Construction and Preliminary Characterization of a Library of “Lethal” Preterminal Protein Mutant Adenoviruses

Abstract

Adenoviruses containing lethal in-frame insertion mutant alleles of the preterminal protein (pTP) gene were constructed with cell lines that express pTP. Thirty in-frame insertion mutant alleles, including 26 alleles previously characterized as lethal and 4 newly constructed mutant alleles, were introduced into the viral chromosome in place of the wild-type pTP gene. The viruses were tested for ability to form plaques at 37°C in HeLa-pTP cells and at 32°C and 39.5°C in HeLa cells. Two of the newly constructed viruses exhibited temperature sensitivity for plaque formation, one virus did not form plaques in the absence of complementation, seven additional mutants exhibited a greater than 10-fold reduction in plaque formation in the absence of complementation, and another eight mutants exhibited stronger phenotypes than did previously characterized in-frame insertion mutants in the plaque assay. These mutant viruses offer promise for analysis of pTP functions.

Adenovirus replication is absolutely dependent on virally encoded DNA binding protein, preterminal protein (pTP), and DNA polymerase (Pol). pTP serves as a primer to which the first base, a dCMP residue, is covalently bound (for a review of adenovirus replication, see references 12 and 29). pTP is proteolytically matured following encapsulation of the DNA to yield terminal protein (TP), the form of the protein found in infectious virions (5). The function of the proteolytic processing is not known.

pTP and TP act to bind the adenovirus chromosome to the nuclear matrix (3, 7, 25) and to facilitate efficient transcription of viral early-region genes (23, 25). Binding of pTP and TP to the nuclear matrix occurs through interaction with the CAD enzyme (2) in a manner that is regulated by tyrosine phosphorylation (3).

Genetic studies of the roles of pTP in the infectious cycle have been limited by difficulties in isolating pTP mutant viruses with strong phenotypes. In two studies, in-frame insertion mutant alleles of pTP were generated in vitro, after which attempts to introduce the mutant alleles into the viral chromosome were made (9, 19, 20). Of the 56 alleles constructed, 26 were successfully introduced into the viral chromosome, of which two caused moderately strong phenotypes, three caused weaker phenotypes, and the remainder caused no phenotype (9, 19, 20). The alleles that could not be inserted into the viral chromosome were termed lethal. It was among the lethal alleles that we expected to find the most interesting and informative phenotypes.

pTP is involved in the rate-limiting step at initiation of adenovirus replication. While the use of virions makes adenovirus DNA extremely infectious, the use of purified DNA introduced into cells by transfection greatly reduces the infectivity of adenovirus DNA. The infectivity of the DNA can be increased by using adenovirus DNA-TP complex (6), but the resulting infectivity is still many orders of magnitude lower than that of DNA within virions. Since standard methods for introduction of mutated pTP alleles into the viral chromosome require the transfection of either naked adenovirus DNA or DNA-TP complex, the reduced infectivity of DNA and DNA-TP complex offers a likely explanation for the difficulty in generating pTP mutant viruses that exhibit strong phenotypes: even a modest effect on the ability of pTP to act in initiating adenovirus replication may make the mutation appear to be lethal by preventing the overlap recombination required to form virus. However, we expected that cell lines that express pTP (1, 13, 15, 21) would permit construction of the pTP mutant viruses, that once the pTP mutation was incorporated into the viral chromosome in place of the wild-type allele and virions were generated, the vastly increased infectivity of virions would overcome the limitations that blocked construction of the pTP mutant viruses by standard means, and that, therefore, the majority of the mutant viruses would grow in the absence of complementation.

Construction of viruses containing lethal pTP in-frame insertion mutant alleles.

A system for construction of viruses containing pTP mutant alleles that permitted screening or selection for the desired viruses was developed (Fig. 1). This procedure was found to efficiently select for the desired viruses.

FIG. 1.

