Codon Usage Bias and Cluster Analysis of the MMP-2 and MMP-9 Genes in Seven Mammals.

Matrix metalloproteinase (MMP)-2 and MMP-9 are a family of Zn2+ and Ca2+-dependent gelatinase MMPs that regulate muscle development and disease treatment, and they are highly conservative during biological evolution. Despite increasing knowledge of MMP genes, their evolutionary mechanism for functional adaption remains unclear. Moreover, analysis of codon usage bias (CUB) is reliable to understand evolutionary associations. However, the distribution of CUB of MMP-2 and MMP-9 genes in mammals has not been revealed clearly. Multiple analytical software was used to study the genetic evolution, phylogeny, and codon usage pattern of these two genes in seven species of mammals. Results showed that the MMP-2 and MMP-9 genes have CUB. By comparing the content of synonymous codon bases amongst seven mammals, we found that MMP-2 and MMP-9 were low-expression genes in mammals with high codon conservation, and their third codon preferred the G/C base. RSCU analysis revealed that these two genes preferred codons encoding delicious amino acids. Analysing what factors influence CUB showed that the third base distributors of these two genes were C/A and C/T, and GC3S had a wide distribution range on the ENC plot reference curve under no selection or mutational pressure. Thus, mutational pressure is an important factor in CUB. This study revealed the usage characteristics of the MMP-2 and MMP-9 gene codons in different mammals and provided basic data for further study towards enhancing meat flavour, treating muscle disease, and optimizing codons.

1. Introduction

Codon usage bias (CUB) is defined as unequal utilisation in the frequency of synonymous codons in coding amino acids (AAs), and it has been used extensively for investigating gene phylogeny [1]. The synonymous codon characteristics include universality, degeneracy, and wobble, and they should be used randomly to encode corresponding AAs with no pressure of interference pressure. However, CUB can be affected by nucleotides composition, translation, hydrophobicity, tRNA abundance, and protein structure [2-6]. Notably, natural selection and mutational pressure, which drive the correct translation process, are the major factors associated with CUB [7, 8]. Natural selection affects the pattern of codon usage in organisms, and mutational pressure may arise whilst the proportion of codon bases changes. CUB greatly increases the variability of genetic information and reflects the genetic drift of codons to a certain extent [9]. Therefore, CUB can reveal the evolution of genes or organisms and environmental adaptation [10].

CUB is assessed by using the effective number of codons (ENC), codon adaptation index (CAI), frequency of optimal codons (FOP), codon bias index (CBI), and relative usage of synonymous codons (RSCU). ENC is calculated by comparing the GC content of synonymous codon positions [11]. CAI is 0-1; the closer the value is to 1, the stronger the nucleotide bases prefer synonymous codons [12]. FOP and CBI are both 0-1. These two indicators are close to 1, and the optimal codon for encoding amino acids is preferred. However, if CBI is negative, the optimal codon usage is less than the average number of codons used [3, 13]. RSCU is the specific value between the actual observation and theoretical observation, amongst which the theoretical observation value is the observation value when the synonymous codon usage frequency is the same, namely, there is no codon bias. If RSCU = 1, there is no CUB. If RSCU > 1, the appearance frequency of the codon is higher than the other synonymous codon. By contrast, it indicates lower genes. If RSCU > 2, then the frequency of CUB is extremely high [14].

Matrix metalloproteinases (MMPs) are a family of Zn2+ and Ca2+-dependent proteolytic enzymes that are widely expressed in animal tissues and highly conservative during biological evolution [15]. MMP-2 and MMP-9 can regulate muscle growth, repair, and some relative processes that affect biochemical reactions for muscle regulation [16]. Although recent research mainly focused on exploring MMP-2 and MMP-9 function for animal skeletal muscle development, healing diseased muscle and even meat [17-23], studies on MMP codons is rare. Therefore, there is an urgent need for exploring mammals' MMP-2 and MMP-9 genetic evolution and codon usage pattern regulating muscle growth.

