Efficiency and Frequency of Translational Coupling between the Bacteriophage T4 Clamp Loader Genes

Abstract

The bacteriophage T4 DNA polymerase holoenzyme is composed of the core polymerase, gene product 43 (gp43), in association with the “sliding clamp” of the T4 system, gp45. Sliding clamps are the processivity factors of DNA replication systems. The T4 sliding clamp comes to encircle DNA via the “clamp loader” activity inherent in two other T4 proteins: 44 and 62. These proteins assemble into a pentameric complex with a precise 4:1 stoichiometry of proteins 44 and 62. Previous work established that T4 genes 44 and 62, which are directly adjacent on polycistronic mRNA molecules, are—to some degree—translationally coupled. In the present study, measurement of the levels (monomers/cell) of the clamp loader subunits during the course of various T4 infections in different host cell backgrounds was accomplished by quantitative immunoblotting. The efficiency of translational coupling was obtained by determining the in vivo levels of gp62 that were synthesized when its translation was either coupled to or uncoupled from the upstream translation of gene 44. Levels of gp44 were also measured to determine the relative stoichiometry of synthesis and the percentage of gp44 translation that was transmitted across the intercistronic junction (coupling frequency). The results indicated a coupling efficiency of ∼85% and a coupling frequency of ∼25% between the 44-62 gene pair during the course of infection. Thus, translational coupling is the major factor in maintaining the 4:1 stoichiometry of synthesis of the clamp loader subunits. However, coupling does not appear to be an absolute requirement for the synthesis of gp62.

Genes 44 and 62 of bacteriophage T4 produce two proteins that are essential for the replication of the phage DNA (1). These proteins assemble into an oligomeric complex, gp44-gp62 (gp44/62), that possesses very low intrinsic ATPase activity. The ATPase activity of the gp44/62 complex is greatly stimulated upon its specific interaction with another T4-encoded protein, gp45 (14, 23). gp45 is the “sliding clamp” of the T4 system and is structurally homologous to the other known sliding clamps, namely, the β subunit of Escherichia coli and the eukaryotic proliferating cell nuclear antigen (20, 21). In vitro studies in a variety of laboratories have led to the consensus view that the role of gp44/62 during a T4 infection is to chaperone the gp45 protein to DNA (16, 30). Upon binding to DNA, the gp44/62 complex imparts a transient discontinuity to the ring-like structure of the T4 sliding clamp, resulting in the gp45 protein being topologically deposited onto double-stranded DNA (3, 22). That is, the gp44/62 complex is a “clamp loader.” Although it is presumed that in vivo this function is driven by the hydrolysis of ATP, the exact molecular mechanism by which this event occurs remains unknown. The role of gp44/62 as a clamp loader is catalytic (16). Once loaded onto DNA, the gp45 sliding clamp subsequently associates with the T4 DNA polymerase (gp43) to form the processive holoenzyme (9, 17, 31, 34).

The T4 clamp loader is assembled upon the tight binding of one gp62 subunit to a gp44 tetramer (15, 35). The in vivo mechanism by which this strict 4:1 stoichiometry of the gp44/62 complex is maintained remains unknown. It is to be appreciated that the gp44/62 complex is similar in both sequence and function to the E. coli γ complex and eukaryotic RF-C (19, 28). Furthermore, the latter two clamp loaders are also composed of five subunits; however, the implications of the conservation in evolution of this pentameric stoichiometry have yet to be determined.