Construction of pTP in-frame insertion mutant viruses. The Ad5dl308_Bstβ-gal-TP (24) complex is represented schematically by the thick horizontal line, with TP indicated by solid ovals. The unique BstBI and XbaI sites within the viral chromosome are indicated. The locations of the LacZ and the E2B mRNAs are indicated by lines with solid arrows, with splicing of the primary E2 transcript to yield the E2B mRNAs (encoding pTP and DNA Pol) indicated by dashed lines above the viral chromosome. The pTP and DNA Pol coding regions are indicated above the chromosome. Mutant pTP alleles that could not be introduced into the virus by standard means (9, 19, 20) as well as newly constructed mutant alleles constructed by partial digestion with HpaII in the presence of ethidium bromide (16, 17) and the two-amino acid insertion method of Barany (4) were cloned into plasmid p0-29.4 (22), which is indicated below Ad5dl308_Bstβ-gal, with the intermediate-thickness line indicating adenovirus sequence, the thickest line indicating the region encoding the main exon of pTP, and the thinnest line indicating the plasmid vector (the entire plasmid vector is not shown). The sites for XmnI, KpnI, and XbaI are indicated. p0-29.4 carrying the various pTP in-frame insertion mutant was digested with XbaI (for ligation with the large arm of XbaI-digested Ad5dl308_Bstβ-gal-TP) plus XmnI (to provide a blunt end outside the adenovirus sequence that would reduce recircularization and concatamerization during ligation), ligated with XbaI-digested Ad5dl308_Bstβ-gal-TP, and used to transfect 293-pTP cells, and the resultant viruses were plaque purified in HeLa-pTP cells where the pTP mutant viruses but not the parental Ad5dl308_Bstβ-gal would grow.

pTP in-frame insertion mutant viruses plaque purified with HeLa-pTP cells were grown in 293-pTP cells. Viral DNA was prepared by the method of Hardy et al. (11) and analyzed for the presence of the inserted restriction site within the pTP gene, the presence of the BstBI site within the E1A gene, and the absence of the KpnI site at bp 2052 within the E1B gene (the mutations in the E1 genes, which do not affect function of the proteins, were introduced to facilitate cloning [22]). In addition, the sizes of the fragments generated by digestion with the restriction enzyme cleaving the introduced sequence within the pTP gene were compared. This analysis demonstrated the successful construction of the library of lethal pTP in-frame insertion mutant viruses (Table 1).

TABLE 1.

In-frame pTP mutant alleles newly introduced into the virus in this study

Codona	Enzymeb	Insertion sitec	Enzymed	Wild-type sequencee	Mutant sequencef	Sourceg
43	NaeI	10470	ClaI	AAGI	AA-PIDG-GI	R43
50	RsaI	10448	SmaI	SRYI	SR-SR-NI	F390
88	RsaI	10334	ClaI	YQYL	YQ-SHRW-DL	R89
88	RsaI	10334	SmaI	YQYL	YQ-SR-NL	F391
97	SnaBI	10308	ClaI	DYVF	DY-PIDG-VF	R97
108	RsaI	10274	SmaI	LRYT	LR-SR-NT	F392
108	RsaI	10274	ClaI	LRYT	LR-SHRW-DT	R109
111	AluI	10266	ClaI	TELS	TE-PIDG-LS	R111
115	HpaII	10254	BamHI	QPGH	QP-DP-GH	H115
120	HincII	10239	ClaI	TVNW	TV-RIDR-NW	F438
125	BalI	10223	ClaI	VMAN	VM-AHRW-AN	R126
234	HaeIII	9896	ClaI	GMAD	GM-AHRW-AD	R235
234	HaeIII	9896	ClaI	GMAD	GM-ADRS-AD	F346
242	HaeIII	9872	ClaI	QQAG	QQ-AHRW-AG	R243
252	HincII	9844	ClaI	LSTI	LS-PSMG-TI	R252
252	HincII	9844	ClaI	LSTI	LS-GSIA-TI	F439
304h	HpaII	9684	BamHI	SDPV	SD-PD-PV	H304
378	HincII	9465	ClaI	FVDR	FV-PIDG-DR	R378
403h	HaeIII	9389	ClaI	GEAL	GE-ADRS-AL	F344
430	AluI	9309	ClaI	AELI	AE-PIDG-LI	R430
478	RsaI	9164	SmaI	VMYF	VM-SR-DF	F394
508	AluI	9075	ClaI	VELN	VE-PIDG-LN	R508
534h	HaeIII	8995	ClaI	EGGL	EG-GGSI-RL	F348
535	StuI	8995	ClaI	GGLN	GG-PSMG-LN	R535
541	AluI	8976	ClaI	SQLM	SQ-PIDG-LM	R541
557	HpaII	8927	BamHI	RAGR	RA-GS-GR	H557
592	HincII	8823	ClaI	AVND	AV-PIDG-ND	R592
609	HpaII	8771	BamHI	LTGP	LT-GS-GP	H609i
610	HaeIII	8767	ClaI	TGPV	TG-RIDR-PV	F342i
639	AluI	8682	ClaI	HQLL	HQ-PIDG-LL	R639i