In this study, seven mammals (Bos grunniens, Bos taurus, and Sus scrofa among Artiodactyla; Macaca mulatta in Primates; Canis lupus familiaris in Carnivora; Oryctolagus cuniculus in Lagomorpha; and Mus musculus in Rodentia) were chosen to analyse CUB and base pair composition dynamics. This study would give insight into the factors affecting CUB for MMP-2 and MMP-9 genes and provide basic data for enhancing the meat flavour and finding a promising gene treatment for muscle disease.

2. Materials and Methods

2.1. Software

MEGA 7.0, CodonW 1.4.2, pheatmap, and ggplot packages based on R 4.4.3 software were used to complete the relevant analysis.

2.2. Base Composition of MMP Genes' CDS in Different Mammals

The coding sequence (CDS) of yak MMP-2 and MMP-9 genes were obtained in our laboratory, and the NCBI accession numbers were MZ476247 and MZ476248, respectively. The CDS of other animals' genes were from NCBI GenBank, and their accession numbers are shown in Figure 1.

Figure 1

NJ trees of MMP-2 (a) and MMP-9 (b) genes of seven mammal species.

CodonW 1.4.2 software developed by J. Peden was used to analyse the MMP-2 and MMP-9 CDS in seven mammals for calculating A/T (A/T base content, the same below), G/C, T3S (third base of the codon is T content, the same below), C3S, A3S, G3S, GC3S, AT3S, ENC, CAI, CBI, FOP, and RSCU [24]. R packages pheatmap and ggplot2 were used to analyse the data.

2.3. PR2 Plot

PR2 plot could analyse the bias amongst ATCG under gene mutation [25]. If the frequency of the third base is A > T, then dots are scattered on the top of the PR2 plot. If the frequency is C > G, then dots are on the left. When the codon does not show usage bias, the dots are in the centre of the graph [26].

2.4. Codon Neutral Analysis

Codon neutral analysis was carried out by the correlation analysis of GC12 (the average of the GC content of the first and second bases) and GC3S to compare the influence of natural selection pressure and mutational pressure on CUB [27]. A significant correlation between GC12 and GC3S indicated that mutational pressure had a strong influence on codon preference; otherwise, natural selection influenced CUB [28].

2.5. ENC Plot

The relationship between ENC and GC3S without environmental selection pressure could be simulated by the following formula (1). The ENC/GC3S reference curve shows the main characteristics of codon usage patterns [24]. If CUB is more affected by natural selection, it should be below the standard curve. By contrast, it should be above the standard curve if it is more affected by other factors such as gene mutation. In general, the ENC is from 35 to 61. If ENC > 35, CUB is weak [11].

3. Results

3.1. Phylogenetic Analysis

Neighbour joining (NJ) trees were established based on the MMP-2 and MMP-9 CDS in seven mammals. The results (Figure 1) showed that the MMP-2 and MMP-9 genes of Bos grunniens were similar to those of B. taurus. These two genes of S. scrofa were similar to those of B. grunniens and B. taurus. Interestingly, the MMP-9 genes of C. lupus familiaris showed closer proximity to those of S. scrofa but those of the MMP-2 gene was farther.

3.2. Nucleotide Composition of MMP-2 and MMP-9 Genes

Compared with the content of codon bases of the MMP-2 and MMP-9 genes in seven mammals, the results showed (Tables 1 and 2) that the G/C content was higher than the A/T content. Most mammals' MMP-2 and MMP-9 GC3S were larger than AT3S, except for the MMP-2 gene of B. taurus and M. musculus. The above findings indicated that the MMP-2 and MMP-9 gene codons preferred GC3S.

Table 1

Nucleotide composition in the sequence of MMP-2 gene.

Species	A/T (%)	G/C (%)	T 3S (%)	C 3S (%)	A 3S (%)	G 3S (%)	GC3S (%)	AT3S (%)
Bos grunniens	0.415	0.585	0.2093	0.372	0.2081	0.3703	0.637	0.363
Bos taurus	0.384	0.616	0.2638	0.3049	0.34	0.2724	0.488	0.512
Macaca mulatta	0.409	0.591	0.2247	0.3128	0.2741	0.3031	0.551	0.449
Mus musculus	0.433	0.567	0.2974	0.3234	0.3212	0.2566	0.482	0.518
Oryctolagus cuniculus	0.4	0.6	0.1997	0.5452	0.1304	0.3621	0.73	0.27
Sus scrofa	0.393	0.607	0.1871	0.5485	0.1404	0.3488	0.731	0.269
Canis lupus familiaris	0.417	0.583	0.284	0.3302	0.3285	0.2415	0.482	0.518

Table 2

Nucleotide composition in the sequence of MMP-9 gene.