The 44 and 62 genes occur in a region of the T4 genome clustered with other genes required for DNA replication. The genes of the T4 replication complex are arranged so that they may be transcribed as a cassette including genes 45, 44, 62, regA, and 43. They are expressed early in the infection cycle, and the transcripts are polycistronic (11). Transcription proceeds from early or middle promoters upstream of gene 45. There are strong consensus translation initiation regions (TIR) serving genes 45 and 44. The polycistronic message is subject to at least two known forms of translational regulation. One is repression due to direct binding of the RegA protein to the TIR of gene 44 (37). The second form of regulation is translational coupling (see references 6, 13, and 25 for reviews). This event occurs when translation of a distal gene on a polycistronic mRNA is strongly, or exclusively, dependent upon translation of a gene immediately upstream. In 1979, the laboratory of Karam initially provided evidence that the expression of gene 44 and that of gene 62 are translationally coupled (18). However, reflecting an inability to detect the low levels of gp62 produced, the precise degree of translational coupling was not determined by those investigators. In the current study, by employing specific antibodies and an extremely sensitive chemiluminescence detection protocol, we have quantitatively determined the levels of gp62 present in vivo during T4 infections. We have also examined the degree of translational coupling as a first step in determining how strict stoichiometry may be established for the gp44/62 complex.

MATERIALS AND METHODS

Strains, media, and growth conditions.

E. coli B strain Nap IV (sup-1 hsdR_K hsdM_K⁺ hsdS_K⁺ thi) (26) was used as the suppressing host for T4 amber mutant 44amN82 (gene 44) or 43amB22 (gene 43). sup-1 is a serine-inserting amber (UAG) suppressor. The wild-type (sup-0) counterpart of the Nap IV strain was used as the nonsuppressing host for T4 amber mutants, as well as for infection by wild-type bacteriophage T4 strain D (T4D⁺). Cells were grown at 37°C in glycerol Casamino Acids medium supplemented with thiamine (10 μg/ml) in a rotary shaker to an A₆₀₀ of ∼0.6, which corresponded to a cell density of ∼1.1 × 10⁹/ml. Standard plate count assays were performed to determine the exact cell density. The cells were then infected (multiplicity of infection of 5) with bacteriophage that had been previously mixed with a small volume of phage diluent buffer (5) containing l-tryptophan (final concentration of 5 μg/ml). The number of uninfected cells at 1 min after phage addition was determined by the standard plate count assay to be 1% or less. This protocol ensured a nearly simultaneous infection round and yielded the actual number of infected cells. At certain time points during the infection round, 0.25-ml aliquots were withdrawn and phage development was immediately arrested by the addition of 50 μl of 6× lysis buffer (0.37 M Tris [pH 6.8], 6 mM dithiothreitol, 6% [wt/vol] sodium dodecyl sulfate [SDS], 30% [vol/vol] glycerol, 0.01% bromophenol blue). The aliquots were boiled for 5 min and then stored frozen at −20°C.

Polyclonal antibody against gp44 and gp62.

Anti-gp62 and anti-gp44 sera were produced in collaboration with Cynthia Sommer (Biological Sciences, University of Wisconsin-Milwaukee) and Berri Forman (Animal Care Facility, University of Wisconsin-Milwaukee) by following standard protocols. The antibodies were affinity purified from serum on a protein A column (5-ml cartridge) in accordance with manufacturer (Bio-Rad) specifications. The resulting antibodies were tested by immunoblotting against cell lysates and found to have high specificity for gp62 or gp44, although some cross-reactivity with E. coli proteins was retained.

Quantitative immunoblotting.

Immunoblotting and detection by chemiluminescence was employed to quantify the gp44 and gp62 levels in the lysed time point aliquots. The aliquots were sonicated to shear DNA, and equal numbers of infected cells (in 6- to 12-μl portions of the aliquots) were then fractionated via discontinuous SDS-polyacrylamide gel electrophoresis on a 15% polyacrylamide minigel. Electrophoresis was performed in 49 mM Tris–384 mM glycine–0.1% SDS (pH 8.5). Fractionated proteins were electroblotted in a high-pH transfer buffer [10 mM 3-(cyclohexylamino)-1-propanesulfonic acid (CAPS), pH 11; 10% methanol] onto a nitrocellulose membrane (Life Technologies, Inc.) by using a Mini Trans-blot apparatus (Bio-Rad). The blots were blocked overnight at 4°C with 0.2% (wt/vol) casein and then incubated with anti-gp62 and/or anti-gp44 antibodies in phosphate-buffered saline. The blots were developed by using a chemiluminescence system with the reagents and alkaline phosphatase-conjugated secondary antibody supplied by the manufacturer (Tropix).