^aWild-type codon immediately 5′ of the in-frame insertion site. 

^bRestriction enzyme used to create the insertion site. 

^cPosition of the insertion (in base pairs numbered starting from the left end of the viral chromosome). 

^dRestriction enzyme recognizing the inserted sequence. 

^eWild-type amino acids encoded at the mutated site. 

^fAmino acids encoded by the mutated allele. The inserted amino acids are indicated between hyphens, and the amino acids mutated by the insertion are underlined. 

^gPlasmid source of the DNA. Plasmids with the prefix R were constructed by Roovers et al. (19, 20), plasmids with the prefix F were constructed by Freimuth and Ginsberg (9), and plasmids with the prefix H were newly constructed for this study. 

^hThe first codon in the inserted sequence recreated the codon immediately 3′ to the insertion site. 

ⁱOne of the three most C-terminal mutations, which lie within the overlapping sequence that encodes the N-terminal region of DNA pol. 

Initial characterization of the pTP in-frame insertion mutants.

Stocks of the pTP in-frame insertion mutant viruses grown in 293-pTP cells were plaque titered with, as controls, the phenotypically wild-type Ad5dl309, the temperature-sensitive (ts) pTP mutant virus Ad5sub100r, and Ad5dl308pTP155 (equivalent to Ad5in425 [9]) in HeLa-pTP cells at 37°C and in HeLa cells at 32°C and 39.5°C. The input virus had either the mutant form of TP encoded by the virus or the wild-type TP encoded by 293-pTP cells covalently attached to the viral DNA. In cases where wild-type TP was associated with the DNA, it is possible that an advantage for virus replication was provided. However, this advantage would have been apparent only for the input viral genome and thus would have provided a linear, rather than exponential, advantage to virus growth. Thus, plaque formation would have been dependent on the ability of the mutated pTP to promote virus formation.

The results (Table 2) demonstrated that virus yields from 293-pTP cells varied over nearly 3 orders of magnitude among the newly constructed pTP in-frame mutant viruses in spite of similar progression of cytopathic effect. Since virus growth was in the presence of complementing pTP supplied by the cells and was not optimized for the individual viruses, the differences in yield, while suggestive of phenotypic differences due to the mutations introduced into the pTP gene, cannot be considered as a primary indicator of phenotype for the purposes of classification until virus growth is examined in detail in noncomplementing cells. Comparison of relative plaquing efficiencies in HeLa and HeLa-pTP cells normalizes for differences in yield and so offers strong evidence of phenotypic differences. Since certain mutant adenoviruses, including viruses deleted in the E1A and E1B genes, are capable of growing more efficiently after infection of noncomplementing cells at high multiplicity (26), plaque formation represents a fairly strict definition of viral replication.

TABLE 2.

