination proteins for activity under our experimental conditions.

Short nascent DNAs accumulate early during infection

If nascent DNA polymers grew continuously from their origin to the ends of the genome, one would expect to see the amount of DNA sequentially increase across the genome over time. So, as a single replication fork moved from an origin to adjacent loci, the amount of DNA should increase by a factor of two, initially at the origin and subsequently at adjacent loci. This dynamic should be easily measured by our assay, and, given a physiological replisome rate of 30–45 kb per minute ⁷, one would expect the two fold increase in DNA associated with each nascent replication fork to travel across the T4 genomic array at the rate of four to six adjacent loci per minute.

In contrast to the expected DNA synthesis dynamic, the fold increases initially observed at the origin loci do not move evenly across the genome with time (Figure 4). Instead, the pattern of T4 DNA synthesis early during infection is one of peaks and valleys, wherein 5 to10 fold increases in the amount of DNA are observed near the origins during the first 12 minutes of infection, but much smaller increases are observed in the regions adjacent to these loci. So it appears that nascent replication forks are moving slower than anticipated through some regions of the T4 genome and, as a result, short nascent DNAs are accumulating near the origins.

To directly determine the size distribution of nascent DNAs, viral DNA harvested early during infection was size fractionated on alkaline agarose gels and blotted to nylon membranes and hybridized to random primed probes derived from full length T4 DNA and a PCR product near oriE (Figure 5). To simplify the analysis, the fraction of DNA less than 27 kb was graphed for each time point, and, as expected from the replication dynamics in Figure 3, most of the DNA detected with either full length T4 (Figure 5B) or oriE probes (Figure 5C) was less than 27 kb, until 12 minutes post infection. After this time most of the DNA detected with both probes was greater than 27 kb.

Figure 5.

Size distribution of nascent T4 DNA synthesized early during infection. (A) Wild type T4D and the recombination mutant uvsX were used to infect E. coli BL21(DE3) cells at a multiplicity of 0.5 viruses per cell. Total viral DNA, harvested from wt and uvsX infections at 2, 6, 8, 10, 12, 15, and 20 minutes post infection, was size fractionated on an alkaline agarose gel, transferred to nylon and probed with full length T4 DNA as described in Materials and Methods. Numbers above lanes indicate minutes post infection and M identifies the PacI digested viral DNA molecular weight standard. (B) The blot in (A) was exposed to a phosphoimager screen and scanned. To determine the fraction of DNA shorter than 27 kb, the PSL (detected radiation) in a given lane associated with 2.7 to 27 kb DNA was divided by the PSL present in the total DNA in the same lane, from 2.7 kb to the top of the well, as described in Materials and Methods. Results from wt infections are graphed with filled bars and uvsX infections with open bars. (C) The blot in (A) was stripped and reprobed with a PCR fragment that overlaps the putative position of oriE, LP16 from Table 1. The fraction of DNA less than 27 kb was then calculated as in (B).

It is worth noting that the smaller DNAs observed early during the wt infection do not simply disappear, as would be expected if a certain number of nascent polymers were initially made, then elongated over time. Rather a certain abundance of short DNAs are apparent throughout the first 20 minutes of wt infection, suggesting that they are produced at a more or less constant rate during this time. This does not occur in uvsX mutant infections, where the abundance of short DNAs increases throughout the time studied (Figure 5A), as if short polymers were still being made, but not efficiently elongated, culminating in a larger percentage of polymers less than 27 kb, compared to wt. This observation is consistent with a previous study ³³, and indicates that uvsX influences the normal size distribution of nascent DNA.

Discussion

The chromosome is the fundamental unit of heredity, a dynamic macro molecular structure that serves as a template for transcription of genes, replication of genetic material, and the various DNA transactions required for genome maintenance. As a first step in characterizing the interplay between these activities, we have developed a unique set of experimental strategies that allowed us to measure the actual accumulation of nascent DNA, across the T4 genome, over the course of infection. Our data indicate that early T4 DNA synthesis is discontinuous, with several origins contributing to a punctate replication pattern, comprised of multiple, short, origin derived polymers, an observation that raises a number of questions regarding the regulation of T4 DNA replication.