For quantitation of gp44 and gp62 bands, serial dilutions of purified gp44/62 were included on every gel. Via this method, gp44 and gp62 could be detected at 50 and 10 pg per lane, respectively. Quantitation of gp62 over a linear range of 20 to 900 pg was achieved by using a standard dilution-response curve constructed for each immunoblot. Quantitation of gp44 and gp62 was performed by comparing the intensities of bands on Hyperfilm ECL (Amersham), using a nonspecific cross-reactive E. coli protein band as an internal reference for the amount of lysate loaded per lane. Due to the narrow linear response range of the film, several exposures were necessary to assign the correct optical density (OD) value to the low and high ends of the gp44 and gp62 standards. The photographic negatives of the immunoblots were scanned on a Color OneScanner (Apple) set to a resolution of 600 dots/in. in the grayscale mode. The scanned images were converted to a histogram with NIH Image 1.62 software, through which the OD could be plotted as a response profile. The total OD of a desired band was obtained by defining a rectangular area around it, summing the OD along the short axis of the rectangle, and then using the wand tool facility to measure the peak area in the resulting one-dimensional plot. Irregular baselines were defined by plotting a section adjacent to the bands. The areas under the histogram peaks were quantitated. The ratios of the ODs of any two neighboring gp44 or gp62 standards at the low and high ends of the range at two (or more) exposures were averaged and the mean value was used to correct the otherwise over- or underexposed standard at the high or low end, respectively.

RESULTS AND DISCUSSION

Synthesis of gp44 and gp62 by 44amN82.

In the course of T4 infection, a heterogeneous collection of mRNAs containing the coding sequences for genes 44 and 62 is produced. In all cases, gene 62 is found directly downstream of gene 44, and translation of gene 62 is known to be linked to translation of upstream gene 44. Previously, translation of gene 62 has not been observed in the absence of complete gene 44 translation. Karam et al. were unable to directly observe production of gp62 with gene 44 amber mutants 44amE4408 and 44amN82, although low levels of gene 62 translation were reasoned to occur (18). Further, gp62 synthesis was not observed when gene 62 was present on an expression vector in the absence of the initial portion of gene 44 (33). In this study, we have determined the levels of gp44 and gp62 produced during infections by 44amN82 under suppressing and nonsuppressing conditions via quantitative immunoblotting. We have also been able to demonstrate the synthesis of gp62 in the absence of complete gene 44 translation.

The results of infections of the E. coli suppressing and nonsuppressing hosts by 44amN82 are shown in Fig. 1. In the infection of the suppressing host (sup-1), gp44 and gp62 were easily detectable by 3.5 min postinfection (p.i.) at levels of 230 and 70 monomers per cell, respectively. At 5 min p.i., these levels increased to 1,400 gp44 and 400 gp62 monomers per cell. A final measurement was taken at 10 min p.i., when early protein synthesis is nearly complete. At that point, the level of gp44 was 6,700 monomers per cell, while there were 1,600 monomers of gp62 per cell.

FIG. 1.

Immunoblot analysis for detection and quantitation of gp44 and gp62 in extracts of gene 44 amber mutant (44amN82) phage T4-infected cells. Aliquots containing equal numbers of infected cells were harvested at different times p.i., and phage development was arrested immediately. Samples of these whole-cell lysates were fractionated on an SDS–15% polyacrylamide gel and then, after transfer by electroblotting to a nitrocellulose membrane, probed with anti-gp44 and -gp62 polyclonal antibodies. Proteins were visualized with a chemiluminescence detection protocol as described in Materials and Methods. Serial dilutions of the purified gp44/62 complex were run on each gel (standards, pg). Arrowheads indicate gp44 and gp62 in the cell lysates at fixed time points (minutes post-infection) produced during the infection of a suppressing (sup-1) E. coli host (“coupled”) or a nonsuppressing (sup-0) host (“uncoupled”) with 44amN82. To account for greater gp44 production in the suppressor host (“coupled”), the 10-min lysate was diluted twofold with control lysate. The control (0 min) was a whole-cell lysate just before phage addition. A portion of an overexposed blot showing gp62 is shown below to emphasize the difference in gp62 translation during infection of suppressor (“coupled”) and nonsuppressor (“uncoupled”) hosts. The higher-molecular-weight band is due to cross-reactivity of the anti-gp62 antibodies with E. coli protein.