Phenotypes of in-frame pTP mutants^a

Virus	Codonb	Plaque titer in:			HeLa-pTP:HeLa plaque titer ratioc		Phenotype
		HeLa-pTP cells at 37°C	HeLa cells		39.5°C	32°C	HeLad	In vitroe
			39.5°C	32°C
pTP43	43	8.4 × 107	3.7 × 107	1.0 × 107	1.3	2.0
pTP50	50	9.4 × 106	7.5 × 105	2.3 × 105	6.9	13
pTP88-1	88	4.9 × 106	1.3 × 106	4.1 × 105	2.1	2.9
pTP88-2	88	5.6 × 106	1.1 × 106	5.0 × 105	2.7	2.3		No pTP
pTP97	97	6.4 × 106	1.5 × 106	5.8 × 105	2.4	2.6
pTP108-1	108	1.1 × 107	3.4 × 104	2.6 × 104	180	100	rd*
pTP108-2	108	3.4 × 108	<2 × 103	>3.4 × 105	>9 × 104	<240	ts, rd*
pTP111	111	4.3 × 106	1.0 × 105	5.2 × 104	24	17	rd
pTP115	115	2.9 × 108	1.8 × 107	8.4 × 106	8.3	8.3
pTP120	120	2.2 × 108	<2 × 103	<2 × 103	>105	>4 × 104	Dead
pTP125	125	8.6 × 106	1.3 × 104	5.3 × 103	370	330	rd*	(−−)
pTP234-1	234	5.0 × 106	1.6 × 106	3.8 × 105	1.7	3.6
pTP234-2	234	1.7 × 107	4.8 × 106	1.8 × 106	1.9	3.1		−−
pTP242	242	4.2 × 105	2.2 × 104	7.6 × 103	11	13	rd
pTP252-1	252	1.4 × 106	1.1 × 105	4.9 × 104	13	29	rd
pTP252-2	252	8.3 × 106	5.6 × 105	2.5 × 105	8.3	6.9		No TP
pTP304	304	4.2 × 106	7.8 × 105	4.1 × 105	5.4	10
pTP378	378	1.4 × 107	5.9 × 106	1.3 × 106	1.3	2.6		(++)
pTP403	403	7.0 × 107	3.2 × 107	1.2 × 107	2.2	5.8		++
pTP430	430	2.7 × 107	2.1 × 107	2.8 × 106	0.7	2.3
pTP478	478	3.5 × 106	2.3 × 105	9.5 × 104	8.5	8.8		−−
pTP508	508	3.0 × 108	7.6 × 104	>6.4 × 105	2,300	<110	ts, rd*
pTP534	534	4.8 × 106	6.5 × 105	2.0 × 105	3.9	6.7		No pTP
pTP535	535	1.5 × 108	1.4 × 107	5.3 × 106	6.0	6.7
pTP541	541	1.3 × 107	1.2 × 107	1.8 × 106	0.6	1.7
pTP557	557	1.4 × 108	2.0 × 107	8.0 × 106	7.0	18
pTP592	592	1.4 × 106	5 × 104	9.0 × 103	16	36	rd
pTP609	609	1.4 × 107	1.1 × 105	4.2 × 104	72	79	rd
pTP610	610	1.7 × 108	9.3 × 107	3.6 × 107	1.1	1.0
pTP639	639	4.1 × 107	2.4 × 107	1.0 × 107	0.9	1.0
pTP115f	155	1.7 × 107	3.1 × 106	1.6 × 106	3.0	2.3		++
Ad5sub100rf	12	4.8 × 107	2.0 × 107	6.3 × 106	1.2	1.6		−
Ad5dl309f		1.0 × 109	5.5 × 108	2.4 × 108	−1.8	−4.2

^aStocks of the pTP mutants (all of which are in the Ad5dl308 background, but which are named here simply by the pTP mutation site) and various control viruses grown in 293-pTP 2C1 cells (except for the Ad5dl309 stock, which was grown in 293 cells [10]) were plaque titered with HeLa-pTP cells at 37°C and HeLa cells at both 32°C and 39.5°C. 