Transition from origin- to recombination-mediated replication

The maintenance of genome integrity during multiple replication cycles is a key priority in the evolution of chromosomes. Lacking defined telomeres, the linear T4 chromosome has evolved a unique replication strategy that includes both origin- and recombination-mediated DNA synthesis. This bipartite strategy ensures the production of full length genomes and allows the virus to rapidly synthesize several hundred progeny genomes during the short time course of a single infection. Although the individual components of this replication strategy have been studied in vitro, little is known about the interplay between these two modes of replication and the dynamics of resultant DNA synthesis during actual infection.

The switch between these two modes of replication was previously thought to depend on the synthesis of proteins made at later times ^{15; 16}. However, we found that the interval prior to the onset of recombination-mediated T4 replication is not a fixed duration. Rather, it varies with the number of viruses per infected cell. In singly infected cells the recA homologue uvsX does not contribute to the total amount of viral DNA produced during the first 15 minutes of infection. In contrast, when multiple viruses infect the same cell, under otherwise identical conditions, the transition to recombination-mediated replication occurs much earlier, and both uvsX and gene59 are required for viral DNA synthesis as early as 7 minutes post infection.

It is not clear how increasing numbers of infecting viruses influence the T4 replication strategy such that recombination is favored, but scarcity of required host factors such as RNA polymerase or altered expression of specific viral genes could cause an imbalance in protein levels, creating an environment wherein origin activation proteins are limited or that recombination proteins are overly abundant. Of course, increasing the number of viral copies may also increase the substrates available for recombination and, thus, may in itself be a factor in the switch to recombination-mediated replication.

It is also not clear why replication slows earlier in the gene 59 mutant compared to the uvsX mutant in singly infected cells. Taken at face value, this observation implies that gene 59 is required prior to the onset of recombination-mediated replication. However, it is not likely that 59 protein is absolutely required during the initial steps of origin- mediated replication, since one would expect no DNA synthesis in gene59 mutants if this was the case. Rather, our observations are consistent with secondary role for 59 protein during origin-mediated replication, as asserted in a previous study, where mutation of gene 59 had no effect on initial leading strand synthesis from oriF but disrupted subsequent retrograde origin-mediated synthesis ¹⁴. Such a role at the T4 origins would presumably cause DNA synthesis to arrest earlier in gene 59 mutants than in uvsX mutants.

Multiple origins of T4 DNA replication

There are clearly multiple origins of T4 DNA replication, including a major origin near oriE and at least four secondary origins located near oriA, oriC, oriF, and oriG. Unlike previous studies ^{22; 24; 25}, we observed synthesis near these origins concomitantly, within a single population of infected cells, raising the possibility that some or all of these origins are simultaneously active within a single viral genome. However, this issue is still unresolved, and it is entirely possible that micro-environmental differences between infected cells influence origin activity such that particular origins are used discretely within individual infected cells.

There is some reason to suspect that environmental factors may influence origin usage, as the characterized T4 origins do not appear to be mechanistically similar. Of the five, the best characterized is oriF. This origin was initially recovered by Kreuzer and Alberts and is able to support replication of episomal plasmids during T4 infection ^{18; 27}. Normal origin activity from this locus is dependent on several cis elements, including a T4 middle mode promoter ^{34; 35}, a downstream unwinding element or DUE ³¹, and a sequence element thought to form a hairpin structure ³⁵. The current model predicts that the T4 middle mode uvsY promoter upstream from oriF produces an RNA, which is then used as primer for DNA synthesis by the T4 replisome. Both oriA and oriG are also located near middle mode promoters, and each is thought to operate in a similar manner to oriF ²¹.