Species	A/T (%)	G/C (%)	T 3S (%)	C 3S (%)	A 3S (%)	G 3S (%)	GC3S (%)	AT3S (%)
Bos grunniens	0.366	0.634	0.1842	0.5263	0.1981	0.2891	0.685	0.315
Bos taurus	0.378	0.622	0.173	0.5554	0.1202	0.3781	0.759	0.241
Macaca mulatta	0.379	0.621	0.1766	0.5135	0.1277	0.3978	0.747	0.253
Mus musculus	0.433	0.567	0.2776	0.3436	0.3056	0.2458	0.503	0.497
Oryctolagus cuniculus	0.345	0.655	0.2237	0.3099	0.1976	0.3431	0.606	0.394
Sus scrofa	0.385	0.615	0.1814	0.5411	0.1207	0.3852	0.752	0.248
Canis lupus familiaris	0.373	0.627	0.1554	0.5702	0.1143	0.39	0.779	0.221

The codon usage results (Tables 3 and 4) showed that ENCs of the MMP-2 and MMP-9 genes in seven mammals were 40–56, indicating that these two genes had low expression and their codon conservation was high.

Table 3

Codon analysis of MMP-2 gene.

Species	ENC	CAI	CBI	Fop	GC1S (%)	GC2S (%)	GC12 (%)
Bos grunniens	48.04	0.193	−0.004	0.416	0.4932	0.5961	0.54465
Bos taurus	53.82	0.129	−0.006	0.386	0.7471	0.576	0.66155
Macaca mulatta	52.31	0.157	0.006	0.413	0.4982	0.6751	0.58665
Mus musculus	55.09	0.163	0.034	0.416	0.6238	0.5439	0.58385
Oryctolagus cuniculus	43.1	0.29	0.193	0.539	0.565	0.491	0.528
Sus scrofa	44.62	0.261	0.17	0.524	0.5579	0.52	0.53895
Canis lupus familiaris	55.29	0.146	0.014	0.4	0.6706	0.5569	0.61375

Table 4

Codon analysis of MMP-9 gene.

Species	ENC	CAI	CBI	Fop	GC1S (%)	GC2S (%)	GC12 (%)
Bos grunniens	44.44	0.256	0.24	0.558	0.6713	0.5295	0.6004
Bos taurus	42.48	0.291	0.249	0.566	0.6015	0.4955	0.5485
Macaca mulatta	44.99	0.252	0.191	0.53	0.5997	0.5045	0.5521
Mus musculus	56.96	0.16	0.042	0.419	0.6105	0.5448	0.57765
Oryctolagus cuniculus	54.05	0.154	0.011	0.419	0.5065	0.8191	0.6628
Sus scrofa	43.79	0.282	0.222	0.549	0.5958	0.4863	0.54105
Canis lupus familiaris	40.95	0.287	0.243	0.563	0.6025	0.4916	0.54705

CAI showed that the preference for synonymous codons of the MMP-9 genes in seven mammals was significantly better than that of MMP-2, but both were lower than 0.3, indicating that it failed to reflect the preference of synonymous codons.

FOP and CBI results of the MMP-2 and MMP-9 genes showed that the optimal codon usage of MMP-2 in B. grunniens and B. taurus was inferior to the five other animals, whilst the optimal codon usage of MMP-9 was better than that of MMP-2.

3.3. RSCU Analysis

The RSCU results of the MMP-2 and MMP-9 genes showed that these two genes had a preference for 27 and 20 codons, respectively (Tables 5 and 6). Amongst them, CUG (encoding leucine, Leu) of MMP-2, CUG (encoding leucine, Leu), ACC (encoding threonine, Thr), and CGC (encoding arginine, Arg) of MMP-9 had strong high CUB (RSCU > 2).