^bThe position of the insertion mutation is as indicated in Table 1. 

^cRatios of plaque titers in HeLa-pTP cells to those in HeLa cells are presented normalized to plaque titers observed for Ad5dl309 for both temperatures examined for HeLa cells. 

^dThe phenotype of the virus determined from relative plaquing efficiency. rd, replication defective (defined here as a 10- to 99-fold reduction in plaquing efficiency in HeLa cells relative to that in HeLa-pTP cells; rd*, indicates severely replication defective (defined here as at least a 100-fold a reduction in plaquing efficiency in HeLa cells relative to that in HeLa-pTP cells; ts, temperature-sensitive behavior (defined here as a greater than 10-fold increase in plaquing efficiency at 32°C relative to that at 39.5°C in HeLa cells). 

^eThe phenotype in in vitro replication assays determined with data derived from Fredman et al. (8). No pTP, pTP synthesis was not directed by the pTP; −−, the protein directed replication activity less than 5% of that of the wild-type pTP in both initiation and elongation assays; −, the protein directed replication activity less than 30% and greater than 5% of that of the wild-type allele; ++, the protein directed replication activity greater than 50% of that of the wild-type pTP. In cases where results were obtained using an allele mutagenized at the same site as a mutant virus created in this study, the results are presented in parentheses. 

^fControl virus. 

Both Ad5sub100r and Ad5in425 exhibited fairly strong phenotypes in 293 cells (9, 18, 21) in contrast to the weaker phenotypes apparent in plaque assays in HeLa cells (Table 2), suggesting that HeLa cells may partially complement growth of adenoviruses mutated in the pTP gene relative to 293 cells. Thus, reliance on significant reductions in plaque-forming efficiency in HeLa cells for phenotypic comparison sets a fairly rigorous standard for the definition of mutant behavior.

All but one of the newly constructed mutant viruses, Ad5dl308pTP120, were able to grow to form plaques in the absence of complementation. The mutation in Ad5dl308pTP120 may truly be lethal because of nonfunctional pTP. Alternatively, it is possible that the pTP allele may not direct expression of pTP.

In the absence of complementation, Ad5dl308pTP111, Ad5dl308pTP243, Ad5dl308pTP252, Ad5dl308pTP590, and Ad5dl308pTP609 formed plaques 10- to 100-fold less efficiently and Ad5dl308pTP108-1, Ad5dl308pTP108-2, Ad5dl308pTP125, and Ad5dl308pTP508 formed plaques at least 100-fold less efficiently than they did in the presence of complementation. Ad5dl308pTP108-2 and Ad5dl308pTPR508 also exhibited ts plaque formation in the absence of complementation. The phenotypes of all of these mutants, particularly the ts mutants, appeared to be much stronger than those of any of the previously constructed pTP mutants, excluding pTP deletion mutants.

The remaining viruses included 8 that exhibited stronger phenotypes than either Ad5sub100r or Ad5dl308pTP155 and 12 that exhibited little or no phenotype in the plaque assays. Since both Ad5sub100r (9, 21, 23, 25) and Ad5in425 (9, 18) exhibit interesting and informative phenotypes, it is likely that at least some of the less-defective pTP mutant viruses will exhibit stronger phenotypes in other assays of virus growth. Such assays will include one-step growth curves, determination of relative affinities of the viral chromosomes for the nuclear matrix, and analysis of the synthesis and fate of pTP in cells infected with the various pTP mutant viruses. The fact that the great majority of the pTP mutant viruses grow, at least to a low level, in the absence of complementation will facilitate studies of pTP function. Any pTP mutant that can be generated is of potential value, but the viruses that can be grown in the absence of complementation have expanded potential, since growth in the absence of complementation ensures that the mutant forms of pTP and TP are associated with all viruses in a stock. This will permit the effects of most of the mutations to be examined throughout the infectious cycle.