The region near oriE appears to be quite different than that near oriF. There is no middle mode promoter in the region, which lead Gisela Mosig’s group to suggest that DNA replication is primed by an RNA produced from an early promoter downstream from six copies of a directly repeated sequence element, 5′-AT(T/C)(T/A)CC(A/T)T(T/C)(A/G)AC-3′ ³². These so called iterons are each separated by 12 bp of non-repeated sequence and are maintained in the anti-sense direction of a large open reading frame encoding the gene 5 base plate lysozyme, a protein necessary for viral entry into host cells. Similar repeats are found in the syntenic gene 5 region regions of other T4-like bacteriophages ³⁶, but it is not clear if these iterons are functionally linked either to gene 5 or to origin activity from oriE.

Thus T4 may have mechanistically distinct classes of origins, and the difference between these classes could contribute to their functional usage during infection. This would make sense on at least two levels. First, as genomes evolve, one would expect them to encounter multiple selective pressures over time, which may lead to the maintenance of several types of origins, allowing the organism to reproduce under a variety of environmental conditions, and, hence, increasing fitness. Second, the individual origins may fulfill specific functions during the replication of a single viral chromosome. Yet, it remains entirely unclear how T4 origin activity is regulated, and how several origin loci, with distinct genetic elements, function during the course of infection.

Punctate patterns of T4 DNA synthesis

Early during infection, short, nascent DNAs accumulate near the T4 origins, suggesting that these loci function as amplicons, producing a number of sub genomic length DNAs that can be subsequently used during recombination-mediated synthesis (Figures 3 and 4). However, the physical state of these origin-derived DNAs is not clear. Perhaps, initially, each nascent DNA is displaced from the origin and later recombines with the parental template where it is used to prime subsequent DNA synthesis. This might explain why the proportion of short nascent DNA remained high in the absence of the RecA homologue uvsX, a protein involved in both recombination and the restart of stalled replication forks ³⁷. Alternatively, the nascent polymers may remain associated with the template, creating an “onion skin” structure wherein the multiple nascent strands are held in place inside the parental duplex. These structures have been seen before in electron micrographs of replicating T4 DNA ³⁸, though not the 5 to 10 layers implied by the data presented here.

It is not clear why replication forks do not efficiently progress into the regions adjacent to the T4 origins early during infection. Fork progression is regulated by both active and passive mechanisms in other systems. Terminator sequences impede E. coli DNA synthesis in a defined polarity through a mechanism mediated by tus proteins ^{1; 39}, but there is reason to suspect that other mechanisms contribute to discontinuous DNA synthesis observed throughout the E coli genome ^{3; 4}. Both replication pause sites and initiation rates influence DNA synthesis patterns in Epstein-Barr virus ⁴⁰. In eukaryotic cells DNA synthesis rates differ from one genomic region to another ⁴¹, and several factors, including local and global transcription patterns appear to influence replication dynamics ^{5; 42}.

No replication terminators have been identified in T4, and though it has been suggested that transcription may impede T4 DNA synthesis ⁴³, the relationship between viral transcription and replication is not established. There has been a genomic study of T4 transcription during infection using a microarray, but, unfortunately, the conditions used in this study were different than ours, preventing any direct comparisons of replication and transcription timing ⁴⁴. Yet, it is worth noting that all of the origins of replication identified in Figure 4 are near T4 genes expressed during late transcription (see Figure 1), suggesting that further investigation into the interplay between T4 transcription and replication is warranted.

Certainly many aspects of the T4 replication dynamics observed in this study await further experimental exploration. Yet, the assays developed here should allow us to directly determine the genetic factors required for the normal pattern of DNA synthesis, and further development of these assays should provide the experimental tools necessary to define the mechanisms of origin function. The amenability of T4 to both the genomic investigations as demonstrated here and biochemical characterization demonstrated by a number of groups should provide a unique system where genomic phenomena observed in vivo can be quickly correlated with biochemical mechanisms defined in vitro. Ultimately, such studies should shed light on the dynamics and underlying mechanics of genomic transactions, and how these contribute to the evolution of chromosomes.