Table 5

RSCU for MMP-2 gene among seven species.

AA	Codon	Frequency	RSCU
Phe	UUU	66	0.675714
	UUC∗	142	1.324286
Leu	UUA	28	0.268571
	UUG	71	0.732857
	CUU∗	127	1.09
	CUC	101	1.011429
	CUA	53	0.468571
	CUG∗∗	239	2.428571
Ile	AUU	34	0.801429
	AUC∗	67	1.28
	AUA	34	0.918571
Met	AUG	110	1
Val	GUU	65	0.861429
	GUC	47	0.715714
	GUA	39	0.522857
	GUG∗	122	1.902857
Ser	UCU∗	89	1.078571
	UCC∗	104	1.27
	UCA	70	0.904286
	UCG	23	0.317143
Pro	CCU	157	0.974286
	CCC∗	256	1.641429
	CCA	124	0.75
	CCG	102	0.635714
Thr	ACU	74	0.795714
	ACC∗	134	1.365714
	ACA∗	108	1.142857
	ACG	64	0.697143
Ala	GCU∗	128	1.047143
	GCC∗	199	1.585714
	GCA	99	0.735714
	GCG	79	0.631429
Tyr	UAU	30	0.678571
	UAC∗	75	1.321429
His	CAU	92	0.74
	CAC∗	148	1.26
Gln	CAA	144	0.864286
	CAG∗	132	1.135714
Asn	AAU	40	0.787143
	AAC∗	75	1.212857
Lys	AAA	58	0.822857
	AAG∗	138	1.178571
Asp	GAU	94	0.835714
	GAC∗	131	1.164286
Glu	GAA	116	0.944286
	GAG∗	121	1.055714
Cys	UGU	73	0.69
	UGC∗	151	1.31
Trp	UGG	206	1
Arg	CGU	34	0.428571
	CGC∗	104	1.327143
	CGA	68	0.772857
	CGG∗	92	1.177143
Ser	AGU	66	0.897143
	AGC∗	112	1.534286
Arg	AGA∗	109	1.212857
	AGG∗	93	1.08
Gly	GGU	71	0.527143
	GGC∗	207	1.371429
	GGA∗	156	1.074286
	GGG∗	149	1.03
TER	UAA	21	0.374286
	UAG	21	0.425714
	UGA∗∗	132	2.201429

Note. RSCU > 1; RSCU > 2; AA. amino acid; TER. termination codon; the same below.

Table 6

RSCU for MMP-9 gene among seven species.

AA	Codon	Frequency	RSCU
Phe	UUU	72	0.584285714
	UUC∗	197	1.415714286
Leu	UUA	17	0.17
	UUG	44	0.621428571
	CUU	76	0.848571429
	CUC∗	107	1.362857143
	CUA	33	0.357142857
	CUG∗∗	195	2.641428571
Ile	AUU	30	0.811428571
	AUC∗	61	1.881428571
	AUA	11	0.308571429
Met	AUG	49	1
Val	GUU	48	0.684285714
	GUC	66	0.957142857
	GUA	29	0.375714286
	GUG∗	147	1.978571429
Ser	UCU	75	0.994285714
	UCC∗	118	1.862857143
	UCA	41	0.494285714
	UCG	55	0.832857143
Pro	CCU	117	0.85
	CCC∗	209	1.644285714
	CCA	108	0.735714286
	CCG	102	0.774285714
Thr	ACU	74	0.688571429
	ACC∗∗	208	2.145714286
	ACA	54	0.484285714
	ACG	82	0.684285714
Ala	GCU	84	0.792857143
	GCC∗	175	1.731428571
	GCA	71	0.655714286
	GCG	87	0.815714286
Tyr	UAU	35	0.54
	UAC	119	1.46
His	CAU	39	0.441428571
	CAC∗	116	1.558571429
Gln	CAA	62	0.537142857
	CAG∗	136	1.462857143∗
Asn	AAU	33	0.664285714
	AAC∗	72	1.335714286
Lys	AAA	41	0.598571429
	AAG∗	105	1.401428571
Asp	GAU	70	0.571428571
	GAC∗	210	1.428571429
Glu	GAA	72	0.592857143
	GAG∗	159	1.407142857
Cys	UGU	46	0.417142857
	UGC	135	1.582857143
Trp	UGG∗	128	1
Arg	CGU	45	0.532857143
	CGC∗∗	134	2.072857143
	CGA	54	0.667142857
	CGG	77	0.965714286
Ser	AGU	31	0.418571429
	AGC∗	91	1.395714286
Arg	AGA	63	0.752857143
	AGG∗	79	1.011428571
Gly	GGU	77	0.615714286
	GGC∗	219	1.738571429
	GGA	86	0.654285714
	GGG	124	0.991428571
TER	UAA	18	0.571428571
	UAG∗	20	1.538571429
	UGA	37	0.89