Viruses that did not display a strong phenotype were retested for the presence of the restriction site introduced within the pTP gene. Aliquots of the stocks used in plaque titration were grown in 293-pTP 2C1 cells and DNA isolated (11). All of the viruses maintained the restriction enzyme site introduced in the mutated pTP allele (data not shown), demonstrating that the lack of a strong phenotype in plaque formation in HeLa cells was characteristic of the mutation and did not result from recombination within the pTP-expressing cells or contamination with the wild-type virus.

Examination of the pTP mutant map (Fig. 2) demonstrates regions of clustering of strong phenotypes near amino acid 120, suggesting that this region is particularly important for pTP function. Stronger conclusions will require further mutagenesis of pTP.

FIG. 2.

Map of in-frame insertion alleles of pTP. A schematic map of the pTP coding sequence is presented as a horizontal line, with the N terminus to the left. Positions of amino acids, numbered from the first amino acid of pTP encoded by the 39.6-39.2 map unit exon (27), are indicated as ticks on the pTP coding sequence, with numbering shown below. The positions of viable in-frame pTP mutations introduced into the virus in earlier studies (9, 19, 20) are indicated by vertical lines below the pTP coding sequence, with the positions of mutations leading to replication-defectiveness (9) indicated by elongated lines. The positions of mutations introduced into the virus in this study are indicated above the pTP coding sequence, with the positions of mutations leading to strongly defective behavior (at least a 10-fold reduction in plaquing efficiency in HeLa cells relative to that in HeLa-pTP cells) indicated by elongated lines. The overlapping coding sequence for DNA Pol is indicated by the dashed line with arrow.

Fredman et al. (8) examined the ability of certain of the pTP proteins encoded by mutant alleles constructed by Freimuth and Ginsberg (9) to direct replication in vitro. The results that pertain to viruses examined in this study are indicated in Table 2. Reasonable agreement between the in vitro results and results of the plaque formation assay were observed for Ad5dl308pTP125 and Ad5dl308pTP378 (in both cases in comparison with paired mutant alleles generated by Freimuth and Ginsberg [9]), Ad5dl308pTP403, and Ad5dl308pTPF155. Ad5sub100r pTP directed poor replication in vitro (8) but had only modest effects on plaque formation in HeLa cells. The in vitro replication data are thus in better agreement with the results of plaque formation by the virus in 293 cells (9, 21) than with the results of plaque formation by the virus in HeLa cells. The F233 mutation has a strong effect on the progression of Ad5in425 through the infectious cycle (9, 18) but relatively little effect on plaquing efficiency, supporting the suggestion that the mutation affects a pTP function separate from its role in replication (8).

The pTP alleles encoded by F391, F439, and F348 did not direct synthesis of pTP in vitro (8). The ability of the viruses generated by using these alleles, Ad5dl308pTP88-2, Ad5dl308pTP252-2, and Ad5dl308pTP534, respectively, to form plaques in HeLa cells indicates that pTP synthesis was directed in vivo. However, pTP encoded by all of these alleles, and by their paired counterparts with insertions at the same sites generated by Roovers et al. (19, 20), may act as dominant negatives, as yields of these viruses grown in 293-pTP cells were reduced 40- to 90-fold relative to the highest yields of pTP in-frame insertion mutants (Table 2) and pTP deletion mutant viruses (data not shown).

F346 and F394 both directed little or no replication in vitro (8). The F346 mutation in Ad5dl308pTP234-2 led to only a modest phenotype in the plaque formation assays. The F394 mutation in Ad5dl308pTP478 led to a stronger phenotype, although possibly not as strong as predicted from the in vitro replication results. The increased effect of the mutations in vitro may reflect the stringency of the assay, since replication in vitro is inefficient relative to that in vivo. Alternatively, the formation of complexes with DNA Pol (14, 28) may help to stabilize the mutant pTPs in vivo.