Genomic Complexity among Strains of the Facultative Photoheterotrophic Bacterium Rhodobacter sphaeroides

Abstract

Pulsed-field gel electrophoresis following the use of rare cutting restriction endonucleases together with Southern hybridization, using markers distributed on chromosomes I and II of Rhodobacter sphaeroides 2.4.1, has been used to examine approximately 25 strains of R. sphaeroides in an effort to assess the occurrence of genome complexity in these strains. The results suggest that genome complexity is widespread and is accompanied by substantial genomic heterogeneity.

In this study, we have examined approximately 25 strains of diverse origins, but otherwise classified as Rhodobacter sphaeroides. We have determined minimum genome size and overall macro-restriction pattern polymorphisms by pulsed-field gel electrophoresis (PFGE). The number of rrn operons present in each strain and their direction of transcription relative to each other were also determined. Gene probes diagnostic for chromosomes CI and CII (11, 22, 23, 27) of R. sphaeroides 2.4.1, together with the above-described results, have been used to assess the likelihood that these strains contain more than one chromosome, i.e., genome complexity. In a recent report Jumas-Bilak et al. (14) have surveyed genomic complexity in chromosome number amongst members of the α-3 subgroup and related subgroups of the Proteobacteria. When several strains of Agrobacterium were examined all had complex genomes, whereas the same was not true for Brucella (14). This observation raises important questions and warrants an examination of the distribution of genome complexity amongst members of the R. sphaeroides group of organisms. Is strain R. sphaeroides 2.4.1 an isolated example, or is the existence of multiple chromosomes within R. sphaeroides more widespread?

Based upon previous work in this laboratory (27, 28), we slightly modified electrophoretic conditions required to optimally separate large, mid-size, and small AseI-generated DNA fragments derived from the strains listed in Table 1, which were grown as previously described (26, 29). From these data (Table 2) we were able to estimate the minimal genome sizes for each of the strains examined. Table 2 is presented so as to enable the reader to visually infer the DNA fragment distributions and to provide fragment sizes. The running conditions which were employed did not permit us to accurately estimate fragment sizes greater than 1,100 kb, thus the indication in some instances of >1,105 kb. Previous work identified which AseI fragments represent plasmid DNA in R. sphaeroides 2.4.1 (24, 29). However, we did not attempt to identify plasmid-derived AseI fragments in any of the other strains examined here, although earlier studies (8) indicate that the 110-kb replicon of strain 2.4.1 was the largest apparent plasmid observed.

TABLE 1.

Bacterial strains and plasmids

Strain of plasmid	Relevant characteristic(s)	Sourcea (reference)
R. sphaeroides
2.4.1	Wild type	W. Sistrom (31)
21286	Wild type	ATCC (20)
21455	Wild type	ATCC (33)
35053	Wild type	ATCC (1)
35054	Wild type	ATCC (1)
IL 106	Wild type	T. Satoh (24)
14690	Wild type	ATCC (31)
RS630	RS602 cured of bacteriophage Rφ6P	J. Pemberton (30)
28/5	Wild type	G. Drews
Geller	Wild type	M. Madigan (32)
17028	Wild type	ATCC (31)
159	Wild type	DSM (5)
160	Wild type	DSM (5)
33575	Wild type	ATCC (9)
RS2	Wild type	S. Harayama (18)
L	Wild type	J. Lascelles (15)
SCJ	Wild type	J. Wall (32)
NCIMB 8253	Wild type	NCIMB (21)
158	Wild type	DSM (5, 31)
17023	Wild type	ATCC (31)
17024	Wild type	ATCC (31)
17025	Wild type	ATCC (31)
17027	Wild type	ATCC (31)
17029	Wild type	ATCC (31)
WS8	Wild type	W. Sistrom (28)
Rhodobacter capsulatus	Wild type
Plasmids
pL110	A 3.0-kb BamHI restriction fragment of R. sphaeroides 2.4.1 containing rbcR cloned into pBR322	E. Muller (19)
pPRK12b	A 3.4-kb EcoRI restriction fragment of R. sphaeroides 2.4.1 containing rbcL cloned into pBR322	P. Hallenbeck (10)
pBRBE47HindIII	A 0.7-kb HindIII-EcoRI restriction fragment of R. sphaeroides 2.4.1 containing the 16S subunit of rrnB	S. Dryden (6)
pUC304L18	A 3.0-kb PvuII restriction fragment of R. sphaeroides 2.4.1 containing rrnA cloned into pUC19	S. Dryden (6)
pUI1015	A 2.0-kb NaeI-BamHI restriction fragment of R. sphaeroides 2.4.1 containing hema cloned into pUC19	E. Neidle (22)
pUI1005	A 1.8-kb BamHI restriction fragment of R. sphaeroides 2.4.1 containing hemT cloned into pUC19	E. Neidle (22)

^aATCC, American Type Culture Collection; DSM, Deutsche Sammlung von Mikroorganismen; NCIMB, National Collection of Industrial and Marine Bacteria. 