Heat map analysis of the correlation between codon base composition and GC3S (Figure 2) showed that most of the codons of the MMP-2 and MMP-9 genes in different mammals were positively correlated with GC3S and in line with AC-, CG-, AT-, TC-, GG-, CC-, GC-, and other codons whose third base was C.

Figure 2

Heat maps of MMP-2 (a) and MMP-9. (b) Correlation coefficient of codons with GC3S.

Cluster analysis of the RSCU of the MMP-2 and MMP-9 genes showed that the MMP-2 gene preferred CUG, GUG, UCC, GAG, AUC, AAC, UAC, GCC, AGA, UUG, and AGG codons, which were mainly involved in encoding Leu (leucine), Val (valine), Ser (serine), Glu (glutamic acid), Iso (isoleucine) Asn (asparagine), Tyr (tyrosine), Gly (glycine), and Arg (arginine), respectively (Figures 3 and 4). In addition to B. taurus and C. lupus familiaris, the five other species had a strong preference for CUG and GUG (RSCU > 2), amongst which the RSCU of O. cuniculus and B. grunniens > 3. The MMP-9 gene preferred UCC, ACC, CGC, CUG, and AUC codons, which are mainly involved in Ser, Thr, Arg, Leu, and Iso, respectively. Except for M. musculus and O. cuniculus, the last five species had strong preferences similar to one another, indicating that the MMP-9 gene was more conservative than MMP-2.

Figure 3

Clustering of RSCU values of each codon in MMP-2 gene.

Figure 4

Clustering of RSCU values of each codon in MMP-9 gene.

3.4. Factors Influenced CUB

The PR2 plot result (Figure 5) showed that the ATCG base distribution of the MMP-2 and MMP-9 genes amongst seven mammals was above 0.5 on the x-axis. The bases distribution of the MMP-2 genes was mainly on the x-axis and the upper right of the y-axis and that of the MMP-9 genes was to the x-axis and the upper right of the y-axis. The above results indicated that the contents of A3S and C3S for the MMP-2 gene and the content of T3S and C3S for the MMP-9 gene were high, respectively.

Figure 5

PR2 plot among seven species.

Neutral analysis (Figure 6 and Table 7) showed that GC3S of these two genes was in the range of 0.44–0.78, whereas GC12 was from 0.52 to 0.67. The difference was that GC12 and GC3S of the MMP-2 gene were strongly negatively correlated (Pearson r = −0.851, p value < 0.05), whilst GC12 and GC3S of the MMP-9 gene were not significantly correlated, indicating that the base composition of the MMP-2 gene codons was susceptible to mutational pressure, but the factor influencing the MMP-9 gene was natural selection.

Figure 6

Codon neutral analysis.

Table 7

Pearson relative analysis with GC3S.

	Pearson r (MMP-2)	P values (MMP-2)	Pearson r (MMP-9)	P values (MMP-9)
ENC	−0.993	0.000∗∗	−0.963	0.000∗∗
GC12	−0.851	0.015∗	−0.589	0.164

Note. P value < 0.05; p value < 0.01; red represents strong correlation, blue represents moderate correlation, and black represents irrelevance.