Materials and Methods

Strains

E. coli BL21(DE3) and SCS110 were obtained from Stratagene. E. coli B and CR63, as well as wt T4D, have been maintained in this laboratory. The T4 recombination mutant T4uvsXam11 was a gift from John W. Drake ⁴⁵.

Growth of bacteria and phages

All bacteria were grown in LB broth at 37°C. All T4 infections were done at 37°C. T4 isolates were plaque purified from stock and expanded in either E. coli CR63 or BL21(DE3) cells. Phage stocks were stored at 4°C in 10mM Tris (pH7.4), 150mM NaCl, 0.03% gelatin.

Construction of T4 phage harboring a deletion of gene 59

The T4 gene 59 deletion was made by site-directed mutagenesis of the plasmid pNN2002 ⁴⁶, which contains wild type gene 59 and flanking T4 DNA. The method of Kunkel et al. ⁴⁷ was modified to include the use of T4 DNA polymerase, T4 44/62 clamp loader, and 45 clamp to copy the ssDNA template as previously described ⁴⁸. Two primers, 59del1 (5′-GCATGCGGAGTTTGATCATAGTAGAAAATC-3′) and 59del2 (5′-CAATACTTGCAAGTGATCACAGTTTCAATG-3′), were used to introduce BclI restriction sites into pNN2002 at the termini of the gene 59 open reading frame. Mutated plasmids grown in the dam methylation deficient host E. coli SCS110, were digested with BclI. The larger fragment was self ligated to give a circular plasmid, pJRB1113, in which all but six codons, of the gene 59 open reading frame had been deleted. The gene 59 deletion thus contains a very small open reading frame that codes for the amino acid sequence MITCKY.

The gene 59 deletion was introduced into the T4D viral background by homologous recombination as described ⁴⁹. Briefly, E. coli B (pJRB1113) was grown in LB supplemented with carbencillin (50 ug/ml) to an OD600 of 0.28 and infected with T4D at a multiplicity of infection of five viruses per cell. Infected cultures were incubated at 37°C with vigorous shaking for 2 hours, and then lysed with chloroform. Resultant phage stocks were titered and plated on E. coli BL21(DE3) cells harboring the gene 59 expression plasmid pNN2859 ⁴⁶. Plaques were lifted onto nitrocellulose, and probed with an oligonucleotide that spans the gene 59 deletion, 59del3 (5′-TAGATTTTCTACTATGATCACTTG-3′), in 6x SSPE (3M NaCl, 0.2 M NaH₂PO₄, 0.05 M EDTA), 10% dextran, 1% SDS at 50°C. Filters were washed twice in 2x SSPE, 1% SDS at 50°C. There was no hybridization of the gene 59 deletion probe with wt T4D plaques under these conditions.

Six plaques harboring gene 59 deleted phage were identified and isolated in this first screen. These plaques were resuspended in LB, titered, and plated again on E. coli BL21(DE3) cells harboring pNN2859. A secondary plaque hybridization screen was then done using the gene 59 deletion oligonucleotide as above, and several plaques were isolated. A stock designated T4gene59Δ was expanded from one of these plaques. The mutant phage yields pin point plaques on E. coli BL21(DE3) cells and normal sized plaques when complimented with pNN2859.

Isolation of T4 DNA from Infected Cells

E. coli BL21(DE3) host cells were grown to a density of 3 × 10⁸ cells per ml at 37°C. and infected in parallel with pre-warmed wt and mutant phage. After addition of phage to bacterial cultures at 37°C and through mixing for 1 min, phages were allowed to absorb for an additional 45 seconds without mixing. Samples were then mixed again for 15 seconds, then mixing was stopped and two minute samples were withdrawn. Mixing was again started and vigorous mixing was then maintained, except for brief stops to withdraw samples. Aliquots of infected cultures were withdrawn and immediately submersed in ½ volume phenol and ½ volume chloroform. Tubes were inverted 5–7 times to completely lyse the cells. After all aliquots had been withdrawn, all phenol/chloroform extractions were inverted another 10–14 times and centrifuged to separate aqueous and organic phases. The recovered aqueous phase was extracted with 1 volume of chloroform and stored at 4°C overnight.