A second round of infections was performed by using 44amN82 and the nonsuppressing E. coli host (sup-0). The amber mutation occurs in the middle portion of gene 44, producing a truncated protein which is about 50% smaller than wild-type gp44 (40). The truncated gp44 was not detected by our assay. Synthesis of gp62 was detected in the absence of complete gene 44 translation, albeit at much lower levels than when gene 44 translation proceeded unhindered. At 3.5 min p.i., 15 molecules of gp62 per cell were detected. At 5 min p.i., there were 100 molecules of gp62 per cell, while the level rose to 360 molecules per cell at 10 min p.i.

Clearly, while gene 44 translation was important for efficient gene 62 translation, it was not an absolute requirement. The gene 44 amber mutant has a compromised ability to replicate DNA in a nonsuppressing host (1). Deficiency in DNA replication will cause a lower number of T4 templates, which could decrease the level of gene 44 and gene 62 encoding mRNAs. This would lead to a lower level of gp62 produced by the uncoupled pathway, yielding an artificially high degree of translational coupling. However, the various mRNA species containing gene 44 and gene 62 are transcribed from either early or middle promoters (11), which are utilized before the onset of DNA replication. It can be postulated that deficiency of DNA replication should not greatly affect the level of the gene 44 and 62 messages or the amounts of gp44 and gp62 produced. To confirm this argument, we carried out infections of suppressor and nonsuppressor hosts with 43amB22.

Synthesis of gp44 and gp62 by 43amB22.

Gene 43 codes for the T4 DNA polymerase. The mutation amB22 causes truncation of the carboxyl-terminal 20% of the protein. The mutant gp43 possesses DNA binding affinity and exonuclease activity but cannot interact with DNA polymerase accessory proteins and exhibits no polymerase activity. Thus, in a nonpermissive host, 43amB22 is deficient in DNA replication (27).

Figure 2 depicts the levels of the clamp loader proteins during infections of E. coli suppressor and nonsuppressor hosts with 43amB22. In the infection of the suppressor host, gp44 and gp62 were detected at 3.5 min p.i. at levels of 600 and 100 monomers per cell, respectively. These levels rose to 2,700 gp44 and 800 gp62 monomers per cell at 5 min p.i. At 10 min p.i., there were 7,700 gp44 and 2,700 gp62 monomers present per cell.

FIG. 2.

Lack of DNA replication does not affect the translation of gene 44-gene 62 mRNA nor the apparent stoichiometry of gp44/62 complex synthesis. The DNA polymerase-deficient T4 mutant (43amB22) produced gp44 and gp62 during the infection of a sup-0 (nonsuppressor) host in amounts similar to those of a sup-1 (suppressor) E. coli host, as determined by quantitative immunoblotting (Fig. 1). The 4:1 stoichiometry was not changed.

Under nonsuppressing conditions, results at 3.5 and 5 min p.i. are nearly identical, within the range of experimental error, to those found under suppressing conditions. For example, at 3.5 min p.i., there were 420 gp44 and 130 gp62 monomers present per cell in the nonsuppressor host while there were 600 and 120 monomers of gp44 and gp62, respectively, per cell under suppressing conditions. At 10 min p.i., the levels in the nonsuppressor host were only slightly lower than those found under suppressing conditions. The observation that the levels of gp44 and gp62 produced are similar in both infections confirms that DNA replication has little or no effect on the amounts of gp44 and gp62 produced, at least during early portions of T4 infections.