TABLE 2.

Summary of AseI fragment sizes and estimate of minimum genome size

Sizes (kb) of AseI fragments in strain:
241	21286	21455	35053	35054	IL106	14690	RS630	28/5	Geller	17028	17023	17024	17025	17027	17029	159	160	33575	RS2	L	SCJ	8253	158
1,105	1,105	>1,105	>1,105	1,105	1,105	1,105			1,105	1,105	1,105	>1,105	>1,105	>1,105	>1,105		>1,105		1,105	>1,105	>1,105	1,105	1,105
910	910	910			910	910			910	910	910			910	910				910	910		910	910
			847	722			847	800		785						847		847
			660				660	598				660		743	755			754
			535				578						660			577		620
												493	493			521		470			540
410		410				410			410											410		410	410
	385			385			377			378	400					370		370	380
				370	370
360	360	360	360			360			360	360									360	360		360	360
				350							350
340	340	340				340					330								340	340		340	340
							318	310	318					330	330	310		310		292	318
	296	296			296		296		296	280								292
275			275	275		275				275						275	275		275			275	275
											260
244	244					244			244					244	244				244		244	244	244
				229							229						230
214	214				214	214		214	214					214				214	214		214	214	214
		179			179		179	179				179	196				188	130			133
								170			112	196	193			162	145			162
					162		150	162	154		162	124	179			145	121
					145		140	145					124		112	135
								133								130				120
110	110	110				110			110	110		110	110						110	110		110	110
											105	100	100		105
97	97	97		97		97		97						97			97		97	97		97	97
			95	95	95			89			90							95
											80
73	73	73	73		73	73		73	73	73									73			73	73
											65	65	65
63	63	63	63		63	63	63		63	63						63	63	63	63	63	63	63	63
		45	45	54	42			42	50	50	60	53	53	59	59
								37	40		45	43		43	49	49
												37	37
31						31		31	31			31	31								31	31	31
											28	23
18						18		18			18	18							18	18		18	18
												14	14	13	13	15	9	16
5						5												5
4,255a	4,197	>3,988	>4,058	3,682	3,654	4,255	3,608	3,098	4,378	4,389	4,349	>3,262	>3,414	>3,776	>3,694	3,550	>2,233	4,181	4,198	>3,987	>3,108	4,250	4,250

^aTotal 

In order to use the distribution of various marker genes (3, 4), namely hemA and rbcL on chromosome I and hemT and rbcR on chromosome II (Table 1) (22, 27, 28) of strain 2.4.1, as representative of the presence of two chromosomes in diverse strains of R. sphaeroides, all of the strains were subjected to PFGE following digestion with the intron-encoded restriction endonuclease I-CeuI (16). Nylon membranes derived from the TAFE gels were prepared as previously described (27), and these were probed at high stringency with radiolabeled (25) purified internal fragments of both the 23S and 16S rrn cistrons (Table 1) (7). In strain 2.4.1 one rrn operon is located on chromosome I and two rrn operons are located on chromosome II, 30 kb apart (7, 27). Sequence analysis confirmed the recognition site for I-CeuI to be present within each 23S cistron, and the 23S rrn probe encompassed this sequence. When 2.4.1 is treated with I-CeuI, the large chromosome is linearized to a 3-Mbp DNA fragment and the small chromosome yields a 0.87-Mbp fragment and a 30-kb fragment, as expected. Hybridization of 2.4.1 with the 23S and 16S rrn probes yields both the expected and identical numbers of signals, confirming the presence of three rrn operons transcribed in the same direction.

Table 3 provides a summary of this analysis. Several features are immediately obvious. The numbers of rrn operons varies from two to five depending upon the strain of R. sphaeroides. Those strains of R. sphaeroides showing fewer signals with the 16S rrn probe than with the 23S rrn probe indicate that one or more operons are transcribed in opposite directions relative to one another. Thus, unlike Escherichia coli (13), Salmonella typhimurium (13, 17), and Bacillus subtilis (12), which appear to possess a constant number of rrn operons, strains of R. sphaeroides are quite heterogeneous. However, Bacillus cereus may also be heterogeneous (2). This heterogeneity could reflect the existence of rrn operons on multiple linkage groups, which over time and through unequal crossing over give rise to variable numbers of such operons. Nonetheless, these observations raise interesting, and for the moment unique, questions regarding the derivation of each of these strains.

TABLE 3.