Quantification of in vivo T4 DNA synthesis

T4 DNA synthesis in vivo was monitored by a dot blot assay. Recovered aqueous phase aliquots from the chloroform extractions detailed above were digested with 40 units/ml RNAse If (New England Biolabs) at 37°C for 30 minutes. Digested samples (100 uL) were denatured by the addition of 1 volume of 0.5M NaOH, 1.5M NaCl at 65°C for 10 minutes and cooled to room temperature. Denatured aliquots were applied to Hybond-XL membranes (Amersham Biosciences) using a Minifold-1 dot blot system (Schleicher and Schuell) in accordance with manufacturer’s instructions. Blots were dried at room temperature for 15–30 minutes, neutralized in 6x SSPE (3M NaCl, 0.2 M NaH₂PO₄, 0.05 M EDTA, pH 7.4), and dried completely.

In preparation for hybridization, blots of viral DNA samples were pretreated in 6x SSPE (pH 7.4), 1% SDS at 62°C for 4 to 6 hours. This buffer was then replaced with fresh hybridization buffer, 6x SSPE (pH 7.4), 1% SDS, 10% dextran. Blots were hybridized to probes generated from full length T4 DNA isolated from purified virions or from PCR fragments amplified from the T4 genome. In either case, probes were made using the Prime-It II random primer labeling kit (Stratagene), cleaned using ProbeQuant G50 spin columns (GE Healthcare), and denatured at 95°C for 10–15 minutes prior to hybidization with blots at 62°C for 12 to 16 hours. Approximately 5 × 10⁵ cpm of probe per ml of hybridization solution was used.

After hybridization, blots were washed 3 times in 2x SSPE (pH7.4), 0.1% SDS at room temperature for 30 minutes each, and once in 2x SSPE (pH7.4), 0.1% SDS at 62°C for 30 minutes. These hybridization and wash conditions were empirically determined to prevent cross-hybridization of T4 DNA with E. coli DNA as judged by dot blotting. Washed blots were exposed to a phosphoimager screen, which was scanned by a FUJIFILM FLA-3000 phosphoimager. Data was analyzed using FUJIFILM Image Guage V3.12 software.

The amount of DNA present at a given time point was calculated as a fold increase over the amount in the same infection at the 2 minute time point, prior to the onset of viral replication. In infections done at an MOI of 0.5, each infected cell contains on average a single infecting virus, and the amount of DNA at 2 minutes post infection is equivalent to one T4 genome per infected cell. Thus the fold increase in DNA over that at 2 minutes is equal to the average number of genomes per infected cell. In contrast, at an MOI of 5, on average each cell is infected with five genomes, and so that the fold increase in DNA is not over a single genome, but five genomes, and the fold increase must be multiplied by 5 to determine the average number of genomes per infected cell.

Amplification of the T4 macro array

The PCR fragments that comprise the T4 macroarray were amplified from full length T4 DNA isolated from purified virions using the Expand Long Template PCR kit (Roche) with buffer 2 as directed by the manufacturers instructions using 150 ng of T4 DNA in each 50 uL reaction. The PCR primers are listed in Table 1 and were designed such that all hybridized to the T4 genomic template at the same temperature, allowing all the reactions to be run concurrently in the same thermal cycler. The thermal cycling was done in a GeneAmp PCR system 9700 (Applied Biosystems) under the following conditions: 10 cycles of 92°C for 2 minutes, 55°C for 30 seconds, and 68°C for 4 minutes followed by 25 cycles of 10 cycles of 92°C, 2 minutes, 55°C for 30 seconds, and 68°C for 4 minutes plus an additional 20 seconds each additional cycle. Under these conditions each primer pair produced a single band when visualized on EtBr stained agarose gels.

Table 1.