The amounts of gp44 and gp62 produced by 43amB22 are similar to the levels found in wild-type infections. Burke et al. measured the amounts of gp44 and gp62 synthesized between 3.5 and 7 min p.i. by using ¹⁴C labeling. For infection by wild-type phage, 2,900 gp44 and 700 gp62 molecules per cell were detected during this time period (4). Our results at 5 min from 43amB22 infections, in both suppressor and nonsuppressor backgrounds, agree well with these values. Infections with wild-type phage yielded 2,500 molecules of gp62 per cell after 10 min (results not shown). Our results at 10 min p.i. with 43amB22 in both backgrounds again correspond to the wild-type value. It appears that the levels of gp44 and gp62 detected in the 43amB22 infections of the suppressor and nonsuppressor hosts represent the amounts of these proteins produced in wild-type infections.

The levels of both proteins are higher in both 43amB22 infections than in the 44amN82 infection of the suppressor host (Table 1). When gene 44 is translated from 44amN82 in the sup-1 background, there is competition between the suppressor tRNA and release factors for the amber codon, leading to less than 100% suppression efficiency (7, 10) and a subsequent decrease in the amount of gp44 synthesized. As gene 62 is translationally coupled to gene 44, the decrease in the level of gp44 produced is reflected as a decrease in the amount of gp62 synthesized compared to wild-type gp62 levels.

TABLE 1.

Intracellular concentration of gp44/62 in vivo

T4 strain	Min p.i.	Mean no. of monomers/cell ± SD
		Suppressor host		Nonsuppressor host
		gp44	gp62	gp44	gp62
44amN82	3.5	230 ± 30a	70 ± 15	NDb	15 ± 5
	5	1,400 ± 110	400 ± 120	ND	100 ± 25
	10	6,700 ± 900	1,600 ± 360	ND	360 ± 130
43amB22	3.5	600 ± 120	120 ± 20	460 ± 130	130 ± 30
	5	2,700 ± 200	800 ± 280	2,100 ± 500	700 ± 100
	10	7,700 ± 260	2,750 ± 100	5,700 ± 700	1,900 ± 600

^aUncertainty is represented by one sample standard deviation. 

^bND, none detected. 

Degree of translational coupling.

Translational coupling occurs when translation of a downstream gene on a polycistronic mRNA is dependent upon translation of a gene immediately upstream (29). Features of the intercistronic region of genes 44 and 62 are illustrated in Fig. 3. Factors which may contribute to the translational coupling of genes 44 and 62 are the facts that (i) the start codon of gene 62 is one nucleotide downstream from the stop codon of gene 44 (33), (ii) the Shine-Dalgarno region from gene 62 is weak by statistical criteria, and (iii) a putative mRNA secondary structure could sequester part of the gene 62 TIR (36). Translation of gene 44 could increase the translation of gene 62 by unfolding local secondary structure, by increasing the frequency of favorable ribosome interaction with the gene 62 TIR via an increase in the local concentration of ribosomal subunits, or by a combination of these two mechanisms (12, 32).

FIG. 3.

Schematic of a portion of the polycistronic mRNA containing coding regions for gp44 and gp62. The expanded portion illustrates features of the intercistronic region, including the stop codon for gene 44, the start codon for gene 62, and the Shine-Dalgarno (SD) region for gene 62. An imperfect inverted repeat implicated in a putative mRNA secondary structure is indicated by the opposing arrows (36).

We define the efficiency of translational coupling by comparing the amounts of gp62 produced in the 44amN82 infection of the nonsuppressor host (uncoupled translation) to levels produced in the 43amB22 infections (coupled translation). As the gp62 levels determined in the 44amN82 infection of the suppressor host were reduced by inefficient suppression, they do not represent the true amounts of gp62 produced by coupled translation. Furthermore, the gp62 levels determined from the 43amB22 infections were not affected by the amber suppression and are, in fact, directly comparable to wild-type levels (see above). For example, at 3.5 min p.i., the amount of gp62 produced by 44amN82 under nonsuppressing conditions was 12.5% of the amount produced by 43amB22 in the infection of the suppressing host (Table 1). This indicates an efficiency of translational coupling of 87.5%. The overall average (3.5, 5, and 10 min p.i.) efficiency of translational coupling was 86% ± 3%.