The ENC plot showed (Figure 7 and Table 7) that all ENC/GC3S dots of the MMP-2 and MMP-9 genes were distributed below the reference line. ENC and GC3S had a strongly negative correlation (MMP-2: Pearson r = −0.993, p value < 0.01; MMP-9: Pearson r = −0.963, p value < 0.01), and the distribution range of GC3S was large, indicating that the CUB of these two genes was affected by mutational pressure.

Figure 7

ENC plot among seven species.

4. Discussion

This study found that gelatinase MMP genes had CUB for encoding amino acids such as Ile, Arg, Glu, and Ser related to muscle development and meat quality. Gly, Arg, and Leu can promote collagen synthesis, and animal muscle is the main way to obtain natural collagen for humans [29, 30]. Delicious amino acids (DAAs), including Glu, Gly, Ser, Asp, Arg, and Ile, are known as precursor substances that determine the flavour of meat and can improve the taste of chicken and keep the meat soft [31]. Recent research found that the quality of chicken improves and the content of DAAs increases [32]. Otherwise, Strecker amino acids (SAAs), including Phe (phenylalanine), Cys (cysteine), Ile (isoleucine), and Leu (leucine), are highly related to the production of flavour. The higher their content, the stronger the fragrance [33]. For the MMP-2 and MMP-9 genes, the RSCUs of AUC encoding Ile; UCC and AGC encoding Ser; CGC encoding Arg; GAC encoding Asp; GAG encoding Glu; UUC encoding Phe; and GGA, GGC, and GGG encoding Gly were > 1. In particular, the RSCUs of CUG encoding Leu and CGC encoding Arg > 2; this value indicated that MMP-2 and MMP-9 demonstrated CUB for DAAs and SAAs. Besides, Leu, Ile, and Val belong to branched-chain amino acids (BCAAs), and they are essential AAs in humans and animals, accounting for about 35% of muscle protein. Previous studies have found that skeletal muscle, as the initial site of BCAAs catabolism, can be activated by branched-chain keto acids (BCKAs) to increase BCAAs synthesis to relieve muscle wasting disorders [34]. Also, Leu supplementation could be the prevention and treatment of sarcopenia with aging [35]. Thus, BCAAs are important regulators of metabolism and metabolic health in in vivo [36]. The gelatinase MMP CUB associated with corresponding AAs can provide basic data for the improvement of meat quality and muscle disease of MMP molecular modification.

Mutational pressure may be the main factor influencing the CUB of MMPs. This study found that the clustering results of the RSCU were different from the NJ trees of the genes, indicating that the MMP genes were highly conserved but maybe subjected to mutations during the evolution of different species. This influence caused a decline in the accuracy of single-gene species classification. Nucleotide AT (U) CG base composition is an important feature of genes, and the GC content can reflect the overall trend of gene mutation which is a decisive factor affecting the frequency of nucleotide use. Changes in the third base of the codon did not affect the encoded AAs, so GC3S could be an important reference for analysing the codon usage pattern. The gene mutation will affect the composition of the synonymous codon third bases with no natural selection, and the stronger the CUB, the more the codon is inclined to GC3S [37, 38]. Novembre et al. also found that the third base distribution of the MMP-2 and MMP-9 genes is mainly AC3S and CT3S, respectively, and the ENC/GC3S dot distribution can reach a wide range compared with the reference curve with gene mutation pressure. Thus, mutational pressure may play an important role in affecting the CUB for MMP-2 and MMP-9 genes, which also explains the difference in RSCU clustering in the seven mammals.

Interestingly, we also found that the clustering results based on the RSCU of the MMP-2 gene were not completely consistent with the phylogenetic results based on the MMP-2 gene's CDS. Given that wild yak and Tibetan antelope grow in harsh environments with low altitudes and oxygen consumption, their EGLN1 gene has mutated changing nucleotide bases and leading to CUB changes [39, 40]. Therefore, we believe that the phylogenetic evolution of MMP-2 genes should not only refer to gene sequence but also CUB, which could be a supplement to species classification.

5. Conclusion

MMP-2 and MMP-9 are low-expression genes in mammals, and their codons are highly conservative. Both have a CUB at GC3S and prefer codons encoding DAAs and SAAs for improving soft meat and muscle disease treatment.