Oligonucleotides used to construct the T4 macro array

PCR Locus	Map Position	Oligonucleotide sequences
LP1	168102 1413	5′-TTACAACTCGGCTGGTTGAACTACC-3′ 5′-ACCGTAACTGGCTAAGCATTCGC-3′
LP2	4033 6512	5′-TACCACGCAATGGATAACCACCG-3′ 5′-GGTGTTGATACTCTTCTGAAACGTATCGC-3′
LP3	9559 11721	5′-GCGCTTTATGGAATGTTGATGCAGG-3′ 5′-CGTTTAGAGCATACTCAGGCATCTGC-3′
LP4	14745 16978	5′-AAGAACTTCCACATGAATCACCTGCC-3′ 5′-GGGCAGTCGAATGTTTTAGAACACCAC-3′
LP5	19392 21773	5′-CACCGACATACTCTGTGCTGCG-3′ 5′-TCTCCGGAAGACCACAGCTGG-3′
LP6	24549 27103	5′-CTCATTGACTCTATCAATGAGTTCAGCACG-3′ 5′-AGATGATTCTGGACCCTTTGGATTCC-3′
LP7	29048 31963	5′-GCTGCTAAATTTGAAGGTGAACACGC-3′ 5′-CAACTGATTCCAAAGAGAATGACGGC-3′
LP8	34525 37124	5′-GCAGGTCTTCCACAAGCTTTCTTCG-3′ 5′-TGGTAAATCCACTGTAATGGAAGGTCTGG-3′
LP9	39786 42055	5′-CACGGCTGTCATAGACATTGTGAACG-3′ 5′-CCAGATGAAGTAGGTCGTTGTCCTGG-3′
LP10	44749 46729	5′-AAGCTAATCACCTCGACCATGACC-3′ 5′-TCATTCGTTCAGTCCAGTCAGTTCC-3′
LP11	49197 52446	5′-TCGTATTATCATCAAAGGATTCAACTACGTCC-3′ 5′-GGTGCTCGATTGGCTAAAGGTCTAG-3′
LP12	55230 57809	5′-TCTCCTTGTGCATCTACCGGAGC-3′ 5′-TGTAATGATGAGCGAGGCGAAACG-3′
LP13	60217 62007	5′-GAGAAACGACTTCACAGACAGAATCGC-3′ 5′-GCATCACGCCTACGGAATGTTCC-3′
LP14	65319 67876	5′-CGATATTCAGCCATTAAGTGTTGGTCAGC-3′ 5′-TGATAGAGTTGATTGGAATGGTTGTTCCG-3′
LP15	71150 74155	5′-GAACCATCAGTCCGACGACTTACC-3′ 5′-CAAGATGGATCGACTCTTGTACTTGTGC-3′
LP16	76349 79504	5′-CTGACGAGCTTTAGGTGGAATATAGTGC-3′ 5′-TTCCTTCAACTAAAGTGGTAGCATCTCC-3′
LP17	82520 85510	5′-GATCATTCTGTCATTGGTTCATCAGCG-3′ 5′-ACTTCTTGACCAGCAATTAATCGTCCC-3′
LP18	88424 91406	5′-CCAGTTACACTAGTAGCTGCTTCCG-3′ 5′-TGACCACCTCTTTGCTGGATAACG-3′
LP19	94422 96976	5′-CCATTCCTCAACCATTTCACGAGTCC-3′ 5′-CGGTCAAATTTTAGCACGCCACC-3′
LP20	100895 103871	5′-TGCTCGTCAGAAGTTCGTTGATTGG-3′ 5′-TTGTGTCAGTCAATGAACCTAATCCACG-3′
LP21	107170 110021	5′-TCGTCAAGAGTAGCATGAGCTCCG-3′ 5′-TCTGGAAAGTCATTCATTGAAATTGCTCG-3′
LP22	112743 115578	5′-CATCGATTGTGAAGAGAGCACGTTCC-3′ 5′-CTTACTGCTGACACGGTTGAACGC-3′
LP23	119325 122020	5′-GTTGAGGATGATACTGCTGAATCTGTGG-3′ 5′-GAACTGTACGTTGACTCATGGCACC-3′
LP24	125094 128001	5′-TAAGAAGGACGTATCTACCACCTGAACG-3′ 5′-TGTTACAGCATTGGGTGATTCTATTGACG-3′
LP25	131764 134628	5′-GGCATAAAGTCTGCATCACGAGTACC-3′ 5′-TAACCTGAATCCTGAACGTCGCC-3′
LP26	138347 141120	5′-CCATCAGTACCGGATTGAAGTTGACG-3′ 5′-GTTACCAACTATGCAGCTTGGCTGG-3′
LP27	144009 146978	5′-ACTTATGGGTTTCTGGATTCCAACGC-3′ 5′-GGTAGTTCATCTAGTGCTGATGACACG-3′
LP28	150898 153249	5′-ATCAACGTAGCTTTAGCTGATCGTACC-3′ 5′-AGTAGGTTCACTCTGCGCAATCC-3′
LP29	156094 159006	5′-TGGCATTCAAGTCCCATATGCTCC-3′ 5′-AATCCAGCAGGAACTGAATCACTCG-3′
LP30	162235 165005	5′-CCACGTTCAATACTTACTTGTCCAGCG-3′ 5′-CACAAACTCGCTTCTATTGCAGGTGG-3′