When the amount of uncoupled gp62 synthesis is subtracted from the overall gp62 amounts, the percentage of gene 44 translation that is transmitted across the intercistronic junction can be calculated (Table 2). Approximately one-quarter, 24% ± 4%, of gene 44 translation events lead to reinitiation at the gene 62 TIR. This coupling frequency appears to be the major factor in controlling the stoichiometry of synthesis of the clamp loader subunits.

TABLE 2.

Apparent stoichiometry^a of gp44 versus gp62 synthesis in vivo

T4 strain	Host genotype	Mean ± SD
		gp44/62 stoichiometryb	% of gp44 translation transmittedd
44amN82	sup-1	3.7 ± 0.5c	21 ± 3e
43amB22	sup-1	3.8 ± 1.1	25 ± 6
43amB22	sup-0	3.2 ± 0.3	27 ± 2

^aThe reported stoichiometry for the gp44/62 complex is 4:1. 

^bObserved mean value of the gp44/62 complex at 3.5, 5, and 10 min p.i. 

^cSample standard deviation obtained from the results of three experiments. 

^dThe percentage of gp44 translation transmitted across the intercistronic region was calculated by subtracting the amount of gp62 synthesized via the uncoupled mechanism from that synthesized by the coupled pathway and then dividing the results by the amount of gp44 produced. The values shown are the means of the three time points examined. 

^eUncertainty is represented by one sample standard deviation. 

An important consideration is the position of the amber mutation in 44amN82. This mutation is known to occur at Gln 171 (40). This would result in a truncated gene 44-encoded protein almost 50% smaller than wild-type gp44 (which has 319 amino acids). As translational coupling is related to the distance between the start and stop codons, the large increase in the distance between the stop and start codons introduced by the amber mutation should abolish translational coupling. That is, termination will occur well before the elongating ribosomes reach the intercistronic region. In the infections of the nonsuppressing host by 44amN82, there is no apparent mechanism for melting of the putative RNA secondary structure in the gene 44-62 intercistronic region, yet gp62 is still produced. This suggests that the positioning of the terminating ribosome near the gene 62 TIR (increase in local concentration) is the major contribution to translational coupling in this system. Further study is necessary to separate secondary-structure contributions from the TIR effect for this gene pair.

Stoichiometry of gp44 and gp62 synthesis.

The T4 clamp loader complex maintains a strict 4:1 stoichiometry. The gene 44 and gene 62 messages are present in a one-to-one ratio; however, the stoichiometry of synthesis observed when all instances of coupled translation were considered was 3.5 ± 0.7. This gives a very slight excess of gp62 subunits produced when considering complex assembly. If the coupling mechanism were not active, approximately eightfold lower amounts of gp62 would be produced. This would lead to the production of a vast excess of gp44 monomers. It appears that this gene pair has evolved such that there is minimal translation of gene 62 unless the partner gp44 subunits are expressed, and the frequency of translational coupling aids stoichiometry of synthesis by allowing one-quarter of all gene 44 translation events to be transmitted across the intercistronic junction.

In the majority of known examples of translational coupling, the coupling exists as a means of regulation where expression of the downstream gene product is maintained in equimolar or greater amounts (2, 8, 29, 38, 39). This clearly is not the case with T4 genes 44 and 62. However, if f1 and the related IKe phage genes V and VII, the downstream gene product is produced in amounts much smaller than those of the upstream gene product (12, 24). It is possible that other phages have evolved downregulation across the intercistronic junction of coupled gene pairs as a means of avoiding unnecessary gene product accumulation.