Summary of results of SoutherN hybridization analysis

Strain	No. of large I-CeuI fragmentsa	Approx. sizes of large I-CeuI fragments (Mbp)	Probeb				No. of 23S signals	No. of 16S signals
			hemA	hem	rbcR	rbcL
2.4.1	2	3.0, 0.9	lg	sm	sm	lg	3	3
21286	2	3.0, 0.9	lg	sm	sm	lg	3	3
21455	2	3.0, 0.9	lg	lgc sm	lg, sm	lg	4	4
35053	2	3.0, 0.7	lg	sm	sm	lg	4	4
35054	2	3.0, 1.5	lg	sm	sm	lg	3	3
IL106	1 (2e)	3.0, ∼0.4	lg	None	smd	lg	4	3
14690	2	3.0, 0.9	lg	sm	sm	lg	3	3
RS630	2	3.0, 1.2	lg	lgc sm	lg, smc	lg	4	3
28/5	1 (2)	3.0, ∼0.4	None	None	None	None	2	2
Geller	3	3.0, 1.2, 0.7	lg	md	lg, md	lg	5	5
17028	2	3.0, 0.9	lg	sm	sm	lg	3	3
17023	2	3.0, 0.9	lg	sm	sm	lg	4	4
17024	2	3.0, 0.4	lg	None	sm	lg	5	5
17025	2	3.0, 0.4	lg	None	sm	lg	5	5
17027	2	3.0, 1.5	lg	sm	sm	lg	3	3
17029	2	3.0, 2.0	lg	sm	sm	lg	4	4
159	2	3.0, 0.9	lg	sm	None	lg	5	5
160	2	3.0, 1.2	lg	sm	smc	lgc	4	4
33575	2	3.0, 0.8	None	None	None	lg	4	3
RS2	2	3.0, 0.9	lg	sm	sm	lg	3	3
L	2	3.0, 0.8	lg	sm	lg, sm	lg	4	4
SCJ	2	3.0, 1.2	lg	sm	sm	lg	4	3
8253	2	3.0, 0.9	lg	sm	sm	lg	3	3
158	2	3.0, 0.8	lg	sm	sm	lg	2	2

^aI-CeuI restriction fragments greater than or equal to 0.7 Mbp. 

^bsm, smallest I-CeuI restriction fragment indicated as a large I-CeuI restriction fragment; md, second largest I-CeuI restriction fragment indicated as a large I-CeuI restriction fragment; lg, largest I-CeuI restriction fragment indicated as a large I-CeuI restriction fragment. 

^cA heterologous signal. 

^dHybridizes to second largest I-CeuI restriction fragment, which is smaller than 0.7 Mbp. 

^eNumbers in parentheses indicate that there could be two chromosomes and not one if rrn operons are in a small linkage group. 

Nearly all strains, with the exception of IL106, 28/5, 17024, and 17025, possess at least two large I-CeuI-generated fragments, as is the case for 2.4.1. Further, the distribution of hemA and rbcL on the large I-CeuI fragment is true for all strains with the exception of 28/5 even when we consider the existence of a heterologous rbcL-generated signal for strain 160 and the absence of a hemA signal in strain 33575. The distribution of hemT and rbcR for the most part is similar to that observed for 2.4.1 (17 of 23 strains), although somewhat greater heterogeneity in their distribution is present (6 of 23). Again 28/5 and 33575 stand out. In fact, if we also consider the AseI-generated restriction pattern, the data for strain 28/5 suggest that it may not be R. sphaeroides, at least by the criteria employed here. Likewise, the results observed for strain 33575 make any conclusions regarding it difficult at the present time. Like 2.4.1, strains 17024 and 17025 possess, in addition to the large 3.0-Mbp I-CeuI-generated DNA fragment, a much smaller 0.40-Mbp second fragment. However, the presence of five rrn operons in these strains could be responsible for the smaller size of the second I-CeuI-generated fragment, if we assume that approximately four of the five rrn operons are present on the smaller of the two linkage groups. Thus, by a combination of criteria, as described here, most strains of R. sphaeroides obtained from culture collections around the world appear to contain at least two chromosomes, using strain 2.4.1 as the prototype. Further, our findings that additional markers, e.g., groEL₁, groEL₂, rpoN₂, rpoN₁, etc., are distributed between the two chromosomes of strain 2.4.1 (23a) add to a growing list of representative genes which might be used to more accurately assess genome complexity within R. sphaeroides. However, only the actual determination of the complete physical maps of the genomes of each isolate reported here represents an iron-clad approach to answering the questions posed here.

Finally, an issue regarding strain identification in the published literature is worth addressing. Strains 14690, NCIMB 8253, and RS2 appear to be identical to 2.4.1. The absence of the 5-kb restriction fragment from the PFGE profile of NCIMB 8253 should not be considered a difference since this fragment is very often difficult to observe because it diffuses so readily in the gel. Strain RS2 has a 380-kb AseI fragment in place of the 410-kb fragment in 2.4.1. We have repeatedly observed that a 30-kb fragment can spontaneously excise from chromosome I of 2.4.1, reducing the 410-kb fragment to 380 kb (28, 29). Once removed, this fragment is lost from the genome. Strain 158, except for the presence of two rrn operons, is otherwise identical to 2.4.1 by the criteria applied here.

However, it should also be noted that some strains described here have been reported to be identical to 2.4.1, e.g., ATCC 17023 and an 8253 derivative obtained from R. Niederman some years ago. Although certainly similar, these strains are not identical to 2.4.1, and therefore we have arbitrarily designated the 2.4.1 strain in our laboratory as R. sphaeroides 2.4.1^T.