The blotting of PCR products to Hybond-XL membranes was done essentially as described above. A total of 50 ng of each PCR product was blotted in 0.5M NaOH, 1.5M NaCl. Blots were hybridized to probes generated from T4 DNA harvested from infections as above. A control probe generated from full length T4 DNA isolated from purified virions was also used in duplicate. The blots were washed and scanned as described above. The relative abundance of DNA at a particular locus was then calculated ^{24; 25}. First, the percent abundance of a given locus in each DNA sample was determined by dividing the number of phosphoimager units (PSLs) associated with an individual PCR fragment probed with that DNA by the sum of PSLs at all PCR fragments in the array when probed with the same DNA. Second, the percent abundance of each experimental data point was standardized by dividing this quotient by the percent abundance of the same PCR locus, when probed by control DNA purified from T4 virions in parallel hybridizations.

Size selection of nascent T4 DNA

DNA recovered from infections as described previously in this section was precipitated with the addition of isopropanol. This DNA was resuspended in 10mM Tris-HCL, 10mM MgCl2, 50mM NaCl, 1mM EDTA (pH 7.9) and digested for 1 hour at 37°C with Hae III, to fragment bacterial DNA, and Rnase If (New England Biolabs). DNA was then ethanol precipitated at room temperature for 15 minutes and loaded onto a 1% alkaline agarose gel in 50mM NaOH, 1mM EDTA. New England Biolabs 1kb ladder was loaded as a size marker. The DNA was resolved at 25 volts (160mA) for 16 hours, and the gel was neutralized and stained in 1x TBE (0.4ug/mL ETBr) for 1hr. DNA was visualized with UV light, excised from the gel, and purified using the Qiagen Qiaex II kit.

T4 DNA replication dynamics

T4 DNA synthesis in vivo at the loci along the T4 macroarray was monitored by a dot blot assay. Recovered aliquots of DNA from infections were processed and blotted to nylon membranes as described above. These blots were then hybridized with random primed probes generated from each of the 30 PCR fragments that comprise the T4 genomic macroarray. The amount of DNA present at a given locus at a given time point was then calculated as a fold increase over the amount at the same locus in the same infection at the 2 minute time point.

Southern blots of DNA from infected cells

DNA for Southern blotting was prepared from infections as described above. Hae III, RNAse If (New England Biolabs) digested DNA samples were then resolved on a 0.7% alkaline agarose gel in 50mM NaOH, 1mM EDTA, at 30 volts (200 mA) for 21 hours. A lane containing Pac I digested T4 DNA was included as a size standard. The gel was equilibrated in 0.5N NaOH, 1.5M NaCl for 1hr, and blotted to Hybond-N+ (Amersham) in the same buffer overnight by capillary action. Southern blots were probed as described previously in this section and stripped according to manufacturer’s instructions.